luke
/
magic
mirror of https://github.com/RTimothyEdwards/magic.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
magic/ext2spice/ext2spice.c

4769 lines
121 KiB
C
Raw Normal View History

Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ext2spice.c --
*
* Program to flatten hierarchical .ext files and produce
* a .spice file, suitable for use as input to simulators
* such as spice and hspice.
*
* Flattens the tree rooted at file.ext, reading in additional .ext
* files as specified by "use" lines in file.ext. The output is left
* in file.spice, unless '-o esSpiceFile' is specified, in which case the
* output is left in 'esSpiceFile'.
*
*/
#ifndef lint
ext2spice: use 'const' with 'rcsid'
2025-07-18 15:12:02 +02:00
static const char rcsid[] __attribute__ ((unused)) = "$Header: /usr/cvsroot/magic-8.0/ext2spice/ext2spice.c,v 1.8 2010/08/25 17:33:56 tim Exp $";
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
#endif /* not lint */
#include <stdio.h>
ext2spice/ext2spice.c: %x => (intmax_t) %lx Wrong type of arguments to formatting function Looks like SPICE comment, change maintains hex without 0x prefix in portable way. Copilot Autofix rejected: TxError("** %s (%p)\n", nsn, node); CodeQL: https://github.com/dlmiles/magic/security/code-scanning/70
2024-09-30 00:00:00 +02:00
#include <stdint.h>
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
#include <stdlib.h> /* for atof() */
#include <string.h>
strings.h: add #include required when strict Due to use of strcasecmp() or similar C API. Maybe HAVE_STRINGS_H is needed ? If so which platforms needs this ? clang18 default warning cleanup (strict)
2024-10-04 19:59:39 +02:00
#include <strings.h>
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
#include <ctype.h>
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
#include <math.h> /* for fabs() */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
#include "tcltk/tclmagic.h"
#include "utils/magic.h"
#include "utils/malloc.h"
#include "utils/geometry.h"
#include "utils/hash.h"
#include "utils/dqueue.h"
#include "utils/utils.h"
#include "tiles/tile.h"
#include "database/database.h"
#include "windows/windows.h"
#include "textio/textio.h"
#include "dbwind/dbwind.h" /* for DBWclientID */
#include "commands/commands.h" /* for module auto-load */
#include "textio/txcommands.h"
#include "extflat/extflat.h"
#include "extflat/EFint.h"
#include "extract/extract.h" /* for extDevTable */
#include "utils/runstats.h"
#include "ext2spice/ext2spice.h"
/* Options specific to ext2spice */
bool esDoExtResis = FALSE;
bool esDoPorts = TRUE;
bool esDoHierarchy = FALSE;
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
bool esDoBlackBox = FALSE;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
bool esDoRenumber = FALSE;
bool esDoResistorTee = FALSE;
int esDoSubckt = AUTO;
bool esDevNodesOnly = FALSE;
bool esMergeNames = TRUE;
bool esNoAttrs = FALSE;
bool esHierAP = FALSE;
char spcesDefaultOut[FNSIZE];
ext2sim.c: constify string globals
2025-07-24 15:20:12 +02:00
const char *esSpiceCapNode;
const char esSpiceDefaultGnd[] = "0";
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
char *spcesOutName = spcesDefaultOut;
FILE *esSpiceF = NULL;
float esScale = -1.0 ; /* negative if hspice the EFScale/100 otherwise */
unsigned short esFormat = SPICE3 ;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
int esCapNum, esDevNum, esResNum, esDiodeNum, esVoltNum;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int esNodeNum; /* just in case we're extracting spice2 */
int esSbckNum; /* used in hspice node name shortening */
int esNoModelType; /* index for device type "None" (model-less device) */
/*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* The following hash table and associated functions are used only if
* the format is hspice, to keep the translation between the hierarchical
* prefix of a node and the x num that we use to output valid hspice
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* which also are meaningful.
*/
HashTable subcktNameTable ; /* the hash table itself */
Corrected an issue with the limited bitfield dedicated to port numbers; this limited ports to 16384, which seemed reasonable at the time. However, the sky130_sram_macro layouts connect power and ground in a way that when coupled with "extract unique" can generate tens of thousands of ports and overrun the bit field, showing that automation can do the unexpected. The solution was to split out the port number from the label record as its own 32-bit value.
2021-12-13 04:09:31 +01:00
DQueue subcktNameQueue ; /* q used to print it sorted at the end */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
fetInfoList esFetInfo[TT_MAXTYPES];
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Record for keeping a list of global names */
typedef struct GLL *globalListPtr;
typedef struct GLL {
globalListPtr gll_next;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
char *gll_name;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
} globalList;
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
TileTypeBitMask initMask; /* Used for device types, not tile types */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
bool esMergeDevsA = FALSE; /* aggressive merging of devs L1=L2 merge them */
bool esMergeDevsC = FALSE; /* conservative merging of devs L1=L2 and W1=W2 */
/* used with the hspice multiplier */
bool esDistrJunct = FALSE;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
/*
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*---------------------------------------------------------
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Variables & macros used for merging parallel devs
* The merging of devs is based on the fact that spcdevVisit
* visits the devs in the same order all the time so the
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* value of esFMult[i] keeps the multiplier for the ith dev
*---------------------------------------------------------
*/
float *esFMult = NULL; /* the array itself */
int esFMIndex = 0; /* current index to it */
int esFMSize = FMULT_SIZE ; /* its current size (growable) */
int esSpiceDevsMerged;
extraction: constify struct devMerge.next, devMergeList Feel free to undo variables as necessary in the future as requirements change.
2025-07-24 15:36:22 +02:00
const devMerge *devMergeList = NULL ;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#define atoCap(s) ((EFCapValue)atof(s))
ext2spice: forward declaration arguments or removal Remove unnecessary forward declaration: extHierSDAttr() as ext2spice.h has prototype
2025-07-18 15:40:53 +02:00
extern void ESGenerateHierarchy(char *inName, int flags); /* forward reference ext2hier.c */
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
Tracked down two more memory leaks coming from ext2spice, due to client data generated by ext2spice and attached to a node's nodeClient record; there is an initNodeClient() routine but no corresponding freeNodeClient() routine. Eventually had to add a callback function passed to EFDone() and EFFlatDone() to clean up these entries. After doing that, valgrind reports clean for all memory allocated within ext2spice (there are other things that are not freed but not related to a specific command, so do not need to be treated as leaks).
2021-12-14 00:05:53 +01:00
/*
* ----------------------------------------------------------------------------
*
* esFreeNodeClient ---
*
* Free the string spiceNodeName associated with the nodeClient
* record that ext2spice allocates per each node structure.
*
* Returns:
* 0 always.
*
* Side effects:
* Frees an allocated string.
*
* ----------------------------------------------------------------------------
*/
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
esFreeNodeClient(
nodeClient *client)
Tracked down two more memory leaks coming from ext2spice, due to client data generated by ext2spice and attached to a node's nodeClient record; there is an initNodeClient() routine but no corresponding freeNodeClient() routine. Eventually had to add a callback function passed to EFDone() and EFFlatDone() to clean up these entries. After doing that, valgrind reports clean for all memory allocated within ext2spice (there are other things that are not freed but not related to a specific command, so do not need to be treated as leaks).
2021-12-14 00:05:53 +01:00
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
if (client != (nodeClient *)NULL)
Tracked down two more memory leaks coming from ext2spice, due to client data generated by ext2spice and attached to a node's nodeClient record; there is an initNodeClient() routine but no corresponding freeNodeClient() routine. Eventually had to add a callback function passed to EFDone() and EFFlatDone() to clean up these entries. After doing that, valgrind reports clean for all memory allocated within ext2spice (there are other things that are not freed but not related to a specific command, so do not need to be treated as leaks).
2021-12-14 00:05:53 +01:00
if (client->spiceNodeName != NULL)
freeMagic((char *)client->spiceNodeName);
return 0;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* Apply modifications to a global node name and output to file "outf"
*
* Results:
* None.
*
* Side effects:
* Output to file
* ----------------------------------------------------------------------------
*/
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
esFormatSubs(
FILE *outf,
char *suf)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
char *specchar;
int l;
if (outf)
{
l = strlen(suf) - 1;
12 x warning: suggest parentheses around '&&' within '||' The '&' has higher precedence, so the expression of the '&' side receive extra parentheses. ResMakeRes.c:671:37: warning: suggest parentheses around '&&' within '||' ResMakeRes.c:942:32: warning: suggest parentheses around '&&' within '||' ResSimple.c:201:70: warning: suggest parentheses around '&&' within '||' ResRex.c:1036:28: warning: suggest parentheses around '&&' within '||' ResRex.c:1038:50: warning: suggest parentheses around '&&' within '||' ext2sim.c:1341:47: warning: suggest parentheses around '&&' within '||' ext2spice.c:172:44: warning: suggest parentheses around '&&' within '||' ext2spice.c:173:49: warning: suggest parentheses around '&&' within '||' ext2spice.c:3655:50: warning: suggest parentheses around '&&' within '||' ext2hier.c:1412:50: warning: suggest parentheses around '&&' within '||' gcrDebug.c:638:55: warning: suggest parentheses around '&&' within '||' selEnum.c:323:28: warning: suggest parentheses around '&&' within '||' GCC14 -Wall cleanup series [-Wparentheses]
2024-10-04 18:24:47 +02:00
if (((EFOutputFlags & EF_TRIMGLOB ) && suf[l] == '!') ||
((EFOutputFlags & EF_TRIMLOCAL) && suf[l] == '#'))
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
suf[l] = '\0' ;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
if (EFOutputFlags & EF_CONVERTCOMMA)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
while ((specchar = strchr(suf, ',')) != NULL)
Added substitution for commas in SPICE output, because commas in instance names and node names are basically fatal to running ngspice.
2019-10-15 03:46:07 +02:00
*specchar = '|';
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
if (EFOutputFlags & EF_CONVERTBRACKETS)
Added substitution for commas in SPICE output, because commas in instance names and node names are basically fatal to running ngspice.
2019-10-15 03:46:07 +02:00
{
while ((specchar = strchr(suf, '[')) != NULL)
*specchar = '_';
while ((specchar = strchr(suf, ']')) != NULL)
*specchar = '_';
}
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
if (EFOutputFlags & EF_CONVERTEQUAL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
while ((specchar = strchr(suf, '=')) != NULL)
*specchar = ':';
fprintf(outf, "%s", suf);
}
}
#ifdef MAGIC_WRAPPER
#ifdef EXT2SPICE_AUTO
/*
* ----------------------------------------------------------------------------
*
* Tcl package initialization function
*
* ----------------------------------------------------------------------------
*/
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
Exttospice_Init(
Tcl_Interp *interp)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
/* Sanity checks! */
if (interp == NULL) return TCL_ERROR;
if (Tcl_PkgRequire(interp, "Tclmagic", MAGIC_VERSION, 0) == NULL)
return TCL_ERROR;
TCL9: Tcl_InitStubs(version="9.0")
2024-10-21 10:19:15 +02:00
if (Tcl_InitStubs(interp, Tclmagic_InitStubsVersion, 0) == NULL) return TCL_ERROR;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
TxPrintf("Auto-loading EXTTOSPICE module\n");
TxFlushOut();
/* Replace the auto-load function with the one defined in */
/* this package in the command functions list. */
if (WindReplaceCommand(DBWclientID, "exttospice", CmdExtToSpice) < 0)
return TCL_ERROR;
/* Add "ext2spice" as an alias for "exttospice" */
if (WindReplaceCommand(DBWclientID, "ext2spice", CmdExtToSpice) < 0)
return TCL_ERROR;
Tcl_PkgProvide(interp, "Exttospice", MAGIC_VERSION);
return TCL_OK;
}
#endif /* EXT2SPICE_AUTO */
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#endif /* MAGIC_WRAPPER */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* Main callback for command "magic::exttospice"
*
* ----------------------------------------------------------------------------
*/
#define EXTTOSPC_RUN 0
#define EXTTOSPC_DEFAULT 1
#define EXTTOSPC_FORMAT 2
#define EXTTOSPC_RTHRESH 3
#define EXTTOSPC_CTHRESH 4
#define EXTTOSPC_MERGE 5
#define EXTTOSPC_EXTRESIST 6
#define EXTTOSPC_RESISTORTEE 7
#define EXTTOSPC_SCALE 8
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
#define EXTTOSPC_SHORT 9
#define EXTTOSPC_SUBCIRCUITS 10
#define EXTTOSPC_HIERARCHY 11
#define EXTTOSPC_BLACKBOX 12
#define EXTTOSPC_RENUMBER 13
#define EXTTOSPC_MERGENAMES 14
#define EXTTOSPC_LVS 15
#define EXTTOSPC_HELP 16
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
CmdExtToSpice(
MagWindow *w,
TxCommand *cmd)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
int i,flatFlags;
char *inName;
int value;
int option = EXTTOSPC_RUN;
int argc = cmd->tx_argc;
char **argv = cmd->tx_argv;
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
const char * const *msg;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
char *resstr = NULL;
char *substr = NULL;
bool err_result, locDoSubckt;
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
short s_rclass, d_rclass, sub_rclass;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
char *devname;
char *subname;
Made a change to the way ext2sim determines if a FET type is an n or p device; this was previously dependent only on the first character of the extracted device model name. Since the tech file has control over what the device layer names are but not the extracted model names, the device layer type name is used as a backup way to determine if the type is n or p, if that cannot be determined from the extracted model name.
2020-10-20 18:22:02 +02:00
TileType devtype;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int idx, idx2;
globalList *glist = NULL;
static EFCapValue LocCapThreshold = 2;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
static int LocResistThreshold = INFINITE_THRESHOLD;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const spiceFormats[] = {
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
"SPICE2", "SPICE3", "HSPICE", "NGSPICE", "CDL", NULL
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
};
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const cmdExtToSpcOption[] = {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"[run] [options] run exttospice on current cell\n"
" use \"run -help\" to get standard options",
"default reset to default values",
"format [<type>] set output format",
"rthresh [<value>] set resistance threshold value",
"cthresh [<value>] set capacitance threshold value",
"merge [<type>] merge parallel transistors",
"extresist [on|off] incorporate information from extresist",
"resistor tee [on|off] model resistor capacitance as a T-network",
"scale [on|off] use .option card for scaling",
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
"short [voltage|resistor|none]\n"
" set method for handling shorted ports",
Fixed a long-standing error in which "ext2spice merge" fails when used with "ext2spice hierarchy on" because the device index is not reset between calls to output cells in the hierarchy, leading to a mismatch of the index for all cells after the first one output.
2019-10-23 15:33:37 +02:00
"subcircuits [top|descend] [on|off|auto]\n"
" standard cells become subcircuit calls",
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"hierarchy [on|off] output hierarchical spice for LVS",
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
"blackbox [on|off] output abstract views as black-box entries",
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"renumber [on|off] on = number instances X1, X2, etc.\n"
" off = keep instance ID names",
"global [on|off] on = merge unconnected global nets by name",
Two improvements: (1) Command extension "ext2spice lvs" sets up all the settings normally used for LVS (hierarchy on, cthresh infinite, subcircuit top auto, etc.). (2) Extract and extract unique ignore cells marked as abstract views (property LEFview is set) when checking for unconnected nets with the same name label.
2018-10-31 19:33:24 +01:00
"lvs apply typical default settings for LVS",
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"help print help information",
NULL
};
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const cmdMergeTypes[] = {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"none don't merge parallel devices",
"conservative merge devices with same L, W",
"aggressive merge devices with same L",
NULL
};
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const cmdShortTypes[] = {
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
"none merge shorted ports",
"resistor separate shorted ports with 0 ohm resistor",
"voltage separate shorted ports with 0 volt source",
NULL
};
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const cmdExtToSpcFormat[] = {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"spice2",
"spice3",
"hspice",
"ngspice",
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
"cdl",
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
NULL
};
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const yesno[] = {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"yes",
"true",
"on",
"no",
"false",
"off",
NULL
};
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const subcktopts[] = {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"yes",
"true",
"on",
"no",
"false",
"off",
"automatic",
"top",
"descend",
NULL
};
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
#if 0 /* -Wunused-variable */
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const shorttypes[] = {
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
"none",
"resistor",
"voltage",
NULL
};
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
#endif
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
ext2spice.c: cleanup -Wunused-local-typedefs yesnoType
2025-07-24 15:36:22 +02:00
enum {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
IDX_YES, IDX_TRUE, IDX_ON, IDX_NO, IDX_FALSE, IDX_OFF,
IDX_AUTO, IDX_TOP, IDX_DESCEND
ext2spice.c: cleanup -Wunused-local-typedefs yesnoType
2025-07-24 15:36:22 +02:00
};
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esNoModelType = -1;
if (cmd->tx_argc > 1)
{
option = Lookup(cmd->tx_argv[1], cmdExtToSpcOption);
if (option < 0) option = EXTTOSPC_RUN;
else argv++;
}
switch (option)
{
case EXTTOSPC_EXTRESIST:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esDoExtResis) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Extresist: %s\n", (esDoExtResis) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else if (cmd->tx_argc != 3)
goto usage;
idx = Lookup(cmd->tx_argv[2], yesno);
if (idx < 0) goto usage;
else if (idx < 3) esDoExtResis = TRUE;
else esDoExtResis = FALSE;
break;
case EXTTOSPC_RESISTORTEE:
if (cmd->tx_argc == 3)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esDoResistorTee) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Resistor tee: %s\n", (esDoResistorTee) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else if (cmd->tx_argc != 4)
goto usage;
idx = Lookup(cmd->tx_argv[3], yesno);
if (idx < 0) goto usage;
else if (idx < 3) esDoResistorTee = TRUE;
else esDoResistorTee = FALSE;
break;
case EXTTOSPC_SCALE:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esScale < 0) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Scale: %s\n", (esScale < 0) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else if (cmd->tx_argc != 3)
goto usage;
idx = Lookup(cmd->tx_argv[2], yesno);
if (idx < 0) goto usage;
else if (idx < 3) esScale = -1.0;
else esScale = 0.0;
break;
case EXTTOSPC_HIERARCHY:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esDoHierarchy) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Hierarchy: %s\n", (esDoHierarchy) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
idx = Lookup(cmd->tx_argv[2], yesno);
if (idx < 0) goto usage;
else if (idx < 3) /* yes */
esDoHierarchy = TRUE;
else /* no */
esDoHierarchy = FALSE;
break;
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
case EXTTOSPC_BLACKBOX:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
Tcl_SetResult(magicinterp, (esDoBlackBox) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Black box: %s\n", (esDoBlackBox) ? "on" : "off");
#endif
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
return;
}
idx = Lookup(cmd->tx_argv[2], yesno);
if (idx < 0) goto usage;
else if (idx < 3) /* yes */
esDoBlackBox = TRUE;
else /* no */
esDoBlackBox = FALSE;
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case EXTTOSPC_RENUMBER:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esDoRenumber) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Renumber: %s\n", (esDoRenumber) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
idx = Lookup(cmd->tx_argv[2], yesno);
if (idx < 0) goto usage;
else if (idx < 3) /* yes */
esDoRenumber = TRUE;
else /* no */
esDoRenumber = FALSE;
break;
case EXTTOSPC_MERGENAMES:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esMergeNames) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Merge names: %s\n", (esMergeNames) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
idx = Lookup(cmd->tx_argv[2], yesno);
if (idx < 0) goto usage;
else if (idx < 3) /* yes */
esMergeNames = TRUE;
else /* no */
esMergeNames = FALSE;
break;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
case EXTTOSPC_SHORT:
if (cmd->tx_argc == 2)
{
if ((EFOutputFlags & EF_SHORT_MASK) == EF_SHORT_NONE)
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
Tcl_SetResult(magicinterp, "none", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Shorts: none\n");
#endif
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
else if ((EFOutputFlags & EF_SHORT_MASK) == EF_SHORT_R)
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
Tcl_SetResult(magicinterp, "resistor", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Shorts: resistor\n");
#endif
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
else if ((EFOutputFlags & EF_SHORT_MASK) == EF_SHORT_V)
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
Tcl_SetResult(magicinterp, "voltage source", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Shorts: voltage source\n");
#endif
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
return;
}
idx = Lookup(cmd->tx_argv[2], cmdShortTypes);
if (idx < 0) goto usage;
else switch (idx)
{
case 0:
EFOutputFlags &= ~EF_SHORT_MASK;
EFOutputFlags |= EF_SHORT_NONE;
break;
case 1:
EFOutputFlags &= ~EF_SHORT_MASK;
EFOutputFlags |= EF_SHORT_R;
break;
case 2:
EFOutputFlags &= ~EF_SHORT_MASK;
EFOutputFlags |= EF_SHORT_V;
break;
}
break;
Two improvements: (1) Command extension "ext2spice lvs" sets up all the settings normally used for LVS (hierarchy on, cthresh infinite, subcircuit top auto, etc.). (2) Extract and extract unique ignore cells marked as abstract views (property LEFview is set) when checking for unconnected nets with the same name label.
2018-10-31 19:33:24 +01:00
case EXTTOSPC_LVS:
/* Apply default command settings for LVS */
/* hierarchy = on */
/* format = ngspice */
/* cthresh = infinite */
/* rthresh = infinite */
/* renumber = off */
/* scale = off */
/* blackbox = on */
Corrected "ext2spice lvs" to add "global off" as a setting, which is normal for top level designs (no implicit label connections should be made at the top level).
2018-10-31 20:41:22 +01:00
/* global = off */
Two improvements: (1) Command extension "ext2spice lvs" sets up all the settings normally used for LVS (hierarchy on, cthresh infinite, subcircuit top auto, etc.). (2) Extract and extract unique ignore cells marked as abstract views (property LEFview is set) when checking for unconnected nets with the same name label.
2018-10-31 19:33:24 +01:00
/* subcircuit top = auto */
esDoHierarchy = TRUE;
esFormat = NGSPICE;
LocCapThreshold = (EFCapValue)INFINITE_THRESHOLD_F;
LocResistThreshold = INFINITE_THRESHOLD;
esDoRenumber = FALSE;
esScale = 0.0;
esDoBlackBox = TRUE;
Corrected "ext2spice lvs" to add "global off" as a setting, which is normal for top level designs (no implicit label connections should be made at the top level).
2018-10-31 20:41:22 +01:00
esMergeNames = FALSE;
Two improvements: (1) Command extension "ext2spice lvs" sets up all the settings normally used for LVS (hierarchy on, cthresh infinite, subcircuit top auto, etc.). (2) Extract and extract unique ignore cells marked as abstract views (property LEFview is set) when checking for unconnected nets with the same name label.
2018-10-31 19:33:24 +01:00
esDoSubckt = 2;
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case EXTTOSPC_SUBCIRCUITS:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esDoPorts) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Ports: %s\n", (esDoPorts) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
idx = Lookup(cmd->tx_argv[2], subcktopts);
switch (idx) {
case IDX_YES: case IDX_TRUE: case IDX_ON:
esDoPorts = TRUE;
return;
break;
case IDX_NO: case IDX_FALSE: case IDX_OFF:
esDoPorts = FALSE;
return;
break;
case IDX_DESCEND:
if (cmd->tx_argc == 3)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, (esDoPorts) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Ports: %s\n", (esDoPorts) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
break;
case IDX_TOP:
if (cmd->tx_argc == 3)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp,
(esDoSubckt == 2) ? "auto" :
(esDoSubckt == 1) ? "on" : "off", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Top subcircuit: %s\n",
(esDoSubckt == 2) ? "auto" :
(esDoSubckt == 1) ? "on" : "off");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
break;
default:
goto usage;
break;
}
if (cmd->tx_argc != 4) goto usage;
idx2 = Lookup(cmd->tx_argv[3], subcktopts);
switch (idx2) {
case IDX_YES: case IDX_TRUE: case IDX_ON:
if (idx == IDX_DESCEND)
esDoPorts = TRUE;
else
esDoSubckt = TRUE;
break;
case IDX_NO: case IDX_FALSE: case IDX_OFF:
if (idx == IDX_DESCEND)
esDoPorts = FALSE;
else
esDoSubckt = FALSE;
break;
case IDX_AUTO:
esDoSubckt = AUTO;
break;
default:
goto usage;
}
break;
case EXTTOSPC_FORMAT:
if (cmd->tx_argc == 2)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
/* FIXME check NULL is ok here, should this be TCL_STATIC ? TCL_STATIC==0 maybe using TCL_STATIC is better? */
Tcl_SetResult(magicinterp, (char*)cmdExtToSpcFormat[esFormat], NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Format: %s\n", cmdExtToSpcFormat[esFormat]);
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else if (cmd->tx_argc < 3) goto usage;
idx = Lookup(cmd->tx_argv[2], cmdExtToSpcFormat);
if (idx < 0)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, "Bad format type. Formats are:"
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
"spice2, spice3, hspice, ngspice, and cdl.", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxError("Bad format type. Formats are:"
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
"spice2, spice3, hspice, ngspice, and cdl.");
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else
{
esFormat = idx;
/* By default, use .option to declare scale in HSPICE mode */
if (esFormat == HSPICE) esScale = -1.0;
}
break;
case EXTTOSPC_CTHRESH:
if (cmd->tx_argc == 2)
{
if (!IS_FINITE_F(LocCapThreshold))
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, "infinite", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Capacitance threshold: infinite\n");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetObjResult(magicinterp,
Tcl_NewDoubleObj((double)LocCapThreshold));
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Capacitance threshold: %lf\n",
(double)LocCapThreshold);
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else if (cmd->tx_argc < 3) goto usage;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Note that strtod() (called by StrIsNumeric()) accepts */
/* "infinite" as a valid numerical value; however, the */
/* conversion to C type INF is *not* INFINITE_THRESHOLD, so */
/* we need to check this case first. . . */
if (!strncmp(cmd->tx_argv[2], "inf", 3))
LocCapThreshold = (EFCapValue)INFINITE_THRESHOLD_F;
else if (StrIsNumeric(cmd->tx_argv[2]))
LocCapThreshold = atoCap(cmd->tx_argv[2]);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
TxError("exttospice: numeric value or \"infinite\" expected.\n");
break;
case EXTTOSPC_RTHRESH:
if (cmd->tx_argc == 2)
{
if (LocResistThreshold == INFINITE_THRESHOLD)
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, "infinite", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Resistance threshold: infinite\n");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetObjResult(magicinterp,
Tcl_NewIntObj(LocResistThreshold));
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Resistance threshold: %lf\n",
(double)LocResistThreshold);
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else if (cmd->tx_argc < 3) goto usage;
if (StrIsInt(cmd->tx_argv[2]))
LocResistThreshold = atoi(cmd->tx_argv[2]);
else if (!strncmp(cmd->tx_argv[2], "inf", 3))
LocResistThreshold = INFINITE_THRESHOLD;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
TxError("exttospice: integer value or \"infinite\" expected.\n");
break;
case EXTTOSPC_MERGE:
if (cmd->tx_argc == 2)
{
if (esMergeDevsA)
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, "aggressive", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Merge: aggressive\n");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else if (esMergeDevsC)
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, "conservative", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Merge: conservative\n");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Tcl_SetResult(magicinterp, "none", NULL);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxPrintf("Merge: none\n");
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
else if (cmd->tx_argc < 3) goto usage;
value = Lookup(cmd->tx_argv[2], cmdMergeTypes);
if (value < 0)
{
TxError("Merge types are:\n");
for (msg = &(cmdMergeTypes[0]); *msg != NULL; msg++)
TxPrintf(" %s\n", *msg);
}
else switch (value) {
case 0:
esMergeDevsA = FALSE;
esMergeDevsC = FALSE;
break;
case 1:
esMergeDevsA = FALSE;
esMergeDevsC = TRUE;
break;
case 2:
esMergeDevsA = TRUE;
esMergeDevsC = FALSE;
break;
}
break;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case EXTTOSPC_DEFAULT:
LocCapThreshold = 2;
LocResistThreshold = INFINITE_THRESHOLD;
Correction to prior commit to remove collision between defined flag bits for name trimming and short handling method in the ext2spice options.
2021-07-11 03:27:19 +02:00
/* Clear EFOutputFlags but preserve "short" behavior */
EFOutputFlags &= ~EF_TRIM_MASK;
EFOutputFlags |= EF_CONVERTCOMMA | EF_CONVERTEQUAL;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFScale = 0.0;
if (EFArgTech)
{
freeMagic(EFArgTech);
EFArgTech = NULL;
}
if (EFSearchPath)
{
freeMagic(EFSearchPath);
EFSearchPath = NULL;
}
break;
case EXTTOSPC_RUN:
goto runexttospice;
break;
case EXTTOSPC_HELP:
usage:
for (msg = &(cmdExtToSpcOption[0]); *msg != NULL; msg++)
{
TxPrintf(" %s\n", *msg);
}
break;
}
return;
runexttospice:
/* Reset the device indices */
esCapNum = 0;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
esVoltNum = 0;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esDevNum = 1000;
esResNum = 0;
esDiodeNum = 0;
esSbckNum = 0;
esNodeNum = 10; /* just in case we're extracting spice2 */
esFMIndex = 0;
esSpiceDevsMerged = 0;
esDevNodesOnly = FALSE; /* so using -F doesn't become permanent */
EFInit();
EFResistThreshold = LocResistThreshold;
EFCapThreshold = LocCapThreshold;
/* Process command line arguments */
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
inName = EFArgs(argc, argv, &err_result, spcParseArgs, (ClientData) NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (err_result == TRUE)
{
ext2spice/ext2spice.c: EFDone() Call to function with fewer arguments than declared parameters Passing NULL is valid in efFreeNodeList() to not perform callback.. extflat/EFdef.c: ^231a299 (2017-04-25 08:41:48 -0400 97) void 5e3c26c9 (2021-12-13 18:05:53 -0500 98) EFDone(func) 5e3c26c9 (2021-12-13 18:05:53 -0500 99) int (*func)(); commit 5e3c26c95afadfd07753bb8126fa6528afce5dd8 (tag: 8.3.241) Date: Mon Dec 13 18:05:53 2021 -0500 extflat/EFdef.c:97 void EFDone(int (*func)()); extflat/EFbuild.c:2180 void efFreeNodeList(EFNode *head, int (*func)()); CodeQL: https://github.com/dlmiles/magic/security/code-scanning/20 https://github.com/dlmiles/magic/security/code-scanning/21 https://github.com/dlmiles/magic/security/code-scanning/22
2024-09-30 00:00:00 +02:00
EFDone(NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
if (inName == NULL)
{
/* Assume that we want to do exttospice on the currently loaded cell */
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (w == (MagWindow *) NULL)
{
windCheckOnlyWindow(&w, DBWclientID);
}
if (w == (MagWindow *) NULL)
{
TxError("Point to a window or specify a cell name.\n");
return;
}
if ((inName = ((CellUse *) w->w_surfaceID)->cu_def->cd_name) == NULL)
{
TxError("No cell present\n");
return;
}
}
/*
* Initializations specific to this program.
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Make output name inName.spice if they weren't
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* explicitly specified
*/
if (spcesOutName == spcesDefaultOut)
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
{
if (esFormat == CDL)
sprintf(spcesDefaultOut, "%s.cdl", inName);
else
sprintf(spcesDefaultOut, "%s.spice", inName);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Read the hierarchical description of the input circuit */
Corrects "ext2spice hierarchy off" when subcircuits with ports are present. Note, however, this fix breaks the use of "ext2spice subcircuit descend off" because subcircuits are always descended into. It's not clear that "subcircuit descend off" worked at all before, anyway. That still needs fixing.
2024-04-03 03:55:57 +02:00
if (EFReadFile(inName, esDoHierarchy, esDoExtResis, FALSE, TRUE)
== FALSE)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
ext2spice/ext2spice.c: EFDone() Call to function with fewer arguments than declared parameters Passing NULL is valid in efFreeNodeList() to not perform callback.. extflat/EFdef.c: ^231a299 (2017-04-25 08:41:48 -0400 97) void 5e3c26c9 (2021-12-13 18:05:53 -0500 98) EFDone(func) 5e3c26c9 (2021-12-13 18:05:53 -0500 99) int (*func)(); commit 5e3c26c95afadfd07753bb8126fa6528afce5dd8 (tag: 8.3.241) Date: Mon Dec 13 18:05:53 2021 -0500 extflat/EFdef.c:97 void EFDone(int (*func)()); extflat/EFbuild.c:2180 void efFreeNodeList(EFNode *head, int (*func)()); CodeQL: https://github.com/dlmiles/magic/security/code-scanning/20 https://github.com/dlmiles/magic/security/code-scanning/21 https://github.com/dlmiles/magic/security/code-scanning/22
2024-09-30 00:00:00 +02:00
EFDone(NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
/* If the .ext file was read without error, then open the output file */
if ((esSpiceF = fopen(spcesOutName, "w")) == NULL)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
char *tclres = Tcl_Alloc(128);
sprintf(tclres, "exttospice: Unable to open file %s for writing\n",
spcesOutName);
Tcl_SetResult(magicinterp, tclres, TCL_DYNAMIC);
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#else
TxError("exttospice: Unable to open file %s for writing\n", spcesOutName);
#endif
ext2spice/ext2spice.c: EFDone() Call to function with fewer arguments than declared parameters Passing NULL is valid in efFreeNodeList() to not perform callback.. extflat/EFdef.c: ^231a299 (2017-04-25 08:41:48 -0400 97) void 5e3c26c9 (2021-12-13 18:05:53 -0500 98) EFDone(func) 5e3c26c9 (2021-12-13 18:05:53 -0500 99) int (*func)(); commit 5e3c26c95afadfd07753bb8126fa6528afce5dd8 (tag: 8.3.241) Date: Mon Dec 13 18:05:53 2021 -0500 extflat/EFdef.c:97 void EFDone(int (*func)()); extflat/EFbuild.c:2180 void efFreeNodeList(EFNode *head, int (*func)()); CodeQL: https://github.com/dlmiles/magic/security/code-scanning/20 https://github.com/dlmiles/magic/security/code-scanning/21 https://github.com/dlmiles/magic/security/code-scanning/22
2024-09-30 00:00:00 +02:00
EFDone(NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return;
}
if (EFStyle == NULL)
{
TxError("Warning: Current extraction style does not match .ext file!\n");
TxError("Area/Perimeter values and parasitic values will be zero.\n");
}
/* create default devinfo entries (MOSIS) which can be overridden by
the command line arguments */
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
for ( i = 0 ; i < TT_MAXTYPES ; i++ ) {
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[i].resClassSource = NO_RESCLASS;
esFetInfo[i].resClassDrain = NO_RESCLASS;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFetInfo[i].resClassSub = NO_RESCLASS;
esFetInfo[i].defSubs = NULL;
}
/* Get esFetInfo information from the current extraction style */
/* (this works only for the Tcl version with the embedded exttospice */
/* command) */
idx = 0;
Made a change to the way ext2sim determines if a FET type is an n or p device; this was previously dependent only on the first character of the extracted device model name. Since the tech file has control over what the device layer names are but not the extracted model names, the device layer type name is used as a backup way to determine if the type is n or p, if that cannot be determined from the extracted model name.
2020-10-20 18:22:02 +02:00
while (ExtGetDevInfo(idx++, &devname, &devtype, &s_rclass, &d_rclass,
&sub_rclass, &subname))
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
if (idx == TT_MAXTYPES)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
TxError("Error: Ran out of space for device types!\n");
break;
}
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
i = efBuildAddStr(EFDevTypes, &EFDevNumTypes, TT_MAXTYPES, devname);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (!strcmp(devname, "None"))
esNoModelType = i;
if (EFStyle != NULL)
{
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[i].resClassSource = s_rclass;
esFetInfo[i].resClassDrain = d_rclass;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFetInfo[i].resClassSub = sub_rclass;
esFetInfo[i].defSubs = subname;
}
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (EFCompat == TRUE)
{
/* Tcl variable substitution for substrate node names */
if (subname && (subname[0] == '$'))
{
resstr = (char *)Tcl_GetVar(magicinterp, &subname[1],
TCL_GLOBAL_ONLY);
if (resstr != NULL) esFetInfo[i].defSubs = resstr;
}
}
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esDoHierarchy && (subname != NULL))
{
globalList *glptr;
char *locsubname, *bangptr;
bool isgood = TRUE;
locsubname = StrDup(NULL, subname);
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
if (esFormat != CDL)
{
bangptr = locsubname + strlen(locsubname) - 1;
if (*bangptr == '!') *bangptr = '\0';
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
// Ad-hoc check: Global names with "Error", "err", etc.
// should be rejected from the list. Also node name
// "None" is a common entry indicating that extracting
// an implicit substrate is disallowed.
if (!strncmp(locsubname, "err", 3)) isgood = FALSE;
else if (strstr(locsubname, "error") != NULL) isgood = FALSE;
else if (strstr(locsubname, "Error") != NULL) isgood = FALSE;
else if (strstr(locsubname, "ERROR") != NULL) isgood = FALSE;
else if (!strcasecmp(locsubname, "None")) isgood = FALSE;
for (glptr = glist; glptr; glptr = glptr->gll_next)
if (!strcmp(glptr->gll_name, locsubname))
break;
if (isgood && (glptr == NULL))
{
glptr = (globalList *)mallocMagic(sizeof(globalList));
glptr->gll_name = locsubname;
glptr->gll_next = glist;
glist = glptr;
}
else
freeMagic(locsubname);
}
}
ext2spice.c: ifdef for !MAGIC_WRAPPER build
2025-10-09 00:12:35 +02:00
#ifdef MAGIC_WRAPPER
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
if (esFormat == CDL)
{
/* In CDL format, if the global substrate ends with "!" then
* add it to the list of globals; likewise for VDD and GND.
*/
char *glbstr;
globalList *glptr;
glbstr = (char *)Tcl_GetVar(magicinterp, "SUB", TCL_GLOBAL_ONLY);
if ((glbstr != NULL) && (*(glbstr + strlen(glbstr) - 1) == '!'))
{
glptr = (globalList *)mallocMagic(sizeof(globalList));
glptr->gll_name = StrDup((char **)NULL, glbstr);
glptr->gll_next = glist;
glist = glptr;
}
glbstr = (char *)Tcl_GetVar(magicinterp, "VDD", TCL_GLOBAL_ONLY);
if ((glbstr != NULL) && (*(glbstr + strlen(glbstr) - 1) == '!'))
{
glptr = (globalList *)mallocMagic(sizeof(globalList));
glptr->gll_name = StrDup((char **)NULL, glbstr);
glptr->gll_next = glist;
glist = glptr;
}
glbstr = (char *)Tcl_GetVar(magicinterp, "GND", TCL_GLOBAL_ONLY);
if ((glbstr != NULL) && (*(glbstr + strlen(glbstr) - 1) == '!'))
{
glptr = (globalList *)mallocMagic(sizeof(globalList));
glptr->gll_name = StrDup((char **)NULL, glbstr);
glptr->gll_next = glist;
glist = glptr;
}
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (EFCompat == TRUE)
{
/* Keep a pointer to the "GND" variable, if it exists. */
resstr = (char *)Tcl_GetVar(magicinterp, "GND", TCL_GLOBAL_ONLY);
if (resstr == NULL) resstr = "GND"; /* default value */
}
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Write the output file */
fprintf(esSpiceF, "* %s file created from %s.ext - technology: %s\n\n",
spiceFormats[esFormat], inName, EFTech);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (esScale < 0)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
fprintf(esSpiceF, ".option scale=");
esSIvalue(esSpiceF, 1.0E-6 * EFScale / 100.0);
fprintf(esSpiceF, "\n\n");
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
esScale = EFScale / 100.0;
/* Set output format flags */
flatFlags = EF_FLATNODES;
if (esMergeNames == FALSE) flatFlags |= EF_NONAMEMERGE;
// This forces options TRIMGLOB and CONVERTEQUAL, not sure that's such a
// good idea. . .
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
EFOutputFlags |= EF_TRIMGLOB | EF_CONVERTEQUAL | EF_CONVERTCOMMA;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (IS_FINITE_F(EFCapThreshold)) flatFlags |= EF_FLATCAPS;
if (esFormat == HSPICE)
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
EFOutputFlags |= EF_TRIMLOCAL;
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
if (esFormat == CDL)
EFOutputFlags &= ~EF_TRIMGLOB;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Write globals under a ".global" card */
if (esDoHierarchy && (glist != NULL))
{
fprintf(esSpiceF, ".global ");
while (glist != NULL)
{
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#ifdef MAGIC_WRAPPER
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (EFCompat == TRUE)
{
/* Handle global names that are TCL variables */
if (glist->gll_name[0] == '$')
{
resstr = (char *)Tcl_GetVar(magicinterp,
&(glist->gll_name[1]), TCL_GLOBAL_ONLY);
if (resstr != NULL)
esFormatSubs(esSpiceF, resstr);
else
esFormatSubs(esSpiceF, glist->gll_name);
}
else
esFormatSubs(esSpiceF, glist->gll_name);
}
else
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
#endif
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFormatSubs(esSpiceF, glist->gll_name);
fprintf(esSpiceF, " ");
freeMagic(glist->gll_name);
freeMagic(glist);
glist = glist->gll_next;
}
fprintf(esSpiceF, "\n\n");
}
/* Convert the hierarchical description to a flat one */
if (esFormat == HSPICE) {
HashInit(&subcktNameTable, 32, HT_STRINGKEYS);
#ifndef UNSORTED_SUBCKT
DQInit(&subcktNameQueue, 64);
#endif
}
locDoSubckt = FALSE;
if (esDoHierarchy)
{
ESGenerateHierarchy(inName, flatFlags);
}
else
{
EFFlatBuild(inName, flatFlags);
/* Determine if this is a subcircuit */
if (esDoSubckt == AUTO) {
if (efFlatRootDef->def_flags & DEF_SUBCIRCUIT)
locDoSubckt = TRUE;
}
if ((esDoSubckt == TRUE) || (locDoSubckt == TRUE))
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
topVisit(efFlatRootDef, FALSE);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* When generating subcircuits, remove the subcircuit */
/* flag from the top level cell. Other than being */
/* used to generate the subcircuit wrapper, it should */
/* not prevent descending into its own hierarchy. */
efFlatRootDef->def_flags &= ~(DEF_SUBCIRCUIT);
/* If we don't want to write subcircuit calls, remove */
/* the subcircuit flag from all cells at this time. */
if (!esDoPorts)
EFVisitSubcircuits(subcktUndef, (ClientData) NULL);
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
TTMaskZero(&initMask);
if (!esDistrJunct)
(1) Added a check for unclosed boundaries when reading GDS. This is diagnostic only and does not change the read-in behavior. (2) ext2spice: Corrected an error that had been introduced into version 8.3.171 that accidentally marks all devices as visited which causes all source/drain areas and perimeters to be output as zero. (3) extract: Sweeping changes to handling of fringe capacitance. Removed the (recently added) "fringeshieldhalo" parameter from the tech file. Reworked the fringe capacitance models based on results from the "capiche" project (github/RTimothyEdwards/capiche). Fringe shielding is now done by clipping fringe at the boundary of a shielding shape, rather than trying to calculate the amount of shielding (as the "capiche" project proved this to be equivalent). Values for partial fringing are modeled by atan(x), which like the sidewall (1/x) curve, extends to infinity and values are limited by the halo but do not otherwise depend on the halo. Because of this, the halo can be made variable and controlled by the user for deciding on the tradeoff between accuracy and run time. A new command option "extract halo" was added to allow this control over the halo distance.
2023-01-27 17:47:37 +01:00
TTMaskSetType(&initMask, efNumResistClasses);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esMergeDevsA || esMergeDevsC)
{
extraction: constify struct devMerge.next, devMergeList Feel free to undo variables as necessary in the future as requirements change.
2025-07-24 15:36:22 +02:00
const devMerge *p;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFVisitDevs(devMergeVisit, (ClientData) NULL);
TxPrintf("Devs merged: %d\n", esSpiceDevsMerged);
esFMIndex = 0;
for (p = devMergeList; p != NULL; p = p->next)
extraction: constify struct devMerge.next, devMergeList Feel free to undo variables as necessary in the future as requirements change.
2025-07-24 15:36:22 +02:00
freeMagic((char *) p);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
devMergeList = NULL;
}
else if (esDistrJunct)
EFVisitDevs(devDistJunctVisit, (ClientData) NULL);
EFVisitDevs(spcdevVisit, (ClientData) NULL);
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
TTMaskZero(&initMask);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (flatFlags & EF_FLATCAPS)
EFVisitCaps(spccapVisit, (ClientData) NULL);
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
extflat: integrate cb_extflat_visitresists_t for EFVisitResists() Additionally the HierName's hierName1 and hierName2 arguments have been made 'const' to help convey the receiver can not modify the referenced data passed.
2025-07-20 17:09:08 +02:00
EFVisitResists(spcresistVisit, PTR2CD(NULL));
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFVisitSubcircuits(subcktVisit, (ClientData) NULL);
/* Visit nodes to find the substrate node */
EFVisitNodes(spcsubVisit, (ClientData)&substr);
if (substr == NULL)
Corrected an issue from the last commit, which is that the default ground node name (which is static) gets put on the node list and is improperly deallocated. Corrected by simply allocating the string for the default substrate node instead of using the static string.
2023-03-14 17:53:22 +01:00
substr = StrDup((char **)NULL, esSpiceDefaultGnd);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
esSpiceCapNode = substr;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFVisitNodes(spcnodeVisit, (ClientData) NULL);
if (EFCompat == FALSE) freeMagic(substr);
if ((esDoSubckt == TRUE) || (locDoSubckt == TRUE))
fprintf(esSpiceF, ".ends\n");
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (esFormat == HSPICE)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
printSubcktDict();
Tracked down two more memory leaks coming from ext2spice, due to client data generated by ext2spice and attached to a node's nodeClient record; there is an initNodeClient() routine but no corresponding freeNodeClient() routine. Eventually had to add a callback function passed to EFDone() and EFFlatDone() to clean up these entries. After doing that, valgrind reports clean for all memory allocated within ext2spice (there are other things that are not freed but not related to a specific command, so do not need to be treated as leaks).
2021-12-14 00:05:53 +01:00
EFFlatDone(esFreeNodeClient);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Tracked down two more memory leaks coming from ext2spice, due to client data generated by ext2spice and attached to a node's nodeClient record; there is an initNodeClient() routine but no corresponding freeNodeClient() routine. Eventually had to add a callback function passed to EFDone() and EFFlatDone() to clean up these entries. After doing that, valgrind reports clean for all memory allocated within ext2spice (there are other things that are not freed but not related to a specific command, so do not need to be treated as leaks).
2021-12-14 00:05:53 +01:00
EFDone(esFreeNodeClient);
Corrected an issue with the limited bitfield dedicated to port numbers; this limited ports to 16384, which seemed reasonable at the time. However, the sky130_sram_macro layouts connect power and ground in a way that when coupled with "extract unique" can generate tens of thousands of ports and overrun the bit field, showing that automation can do the unexpected. The solution was to split out the port number from the label record as its own 32-bit value.
2021-12-13 04:09:31 +01:00
if (esFormat == HSPICE) {
HashKill(&subcktNameTable);
#ifndef UNSORTED_SUBCKT
DQFree(&subcktNameQueue);
#endif
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esSpiceF) fclose(esSpiceF);
TxPrintf("exttospice finished.\n");
return;
}
ext2spice: Makefile use EXT2SPICE_MAIN for deprecated standalone
2025-07-18 15:12:51 +02:00
#ifdef EXT2SPICE_MAIN /* Independent program "ext2spice" has been deprecated */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* main --
*
* Top level of ext2spice.
*
* ----------------------------------------------------------------------------
*/
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
main(
int argc,
char *argv[])
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
int i,flatFlags;
char *inName;
FILE *f;
bool locDoSubckt;
esSpiceDevsMerged = 0;
Lookup() constify call-site with Tcl_SetResult() ext2spice/ext2spice.c: Lookup() constify call-site ext2sim/ext2sim.c: Lookup() constify call-site windows/windCmdAM.c: Lookup() constify call-site Tcl_SetResult() uses cast to remove 'const' from type, the pointer is only used to take a copy of the data, the lack of 'const' is due to Tcl heritage when supporting C89 era compilers. TCL9 appears to fix this, in that the macro used ends up at Tcl_NewStringObj() which has 'const' here.
2024-10-10 21:16:09 +02:00
static const char * const spiceFormats[] = {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"SPICE2", "SPICE3", "HSPICE", "NGSPICE", NULL
};
EFInit();
EFResistThreshold = INFINITE_THRESHOLD ;
/* create default devinfo entries (MOSIS) which can be overriden by
the command line arguments */
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
for ( i = 0 ; i < TT_MAXTYPES ; i++ ) {
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[i].resClassSource = NO_RESCLASS;
esFetInfo[i].resClassDrain = NO_RESCLASS;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFetInfo[i].resClassSub = NO_RESCLASS;
esFetInfo[i].defSubs = NULL;
}
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
i = efBuildAddStr(EFDevTypes, &EFDevNumTypes, TT_MAXTYPES, "ndev");
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[i].resClassSource = esFetInfo[i].resClassDrain = 0 ;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFetInfo[i].resClassSub = NO_RESCLASS ;
esFetInfo[i].defSubs = "Gnd!";
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
i = efBuildAddStr(EFDevTypes, &EFDevNumTypes, TT_MAXTYPES, "pdev");
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[i].resClassSource = esFetInfo[i].resClassDrain = 1 ;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFetInfo[i].resClassSub = 8 ;
esFetInfo[i].defSubs = "Vdd!";
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
i = efBuildAddStr(EFDevTypes, &EFDevNumTypes, TT_MAXTYPES, "nmos");
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[i].resClassSource = esFetInfo[i].resClassDrain = 0 ;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFetInfo[i].resClassSub = NO_RESCLASS ;
esFetInfo[i].defSubs = "Gnd!";
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
i = efBuildAddStr(EFDevTypes, &EFDevNumTypes, TT_MAXTYPES, "pmos");
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[i].resClassSource = esFetInfo[i].resClassDrain = 1 ;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFetInfo[i].resClassSub = 8 ;
esFetInfo[i].defSubs = "Vdd!";
/* Process command line arguments */
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
inName = EFArgs(argc, argv, NULL, spcParseArgs, (ClientData) NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (inName == NULL)
exit (1);
/*
* Initializations specific to this program.
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Make output name inName.spice if they weren't
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* explicitly specified
*/
if (spcesOutName == spcesDefaultOut)
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
{
if (esFormat == CDL)
sprintf(spcesDefaultOut, "%s.cdl", inName);
else
sprintf(spcesDefaultOut, "%s.spice", inName);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if ((esSpiceF = fopen(spcesOutName, "w")) == NULL)
{
perror(spcesOutName);
exit (1);
}
/* Read the hierarchical description of the input circuit */
Corrected a previous implementation that was supposed to not print warning messages about ports being electrically connected when those ports have names that match under rules of case-insensitivity, and the .ext file is being read for the purpose of generating a SPICE netlist, which is case-insensitive. Also: Corrected a crash condition when using "extract path <name>" when directory <name> does not exist.
2023-07-28 15:40:41 +02:00
if (EFReadFile(inName, TRUE, esDoExtResis, FALSE, TRUE) == FALSE)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
exit (1);
}
fprintf(esSpiceF, "* %s file created from %s.ext - technology: %s\n\n",
spiceFormats[esFormat], inName, EFTech);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (esScale < 0)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
fprintf(esSpiceF,".option scale=");
esSIvalue(esSpiceF, 1.0E-6 * EFScale / 100.0);
fprintf(esSpiceF, "\n\n");
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
esScale = EFScale / 100.0;
/* Convert the hierarchical description to a flat one */
flatFlags = EF_FLATNODES;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
EFOutputFlags |= EF_TRIMGLOB ;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (IS_FINITE_F(EFCapThreshold)) flatFlags |= EF_FLATCAPS;
if (esFormat == HSPICE) {
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
EFOutputFlags |= EF_TRIMLOCAL ;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
HashInit(&subcktNameTable, 32, HT_STRINGKEYS);
#ifndef UNSORTED_SUBCKT
DQInit(&subcktNameQueue, 64);
#endif
}
EFFlatBuild(inName, flatFlags);
/* Determine if this is a subcircuit */
locDoSubckt = FALSE;
if (esDoSubckt == AUTO) {
if (efFlatRootDef->def_flags & DEF_SUBCIRCUIT)
locDoSubckt = TRUE;
}
if ((esDoSubckt == TRUE) || (locDoSubckt == TRUE))
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
topVisit(efFlatRootDef, FALSE);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* If we don't want to write subcircuit calls, remove the */
/* subcircuit flag from all cells at this time. */
if (!esDoPorts)
EFVisitSubcircuits(subcktUndef, (ClientData) NULL);
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
TTMaskZero(&initMask);
if (!esDistrJunct)
(1) Added a check for unclosed boundaries when reading GDS. This is diagnostic only and does not change the read-in behavior. (2) ext2spice: Corrected an error that had been introduced into version 8.3.171 that accidentally marks all devices as visited which causes all source/drain areas and perimeters to be output as zero. (3) extract: Sweeping changes to handling of fringe capacitance. Removed the (recently added) "fringeshieldhalo" parameter from the tech file. Reworked the fringe capacitance models based on results from the "capiche" project (github/RTimothyEdwards/capiche). Fringe shielding is now done by clipping fringe at the boundary of a shielding shape, rather than trying to calculate the amount of shielding (as the "capiche" project proved this to be equivalent). Values for partial fringing are modeled by atan(x), which like the sidewall (1/x) curve, extends to infinity and values are limited by the halo but do not otherwise depend on the halo. Because of this, the halo can be made variable and controlled by the user for deciding on the tradeoff between accuracy and run time. A new command option "extract halo" was added to allow this control over the halo distance.
2023-01-27 17:47:37 +01:00
TTMaskSetType(&initMask, efNumResistClasses);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if ( esMergeDevsA || esMergeDevsC ) {
EFVisitDevs(devMergeVisit, (ClientData) NULL);
TxPrintf("Devs merged: %d\n", esSpiceDevsMerged);
esFMIndex = 0 ;
{
extraction: constify struct devMerge.next, devMergeList Feel free to undo variables as necessary in the future as requirements change.
2025-07-24 15:36:22 +02:00
const devMerge *p;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
extraction: constify struct devMerge.next, devMergeList Feel free to undo variables as necessary in the future as requirements change.
2025-07-24 15:36:22 +02:00
for ( p = devMergeList ; p != NULL ; p=p->next ) freeMagic((char *)p);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
} else if ( esDistrJunct )
EFVisitDevs(devDistJunctVisit, (ClientData) NULL);
EFVisitDevs(spcdevVisit, (ClientData) NULL);
Recast all uses of resist-class and device class in routines using code from extflat from type unsigned long to type TileTypeBitMask. This increases the number of types of each to 256 and tracks the number of types, so it should be difficult to exceed this amount.
2021-05-27 22:13:06 +02:00
TTMaskZero(&initMask);
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
if (flatFlags & EF_FLATCAPS)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFVisitCaps(spccapVisit, (ClientData) NULL);
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
extflat: integrate cb_extflat_visitresists_t for EFVisitResists() Additionally the HierName's hierName1 and hierName2 arguments have been made 'const' to help convey the receiver can not modify the referenced data passed.
2025-07-20 17:09:08 +02:00
EFVisitResists(spcresistVisit, PTR2CD(NULL));
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFVisitSubcircuits(subcktVisit, (ClientData) NULL);
Corrected an issue from the last commit, which is that the default ground node name (which is static) gets put on the node list and is improperly deallocated. Corrected by simply allocating the string for the default substrate node instead of using the static string.
2023-03-14 17:53:22 +01:00
esSpiceCapNode = StrDup((char **)NULL, esSpiceDefaultGnd);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFVisitNodes(spcnodeVisit, (ClientData) NULL);
if ((esDoSubckt == TRUE) || (locDoSubckt == TRUE))
fprintf(esSpiceF, ".ends\n");
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (esFormat == HSPICE)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
printSubcktDict();
Tracked down two more memory leaks coming from ext2spice, due to client data generated by ext2spice and attached to a node's nodeClient record; there is an initNodeClient() routine but no corresponding freeNodeClient() routine. Eventually had to add a callback function passed to EFDone() and EFFlatDone() to clean up these entries. After doing that, valgrind reports clean for all memory allocated within ext2spice (there are other things that are not freed but not related to a specific command, so do not need to be treated as leaks).
2021-12-14 00:05:53 +01:00
EFFlatDone(esFreeNodeClient);
EFDone(esFreeNodeClient);
Corrected an issue with the limited bitfield dedicated to port numbers; this limited ports to 16384, which seemed reasonable at the time. However, the sky130_sram_macro layouts connect power and ground in a way that when coupled with "extract unique" can generate tens of thousands of ports and overrun the bit field, showing that automation can do the unexpected. The solution was to split out the port number from the label record as its own 32-bit value.
2021-12-13 04:09:31 +01:00
if (esFormat == HSPICE) {
HashKill(&subcktNameTable);
#ifndef UNSORTED_SUBCKT
DQFree(&subcktNameQueue);
#endif
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esSpiceF) fclose(esSpiceF);
TxPrintf("Memory used: %s\n", RunStats(RS_MEM, NULL, NULL));
exit (0);
}
ext2spice: Makefile use EXT2SPICE_MAIN for deprecated standalone
2025-07-18 15:12:51 +02:00
#endif /* EXT2SPICE_MAIN Deprecated */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
* spcParseArgs --
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
* Process those arguments that are specific to ext2spice.
* Assumes that *pargv[0][0] is '-', indicating a flag
* argument.
*
* Results:
* None.
*
* Side effects:
* After processing an argument, updates *pargc and *pargv
* to point to after the argument.
*
* May initialize various global variables based on the
* arguments given to us.
*
* Exits in the event of an improper argument.
*
* ----------------------------------------------------------------------------
*/
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spcParseArgs(
int *pargc,
char ***pargv)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
char **argv = *pargv;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int argc = *pargc;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
char *ftmp, *t;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
gcc: support -Werror=format-security Support gcc -Werror=format-security which is used by default on Fedora. ext2spice/ext2spice.c: format not a string literal and no format arguments extflat/EFargs.c: likewise windows/windCmdNR.c: likewise
2025-01-10 06:19:52 +01:00
const char usage_text[] = "Usage: ext2spice "
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
"[-B] [-o spicefile] [-M|-m] [-J flat|hier]\n"
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
"[-f spice2|spice3|hspice|ngspice|cdl] [-M] [-m] "
Corrected EFArgs() so that "ext2spice run -help" works as advertised; e.g., "-h" or "-help" is now recognized as a valid option instead of printing a usage message by way of throwing an error.
2019-10-16 15:17:58 +02:00
"[file]\n";
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
switch (argv[0][1])
{
case 'd':
esDistrJunct = TRUE;
break;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case 'M':
esMergeDevsA = TRUE;
break;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case 'm':
esMergeDevsC = TRUE;
break;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case 'B':
esNoAttrs = TRUE;
break;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case 'F':
esDevNodesOnly = TRUE;
break;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case 'o':
if ((spcesOutName = ArgStr(&argc, &argv, "filename")) == NULL)
goto usage;
break;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
case 'f':
if ((ftmp = ArgStr(&argc, &argv, "format")) == NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
goto usage;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
if (strcasecmp(ftmp, "SPICE2") == 0)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFormat = SPICE2;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
else if (strcasecmp(ftmp, "SPICE3") == 0)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFormat = SPICE3;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
else if (strcasecmp(ftmp, "HSPICE") == 0)
{
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFormat = HSPICE;
esScale = -1.0;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
}
else if (strcasecmp(ftmp, "NGSPICE") == 0)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esFormat = NGSPICE;
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
else if (strcasecmp(ftmp, "CDL") == 0)
esFormat = CDL;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
else goto usage;
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
case 'J':
if ((ftmp = ArgStr(&argc, &argv, "hierAP_SD")) == NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
goto usage;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
if ( strcasecmp(ftmp, "HIER") == 0 )
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esHierAP = TRUE ;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
else if (strcasecmp(ftmp, "FLAT") == 0 )
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
esHierAP = FALSE ;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
else goto usage;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
case 'y':
if ((t = ArgStr(&argc, &argv, "cap-accuracy")) == NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
goto usage;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
TxPrintf("The cap accuracy flag 'y' is deprecated.\n");
break;
Corrected EFArgs() so that "ext2spice run -help" works as advertised; e.g., "-h" or "-help" is now recognized as a valid option instead of printing a usage message by way of throwing an error.
2019-10-16 15:17:58 +02:00
case 'h': /* -h or -help, as suggested by "ext2spice help" */
TxPrintf(usage_text);
break;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
default:
TxError("Unrecognized flag: %s\n", argv[0]);
goto usage;
}
*pargv = argv;
*pargc = argc;
return 0;
usage:
Corrected EFArgs() so that "ext2spice run -help" works as advertised; e.g., "-h" or "-help" is now recognized as a valid option instead of printing a usage message by way of throwing an error.
2019-10-16 15:17:58 +02:00
TxError(usage_text);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 1;
}
/*
* ----------------------------------------------------------------------------
*
* SpiceGetNode --
*
* function to find a node given its hierarchical prefix and suffix
*
* Results:
* a pointer to the node struct or NULL
*
* ----------------------------------------------------------------------------
*/
EFNode *
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
SpiceGetNode(
HierName *prefix,
HierName *suffix)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
HashEntry *he;
EFNodeName *nn;
he = EFHNConcatLook(prefix, suffix, "output");
if (he == NULL) return NULL;
nn = (EFNodeName *) HashGetValue(he);
if (nn == NULL) return NULL;
return(nn->efnn_node);
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* extHierSDAttr --
* Check if the attribute of the argument dev_terminal or the global
* settings are such that we want a hierarchical extraction of its S/D
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
*
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
* Results:
* TRUE or FALSE
*
* Side effects:
* None.
*
* ----------------------------------------------------------------------------
*/
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
bool
extHierSDAttr(
DevTerm *term)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
bool r = esHierAP;
if (term->dterm_attrs)
{
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (Match(ATTR_HIERAP, term->dterm_attrs) != FALSE)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
r = TRUE;
else if (Match(ATTR_FLATAP, term->dterm_attrs) != FALSE)
r = FALSE;
}
return r;
}
/*
* ----------------------------------------------------------------------------
*
* subcktVisit --
*
* Procedure to output a subcircuit definition to the .spice file.
* Called by EFVisitSubcircuits().
*
* Results:
* Returns 0 always.
*
* Side effects:
* Writes to the file esSpiceF.
*
* Format of a .spice subcircuit call:
*
* X%d node1 node2 ... noden name
*
* where
* node1 node2 ... noden are the nodes connecting to the ports of
* the subcircuit. "name" is the name of the subcircuit. It is
* assumed that the definition of "name" (.defs name ... .ends)
* exists elsewhere and will be appended to the SPICE deck prior
* to simulation (as is also assumed for device models).
*
* ----------------------------------------------------------------------------
*/
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitsubcircuits_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
subcktVisit(
Use *use,
HierName *hierName,
bool is_top) /* TRUE if this is the top-level cell */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
EFNode *snode;
Def *def = use->use_def;
There being a specific situation where the enumeration of ports used by "topVisit" and "subcktVisit" in ext2spice.c, probably caused by having different names on the same port number, the subcktVisit() routine was modified to use exactly the same enumeration as topVisit() so that they are guaranteed to have the same result.
2021-12-13 17:33:02 +01:00
EFNodeName *sname, *nodeName;
HashSearch hs;
HashEntry *he;
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
int portmax, portidx, tchars;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
char stmp[MAX_STR_SIZE];
char *instname, *subcktname;
DevParam *plist, *pptr;
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
EFNodeName **nodeList;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (is_top == TRUE) return 0; /* Ignore the top-level cell */
/* Retain instance name unless esDoRenumber is set, or format is Spice2 */
if (use->use_id == NULL || esDoRenumber == TRUE || esFormat == SPICE2)
{
Corrected hierarchical SPICE extraction to reset the subcircuit name-to-number mapping used for the HSPICE format between subcircuits. Otherwise, subcircuits with the same instance ID remain in the table and may cause nodes to be output with a name that collides with other names in the same subcircuit. This only affects output in HSPICE format.
2018-12-12 23:23:35 +01:00
/* NOTE: This really needs to update subcktNameTable so that */
/* it tracks between instance names and node names, when using */
/* HSPICE format + esDoRenumber. */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, "X%d", esSbckNum++);
tchars = 5;
}
else
{
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
int savflags = EFOutputFlags;
Correction to prior commit to remove collision between defined flag bits for name trimming and short handling method in the ext2spice options.
2021-07-11 03:27:19 +02:00
EFOutputFlags &= ~EF_TRIM_MASK;
EFOutputFlags |= EF_CONVERTCOMMA; // Only substitute commas on subcircuit names
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Use full hierarchical decomposition for name */
/* (not just use->use_id. hierName already has use->use_id at end) */
EFHNSprintf(stmp, hierName);
fprintf(esSpiceF, "X%s", stmp);
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
EFOutputFlags = savflags;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
tchars = 1 + strlen(stmp);
}
/* This is not a DEV, but "spcdevOutNode" is a general-purpose routine that */
/* turns a local name in the use's def to a hierarchical name in the */
/* calling def. */
/* Note that the ports of the subcircuit will not necessarily be */
/* ALL the entries in the hash table, so we have to check. */
Removed code from ext2spice that is no longer functional. Implemented a separate check for ports when writing a subcircuit that cross-checks against the port list in the flattened extraction. This allows ports that were optimized out during flattening of the hierarchy to be removed from the cell's port list, which cuts down on disconnected nodes in the output port list.
2022-01-12 21:30:07 +01:00
portmax = EFGetPortMax(def);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (portmax < 0)
{
/* No port order declared; print them as we see them. */
/* This shouldn't happen for proper .ext files written */
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
/* by the magic extractor, since explicit port order is */
/* generated during topVisit(). */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
There being a specific situation where the enumeration of ports used by "topVisit" and "subcktVisit" in ext2spice.c, probably caused by having different names on the same port number, the subcktVisit() routine was modified to use exactly the same enumeration as topVisit() so that they are guaranteed to have the same result.
2021-12-13 17:33:02 +01:00
HashStartSearch(&hs);
35 x warning: suggest parentheses around assignment used as truth value PlowRules1.c:439:9: warning: suggest parentheses around assignment used as truth value PlowTech.c:645:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:652:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:1019:17: warning: suggest parentheses around assignment used as truth value ResReadSim.c:270:13: warning: suggest parentheses around assignment used as truth value ResReadSim.c:871:9: warning: suggest parentheses around assignment used as truth value ResRex.c:1840:17: warning: suggest parentheses around assignment used as truth value getrect.c:72:9: warning: suggest parentheses around assignment used as truth value getrect.c:79:9: warning: suggest parentheses around assignment used as truth value getrect.c:86:9: warning: suggest parentheses around assignment used as truth value getrect.c:93:9: warning: suggest parentheses around assignment used as truth value hash.c:732:16: warning: suggest parentheses around assignment used as truth value heap.c:328:17: warning: suggest parentheses around assignment used as truth value heap.c:344:17: warning: suggest parentheses around assignment used as truth value netlist.c:323:17: warning: suggest parentheses around assignment used as truth value niceabort.c:121:9: warning: suggest parentheses around assignment used as truth value path.c:1102:12: warning: suggest parentheses around assignment used as truth value pathvisit.c:245:13: warning: suggest parentheses around assignment used as truth value pathvisit.c:295:17: warning: suggest parentheses around assignment used as truth value tech.c:656:17: warning: suggest parentheses around assignment used as truth value ext2spice.c:1591:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1622:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1813:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:1862:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:3808:16: warning: suggest parentheses around assignment used as truth value CalmaRdio.c:437:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:396:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:1772:29: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:372:9: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:1608:29: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:209:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:214:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:220:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:226:9: warning: suggest parentheses around assignment used as truth value CIFrdutils.c:1258:12: warning: suggest parentheses around assignment used as truth value GCC14 -Wall cleanup series [-Wparentheses]
2024-10-04 18:21:15 +02:00
while ((he = HashNext(&def->def_nodes, &hs)))
There being a specific situation where the enumeration of ports used by "topVisit" and "subcktVisit" in ext2spice.c, probably caused by having different names on the same port number, the subcktVisit() routine was modified to use exactly the same enumeration as topVisit() so that they are guaranteed to have the same result.
2021-12-13 17:33:02 +01:00
{
sname = (EFNodeName *) HashGetValue(he);
if (sname == NULL) continue;
snode = sname->efnn_node;
if (snode && (snode->efnode_flags & EF_PORT))
for (nodeName = sname; nodeName != NULL; nodeName = nodeName->efnn_next)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (nodeName->efnn_port >= 0)
{
portmax++;
if (tchars > 80)
{
fprintf(esSpiceF, "\n+");
tchars = 1;
}
tchars += spcdevOutNode(hierName, nodeName->efnn_hier,
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
"subcircuit", esSpiceF);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
}
}
else
{
/* Port numbers need not start at zero or be contiguous. */
/* They will be printed in numerical order. */
Reverted most of yesterday's modifications. Instead located the issue at the change made in revision 214. This was done incorrectly in two ways, one being a set of statements inside an if() block that should have been executed always, and the other an incorrect use of the EF_DEVTERM flag, setting it when it should not have been set.
2021-12-03 18:29:02 +01:00
nodeList = (EFNodeName **)mallocMagic((portmax + 1) * sizeof(EFNodeName *));
for (portidx = 0; portidx <= portmax; portidx++)
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
nodeList[portidx] = (EFNodeName *)NULL;
There being a specific situation where the enumeration of ports used by "topVisit" and "subcktVisit" in ext2spice.c, probably caused by having different names on the same port number, the subcktVisit() routine was modified to use exactly the same enumeration as topVisit() so that they are guaranteed to have the same result.
2021-12-13 17:33:02 +01:00
HashStartSearch(&hs);
35 x warning: suggest parentheses around assignment used as truth value PlowRules1.c:439:9: warning: suggest parentheses around assignment used as truth value PlowTech.c:645:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:652:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:1019:17: warning: suggest parentheses around assignment used as truth value ResReadSim.c:270:13: warning: suggest parentheses around assignment used as truth value ResReadSim.c:871:9: warning: suggest parentheses around assignment used as truth value ResRex.c:1840:17: warning: suggest parentheses around assignment used as truth value getrect.c:72:9: warning: suggest parentheses around assignment used as truth value getrect.c:79:9: warning: suggest parentheses around assignment used as truth value getrect.c:86:9: warning: suggest parentheses around assignment used as truth value getrect.c:93:9: warning: suggest parentheses around assignment used as truth value hash.c:732:16: warning: suggest parentheses around assignment used as truth value heap.c:328:17: warning: suggest parentheses around assignment used as truth value heap.c:344:17: warning: suggest parentheses around assignment used as truth value netlist.c:323:17: warning: suggest parentheses around assignment used as truth value niceabort.c:121:9: warning: suggest parentheses around assignment used as truth value path.c:1102:12: warning: suggest parentheses around assignment used as truth value pathvisit.c:245:13: warning: suggest parentheses around assignment used as truth value pathvisit.c:295:17: warning: suggest parentheses around assignment used as truth value tech.c:656:17: warning: suggest parentheses around assignment used as truth value ext2spice.c:1591:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1622:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1813:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:1862:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:3808:16: warning: suggest parentheses around assignment used as truth value CalmaRdio.c:437:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:396:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:1772:29: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:372:9: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:1608:29: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:209:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:214:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:220:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:226:9: warning: suggest parentheses around assignment used as truth value CIFrdutils.c:1258:12: warning: suggest parentheses around assignment used as truth value GCC14 -Wall cleanup series [-Wparentheses]
2024-10-04 18:21:15 +02:00
while ((he = HashNext(&def->def_nodes, &hs)))
There being a specific situation where the enumeration of ports used by "topVisit" and "subcktVisit" in ext2spice.c, probably caused by having different names on the same port number, the subcktVisit() routine was modified to use exactly the same enumeration as topVisit() so that they are guaranteed to have the same result.
2021-12-13 17:33:02 +01:00
{
sname = (EFNodeName *) HashGetValue(he);
if (sname == NULL) continue;
snode = sname->efnn_node;
if ((snode == NULL) || !(snode->efnode_flags & EF_PORT)) continue;
for (nodeName = sname; nodeName != NULL; nodeName = nodeName->efnn_next)
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
{
portidx = nodeName->efnn_port;
Corrected one error in the recent fix for port enumeration in ext2spice (hierarchy on), where the port index was not checked for value -1 (no port) before indexing into the port list.
2019-10-31 20:19:30 +01:00
if (portidx < 0) continue;
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
if (nodeList[portidx] == NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
nodeList[portidx] = nodeName;
}
else if (EFHNBest(nodeName->efnn_hier, nodeList[portidx]->efnn_hier))
{
nodeList[portidx] = nodeName;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
}
}
Reverted most of yesterday's modifications. Instead located the issue at the change made in revision 214. This was done incorrectly in two ways, one being a set of statements inside an if() block that should have been executed always, and the other an incorrect use of the EF_DEVTERM flag, setting it when it should not have been set.
2021-12-03 18:29:02 +01:00
for (portidx = 0; portidx <= portmax; portidx++)
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
{
nodeName = nodeList[portidx];
Resolved some issues with working with ext2spice using abstract views. Because the abstract view does not necessarily represent actual connectivity, rely on the port indexes in the .ext file to determine the number of ports and port order. Do not use SpiceNodeName() to look up the node name, or unique ports that are deemed shorted will go missing. Also: Modified the read-in of .ext files so that use names may contain backslashes. Only backslashes that end a line will be handled differently.
2020-01-13 18:58:04 +01:00
if (nodeName != NULL)
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
{
if (tchars > 80)
{
fprintf(esSpiceF, "\n+");
tchars = 1;
}
tchars += spcdevOutNode(hierName, nodeName->efnn_hier,
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
"subcircuit", esSpiceF);
Again revisited the problem of duplicate nodes. This time, I think I understand the problem, which is that nodes are ordered according to precedence of EFHNBest() within a circuit, but there is no concept of ordering between circuits. So ports end up listing nodes in arbitrary order, and the only way to resolve the order is to use EFHNBest() as is done within a subcircuit. Appears to work for different edge cases tested.
2019-10-29 18:55:28 +01:00
}
}
freeMagic(nodeList);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Modified the handling of subcircuit names beginning with non-alphanumeric characters. Instead of removing the non-alphanumeric characters, magic now prepends an "x" to the name. Since this naming restriction does not necessarily impact, say, LVS, it would probably be better to let this behavior be enabled or disabled by a command.
2021-03-07 01:39:34 +01:00
/* SPICE subcircuit names must begin with A-Z. */
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
subcktname = def->def_name;
Modified the handling of subcircuit names beginning with non-alphanumeric characters. Instead of removing the non-alphanumeric characters, magic now prepends an "x" to the name. Since this naming restriction does not necessarily impact, say, LVS, it would probably be better to let this behavior be enabled or disabled by a command.
2021-03-07 01:39:34 +01:00
if (!isalpha(*subcktname))
{
subcktname = mallocMagic(2 + strlen(def->def_name));
sprintf(subcktname, "x%s", def->def_name);
freeMagic(def->def_name);
def->def_name = subcktname;
}
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
if (tchars > 80) fprintf(esSpiceF, "\n+");
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
if (esFormat == CDL) fprintf(esSpiceF, " /");
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
fprintf(esSpiceF, " %s", subcktname); /* subcircuit model name */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
// Check for a "device parameter" defined with the name of the cell.
// This contains a list of parameter strings to be passed to the
// cell instance.
instname = mallocMagic(2 + strlen(def->def_name));
sprintf(instname, ":%s", def->def_name);
plist = efGetDeviceParams(instname);
for (pptr = plist; pptr; pptr = pptr->parm_next)
{
if (tchars > 80)
{
fprintf(esSpiceF, "\n+");
tchars = 1;
}
fprintf(esSpiceF, " %s", pptr->parm_name);
tchars += (1 + strlen(pptr->parm_name));
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
freeMagic(instname);
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
fprintf(esSpiceF, "\n");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* subcktUndef --
*
* Procedure to remove the DEF_SUBCIRCUIT flag from all subcells.
* Called by EFVisitSubcircuits().
*
* Results:
* Returns 0 always.
*
* Side effects:
* Undefines the DEF_SUBCIRCUIT flag in each encountered Def.
*
* ----------------------------------------------------------------------------
*/
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitsubcircuits_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
subcktUndef(
Use *use,
HierName *hierName,
bool is_top) /* TRUE if this is the top-level cell */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Def *def = use->use_def;
def->def_flags &= ~(DEF_SUBCIRCUIT);
return 0;
}
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
/* Define a linked node name list */
typedef struct _lnn {
EFNodeName *lnn_nodeName;
struct _lnn *lnn_next;
} linkedNodeName;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* topVisit --
*
* Procedure to output a subcircuit definition to the .spice file.
*
* Results:
* None.
*
* Side effects:
* Writes to the file esSpiceF.
*
* Format of a .spice subcircuit definition:
*
* .subckt name node1 node2 ... noden
*
* where
* node1 node2 ... noden are the nodes connecting to the ports of
Set of changes updating version 8.2 to the level of 8.1, since 8.2 development had been halted since it was first created back in April. Version 8.2 is now the official development version, with the first development push to create a Cairo graphics interface.
2017-08-02 04:14:42 +02:00
* the subcircuit. "name" is the name of the cell def. If "doStub"
* is TRUE, then the subcircuit is a stub (empty declaration) for a
* subcircuit, and implicit substrate connections should not be
* output.
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
Corrected two annoying errors: (1) SPICE netlist output subcircuits could have multiple ports of the same name. This problem had been worked over before, but there was an indpendent mechanism producing the same result for a completely different reason, caused by subcells being much larger than the cookie-cutter extraction method's extraction regions. Solved by tracking port names in a hash table and preventing re-use. (2) ext2spice was producing "no such node" errors; like (1) this had been previously worked on, and like (1) this mechanism was independent. Problem came from not passing -1 to extHierSubstrate for the non-arrayed dimension of a 1-dimensional array. Also: Removed the word "fatal" from extraction error reporting, as nearly all extraction errors are entirely benign. This should clear up confusion among alarmed end-users.
2020-11-24 21:30:49 +01:00
* NOTE: The cookie-cutter method for extraction can result in multiple
* connections to the same port if the net spans multiple extraction regions.
* Because of this, it is necessary to make sure that the same port name is
* not output twice.
*
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* ----------------------------------------------------------------------------
*/
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
topVisit(
Def *def,
bool doStub)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
EFNode *snode;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFNodeName *sname, *nodeName;
HashSearch hs;
(1) Modified the method used for the previous commit, as it was found to make ext2spice runtimes very long for large layouts. The new method is equivalent but doesn't incur the overhead. Also: Changed a flag check which was causing the substrate node to be output as a port for certain layouts where the substrate node connects to no devices, and so should be optimized out.
2022-01-13 23:00:07 +01:00
HashEntry *he, *hep, *heh;
Corrected two annoying errors: (1) SPICE netlist output subcircuits could have multiple ports of the same name. This problem had been worked over before, but there was an indpendent mechanism producing the same result for a completely different reason, caused by subcells being much larger than the cookie-cutter extraction method's extraction regions. Solved by tracking port names in a hash table and preventing re-use. (2) ext2spice was producing "no such node" errors; like (1) this had been previously worked on, and like (1) this mechanism was independent. Problem came from not passing -1 to extHierSubstrate for the non-arrayed dimension of a 1-dimensional array. Also: Removed the word "fatal" from extraction error reporting, as nearly all extraction errors are entirely benign. This should clear up confusion among alarmed end-users.
2020-11-24 21:30:49 +01:00
HashTable portNameTable;
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
int portorder, portmax, tchars;
Corrected a problem in ext2spice which has been in the code for a very long time but never discovered; in which any implicit port connection into a subcell (that is otherwise labeled with ports) that appears at the end of the node list (i.e., after all the declared ports), will not be output, either in the subcircuit definition or calls.
2021-07-25 19:49:13 +02:00
bool explicit;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
DevParam *plist, *pptr;
char *instname;
char *subcktname;
extflat: DevParam.(const char*)param_name consumer, nodeSpiceName()
2025-07-23 21:58:41 +02:00
const char *pname;
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
char **sorted_ports;
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
linkedNodeName *lnn = NULL;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Added a 2nd use of the "device" property, which has long been needed, which instead of defining a device or subcircuit that exists inside the cell, instead redefines the cell itself as a device or subcircuit model that exists in the PDK. This is used where a specific layout subcell has its own associated device definition in the PDK. Instead of the "device" property value being the line that gets generated for a device in the subcells .ext file, the property value should be the word "primitive" optionally followed by any parameters that need to be passed to the subcircuit call.
2022-05-04 22:43:38 +02:00
/* Primitive devices are not output at all */
if (def->def_flags & DEF_PRIMITIVE) return;
Corrected two annoying errors: (1) SPICE netlist output subcircuits could have multiple ports of the same name. This problem had been worked over before, but there was an indpendent mechanism producing the same result for a completely different reason, caused by subcells being much larger than the cookie-cutter extraction method's extraction regions. Solved by tracking port names in a hash table and preventing re-use. (2) ext2spice was producing "no such node" errors; like (1) this had been previously worked on, and like (1) this mechanism was independent. Problem came from not passing -1 to extHierSubstrate for the non-arrayed dimension of a 1-dimensional array. Also: Removed the word "fatal" from extraction error reporting, as nearly all extraction errors are entirely benign. This should clear up confusion among alarmed end-users.
2020-11-24 21:30:49 +01:00
HashInit(&portNameTable, 32, HT_STRINGKEYS);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* SPICE subcircuit names must begin with A-Z. This will also be */
/* enforced when writing X subcircuit calls. */
subcktname = def->def_name;
Modified the handling of subcircuit names beginning with non-alphanumeric characters. Instead of removing the non-alphanumeric characters, magic now prepends an "x" to the name. Since this naming restriction does not necessarily impact, say, LVS, it would probably be better to let this behavior be enabled or disabled by a command.
2021-03-07 01:39:34 +01:00
if (!isalpha(*subcktname))
{
subcktname = mallocMagic(2 + strlen(def->def_name));
sprintf(subcktname, "x%s", def->def_name);
freeMagic(def->def_name);
def->def_name = subcktname;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, ".subckt %s", subcktname);
tchars = 8 + strlen(subcktname);
/* Note that the ports of the subcircuit will not necessarily be */
/* ALL the entries in the hash table, so we have to check. */
HashStartSearch(&hs);
portmax = -1;
35 x warning: suggest parentheses around assignment used as truth value PlowRules1.c:439:9: warning: suggest parentheses around assignment used as truth value PlowTech.c:645:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:652:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:1019:17: warning: suggest parentheses around assignment used as truth value ResReadSim.c:270:13: warning: suggest parentheses around assignment used as truth value ResReadSim.c:871:9: warning: suggest parentheses around assignment used as truth value ResRex.c:1840:17: warning: suggest parentheses around assignment used as truth value getrect.c:72:9: warning: suggest parentheses around assignment used as truth value getrect.c:79:9: warning: suggest parentheses around assignment used as truth value getrect.c:86:9: warning: suggest parentheses around assignment used as truth value getrect.c:93:9: warning: suggest parentheses around assignment used as truth value hash.c:732:16: warning: suggest parentheses around assignment used as truth value heap.c:328:17: warning: suggest parentheses around assignment used as truth value heap.c:344:17: warning: suggest parentheses around assignment used as truth value netlist.c:323:17: warning: suggest parentheses around assignment used as truth value niceabort.c:121:9: warning: suggest parentheses around assignment used as truth value path.c:1102:12: warning: suggest parentheses around assignment used as truth value pathvisit.c:245:13: warning: suggest parentheses around assignment used as truth value pathvisit.c:295:17: warning: suggest parentheses around assignment used as truth value tech.c:656:17: warning: suggest parentheses around assignment used as truth value ext2spice.c:1591:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1622:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1813:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:1862:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:3808:16: warning: suggest parentheses around assignment used as truth value CalmaRdio.c:437:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:396:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:1772:29: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:372:9: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:1608:29: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:209:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:214:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:220:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:226:9: warning: suggest parentheses around assignment used as truth value CIFrdutils.c:1258:12: warning: suggest parentheses around assignment used as truth value GCC14 -Wall cleanup series [-Wparentheses]
2024-10-04 18:21:15 +02:00
while ((he = HashNext(&def->def_nodes, &hs)))
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
sname = (EFNodeName *) HashGetValue(he);
if (sname == NULL) continue;
snode = sname->efnn_node;
Corrected ext2spice for the case where a port has been given more than one name, because in that case one of the port records ends up with a null pointer to a node, and causes a crash condition. This can happen inadvertently, as when a connected node is not specifically designated a port, but is forced to be a port because of the connection.
2020-03-20 19:50:56 +01:00
if ((!snode) || (!(snode->efnode_flags & EF_PORT))) continue;
Corrected a problem in ext2spice which has been in the code for a very long time but never discovered; in which any implicit port connection into a subcell (that is otherwise labeled with ports) that appears at the end of the node list (i.e., after all the declared ports), will not be output, either in the subcircuit definition or calls.
2021-07-25 19:49:13 +02:00
explicit = FALSE;
Change to ext2spice which should solve the problem of ports no longer being output in order after flattening a circuit.
2024-08-06 17:17:27 +02:00
portorder = snode->efnode_name->efnn_port;
if (portorder > portmax) portmax = portorder;
if (portorder != -1) explicit = TRUE;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
for (nodeName = sname; nodeName != NULL; nodeName = nodeName->efnn_next)
{
portorder = nodeName->efnn_port;
if (portorder > portmax) portmax = portorder;
Corrected a problem in ext2spice which has been in the code for a very long time but never discovered; in which any implicit port connection into a subcell (that is otherwise labeled with ports) that appears at the end of the node list (i.e., after all the declared ports), will not be output, either in the subcircuit definition or calls.
2021-07-25 19:49:13 +02:00
if (portorder != -1) explicit = TRUE;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
if (explicit == FALSE)
{
/* Tag this an an implicit port (port without an assigned index) */
linkedNodeName *newlnn =
(linkedNodeName *)mallocMagic(sizeof(linkedNodeName));
newlnn->lnn_next = lnn;
newlnn->lnn_nodeName = sname;
lnn = newlnn;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
/* Make all port numbers explicit (unless this is a black-box */
/* circuit "ext2spice blackbox on" is in effect). */
while (lnn != NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
sname = lnn->lnn_nodeName;
if (esDoBlackBox == FALSE || !(def->def_flags & DEF_ABSTRACT))
sname->efnn_port = ++portmax;
freeMagic(lnn);
lnn = lnn->lnn_next;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
/* Port numbers need not start at zero or be contiguous. They will be */
/* printed in numerical order. This is done by allocating space for */
/* the output first and generating text into the allocated array */
/* indexed by port, to avoid multiple scans through the hash table. */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
sorted_ports = (char **)mallocMagic((portmax + 1) * sizeof(char *));
Fix to the previous commit, which was implemented incorrectly. The current implementation works, but it is not clear that it does anything differently than before the start of today's modifications.
2022-01-12 23:19:29 +01:00
for (portorder = 0; portorder <= portmax; portorder++)
sorted_ports[portorder] = NULL;
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
HashStartSearch(&hs);
35 x warning: suggest parentheses around assignment used as truth value PlowRules1.c:439:9: warning: suggest parentheses around assignment used as truth value PlowTech.c:645:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:652:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:1019:17: warning: suggest parentheses around assignment used as truth value ResReadSim.c:270:13: warning: suggest parentheses around assignment used as truth value ResReadSim.c:871:9: warning: suggest parentheses around assignment used as truth value ResRex.c:1840:17: warning: suggest parentheses around assignment used as truth value getrect.c:72:9: warning: suggest parentheses around assignment used as truth value getrect.c:79:9: warning: suggest parentheses around assignment used as truth value getrect.c:86:9: warning: suggest parentheses around assignment used as truth value getrect.c:93:9: warning: suggest parentheses around assignment used as truth value hash.c:732:16: warning: suggest parentheses around assignment used as truth value heap.c:328:17: warning: suggest parentheses around assignment used as truth value heap.c:344:17: warning: suggest parentheses around assignment used as truth value netlist.c:323:17: warning: suggest parentheses around assignment used as truth value niceabort.c:121:9: warning: suggest parentheses around assignment used as truth value path.c:1102:12: warning: suggest parentheses around assignment used as truth value pathvisit.c:245:13: warning: suggest parentheses around assignment used as truth value pathvisit.c:295:17: warning: suggest parentheses around assignment used as truth value tech.c:656:17: warning: suggest parentheses around assignment used as truth value ext2spice.c:1591:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1622:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1813:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:1862:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:3808:16: warning: suggest parentheses around assignment used as truth value CalmaRdio.c:437:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:396:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:1772:29: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:372:9: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:1608:29: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:209:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:214:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:220:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:226:9: warning: suggest parentheses around assignment used as truth value CIFrdutils.c:1258:12: warning: suggest parentheses around assignment used as truth value GCC14 -Wall cleanup series [-Wparentheses]
2024-10-04 18:21:15 +02:00
while ((he = HashNext(&def->def_nodes, &hs)))
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
{
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
char stmp[MAX_STR_SIZE];
int portidx;
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
sname = (EFNodeName *) HashGetValue(he);
if (sname == NULL) continue; /* Should not happen */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Removed code from ext2spice that is no longer functional. Implemented a separate check for ports when writing a subcircuit that cross-checks against the port list in the flattened extraction. This allows ports that were optimized out during flattening of the hierarchy to be removed from the cell's port list, which cuts down on disconnected nodes in the output port list.
2022-01-12 21:30:07 +01:00
snode = sname->efnn_node;
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
if ((!snode) || (!(snode->efnode_flags & EF_PORT))) continue;
Corrected an issue caused by the commit to version 8.3.190. Although that commit claimed to correct an issue with implicit ports not being output, the solution often failed to properly assign the port number, so while the implicit ports were added to the subcircuit definition, they were often missing from the subcircuit call.
2021-08-04 19:54:08 +02:00
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
for (nodeName = sname; nodeName != NULL; nodeName = nodeName->efnn_next)
{
portidx = nodeName->efnn_port;
Made another correction to the recent change in ext2spice.c where an out-of-bounds array access occurs that can cause bad output at best and a crash condition at worst.
2021-10-07 16:58:21 +02:00
if (portidx < 0) continue;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
(1) Modified the method used for the previous commit, as it was found to make ext2spice runtimes very long for large layouts. The new method is equivalent but doesn't incur the overhead. Also: Changed a flag check which was causing the substrate node to be output as a port for certain layouts where the substrate node connects to no devices, and so should be optimized out.
2022-01-13 23:00:07 +01:00
/* Check if the same hierName is recorded in the flattened/optimized
* def's efNodeHashTable. If not, then it has been optimized out
* and should be removed from the port list.
*/
if (def->def_flags & DEF_ABSTRACT)
Cleaned up a bit of confusing diagnostic output when reading GDS. If cells are instanced before being defined, causing the GDS parser to rewind the cell from the top, then the "already defined" error messages will be suppressed, since it is to be expected that cells will be seen twice (and ignored the 2nd time). When rewinding, an output message is issued so that it is clear that the file contains instances that are used before they are defined, and recommends the "gds ordering on" setting. Also: Fixed the "gds ordering" command code so that the command with no third argument returns the state of the "gds ordering" setting instead of generating a parser error.
2022-01-14 17:07:08 +01:00
heh = HashLookOnly(&efNodeHashTable, (char *)nodeName->efnn_hier);
(1) Modified the method used for the previous commit, as it was found to make ext2spice runtimes very long for large layouts. The new method is equivalent but doesn't incur the overhead. Also: Changed a flag check which was causing the substrate node to be output as a port for certain layouts where the substrate node connects to no devices, and so should be optimized out.
2022-01-13 23:00:07 +01:00
else
Cleaned up a bit of confusing diagnostic output when reading GDS. If cells are instanced before being defined, causing the GDS parser to rewind the cell from the top, then the "already defined" error messages will be suppressed, since it is to be expected that cells will be seen twice (and ignored the 2nd time). When rewinding, an output message is issued so that it is clear that the file contains instances that are used before they are defined, and recommends the "gds ordering on" setting. Also: Fixed the "gds ordering" command code so that the command with no third argument returns the state of the "gds ordering" setting instead of generating a parser error.
2022-01-14 17:07:08 +01:00
heh = HashLookOnly(&efNodeHashTable,
(char *)snode->efnode_name->efnn_hier);
(1) Modified the method used for the previous commit, as it was found to make ext2spice runtimes very long for large layouts. The new method is equivalent but doesn't incur the overhead. Also: Changed a flag check which was causing the substrate node to be output as a port for certain layouts where the substrate node connects to no devices, and so should be optimized out.
2022-01-13 23:00:07 +01:00
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
/* If view is abstract, rely on the given port name, not
* the node. Otherwise, artifacts of the abstract view
* may cause nodes to be merged and the names lost.
*/
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
if (def->def_flags & DEF_ABSTRACT)
{
EFHNSprintf(stmp, nodeName->efnn_hier);
pname = stmp;
}
else
Corrected an issue that produces potentially incorrect output from ext2spice due to "equiv" statements in the .ext file. The algorithm was not properly keeping the port as the preferred name of the node, resulting in the non-preferred name being used instead of the port name in the output. This would happen only if there was a label on a net that had a different name than the port name. The error became much more prevalent after changing the extract behavior to make "extract do aliases" default. Also fixed a somewhat related minor error in which magic would print an error message about ports being shorted together on the same kind of net where both a port and a (different) label were attached to the net. Since the non-port label is not a preferred name, then there should be no warning message. The warning is only intended for cases where two (or more) ports are truly shorted together.
2023-06-13 23:14:36 +02:00
// pname = nodeSpiceName(snode->efnode_name->efnn_hier, NULL);
pname = nodeSpiceName(nodeName->efnn_hier, NULL);
Corrected two annoying errors: (1) SPICE netlist output subcircuits could have multiple ports of the same name. This problem had been worked over before, but there was an indpendent mechanism producing the same result for a completely different reason, caused by subcells being much larger than the cookie-cutter extraction method's extraction regions. Solved by tracking port names in a hash table and preventing re-use. (2) ext2spice was producing "no such node" errors; like (1) this had been previously worked on, and like (1) this mechanism was independent. Problem came from not passing -1 to extHierSubstrate for the non-arrayed dimension of a 1-dimensional array. Also: Removed the word "fatal" from extraction error reporting, as nearly all extraction errors are entirely benign. This should clear up confusion among alarmed end-users.
2020-11-24 21:30:49 +01:00
ext2spice.c: pname used before initialization In order for pname to be assigned a value the loop must have performed an interation. The order of assignment and use of 'pname' is not the natural order you'd expect. It also looks like the TxPrint message may display the incorrect port name from the previous loop iteration. The code block looks like the resolution of 'pname' has no side-effects. So should be done first so the correct port name is output in the log message. SonarCloud 2nd function call argument is an uninitialized value https://sonarcloud.io/project/issues?open=AZJB17lqNGfDNup0Rk6f&id=dlmiles_magic
2024-10-04 21:09:28 +02:00
if (heh == (HashEntry *)NULL) /* pname now resolved for log output */
{
/* Port was optimized out */
snode->efnode_flags &= ~EF_PORT;
TxPrintf("Note: Port %s was optimized out of %s\n",
pname, def->def_name);
continue;
}
There being a specific situation where the enumeration of ports used by "topVisit" and "subcktVisit" in ext2spice.c, probably caused by having different names on the same port number, the subcktVisit() routine was modified to use exactly the same enumeration as topVisit() so that they are guaranteed to have the same result.
2021-12-13 17:33:02 +01:00
hep = HashLookOnly(&portNameTable, pname);
(1) Modified the method used for the previous commit, as it was found to make ext2spice runtimes very long for large layouts. The new method is equivalent but doesn't incur the overhead. Also: Changed a flag check which was causing the substrate node to be output as a port for certain layouts where the substrate node connects to no devices, and so should be optimized out.
2022-01-13 23:00:07 +01:00
if (hep == (HashEntry *)NULL)
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
{
(1) Modified the method used for the previous commit, as it was found to make ext2spice runtimes very long for large layouts. The new method is equivalent but doesn't incur the overhead. Also: Changed a flag check which was causing the substrate node to be output as a port for certain layouts where the substrate node connects to no devices, and so should be optimized out.
2022-01-13 23:00:07 +01:00
hep = HashFind(&portNameTable, pname);
HashSetValue(hep, (ClientData)(pointertype)nodeName->efnn_port);
if (sorted_ports[portidx] == NULL)
sorted_ports[portidx] = StrDup((char **)NULL, pname);
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
}
else
{
(1) Modified the method used for the previous commit, as it was found to make ext2spice runtimes very long for large layouts. The new method is equivalent but doesn't incur the overhead. Also: Changed a flag check which was causing the substrate node to be output as a port for certain layouts where the substrate node connects to no devices, and so should be optimized out.
2022-01-13 23:00:07 +01:00
/* Node that was unassigned has been found to be
* a repeat (see NOTE at top), so make sure its
* port number is set correctly.
*/
nodeName->efnn_port = (int)(pointertype)HashGetValue(hep);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
}
}
Corrected two annoying errors: (1) SPICE netlist output subcircuits could have multiple ports of the same name. This problem had been worked over before, but there was an indpendent mechanism producing the same result for a completely different reason, caused by subcells being much larger than the cookie-cutter extraction method's extraction regions. Solved by tracking port names in a hash table and preventing re-use. (2) ext2spice was producing "no such node" errors; like (1) this had been previously worked on, and like (1) this mechanism was independent. Problem came from not passing -1 to extHierSubstrate for the non-arrayed dimension of a 1-dimensional array. Also: Removed the word "fatal" from extraction error reporting, as nearly all extraction errors are entirely benign. This should clear up confusion among alarmed end-users.
2020-11-24 21:30:49 +01:00
HashKill(&portNameTable);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
/* Output all ports, in order */
for (portorder = 0; portorder <= portmax; portorder++)
{
if (sorted_ports[portorder] != NULL)
{
if (tchars > 80)
{
/* Line continuation */
fprintf(esSpiceF, "\n+");
tchars = 1;
}
fprintf(esSpiceF, " %s", sorted_ports[portorder]);
Correction to yesterday's commit (minor issue with line length in the SPICE netlist output of ext2spice).
2021-10-05 14:50:45 +02:00
tchars += strlen(sorted_ports[portorder]) + 1;
Added a check for interrupt in "extresist". Modified the output of ports in SPICE so that ports are first ordered by index into a string array, and then output. This avoids executing an expensive double loop.
2021-10-05 01:52:22 +02:00
freeMagic(sorted_ports[portorder]);
}
}
freeMagic(sorted_ports);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
// Add any parameters defined by "property parameter" in the cell
instname = mallocMagic(2 + strlen(def->def_name));
sprintf(instname, ":%s", def->def_name);
plist = efGetDeviceParams(instname);
for (pptr = plist; pptr; pptr = pptr->parm_next)
{
if (tchars > 80)
{
/* Line continuation */
fprintf(esSpiceF, "\n+");
tchars = 1;
}
pname = pptr->parm_name;
fprintf(esSpiceF, " %s", pname);
tchars += strlen(pname) + 1;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
freeMagic(instname);
fprintf(esSpiceF, "\n");
}
Additional modification to write out a resistor or capacitor value as-is, without a parameter name and before any device model name or substrate pin, according to CDL syntax, if the parameter has been specified without a parameter name (e.g., "r=" instead of "r=r"). Corrected an error in the extract code which put the substrate node name in front of parameters instead of after. This was previously unexercised because only in CDL format does a resistor or capacitor model have parameters listed by name.
2025-10-06 23:42:34 +02:00
/*
* ----------------------------------------------------------------------------
*
* spcWriteValue ---
*
* Special handling for CDL format: Output any resistor or capacitor value,
* value only, no parameter name.
*
* ----------------------------------------------------------------------------
*/
Corrected the return value of spcWriteValue() (recently added) from "bool" to "void" (does not return a value).
2025-10-21 15:15:05 +02:00
void
Additional modification to write out a resistor or capacitor value as-is, without a parameter name and before any device model name or substrate pin, according to CDL syntax, if the parameter has been specified without a parameter name (e.g., "r=" instead of "r=r"). Corrected an error in the extract code which put the substrate node name in front of parameters instead of after. This was previously unexercised because only in CDL format does a resistor or capacitor model have parameters listed by name.
2025-10-06 23:42:34 +02:00
spcWriteValue(
Dev *dev, /* Dev being output */
HierName *hierName) /* Hierarchical path down to this dev */
{
DevParam *plist;
plist = efGetDeviceParams(EFDevTypes[dev->dev_type]);
while (plist != NULL)
{
switch (plist->parm_type[0])
{
case 'r':
if (*(plist->parm_name) == '\0')
{
fprintf(esSpiceF, " ");
esSIvalue(esSpiceF, (double)dev->dev_res);
}
break;
case 'c':
if (*(plist->parm_name) == '\0')
{
fprintf(esSpiceF, " ");
esSIvalue(esSpiceF, (double)dev->dev_res);
}
break;
}
plist = plist->parm_next;
}
}
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
/*
* ----------------------------------------------------------------------------
*
* spcWriteSubParam ---
*
* Special handling for CDL format: Output any substrate parameter before
* the device model. Why this makes sense, I have no idea.
*
* ----------------------------------------------------------------------------
*/
bool
spcWriteSubParam(
Dev *dev, /* Dev being output */
HierName *hierName) /* Hierarchical path down to this dev */
{
bool retval; /* True if substrate parameter was output */
DevParam *plist;
retval = FALSE;
plist = efGetDeviceParams(EFDevTypes[dev->dev_type]);
while (plist != NULL)
{
switch (plist->parm_type[0])
{
case 's':
if (dev->dev_subsnode == NULL)
TxError("Error: No substrate definition for device %s\n",
EFDevTypes[dev->dev_type]);
else
{
fprintf(esSpiceF, " %s=", plist->parm_name);
spcdevSubstrate(hierName,
dev->dev_subsnode->efnode_name->efnn_hier,
dev->dev_type, esSpiceF);
retval = TRUE;
}
break;
}
plist = plist->parm_next;
}
return retval;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* spcWriteParams ---
*
* Write parameters to a device line in SPICE output. This is normally
* restricted to subcircuit devices but may include other devices to
* accomodate various extensions to the basic SPICE format.
*
* ----------------------------------------------------------------------------
*/
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spcWriteParams(
Dev *dev, /* Dev being output */
HierName *hierName, /* Hierarchical path down to this dev */
float scale, /* Scale transform for output */
int l, /* Device length, in internal units */
int w, /* Device width, in internal units */
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
float sdM, /* Device multiplier */
bool subdone) /* If TRUE, substrate parameter was already output */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
bool hierD;
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
DevParam *plist, *dparam;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int parmval;
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
EFNode *dnode;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
plist = efGetDeviceParams(EFDevTypes[dev->dev_type]);
while (plist != NULL)
{
switch (plist->parm_type[0])
{
case 'a':
// Check for area of terminal node vs. device area
if (plist->parm_type[1] == '\0' || plist->parm_type[1] == '0')
{
fprintf(esSpiceF, " %s=", plist->parm_name);
parmval = dev->dev_area;
if (esScale < 0)
fprintf(esSpiceF, "%g", parmval * scale * scale);
else if (plist->parm_scale != 1.0)
ext2spice/ext2spice.c: Multiplication result converted to larger type Fix code scanning alert no. 43: Multiplication result converted to larger type (#59) Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
2024-09-30 08:46:35 +02:00
fprintf(esSpiceF, "%g", (double)parmval * (double)scale * (double)scale
* (double)esScale * (double)esScale * plist->parm_scale
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* 1E-12);
else
Extended the device parameter notation to include offsets using '+' and '-' in the same way that '*' is currently used for specifying a parameter scaling. The combination of a scale and offset for the same parameter has not (yet) been implemented.
2023-06-22 02:44:38 +02:00
esSIvalue(esSpiceF, 1.0E-12 * (parmval + plist->parm_offset)
* scale * scale * esScale * esScale);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
int pn, resclass;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
pn = plist->parm_type[1] - '0';
if (pn >= dev->dev_nterm) pn = dev->dev_nterm - 1;
hierD = extHierSDAttr(&dev->dev_terms[pn]);
Corrected a bad typo, "==" instead of "="; can cause a crash condition when doing ext2spice.
2021-03-18 16:52:56 +01:00
resclass = (pn > 1) ? esFetInfo[dev->dev_type].resClassDrain :
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[dev->dev_type].resClassSource;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
// For parameter a<n> followed by parameter p<n>,
// process both at the same time.
if (plist->parm_next && plist->parm_next->parm_type[0]
== 'p' && plist->parm_next->parm_type[1]
== plist->parm_type[1])
{
if (hierD)
spcnAPHier(&dev->dev_terms[pn], hierName,
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
resclass, scale, plist->parm_type,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
plist->parm_next->parm_type,
sdM, esSpiceF);
else
{
dnode = SpiceGetNode(hierName,
dev->dev_terms[pn].dterm_node->efnode_name->efnn_hier);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
spcnAP(&dev->dev_terms[pn], dnode, resclass, scale,
plist->parm_name, plist->parm_next->parm_name,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
sdM, esSpiceF, w);
}
plist = plist->parm_next;
}
else
{
if (hierD)
spcnAPHier(&dev->dev_terms[pn], hierName,
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
resclass, scale, plist->parm_type, NULL,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
sdM, esSpiceF);
else
{
dnode = SpiceGetNode(hierName,
dev->dev_terms[pn].dterm_node->efnode_name->efnn_hier);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
spcnAP(&dev->dev_terms[pn], dnode, resclass, scale,
plist->parm_name, NULL, sdM, esSpiceF, w);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
}
}
break;
case 'p':
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
// Check for perimeter of terminal node vs. device perimeter
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (plist->parm_type[1] == '\0' || plist->parm_type[1] == '0')
{
fprintf(esSpiceF, " %s=", plist->parm_name);
parmval = dev->dev_perim;
if (esScale < 0)
fprintf(esSpiceF, "%g", parmval * scale);
else if (plist->parm_scale != 1.0)
ext2spice/ext2spice.c: Multiplication result converted to larger type Fix code scanning alert no. 44: Multiplication result converted to larger type (#60) Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
2024-09-30 08:46:50 +02:00
fprintf(esSpiceF, "%g", (double)parmval * scale
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* esScale * plist->parm_scale * 1E-6);
else
Extended the device parameter notation to include offsets using '+' and '-' in the same way that '*' is currently used for specifying a parameter scaling. The combination of a scale and offset for the same parameter has not (yet) been implemented.
2023-06-22 02:44:38 +02:00
esSIvalue(esSpiceF, 1.0E-12 * (parmval + plist->parm_offset)
* scale * esScale);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
int pn, resclass;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
pn = plist->parm_type[1] - '0';
if (pn >= dev->dev_nterm) pn = dev->dev_nterm - 1;
Corrected a bad typo, "==" instead of "="; can cause a crash condition when doing ext2spice.
2021-03-18 16:52:56 +01:00
resclass = (pn > 1) ? esFetInfo[dev->dev_type].resClassDrain :
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
esFetInfo[dev->dev_type].resClassSource;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
hierD = extHierSDAttr(&dev->dev_terms[pn]);
// For parameter p<n> followed by parameter a<n>,
// process both at the same time.
if (plist->parm_next && plist->parm_next->parm_type[0]
== 'a' && plist->parm_next->parm_type[1]
== plist->parm_type[1])
{
if (hierD)
spcnAPHier(&dev->dev_terms[pn], hierName,
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
resclass, scale, plist->parm_next->parm_type,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
plist->parm_type, sdM, esSpiceF);
else
{
dnode = SpiceGetNode(hierName,
dev->dev_terms[pn].dterm_node->efnode_name->efnn_hier);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
spcnAP(&dev->dev_terms[pn], dnode, resclass, scale,
plist->parm_next->parm_name,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
plist->parm_name, sdM, esSpiceF, w);
}
plist = plist->parm_next;
}
else
{
if (hierD)
spcnAPHier(&dev->dev_terms[pn], hierName,
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
resclass, scale, NULL, plist->parm_type,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
sdM, esSpiceF);
else
{
dnode = SpiceGetNode(hierName,
dev->dev_terms[pn].dterm_node->efnode_name->efnn_hier);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
spcnAP(&dev->dev_terms[pn], dnode, resclass, scale,
NULL, plist->parm_name, sdM, esSpiceF, w);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
}
}
break;
case 'l':
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
// Check for length of device vs. depth of terminal
if (plist->parm_type[1] == '\0' || plist->parm_type[1] == '0')
{
fprintf(esSpiceF, " %s=", plist->parm_name);
if (esScale < 0)
fprintf(esSpiceF, "%g", l * scale);
else if (plist->parm_scale != 1.0)
ext2spice/ext2spice.c: Multiplication result converted to larger type Fix code scanning alert no. 45: Multiplication result converted to larger type (#61) Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
2024-09-30 08:47:05 +02:00
fprintf(esSpiceF, "%g", (double)l * scale * esScale
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* plist->parm_scale * 1E-6);
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
else
Extended the device parameter notation to include offsets using '+' and '-' in the same way that '*' is currently used for specifying a parameter scaling. The combination of a scale and offset for the same parameter has not (yet) been implemented.
2023-06-22 02:44:38 +02:00
esSIvalue(esSpiceF, 1.0E-6 * (l + plist->parm_offset)
* scale * esScale);
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
{
/* l1, l2, etc. used to indicate the length of the terminal */
/* Find the value in dev_params */
for (dparam = dev->dev_params; dparam; dparam = dparam->parm_next)
{
if ((strlen(dparam->parm_name) > 2) &&
(dparam->parm_name[0] == 'l') &&
(dparam->parm_name[1] == plist->parm_type[1]) &&
(dparam->parm_name[2] == '='))
{
int dval;
if (sscanf(&dparam->parm_name[3], "%d", &dval) == 1)
{
fprintf(esSpiceF, " %s=", plist->parm_name);
if (esScale < 0)
fprintf(esSpiceF, "%g", dval * scale);
else if (plist->parm_scale != 1.0)
ext2spice/ext2spice.c: Multiplication result converted to larger type Applying similat fix to others in the file, promote to double ASAP. Copilot Autofix: Retry and if the problem persists contact support. CodeQL: https://github.com/dlmiles/magic/security/code-scanning/46
2024-09-30 00:00:00 +02:00
fprintf(esSpiceF, "%g", (double)dval * scale * esScale
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
* plist->parm_scale * 1E-6);
else
Extended the device parameter notation to include offsets using '+' and '-' in the same way that '*' is currently used for specifying a parameter scaling. The combination of a scale and offset for the same parameter has not (yet) been implemented.
2023-06-22 02:44:38 +02:00
esSIvalue(esSpiceF, (dval + plist->parm_offset)
* scale * esScale * 1.0E-6);
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
dparam->parm_name[0] = '\0';
break;
}
}
}
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
Implemented simple drain and source length calculations as device parameters l1 and l2. Provides a way to pass the source or drain length as a parameter for, for example, an extended FET drain implemented as a resistor abutting the FET gate. Could potentially be used as a way to determine source/drain area and perimeter without resorting to measurements of a shared node.
2022-12-15 03:40:24 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case 'w':
Extended the device parameters to allow a terminal width that is different from the device (i.e., gate) width, for devices that do not define a MOS-like gate spanning the width of the device. This is restricted to the assumption that the terminal is rectangular and therefore a simple width and length can be derived from the area and perimeter. Also, length is defined as the smaller dimension and width as the larger dimension. For additional restrictions, see the updated documentation. This was added to allow correct width and length extraction of a bipolar emitter window, but may be more generally useful.
2024-12-06 22:42:43 +01:00
// Check for width of device vs. width of terminal
if (plist->parm_type[1] == '\0' || plist->parm_type[1] == '0')
{
fprintf(esSpiceF, " %s=", plist->parm_name);
if (esScale < 0)
fprintf(esSpiceF, "%g", w * scale);
else if (plist->parm_scale != 1.0)
fprintf(esSpiceF, "%g", (double)w * scale * esScale
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* plist->parm_scale * 1E-6);
Extended the device parameters to allow a terminal width that is different from the device (i.e., gate) width, for devices that do not define a MOS-like gate spanning the width of the device. This is restricted to the assumption that the terminal is rectangular and therefore a simple width and length can be derived from the area and perimeter. Also, length is defined as the smaller dimension and width as the larger dimension. For additional restrictions, see the updated documentation. This was added to allow correct width and length extraction of a bipolar emitter window, but may be more generally useful.
2024-12-06 22:42:43 +01:00
else
esSIvalue(esSpiceF, 1.0E-6 * (w + plist->parm_offset)
Extended the device parameter notation to include offsets using '+' and '-' in the same way that '*' is currently used for specifying a parameter scaling. The combination of a scale and offset for the same parameter has not (yet) been implemented.
2023-06-22 02:44:38 +02:00
* scale * esScale);
Extended the device parameters to allow a terminal width that is different from the device (i.e., gate) width, for devices that do not define a MOS-like gate spanning the width of the device. This is restricted to the assumption that the terminal is rectangular and therefore a simple width and length can be derived from the area and perimeter. Also, length is defined as the smaller dimension and width as the larger dimension. For additional restrictions, see the updated documentation. This was added to allow correct width and length extraction of a bipolar emitter window, but may be more generally useful.
2024-12-06 22:42:43 +01:00
}
else
{
/* w1, w2, etc. used to indicate the width of the terminal */
/* Find the value in dev_params */
for (dparam = dev->dev_params; dparam; dparam = dparam->parm_next)
{
if ((strlen(dparam->parm_name) > 2) &&
(dparam->parm_name[0] == 'w') &&
(dparam->parm_name[1] == plist->parm_type[1]) &&
(dparam->parm_name[2] == '='))
{
int dval;
if (sscanf(&dparam->parm_name[3], "%d", &dval) == 1)
{
fprintf(esSpiceF, " %s=", plist->parm_name);
if (esScale < 0)
fprintf(esSpiceF, "%g", dval * scale);
else if (plist->parm_scale != 1.0)
fprintf(esSpiceF, "%g", (double)dval * scale * esScale
* plist->parm_scale * 1E-6);
else
esSIvalue(esSpiceF, (dval + plist->parm_offset)
* scale * esScale * 1.0E-6);
dparam->parm_name[0] = '\0';
break;
}
}
}
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case 's':
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
if (!subdone)
{
if (dev->dev_subsnode == NULL)
TxError("Error: No substrate definition for device %s\n",
EFDevTypes[dev->dev_type]);
else
{
fprintf(esSpiceF, " %s=", plist->parm_name);
spcdevSubstrate(hierName,
dev->dev_subsnode->efnode_name->efnn_hier,
dev->dev_type, esSpiceF);
}
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case 'x':
fprintf(esSpiceF, " %s=", plist->parm_name);
if (esScale < 0)
fprintf(esSpiceF, "%g", dev->dev_rect.r_xbot * scale);
else if (plist->parm_scale != 1.0)
ext2spice/ext2spice.c: Multiplication result converted to larger type Fix code scanning alert no. 48: Multiplication result converted to larger type (#63) Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
2024-09-30 08:47:40 +02:00
fprintf(esSpiceF, "%g", (double)dev->dev_rect.r_xbot * scale
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* esScale * plist->parm_scale * 1E-6);
else
Extended the device parameter notation to include offsets using '+' and '-' in the same way that '*' is currently used for specifying a parameter scaling. The combination of a scale and offset for the same parameter has not (yet) been implemented.
2023-06-22 02:44:38 +02:00
esSIvalue(esSpiceF, (dev->dev_rect.r_xbot + plist->parm_offset)
* scale * esScale * 1.0E-6);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case 'y':
fprintf(esSpiceF, " %s=", plist->parm_name);
if (esScale < 0)
fprintf(esSpiceF, "%g", dev->dev_rect.r_ybot * scale);
else if (plist->parm_scale != 1.0)
ext2spice/ext2spice.c: Multiplication result converted to larger type Fix code scanning alert no. 49: Multiplication result converted to larger type (#52) Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
2024-09-30 08:11:16 +02:00
fprintf(esSpiceF, "%g", (double)dev->dev_rect.r_ybot * (double)scale
* (double)esScale * plist->parm_scale * 1E-6);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
Extended the device parameter notation to include offsets using '+' and '-' in the same way that '*' is currently used for specifying a parameter scaling. The combination of a scale and offset for the same parameter has not (yet) been implemented.
2023-06-22 02:44:38 +02:00
esSIvalue(esSpiceF, (dev->dev_rect.r_ybot + plist->parm_offset)
* scale * esScale * 1.0E-6);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case 'r':
Additional modification to write out a resistor or capacitor value as-is, without a parameter name and before any device model name or substrate pin, according to CDL syntax, if the parameter has been specified without a parameter name (e.g., "r=" instead of "r=r"). Corrected an error in the extract code which put the substrate node name in front of parameters instead of after. This was previously unexercised because only in CDL format does a resistor or capacitor model have parameters listed by name.
2025-10-06 23:42:34 +02:00
if (*(plist->parm_name) != '\0')
{
fprintf(esSpiceF, " %s=", plist->parm_name);
esSIvalue(esSpiceF, (double)dev->dev_res);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case 'c':
Additional modification to write out a resistor or capacitor value as-is, without a parameter name and before any device model name or substrate pin, according to CDL syntax, if the parameter has been specified without a parameter name (e.g., "r=" instead of "r=r"). Corrected an error in the extract code which put the substrate node name in front of parameters instead of after. This was previously unexercised because only in CDL format does a resistor or capacitor model have parameters listed by name.
2025-10-06 23:42:34 +02:00
if (*(plist->parm_name) != '\0')
{
fprintf(esSpiceF, " %s=", plist->parm_name);
esSIvalue(esSpiceF, (double)dev->dev_cap);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
}
plist = plist->parm_next;
}
/* Add parameters that are to be copied verbatim */
for (plist = dev->dev_params; plist; plist = plist->parm_next)
fprintf(esSpiceF, " %s", plist->parm_name);
}
/*
* ----------------------------------------------------------------------------
*
* esOutputResistor ---
*
* Routine used by spcdevVisit to print a resistor device. This
* is broken out into a separate routine so that each resistor
* device may be represented (if the option is selected) by a
* "tee" network of two resistors on either side of the central
* node, which then has a capacitance to ground.
*
* Results:
* None.
*
* Side effects:
* Output to the SPICE deck.
*
* ----------------------------------------------------------------------------
*/
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
esOutputResistor(
Dev *dev, /* Dev being output */
HierName *hierName, /* Hierarchical path down to this dev */
float scale, /* Scale transform for output */
DevTerm *term1,
DevTerm *term2, /* Terminals of the device */
bool has_model, /* Is this a modeled resistor? */
int l,
int w, /* Device length and width */
int dscale) /* Device scaling (for split resistors) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
float sdM;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Resistor is "Rnnn term1 term2 value" */
/* extraction sets two terminals, which are assigned */
/* term1=gate term2=source by the above code. */
/* extracted units are Ohms; output is in Ohms */
spcdevOutNode(hierName, term1->dterm_node->efnode_name->efnn_hier,
ext2spice.c: bugfix esOutputResistor() uses name without initialization Constification of APIs has shown up this bug. spcdevOutNode() uses 'name' as a read-only string, used for EFHNConcatLook() as 'errorStr' which is used to prefix error log messages. EFHNConcatLook() will already defer resolving hierName for the message. So the only useful thing to indicate to the user is the procedure that was trying to do the lookup.
2025-07-24 15:36:22 +02:00
"esOutputResistor", esSpiceF);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
spcdevOutNode(hierName, term2->dterm_node->efnode_name->efnn_hier,
ext2spice.c: bugfix esOutputResistor() uses name without initialization Constification of APIs has shown up this bug. spcdevOutNode() uses 'name' as a read-only string, used for EFHNConcatLook() as 'errorStr' which is used to prefix error log messages. EFHNConcatLook() will already defer resolving hierName for the message. So the only useful thing to indicate to the user is the procedure that was trying to do the lookup.
2025-07-24 15:36:22 +02:00
"esOutputResistor", esSpiceF);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
sdM = getCurDevMult();
/* SPICE has two resistor types. If the "name" (EFDevTypes) is */
/* "None", the simple resistor type is used, and a value given. */
/* If not, the "semiconductor resistor" is used, and L and W */
/* and the device name are output. */
if (!has_model)
{
fprintf(esSpiceF, " %f", ((double)(dev->dev_res)
/ (double)(dscale)) / (double)sdM);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, FALSE);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
else
{
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
bool subdone = FALSE;
One more minor change to the behavior just implemented: A value for a resistor or capacitor is optional but is valid SPICE and not just limited to CDL format, so it is now allowed to add "r=" at the end of a resistor or capacitor component (not subcircuit) to generate the optional component value.
2025-10-07 02:37:30 +02:00
spcWriteValue(dev, hierName);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
if (esFormat == CDL)
subdone = spcWriteSubParam(dev, hierName);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (esScale < 0)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " w=%g l=%g", w * scale, (l * scale) / dscale);
else
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
fprintf(esSpiceF, " w=");
esSIvalue(esSpiceF, 1.0E-6 * w * scale * esScale);
fprintf(esSpiceF, " l=");
esSIvalue(esSpiceF, 1.0E-6 * (l * scale * esScale) / dscale);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, subdone);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (sdM != 1.0)
fprintf(esSpiceF, " M=%g", sdM);
}
}
Fixed a long-standing error in which "ext2spice merge" fails when used with "ext2spice hierarchy on" because the device index is not reset between calls to output cells in the hierarchy, leading to a mismatch of the index for all cells after the first one output.
2019-10-23 15:33:37 +02:00
/* Report if device at index n has been deleted due to merging */
bool
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
devIsKilled(
int n)
Fixed a long-standing error in which "ext2spice merge" fails when used with "ext2spice hierarchy on" because the device index is not reset between calls to output cells in the hierarchy, leading to a mismatch of the index for all cells after the first one output.
2019-10-23 15:33:37 +02:00
{
return (esFMult[(n)] <= (float)0.0) ? TRUE : FALSE;
}
/* Add a dev's multiplier to the table and grow it if necessary */
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
addDevMult(
float f)
Fixed a long-standing error in which "ext2spice merge" fails when used with "ext2spice hierarchy on" because the device index is not reset between calls to output cells in the hierarchy, leading to a mismatch of the index for all cells after the first one output.
2019-10-23 15:33:37 +02:00
{
int i;
float *op;
if (esFMult == NULL) {
esFMult = (float *) mallocMagic((unsigned) (esFMSize*sizeof(float)));
}
else if (esFMIndex >= esFMSize)
{
op = esFMult;
esFMSize *= 2;
esFMult = (float *)mallocMagic((unsigned)(esFMSize * sizeof(float)));
for (i = 0; i < esFMSize / 2; i++) esFMult[i] = op[i];
if (op) freeMagic(op);
}
esFMult[esFMIndex++] = f;
}
/* Set the multiplier value f of device at index i */
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
setDevMult(
int i,
float f)
Fixed a long-standing error in which "ext2spice merge" fails when used with "ext2spice hierarchy on" because the device index is not reset between calls to output cells in the hierarchy, leading to a mismatch of the index for all cells after the first one output.
2019-10-23 15:33:37 +02:00
{
esFMult[i] = f;
}
/* Get the multiplier value of the device at the current index esFMIndex */
float
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
getCurDevMult(void)
Fixed a long-standing error in which "ext2spice merge" fails when used with "ext2spice hierarchy on" because the device index is not reset between calls to output cells in the hierarchy, leading to a mismatch of the index for all cells after the first one output.
2019-10-23 15:33:37 +02:00
{
return (esFMult && (esFMIndex > 0)) ? esFMult[esFMIndex-1] : (float)1.0;
}
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
/*
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
*-----------------------------------------------------------------------------
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
* swapDrainSource
*
* Swap drain and source ordering and the related stuff
* including the drain/source area parameters
*
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
* This is typically called if any terminal is marked with attribute "D" or "S"
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
* (label "D$" or "S$" at poly-diffusion interface),
* then swap order of source and drain compared to the default ordering.
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
*
* Note:
* Before calling this function, ensure that dev->dev_nterm >= 3
*
* Results:
* None
*
* Side effects:
* Soure (dev->dev_terms[1]) and drain (dev->dev_terms[2]) terminals
* are swapped.
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
*
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
*-----------------------------------------------------------------------------
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
*/
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
swapDrainSource(
Dev *dev)
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
{
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
DevTerm tmpTerm;
Pulled merge request from Jan Belohoubek, but corrected the problem of needing to revert parameter definitions after swapping them during a device source/drain swap.
2020-08-09 19:11:38 +02:00
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
/* swap original terminals */
memcpy(&tmpTerm, &(dev->dev_terms[1]), sizeof(DevTerm));
memcpy(&(dev->dev_terms[1]), &(dev->dev_terms[2]), sizeof(DevTerm));
memcpy(&(dev->dev_terms[2]), &tmpTerm, sizeof(DevTerm));
swapDrainSource moved into a function to encapsulate swap-related tasks; area parameters swapped according to the D/S attributes; SPICE attribute comment printf moved to the end of export function as it allows to add more device types
2020-08-09 17:34:24 +02:00
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* spcdevVisit --
*
* Procedure to output a single dev to the .spice file.
* Called by EFVisitDevs().
*
* Results:
* Returns 0 always.
*
* Side effects:
* Writes to the file esSpiceF.
*
* Format of a .spice dev line:
*
* M%d drain gate source substrate type w=w l=l * x y
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* + ad= pd= as= ps= * asub= psub=
* **devattr g= s= d=
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
* where
* type is a name identifying this type of transistor
* other types of transistors are extracted with
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* an M card but it should be easy to turn them to whatever
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* you want.
* gate, source, and drain are the nodes to which these three
* terminals connect
* l, w are the length and width of the channel
* x, y are the x, y coordinates of a point within the channel.
* g=, s=, d= are the (optional) attributes; if present, each
* is followed by a comma-separated list of attributes.
*
* ----------------------------------------------------------------------------
*/
extflat: integrate cb_extflat_visitdevs_t for EFVisitDevs()
2025-07-20 01:03:14 +02:00
/* ARGSUSED */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitdevs_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spcdevVisit(
Dev *dev, /* Dev being output */
HierContext *hc, /* Hierarchical context down to this dev */
float scale, /* Scale transform for output */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
Transform *trans, /* (unused) */
ClientData cdata) /* (unused) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
DevParam *plist, *pptr;
DevTerm *gate, *source, *drain;
EFNode *subnode, *snode, *dnode, *subnodeFlat = NULL;
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
int l, w, i;
ext2spice: forward declaration arguments or removal Remove unnecessary forward declaration: extHierSDAttr() as ext2spice.h has prototype
2025-07-18 15:40:53 +02:00
bool subAP= FALSE, hierS, hierD;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
float sdM;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
char name[12], devchar;
bool has_model = TRUE;
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
bool subdone = FALSE;
Additional modification to the devVisit routine in extflat to change from passing the HierName to passing the HierContext (of which HierName is a part) so that more information from HierContext (such as the cell use being visited) can be passed to the callback procedure (largely for diagnostic purposes).
2020-06-01 23:14:22 +02:00
HierName *hierName = hc->hc_hierName;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
sprintf(name, "output");
/* If no terminals, can't do much of anything */
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (dev->dev_nterm < 1 )
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
if ( (esMergeDevsA || esMergeDevsC) && devIsKilled(esFMIndex++) )
return 0;
/* Get L and W of device */
EFGetLengthAndWidth(dev, &l, &w);
/* If only two terminals, connect the source to the drain */
gate = &dev->dev_terms[0];
if (dev->dev_nterm >= 2)
source = drain = &dev->dev_terms[1];
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (dev->dev_nterm >= 3)
{
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
drain = &dev->dev_terms[2];
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* If any terminal is marked with attribute "D" or "S" */
/* (label "D$" or "S$" at poly-diffusion interface), */
/* then force order of source and drain accordingly. */
if ((dev->dev_terms[1].dterm_attrs &&
!strcmp(dev->dev_terms[1].dterm_attrs, "D")) ||
(dev->dev_terms[2].dterm_attrs &&
!strcmp(dev->dev_terms[2].dterm_attrs, "S")))
{
Merged pull request #35 from Jan Belohoubek
2021-02-18 03:08:10 +01:00
swapDrainSource(dev);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
}
subnode = dev->dev_subsnode;
/* Check for minimum number of terminals. */
switch(dev->dev_class)
{
case DEV_SUBCKT:
Updated the version to go along with the merge of pull request version 8.3.521 that, due to an argument size mismatch, causes device parameters in netlist output to be printed as zero. Also added a small extension to the list of extraction types to include "device veriloga", which has the same syntax as "device subcircuit" but generates a component type "N" (Verilog-A component) in the netlist output.
2025-08-07 17:54:49 +02:00
case DEV_VERILOGA:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_RSUBCKT:
Added extraction device type "csubcircuit", which should have been added a long time ago, since capacitors use a different method for calculating width and length than either transistors or resistors, so subcircuits need a special class designator or else the extraction may calculate the wrong dimensions for device width by totalling the perimeter between the device and terminal types, as it would for a MOSFET.
2018-10-30 21:19:20 +01:00
case DEV_CSUBCKT:
Added a new device type "dsubcircuit" which basically behaves like "csubcircuit" but swaps the first two pins (with the device identifier layer becoming the 2nd pin and the other terminal the first), which is needed for n-type diodes modeled as subcircuits where the subcircuit pin order matches the order of pins for a SPICE n-type diode component. Previously "msubcircuit" was used for this purpose, but will calculate the wrong L and W. While use of L and W for diodes is rare, this device type also works for reversed capacitors (where the bottom or non-identifying layer terminal is in the first pin position of the subcircuit).
2025-10-07 22:45:22 +02:00
case DEV_DSUBCKT:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_MSUBCKT:
break;
case DEV_DIODE:
case DEV_PDIODE:
case DEV_NDIODE:
if ((dev->dev_nterm < 2) && (subnode == NULL))
{
TxError("Diode has only one terminal\n");
return 0;
}
break;
default:
if (dev->dev_nterm < 2)
{
TxError("Device other than subcircuit has only "
"one terminal\n");
return 0;
}
break;
}
/* Original hack for BiCMOS, Tim 10/4/97, is deprecated. */
/* Use of "device bjt" preferred to "fet" with model="npn". */
if (!strcmp(EFDevTypes[dev->dev_type], "npn")) dev->dev_class = DEV_BJT;
/* For resistor and capacitor classes, set a boolean to */
/* denote whether the device has a model or not, so we */
/* don't have to keep doing a string compare on EFDevTypes. */
switch(dev->dev_class)
{
case DEV_RES:
case DEV_CAP:
case DEV_CAPREV:
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (dev->dev_nterm < 1)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
if ((dev->dev_type == esNoModelType) ||
!strcmp(EFDevTypes[dev->dev_type], "None"))
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
has_model = FALSE;
break;
}
/* Flag shorted devices---this should probably be an option */
switch(dev->dev_class)
{
case DEV_MOSFET:
case DEV_ASYMMETRIC:
case DEV_FET:
if (source == drain)
{
Changed the end-of-line comment character from ';' to '$' when formatting a SPICE netlist for NGSPICE. This is the documented character as described in the NGSPICE User Manual, section 2.2.4 End-of-line comments
2020-06-24 21:22:04 +02:00
if (esFormat == NGSPICE) fprintf(esSpiceF, "$ ");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, "** SOURCE/DRAIN TIED\n");
}
break;
default:
if (gate == source)
{
Changed the end-of-line comment character from ';' to '$' when formatting a SPICE netlist for NGSPICE. This is the documented character as described in the NGSPICE User Manual, section 2.2.4 End-of-line comments
2020-06-24 21:22:04 +02:00
if (esFormat == NGSPICE) fprintf(esSpiceF, "$ ");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, "** SHORTED DEVICE\n");
}
break;
}
/* Generate SPICE device name */
switch(dev->dev_class)
{
case DEV_MOSFET:
case DEV_ASYMMETRIC:
case DEV_FET:
devchar = 'M';
break;
case DEV_BJT:
devchar = 'Q';
break;
case DEV_DIODE:
case DEV_NDIODE:
case DEV_PDIODE:
devchar = 'D';
break;
case DEV_RES:
devchar = 'R';
break;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
case DEV_VOLT:
devchar = 'V';
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_CAP:
case DEV_CAPREV:
devchar = 'C';
break;
Updated the version to go along with the merge of pull request version 8.3.521 that, due to an argument size mismatch, causes device parameters in netlist output to be printed as zero. Also added a small extension to the list of extraction types to include "device veriloga", which has the same syntax as "device subcircuit" but generates a component type "N" (Verilog-A component) in the netlist output.
2025-08-07 17:54:49 +02:00
case DEV_VERILOGA:
devchar = 'N';
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_SUBCKT:
case DEV_RSUBCKT:
Added extraction device type "csubcircuit", which should have been added a long time ago, since capacitors use a different method for calculating width and length than either transistors or resistors, so subcircuits need a special class designator or else the extraction may calculate the wrong dimensions for device width by totalling the perimeter between the device and terminal types, as it would for a MOSFET.
2018-10-30 21:19:20 +01:00
case DEV_CSUBCKT:
Added a new device type "dsubcircuit" which basically behaves like "csubcircuit" but swaps the first two pins (with the device identifier layer becoming the 2nd pin and the other terminal the first), which is needed for n-type diodes modeled as subcircuits where the subcircuit pin order matches the order of pins for a SPICE n-type diode component. Previously "msubcircuit" was used for this purpose, but will calculate the wrong L and W. While use of L and W for diodes is rare, this device type also works for reversed capacitors (where the bottom or non-identifying layer terminal is in the first pin position of the subcircuit).
2025-10-07 22:45:22 +02:00
case DEV_DSUBCKT:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_MSUBCKT:
devchar = 'X';
break;
}
fprintf(esSpiceF, "%c", devchar);
/* Device index is taken from gate attributes if attached; */
/* otherwise, the device is numbered in sequence. */
if (gate->dterm_attrs)
{
/* Output the name found in the gate attributes */
/* prefixed by the hierarchical name. */
fprintf(esSpiceF, "%s%s", EFHNToStr(hierName), gate->dterm_attrs);
}
else
{
switch (dev->dev_class)
{
case DEV_RES:
fprintf(esSpiceF, "%d", esResNum++);
/* For resistor tee networks, use, e.g., */
/* "R1A" and "R1B", for clarity */
if (esDoResistorTee) fprintf(esSpiceF, "A");
break;
case DEV_DIODE:
case DEV_NDIODE:
case DEV_PDIODE:
fprintf(esSpiceF, "%d", esDiodeNum++);
break;
case DEV_CAP:
case DEV_CAPREV:
fprintf(esSpiceF, "%d", esCapNum++);
break;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
case DEV_VOLT:
fprintf(esSpiceF, "%d", esVoltNum++);
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_SUBCKT:
Updated the version to go along with the merge of pull request version 8.3.521 that, due to an argument size mismatch, causes device parameters in netlist output to be printed as zero. Also added a small extension to the list of extraction types to include "device veriloga", which has the same syntax as "device subcircuit" but generates a component type "N" (Verilog-A component) in the netlist output.
2025-08-07 17:54:49 +02:00
case DEV_VERILOGA:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_RSUBCKT:
Added extraction device type "csubcircuit", which should have been added a long time ago, since capacitors use a different method for calculating width and length than either transistors or resistors, so subcircuits need a special class designator or else the extraction may calculate the wrong dimensions for device width by totalling the perimeter between the device and terminal types, as it would for a MOSFET.
2018-10-30 21:19:20 +01:00
case DEV_CSUBCKT:
Added a new device type "dsubcircuit" which basically behaves like "csubcircuit" but swaps the first two pins (with the device identifier layer becoming the 2nd pin and the other terminal the first), which is needed for n-type diodes modeled as subcircuits where the subcircuit pin order matches the order of pins for a SPICE n-type diode component. Previously "msubcircuit" was used for this purpose, but will calculate the wrong L and W. While use of L and W for diodes is rare, this device type also works for reversed capacitors (where the bottom or non-identifying layer terminal is in the first pin position of the subcircuit).
2025-10-07 22:45:22 +02:00
case DEV_DSUBCKT:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_MSUBCKT:
fprintf(esSpiceF, "%d", esSbckNum++);
break;
default:
fprintf(esSpiceF, "%d", esDevNum++);
break;
}
}
/* Order and number of nodes in the output depends on the device class */
switch (dev->dev_class)
{
case DEV_BJT:
/* BJT is "Qnnn collector emitter base model" */
/* extraction sets collector=subnode, emitter=gate, base=drain */
sprintf(name, "fet");
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
spcdevOutNode(hierName, subnode->efnode_name->efnn_hier, name, esSpiceF);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
spcdevOutNode(hierName, gate->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
/* fix mixed up drain/source for bjts hace 2/2/99 */
if (gate->dterm_node->efnode_name->efnn_hier ==
source->dterm_node->efnode_name->efnn_hier)
spcdevOutNode(hierName, drain->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
else
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
if (esFormat == CDL)
subdone = spcWriteSubParam(dev, hierName);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
Cleanup of a number of uninitialized variable issues flagged by the compiler. Some are obscure functions (plot verstatec hasn't been used in years) but others (like SPICE distributed junctions) are potentially significant sources of unexpected crashes on systems that don't zero uninitialized memory.
2019-06-07 20:13:50 +02:00
sdM = getCurDevMult();
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, subdone);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case DEV_MSUBCKT:
Added a new device type "dsubcircuit" which basically behaves like "csubcircuit" but swaps the first two pins (with the device identifier layer becoming the 2nd pin and the other terminal the first), which is needed for n-type diodes modeled as subcircuits where the subcircuit pin order matches the order of pins for a SPICE n-type diode component. Previously "msubcircuit" was used for this purpose, but will calculate the wrong L and W. While use of L and W for diodes is rare, this device type also works for reversed capacitors (where the bottom or non-identifying layer terminal is in the first pin position of the subcircuit).
2025-10-07 22:45:22 +02:00
case DEV_DSUBCKT:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Corrected ext2spice, which was reversing source and drain for the "msubcircuit" extraction model, which would normally not make any difference except that when source and/or drain are tagged with terminal attributes, then the source and drain are swapped with respect to what is expected in the output SPICE netlist.
2020-07-29 20:13:23 +02:00
/* MOS-like subcircuit is "Xnnn drain gate [source [sub]]" */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* to more conveniently handle cases where MOS devices are */
/* modeled by subcircuits with the same pin ordering. */
Corrected ext2spice, which was reversing source and drain for the "msubcircuit" extraction model, which would normally not make any difference except that when source and/or drain are tagged with terminal attributes, then the source and drain are swapped with respect to what is expected in the output SPICE netlist.
2020-07-29 20:13:23 +02:00
spcdevOutNode(hierName, drain->dterm_node->efnode_name->efnn_hier,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
name, esSpiceF);
/* Drop through to below (no break statement) */
case DEV_SUBCKT:
Corrected a few errors in the implementation of "csubcircuit", now fully tested and verified.
2018-10-31 02:56:05 +01:00
case DEV_CSUBCKT:
Updated the version to go along with the merge of pull request version 8.3.521 that, due to an argument size mismatch, causes device parameters in netlist output to be printed as zero. Also added a small extension to the list of extraction types to include "device veriloga", which has the same syntax as "device subcircuit" but generates a component type "N" (Verilog-A component) in the netlist output.
2025-08-07 17:54:49 +02:00
case DEV_VERILOGA:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/* Subcircuit is "Xnnn gate [source [drain [sub]]]" */
/* Subcircuit .subckt record must be ordered to match! */
spcdevOutNode(hierName, gate->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
/* Drop through to below (no break statement) */
case DEV_RSUBCKT:
/* RC-like subcircuits are exactly like other subcircuits */
/* except that the "gate" node is treated as an identifier */
/* only and is not output. */
Corrected the output of the new "dsubcircuit" device which was generating output for one of the device nodes twice in the netlist.
2025-10-08 16:04:38 +02:00
if (dev->dev_class == DEV_DSUBCKT)
{
/* Do nothing: Both terminals have already been output */
}
else if (dev->dev_class != DEV_MSUBCKT)
Corrected ext2spice, which was reversing source and drain for the "msubcircuit" extraction model, which would normally not make any difference except that when source and/or drain are tagged with terminal attributes, then the source and drain are swapped with respect to what is expected in the output SPICE netlist.
2020-07-29 20:13:23 +02:00
{
if (dev->dev_nterm > 1)
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
if (dev->dev_nterm > 2)
spcdevOutNode(hierName, drain->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
}
else /* class DEV_MSUBCKT */
{
if (dev->dev_nterm > 2)
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Updated the version to go along with the merge of pull request version 8.3.521 that, due to an argument size mismatch, causes device parameters in netlist output to be printed as zero. Also added a small extension to the list of extraction types to include "device veriloga", which has the same syntax as "device subcircuit" but generates a component type "N" (Verilog-A component) in the netlist output.
2025-08-07 17:54:49 +02:00
/* The following only applies to DEV_SUBCKT and DEV_VERILOGA, which */
/* may define as many terminal types as they want. */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
for (i = 4; i < dev->dev_nterm; i++)
{
drain = &dev->dev_terms[i - 1];
spcdevOutNode(hierName, drain->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
}
/* Get the device parameters now, and check if the substrate is */
/* passed as a parameter rather than as a node. */
plist = efGetDeviceParams(EFDevTypes[dev->dev_type]);
for (pptr = plist; pptr != NULL; pptr = pptr->parm_next)
if (pptr->parm_type[0] == 's')
break;
if ((pptr == NULL) && subnode)
{
fprintf(esSpiceF, " ");
subnodeFlat = spcdevSubstrate(hierName,
subnode->efnode_name->efnn_hier,
dev->dev_type, esSpiceF);
}
A number of changes: 1) Corrected spurious error messages about cells already existing in GDS when using "flatten" or "flatglob". 2) Fixed handling of resistance as a subcircuit parameter 3) Added area and perimeter resistance for a device; this is done through the "devresist" statement in the tech file, which is an extension of the original "fetresist" statement. Where "fetresist" only supported type "linear", "devresist" supports types "area" and "perimeter". 4) Support for CDL syntax, including generating subcircuit-like parameters for components starting with SPICE-standard prefixes like M, R, C, etc., adding "/" between pins and subcircuit name, and saving the file as ".cdl" instead of ".spice". 5) Estimated L and W for devices whose geometry is complex and do not reduce to a simple rectangle. L and W are estimated as the square root of the area. 6) Changed the method of extracting L and W for diodes to use the same method as capacitors. Note that diodes are not usually specified by L and W, but if they are, this will produce the right result. 7) Corrected the reported filename and line number when printing error messages related to errors inside a technology file, when the technology file uses "include" to combine multiple files.
2025-10-01 21:17:49 +02:00
/* CDL format support: Output a slash followed by a space. */
if (esFormat == CDL) fprintf(esSpiceF, " /");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
/* Write all requested parameters to the subcircuit call. */
sdM = getCurDevMult();
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, FALSE);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (sdM != 1.0)
fprintf(esSpiceF, " M=%g", sdM);
break;
case DEV_RES:
if (esDoResistorTee)
{
/* There are three ways of handling capacitance */
/* on resistor networks. One is to ignore it */
/* (the default; generates "floating" nodes in */
/* the SPICE output) which is okay for LVS. */
/* Another way is the Pi network, in which the */
/* capacitance is split evenly between the */
/* terminals. Again, the resistor node is left */
/* floating. The third is the Tee network, in */
/* which the resistance is split in two parts, */
/* connecting to a capacitor to ground in the */
/* middle. This is the best solution but plays */
/* havoc with LVS. So, the choice is a command */
/* line option. */
esOutputResistor(dev, hierName, scale, gate, source, has_model,
l, w, 2);
fprintf(esSpiceF, "\n%c", devchar);
if (gate->dterm_attrs)
fprintf(esSpiceF, "%s%sB", EFHNToStr(hierName), gate->dterm_attrs);
else
fprintf(esSpiceF, "%dB", esResNum - 1);
esOutputResistor(dev, hierName, scale, gate, drain, has_model,
l, w, 2);
}
else
{
esOutputResistor(dev, hierName, scale, source, drain, has_model,
l, w, 1);
}
break;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
case DEV_VOLT:
/* The voltage source is "Vnnn term1 term2 0.0" and is used
* only to separate shorted port names.
*/
if (dev->dev_nterm > 1)
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
One more round of edits on that issue. . .
2021-09-21 02:14:25 +02:00
if (dev->dev_nterm > 2)
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
spcdevOutNode(hierName, drain->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
fprintf(esSpiceF, " 0.0");
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_DIODE:
case DEV_PDIODE:
/* Diode is "Dnnn top bottom model" */
spcdevOutNode(hierName, gate->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
if (dev->dev_nterm > 1)
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
else if (subnode)
spcdevOutNode(hierName, subnode->efnode_name->efnn_hier,
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
name, esSpiceF);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
if (esFormat == CDL)
subdone = spcWriteSubParam(dev, hierName);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
Corrected a few additional uninitialized variable situations that had been shadowed by the previous set of fixes.
2019-06-07 21:00:39 +02:00
sdM = getCurDevMult();
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, subdone);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case DEV_NDIODE:
/* Diode is "Dnnn bottom top model" */
if (dev->dev_nterm > 1)
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
else if (subnode)
spcdevOutNode(hierName, subnode->efnode_name->efnn_hier,
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
name, esSpiceF);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
spcdevOutNode(hierName, gate->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
if (esFormat == CDL)
subdone = spcWriteSubParam(dev, hierName);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
Corrected a few additional uninitialized variable situations that had been shadowed by the previous set of fixes.
2019-06-07 21:00:39 +02:00
sdM = getCurDevMult();
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, subdone);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case DEV_CAP:
/* Capacitor is "Cnnn top bottom value" */
/* extraction sets top=gate bottom=source */
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
/* extracted units are fF. */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
spcdevOutNode(hierName, gate->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
sdM = getCurDevMult();
/* SPICE has two capacitor types. If the "name" (EFDevTypes) is */
/* "None", the simple capacitor type is used, and a value given. */
/* If not, the "semiconductor capacitor" is used, and L and W */
/* and the device name are output. */
if (!has_model)
{
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
esSIvalue(esSpiceF, 1.0E-15 * (double)sdM * (double)dev->dev_cap);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, FALSE);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
else
{
One more minor change to the behavior just implemented: A value for a resistor or capacitor is optional but is valid SPICE and not just limited to CDL format, so it is now allowed to add "r=" at the end of a resistor or capacitor component (not subcircuit) to generate the optional component value.
2025-10-07 02:37:30 +02:00
spcWriteValue(dev, hierName);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
if (esFormat == CDL)
subdone = spcWriteSubParam(dev, hierName);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
if (esScale < 0)
fprintf(esSpiceF, " w=%g l=%g", w*scale, l*scale);
else
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
fprintf(esSpiceF, " w=");
esSIvalue(esSpiceF, 1.0E-6 * w * scale * esScale);
fprintf(esSpiceF, " l=");
esSIvalue(esSpiceF, 1.0E-6 * l * scale * esScale);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, subdone);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (sdM != 1.0)
fprintf(esSpiceF, " M=%g", sdM);
}
break;
case DEV_CAPREV:
/* Capacitor is "Cnnn bottom top value" */
/* extraction sets top=source bottom=gate */
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
/* extracted units are fF. */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
spcdevOutNode(hierName, gate->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
sdM = getCurDevMult();
/* SPICE has two capacitor types. If the "name" (EFDevTypes) is */
/* "None", the simple capacitor type is used, and a value given. */
/* If not, the "semiconductor capacitor" is used, and L and W */
/* and the device name are output. */
if (!has_model)
{
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
esSIvalue(esSpiceF, 1.0E-15 * (double)sdM * (double)dev->dev_cap);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, FALSE);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
else
{
One more minor change to the behavior just implemented: A value for a resistor or capacitor is optional but is valid SPICE and not just limited to CDL format, so it is now allowed to add "r=" at the end of a resistor or capacitor component (not subcircuit) to generate the optional component value.
2025-10-07 02:37:30 +02:00
spcWriteValue(dev, hierName);
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
if (esFormat == CDL)
subdone = spcWriteSubParam(dev, hierName);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
if (esScale < 0)
fprintf(esSpiceF, " w=%g l=%g", w*scale, l*scale);
else
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
fprintf(esSpiceF, " w=");
esSIvalue(esSpiceF, 1.0E-6 * w * scale * esScale);
fprintf(esSpiceF, " l=");
esSIvalue(esSpiceF, 1.0E-6 * l * scale * esScale);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, subdone);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (sdM != 1.0)
fprintf(esSpiceF, " M=%g", sdM);
}
break;
case DEV_FET:
case DEV_MOSFET:
case DEV_ASYMMETRIC:
/* MOSFET is "Mnnn drain gate source [L=x W=x [attributes]]" */
spcdevOutNode(hierName, drain->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
spcdevOutNode(hierName, gate->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
spcdevOutNode(hierName, source->dterm_node->efnode_name->efnn_hier,
name, esSpiceF);
if (subnode)
{
fprintf(esSpiceF, " ");
subnodeFlat = spcdevSubstrate(hierName,
subnode->efnode_name->efnn_hier,
dev->dev_type, esSpiceF);
}
fprintf(esSpiceF, " %s", EFDevTypes[dev->dev_type]);
/*
* Scale L and W appropriately by the same amount as distance
* values in the transform. The transform will have a scale
* different from 1 only in the case when the scale factors of
* some of the .ext files differed, making it necessary to scale
* all dimensions explicitly instead of having a single scale
* factor at the beginning of the .spice file.
*/
sdM = getCurDevMult();
if (esScale < 0)
fprintf(esSpiceF, " w=%g l=%g", w*scale, l*scale);
else
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
fprintf(esSpiceF, " w=");
esSIvalue(esSpiceF, 1.0E-6 * w * scale * esScale);
fprintf(esSpiceF, " l=");
esSIvalue(esSpiceF, 1.0E-6 * l * scale * esScale);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Made a change to CDL format output to place the substrate parameter in front of the device name, which is a very inconsistent way to make the syntax, but whatever.
2025-10-06 22:40:19 +02:00
spcWriteParams(dev, hierName, scale, l, w, sdM, FALSE);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (sdM != 1.0)
fprintf(esSpiceF, " M=%g", sdM);
/*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Check controlling attributes and output area and perimeter.
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*/
hierS = extHierSDAttr(source);
hierD = extHierSDAttr(drain);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if ( gate->dterm_attrs )
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
subAP = Match(ATTR_SUBSAP, gate->dterm_attrs ) ;
fprintf(esSpiceF, "\n+ ");
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (hierD)
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
spcnAPHier(drain, hierName, esFetInfo[dev->dev_type].resClassDrain,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
scale, "ad", "pd", sdM, esSpiceF);
else
{
dnode = SpiceGetNode(hierName, drain->dterm_node->efnode_name->efnn_hier);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
spcnAP(drain, dnode, esFetInfo[dev->dev_type].resClassDrain, scale,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"ad", "pd", sdM, esSpiceF, w);
}
if (hierS)
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
spcnAPHier(source, hierName, esFetInfo[dev->dev_type].resClassSource,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
scale, "as", "ps", sdM, esSpiceF);
else {
snode= SpiceGetNode(hierName, source->dterm_node->efnode_name->efnn_hier);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
spcnAP(source, snode, esFetInfo[dev->dev_type].resClassSource, scale,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
"as", "ps", sdM, esSpiceF, w);
}
if (subAP)
{
fprintf(esSpiceF, " * ");
if (esFetInfo[dev->dev_type].resClassSub < 0)
{
TxError("error: subap for devtype %d unspecified\n",
dev->dev_type);
fprintf(esSpiceF, "asub=0 psub=0");
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else if (subnodeFlat)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
spcnAP(NULL, subnodeFlat, esFetInfo[dev->dev_type].resClassSub,
scale, "asub", "psub", sdM, esSpiceF, -1);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
fprintf(esSpiceF, "asub=0 psub=0");
}
/* Now output attributes, if present */
if (!esNoAttrs)
{
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
bool haveSattr = FALSE;
bool haveDattr = FALSE;
if (source->dterm_attrs && (*source->dterm_attrs))
haveSattr = TRUE;
if (drain->dterm_attrs && (*drain->dterm_attrs))
haveDattr = TRUE;
if (gate->dterm_attrs || haveSattr || haveDattr)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF,"\n**devattr");
if (gate->dterm_attrs)
fprintf(esSpiceF, " g=%s", gate->dterm_attrs);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
if (haveSattr)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " s=%s", source->dterm_attrs);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
if (haveDattr)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, " d=%s", drain->dterm_attrs);
}
break;
}
fprintf(esSpiceF, "\n");
Pulled merge request from Jan Belohoubek, but corrected the problem of needing to revert parameter definitions after swapping them during a device source/drain swap.
2020-08-09 19:11:38 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* spcdevSubstrate -
*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Output the node name of the substrate of a dev. If the suffix is the
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* same as the default dont go looking for it just output the default
* (trimmed appropriately). Otherwise look it up ..
*
* Results:
* NULL if not found or the default substrate or the node pointer
* otherwise (might be reused to output area and perimeter of
* the substrate).
*
* Side effects:
* Might allocate the nodeClient for the node through nodeSpiceName.
*
* ----------------------------------------------------------------------------
*/
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
EFNode *
spcdevSubstrate(
HierName *prefix,
HierName *suffix,
int type,
FILE *outf)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
HashEntry *he;
EFNodeName *nn;
char *suf;
suf = EFHNToStr(suffix);
if (esFetInfo[type].defSubs && strcasecmp(suf,esFetInfo[type].defSubs) == 0) {
esFormatSubs(outf, suf);
return NULL;
}
else {
he = EFHNConcatLook(prefix, suffix, "substrate");
if (he == NULL)
{
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (outf)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(outf, "errGnd!");
return NULL;
}
/* Canonical name */
nn = (EFNodeName *) HashGetValue(he);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (outf)
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
fprintf(outf, "%s", nodeSpiceName(nn->efnn_node->efnode_name->efnn_hier,
NULL));
Possible fix to problem with substrate node being unitialized when called after "merge conservative" is selected in ext2spice.
2017-10-20 22:07:52 +02:00
Got to the bottom of why area and perimeter values are no longer output for transistors. The problem came from a change made to fix an issue with capacitors marked as floating nodes because some nodes are not output as source or drain. But those nodes are output before the parameters, so when generating parameter output, all nodes appear to have already been output. Solution: Specify an additional bit in the "visited" mask for the node having been output that is separate from the mask for resist classes used by the code that writes parameter values, and use that bit as a test for whether the node is connected to some device (not necessarily a FET source or drain).
2021-12-20 19:12:49 +01:00
/* Create node client if it doesn't exist */
Restored compiling of the non-Tcl/Tk version of magic. This has traditionally been kept for backwards compatibility. However, the operation of "ext2spice" and "ext2sim" as separate programs has become extremely difficult to maintain, and so it has been dropped in favor of folding both into the program as commands, as was done a long time ago in the Tcl/Tk version.
2022-02-08 22:12:07 +01:00
if ((nodeClient *)nn->efnn_node->efnode_client == (nodeClient *)NULL)
Possible fix to problem with substrate node being unitialized when called after "merge conservative" is selected in ext2spice.
2017-10-20 22:07:52 +02:00
initNodeClientHier(nn->efnn_node);
Got to the bottom of why area and perimeter values are no longer output for transistors. The problem came from a change made to fix an issue with capacitors marked as floating nodes because some nodes are not output as source or drain. But those nodes are output before the parameters, so when generating parameter output, all nodes appear to have already been output. Solution: Specify an additional bit in the "visited" mask for the node having been output that is separate from the mask for resist classes used by the code that writes parameter values, and use that bit as a test for whether the node is connected to some device (not necessarily a FET source or drain).
2021-12-20 19:12:49 +01:00
/* Mark node as visited (set bit one higher than number of resist classes) */
if (esDistrJunct)
update_w(efNumResistClasses, 1, nn->efnn_node);
else
markVisited((nodeClientHier *)nn->efnn_node->efnode_client,
efNumResistClasses);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return nn->efnn_node;
}
}
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
/*
* ----------------------------------------------------------------------------
*
* nDecimals() ---
*
* Code taken from discussion at
* https://stackoverflow.com/questions/277772/avoid-trailing-zeroes-in-printf
*
* Modifies value "d" by truncating it to precision "n", writing the result
* into string "s".
*
* ----------------------------------------------------------------------------
*/
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
nDecimals(
char *s,
double d,
int n)
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
{
int sz;
double d2;
// Allow for negative.
d2 = (d >= 0) ? d : -d;
sz = (d >= 0) ? 0 : 1;
// Add one for each whole digit (0.xx special case).
if (d2 < 1) sz++;
while (d2 >= 1)
{
d2 /= 10.0;
sz++;
}
// Adjust for decimal point and fractionals.
sz += 1 + n;
// Create formatted string then use it.
sprintf(s, "%*.*f", sz, n, d);
}
/*
* ----------------------------------------------------------------------------
*
* morphNumericString() ---
*
* Code taken from discussion at
* https://stackoverflow.com/questions/277772/avoid-trailing-zeroes-in-printf
*
* Remove trailing zeros from the number represented in string "s".
* Assume a precision (defined as precision after the decimal point) of "n".
*
* ----------------------------------------------------------------------------
*/
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
morphNumericString(
char *s,
int n)
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
{
char *p;
int count;
p = strchr (s, '.'); // Find decimal point, if any.
if (p != NULL)
{
count = n; // Adjust for more or less decimals.
while (count >= 0) // Maximum decimals allowed.
{
count--;
if (*p == '\0') // If there's less than desired.
break;
p++; // Next character.
}
*p-- = '\0'; // Truncate string.
while (*p == '0') // Remove trailing zeros.
*p-- = '\0';
if (*p == '.') // If all decimals were zeros, remove ".".
*p = '\0';
}
}
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
/*
* ----------------------------------------------------------------------------
*
* esSIvalue --
*
* Print an output in appropriate SI units used by SPICE. e.g., 1.0e-6
* will be printed as "1.0u"; 1.0e-12 will be printed as "1.0p", etc.
*
* Return value:
* None.
*
* Side effects:
* Generates output to stream "file".
*
* ----------------------------------------------------------------------------
*/
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
esSIvalue(
FILE *file,
float value)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
char vstr[32];
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
char suffix = '\0';
float avalue;
avalue = fabsf(value);
if (avalue < 1.0E-18)
{
/* Do nothing---value is probably zero */
}
Minor adjustment to output from esSIvalue() to avoid round-off error causing the SI suffix not to be the expected one. Adjusted the bounds where each SI suffix is used to keep output in the range of 0.1 to 100, although the boundary is very subjective. Made a correction to the extresist code to scan through all device records for a tile type. However, (1) there are cases being missed, and (2) this should not be necessary since all device types should be able to be known exactly from the contents of the .sim file. Needs more work.
2023-03-23 20:38:31 +01:00
else if (avalue < 0.9999E-13)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
/* NOTE: ngspice does not support "a" for "atto" */
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
suffix = 'f';
value *= 1.0E15;
}
Minor adjustment to output from esSIvalue() to avoid round-off error causing the SI suffix not to be the expected one. Adjusted the bounds where each SI suffix is used to keep output in the range of 0.1 to 100, although the boundary is very subjective. Made a correction to the extresist code to scan through all device records for a tile type. However, (1) there are cases being missed, and (2) this should not be necessary since all device types should be able to be known exactly from the contents of the .sim file. Needs more work.
2023-03-23 20:38:31 +01:00
else if (avalue < 1.0001E-10)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
suffix = 'p';
value *= 1.0E12;
}
Minor adjustment to output from esSIvalue() to avoid round-off error causing the SI suffix not to be the expected one. Adjusted the bounds where each SI suffix is used to keep output in the range of 0.1 to 100, although the boundary is very subjective. Made a correction to the extresist code to scan through all device records for a tile type. However, (1) there are cases being missed, and (2) this should not be necessary since all device types should be able to be known exactly from the contents of the .sim file. Needs more work.
2023-03-23 20:38:31 +01:00
else if (avalue < 1.0001E-7)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
suffix = 'n';
value *= 1.0E9;
}
Minor adjustment to output from esSIvalue() to avoid round-off error causing the SI suffix not to be the expected one. Adjusted the bounds where each SI suffix is used to keep output in the range of 0.1 to 100, although the boundary is very subjective. Made a correction to the extresist code to scan through all device records for a tile type. However, (1) there are cases being missed, and (2) this should not be necessary since all device types should be able to be known exactly from the contents of the .sim file. Needs more work.
2023-03-23 20:38:31 +01:00
else if (avalue < 1.0001E-4)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
suffix = 'u';
value *= 1.0E6;
}
Minor adjustment to output from esSIvalue() to avoid round-off error causing the SI suffix not to be the expected one. Adjusted the bounds where each SI suffix is used to keep output in the range of 0.1 to 100, although the boundary is very subjective. Made a correction to the extresist code to scan through all device records for a tile type. However, (1) there are cases being missed, and (2) this should not be necessary since all device types should be able to be known exactly from the contents of the .sim file. Needs more work.
2023-03-23 20:38:31 +01:00
else if (avalue < 1.0001E-2)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
suffix = 'm';
value *= 1.0E3;
}
Minor adjustment to output from esSIvalue() to avoid round-off error causing the SI suffix not to be the expected one. Adjusted the bounds where each SI suffix is used to keep output in the range of 0.1 to 100, although the boundary is very subjective. Made a correction to the extresist code to scan through all device records for a tile type. However, (1) there are cases being missed, and (2) this should not be necessary since all device types should be able to be known exactly from the contents of the .sim file. Needs more work.
2023-03-23 20:38:31 +01:00
else if (avalue > 0.9999E9)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
suffix = 'G';
value /= 1.0E9;
}
Minor adjustment to output from esSIvalue() to avoid round-off error causing the SI suffix not to be the expected one. Adjusted the bounds where each SI suffix is used to keep output in the range of 0.1 to 100, although the boundary is very subjective. Made a correction to the extresist code to scan through all device records for a tile type. However, (1) there are cases being missed, and (2) this should not be necessary since all device types should be able to be known exactly from the contents of the .sim file. Needs more work.
2023-03-23 20:38:31 +01:00
else if (avalue > 0.9999E3)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
suffix = 'k';
value /= 1.0E3;
}
Corrected ext2spice to avoid incorrect precision when writing out values. The "%.3g" format was changed to "%.*g" with a variable precision that accounts for the actual precision required by the manufacturing grid.
2023-11-05 17:00:30 +01:00
/* Note that "%g" is preferred because it produces more readable
* output. However, it changes the definition of the precision
* from significant digits after the radix to total significant
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
* digits. Using a solution provided in StackOverflow (see above).
Corrected ext2spice to avoid incorrect precision when writing out values. The "%.3g" format was changed to "%.*g" with a variable precision that accounts for the actual precision required by the manufacturing grid.
2023-11-05 17:00:30 +01:00
*/
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
nDecimals(vstr, (double)value, 5);
morphNumericString(vstr, 5);
Corrected ext2spice to avoid incorrect precision when writing out values. The "%.3g" format was changed to "%.*g" with a variable precision that accounts for the actual precision required by the manufacturing grid.
2023-11-05 17:00:30 +01:00
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
if (suffix == '\0')
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
fprintf(file, "%s", vstr);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
else
Modified the code in ext2spice.c that prints parameter values; the former code attempted to determine the precision and generate an output without unnecessary trailing zeros. Unfortunately there were counterexamples that fail to be formatted correctly, as found by Mark Martin, and which generate output that has roundoff error. Reimplemented the method using code found on StackOverflow, which appears to solve the problem more robustly.
2025-01-06 17:49:29 +01:00
fprintf(file, "%s%c", vstr, suffix);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* spcnAP, spcnAPHier --
*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Output the area perimeter of the node with type type if it has not
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* been visited.
* The spcnAPHier version outputs the area and perimeter only within the
* local subcell with hierarchical name hierName.
*
* Return:
* 0 on success, 1 on error
*
* Side effects:
* Set the visited flags so that the node A/P will not be output multiple
* times
*
* ----------------------------------------------------------------------------
*/
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
int
spcnAP(
DevTerm *dterm,
EFNode *node,
int resClass,
float scale,
char *asterm,
char *psterm,
float m,
FILE *outf,
int w)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
char afmt[15], pfmt[15];
float dsc;
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
int area, perim;
char *cptr;
bool haveAttrs = FALSE;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if ((node == NULL) || (node->efnode_client == (ClientData)NULL))
{
TxError("spcnAP: major internal inconsistency\n");
return 1;
}
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
if (asterm) snprintf(afmt, sizeof afmt, " %s=", asterm);
if (psterm) snprintf(pfmt, sizeof pfmt, " %s=", psterm);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
if (!esDistrJunct || w == -1) goto newFmt;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Modified the LEF VIAGEN reading such that it can accomodate a different order of layers than specified in the LEF/DEF spec. It is not clear whether this is common practice, or a bug in the tool that produced the DEF file that prompted this change. NOTE: The "grow" function applied in this case should be replaced by the actual GDS input rule sequence, that includes the grow and shrink merge. Otherwise, vias read from DEF files do not match the layout from those read from GDS, even though the mask layers represented by the layouts are the same.
2020-03-03 23:13:37 +01:00
if (((nodeClient*)node->efnode_client)->m_w.widths != NULL)
dsc = w / ((nodeClient*)node->efnode_client)->m_w.widths[resClass];
else
{
TxError("Device missing records for source/drain area/perim.\n");
dsc = w;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esScale < 0)
{
if (asterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(afmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-12 * node->efnode_pa[resClass].pa_area
* scale * scale * dsc);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (psterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(pfmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-6 * node->efnode_pa[resClass].pa_perim * scale * dsc);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
if (asterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(afmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-12 * ((float)node->efnode_pa[resClass].pa_area
* scale * scale) * esScale * esScale * dsc);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (psterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(pfmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-6 * ((float)node->efnode_pa[resClass].pa_perim
* scale) * esScale * dsc);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
return 0;
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
newFmt:
/* New format introduced 2/15/2023: Area and perimeter of each terminal
* are maintained in the terminal attributes for "fet" or "device mosfet"
* type devices (otherwise, for subcircuit device types, this routine is
* not used and the same values are found in the device parameters).
*
* Values are the last two values in a comma-separated list in
* dterm_attrs. If not found, then default to the
*/
cptr = (dterm) ? dterm->dterm_attrs : NULL;
while (cptr)
{
if (*cptr == ',') cptr++;
if (sscanf(cptr, "%d,%d", &area, &perim) != 2)
cptr = strchr(cptr, ',');
else
{
haveAttrs = TRUE;
*cptr = '\0';
break;
}
}
if (!haveAttrs)
{
area = node->efnode_pa[resClass].pa_area;
perim = node->efnode_pa[resClass].pa_perim;
if (resClass == NO_RESCLASS ||
beenVisited((nodeClient *)node->efnode_client, resClass))
scale = 0;
else
markVisited((nodeClient *)node->efnode_client, resClass);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esScale < 0)
{
if (asterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(afmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-12 * area * scale * scale / m);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (psterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(pfmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-6 * perim * scale / m);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
if (asterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(afmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-12 * ((float)area * scale * scale) * esScale * esScale);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (psterm)
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(pfmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-6 * ((float)perim * scale) * esScale);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
return 0;
}
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
int
spcnAPHier(
DevTerm *dterm,
HierName *hierName,
int resClass,
float scale,
char *asterm,
char *psterm,
float m,
FILE *outf)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
EFNode *node = dterm->dterm_node;
nodeClientHier *nc;
char afmt[15], pfmt[15];
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
int area, perim;
char *cptr;
bool haveAttrs = FALSE;
sprintf(afmt," %s=", asterm);
sprintf(pfmt," %s=", psterm);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (node->efnode_client == (ClientData) NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
initNodeClientHier(node);
nc = (nodeClientHier *)node->efnode_client;
if (nc->lastPrefix != hierName)
{
clearVisited(nc);
nc->lastPrefix = hierName;
}
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
/* Check for area and perim values in dterm_attrs */
cptr = dterm->dterm_attrs;
while (cptr)
{
if (*cptr == ',') cptr++;
if (sscanf(cptr, "%d,%d", &area, &perim) != 2)
cptr = strchr(cptr, ',');
else
{
haveAttrs = TRUE;
*cptr = '\0';
break;
}
}
if (!haveAttrs)
{
area = node->efnode_pa[resClass].pa_area;
perim = node->efnode_pa[resClass].pa_perim;
if (resClass == NO_RESCLASS ||
beenVisited((nodeClient *)node->efnode_client, resClass))
scale = 0;
else
markVisited((nodeClient *)node->efnode_client, resClass);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esScale < 0)
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(afmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-12 * area * scale * scale / m);
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(pfmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-6 * perim * scale / m);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
else
{
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(afmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-12 * ((float)area * scale) * esScale * esScale);
Applied patch from Donn that converts strcpy() calls in ext2spice to "safer" strncpy() calls to prevent string buffer overflow. Also: Reimplemented the loop in the GDS write routine that counts ports and then outputs them in order. It was possible to hang magic for a long time by giving a port a very, very large index number. The new implementation uses qsort() to sort the ports by index, which is obviously much more efficient for the worst case (and efficient enough for all normal cases).
2023-03-22 03:04:30 +01:00
fputs(pfmt, outf);
Important update: Reworked the extraction method to properly isolate the terminal areas of a device (e.g., source and drain) and calculate their area and perimeter individually for the device (in addition to the traditional method of calculating area and perimeter of each resistance class for the entire node). Also: Reworked the SPICE syntax output to generate SI values in the range 1-1000 with the appropriate suffix (e.g., "20u") instead of defaulting to "u" for lengths and "p" for areas. This prevents it from producing weird units like "150000u" when a process definition already includes a scalefactor. Reworked the "extresist" code to use the device terminal area and perimeter. This fixes an error in which "extresist" would lose these values and "ext2spice" with option "extresist on" would generate a new netlist output with zero terminal areas and perimeters.
2023-02-16 17:59:13 +01:00
esSIvalue(outf, 1.0E-6 * ((float)perim * scale) * esScale);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* spcdevOutNode --
*
* Output the name of the node whose hierarchical prefix down to this
* point is 'prefix' and whose name from the end of suffix down to the
* leaves is 'suffix', just as in the arguments to EFHNConcat().
*
*
* Results:
* Return number of characters printed on success, 0 on error.
*
* Side effects:
* Writes to the file 'outf'.
* Sets the efnode_client field as described above.
*
* ----------------------------------------------------------------------------
*/
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spcdevOutNode(
extflat: spcdevOutNode() constify API
2025-07-23 21:10:03 +02:00
const HierName *prefix,
const HierName *suffix,
const char *name,
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
FILE *outf)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
HashEntry *he;
EFNodeName *nn;
extflat: spcdevOutNode() constify API
2025-07-23 21:10:03 +02:00
const char *nname;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
he = EFHNConcatLook(prefix, suffix, name);
if (he == NULL)
{
fprintf(outf, " errGnd!");
return 0;
}
nn = (EFNodeName *) HashGetValue(he);
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
nname = nodeSpiceName(nn->efnn_node->efnode_name->efnn_hier, NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(outf, " %s", nname);
Corrected problem in which using "ext2spice -d" (distributed junctions) causes magic to crash.
2019-10-14 19:49:41 +02:00
Got to the bottom of why area and perimeter values are no longer output for transistors. The problem came from a change made to fix an issue with capacitors marked as floating nodes because some nodes are not output as source or drain. But those nodes are output before the parameters, so when generating parameter output, all nodes appear to have already been output. Solution: Specify an additional bit in the "visited" mask for the node having been output that is separate from the mask for resist classes used by the code that writes parameter values, and use that bit as a test for whether the node is connected to some device (not necessarily a FET source or drain).
2021-12-20 19:12:49 +01:00
/* Mark node as visited (set bit one higher than number of resist classes) */
if (esDistrJunct)
update_w(efNumResistClasses, 1, nn->efnn_node);
else
markVisited((nodeClientHier *)nn->efnn_node->efnode_client,
efNumResistClasses);
Corrected problem in which using "ext2spice -d" (distributed junctions) causes magic to crash.
2019-10-14 19:49:41 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return (1 + strlen(nname));
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* spccapVisit --
*
* Procedure to output a single capacitor to the .spice file.
* Called by EFVisitCaps().
*
* Results:
* Returns 0 always.
*
* Side effects:
* Writes to the file esSpiceF. Increments esCapNum.
*
* Format of a .spice cap line:
*
* C%d node1 node2 cap
*
* where
* node1, node2 are the terminals of the capacitor
* cap is the capacitance in femtofarads (NOT attofarads).
*
* ----------------------------------------------------------------------------
*/
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitcaps_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spccapVisit(
HierName *hierName1,
HierName *hierName2,
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
double cap,
ClientData cdata) /* unused */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
cap = cap / 1000;
if (cap <= EFCapThreshold)
return 0;
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
fprintf(esSpiceF, "C%d %s %s ", esCapNum++, nodeSpiceName(hierName1, NULL),
nodeSpiceName(hierName2, NULL));
esSIvalue(esSpiceF, 1.0E-15 * cap);
fprintf(esSpiceF, "\n");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* spcresistVisit --
*
* Procedure to output a single resistor to the .spice file.
* Called by EFVisitResists().
*
* Results:
* Returns 0 always.
*
* Side effects:
* Writes to the file esSpiceF. Increments esResNum.
*
* Format of a .spice resistor line:
*
* R%d node1 node2 res
*
* where
* node1, node2 are the terminals of the resistor
* res is the resistance in ohms (NOT milliohms)
*
*
* ----------------------------------------------------------------------------
*/
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
extflat: integrate cb_extflat_visitresists_t for EFVisitResists() Additionally the HierName's hierName1 and hierName2 arguments have been made 'const' to help convey the receiver can not modify the referenced data passed.
2025-07-20 17:09:08 +02:00
/* ARGSUSED */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitresists_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spcresistVisit(
extflat: integrate cb_extflat_visitresists_t for EFVisitResists() Additionally the HierName's hierName1 and hierName2 arguments have been made 'const' to help convey the receiver can not modify the referenced data passed.
2025-07-20 17:09:08 +02:00
const HierName *hierName1,
const HierName *hierName2,
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
float res,
ClientData cdata) /* unused */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Corrected "ext2spice" to mark nodes on resistors from "extresist" so that they do not get marked as "FLOATING" nodes when writing out the parasitic capacitances at the end.
2023-02-28 19:11:32 +01:00
HashEntry *he;
EFNodeName *nn;
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
fprintf(esSpiceF, "R%d %s %s %g\n", esResNum++, nodeSpiceName(hierName1, NULL),
nodeSpiceName(hierName2, NULL), res / 1000.);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Corrected "ext2spice" to mark nodes on resistors from "extresist" so that they do not get marked as "FLOATING" nodes when writing out the parasitic capacitances at the end.
2023-02-28 19:11:32 +01:00
/* Mark nodes as visited so that associated capacitances won't be marked
* as "floating". This is inefficient since nodeSpiceName() already does
* a hash lookup of the EFNodeName. Could be improved, but is not a big
* performance issue.
*/
he = EFHNLook(hierName1, (char *)NULL, "nodeName");
if (he != NULL)
{
nn = (EFNodeName *) HashGetValue(he);
/* Mark node as visited (set bit one higher than number of resist classes) */
if (esDistrJunct)
update_w(efNumResistClasses, 1, nn->efnn_node);
else
markVisited((nodeClientHier *)nn->efnn_node->efnode_client,
efNumResistClasses);
}
he = EFHNLook(hierName2, (char *)NULL, "nodeName");
if (he != NULL)
{
nn = (EFNodeName *) HashGetValue(he);
/* Mark node as visited (set bit one higher than number of resist classes) */
if (esDistrJunct)
update_w(efNumResistClasses, 1, nn->efnn_node);
else
markVisited((nodeClientHier *)nn->efnn_node->efnode_client,
efNumResistClasses);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* spcsubVisit --
*
* Routine to find the node that connects to substrate. Copy the
* string name of this node into "resstr" to be returned to the
* caller.
*
* Results:
* Return 1 if the substrate node has been found, to stop the search.
* Otherwise return 0 to keep the search going.
*
* ----------------------------------------------------------------------------
*/
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spcsubVisit(
EFNode *node,
int res, // Unused
double cap, // Unused
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
ClientData cdata)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
char **resstr = (char **)CD2PTR(cdata);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
HierName *hierName;
extflat: nodeSpiceName() consumer
2025-07-23 22:06:12 +02:00
const char *nsn;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
Corrected the use of EF_SUBS_NODE to distinguish between device substrate (bulk terminal) and global substrate. Otherwise, the routine in ext2hier.c that finds the substrate node will find the first device bulk connection, not the default substrate.
2021-12-30 19:27:42 +01:00
if (node->efnode_flags & EF_GLOB_SUBS_NODE)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
hierName = (HierName *) node->efnode_name->efnn_hier;
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
nsn = nodeSpiceName(hierName, NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*resstr = StrDup((char **)NULL, nsn);
return 1;
}
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* spcnodeVisit --
*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Procedure to output a single node to the .spice file along with its
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* attributes and its dictionary (if present). Called by EFVisitNodes().
*
* Results:
* Returns 0 always.
*
* Side effects:
* Writes to the files esSpiceF
*
* ----------------------------------------------------------------------------
*/
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitnodes_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
spcnodeVisit(
EFNode *node,
int res,
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
double cap,
ClientData cdata) /* unused */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
HierName *hierName;
bool isConnected = FALSE;
ext2spice.c: printf style constify fmt call-site
2024-10-20 14:46:02 +02:00
const char *fmt;
extflat: nodeSpiceName() consumer
2025-07-23 22:06:12 +02:00
const char *nsn;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
EFAttr *ap;
if (node->efnode_client)
{
isConnected = (esDistrJunct) ?
(((nodeClient *)node->efnode_client)->m_w.widths != NULL) :
Got to the bottom of why area and perimeter values are no longer output for transistors. The problem came from a change made to fix an issue with capacitors marked as floating nodes because some nodes are not output as source or drain. But those nodes are output before the parameters, so when generating parameter output, all nodes appear to have already been output. Solution: Specify an additional bit in the "visited" mask for the node having been output that is separate from the mask for resist classes used by the code that writes parameter values, and use that bit as a test for whether the node is connected to some device (not necessarily a FET source or drain).
2021-12-20 19:12:49 +01:00
(!TTMaskHasType(&((nodeClient *)node->efnode_client)->m_w.visitMask,
efNumResistClasses));
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (!isConnected && esDevNodesOnly)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
/* Don't mark known ports as "FLOATING" nodes */
if (!isConnected && node->efnode_flags & EF_PORT) isConnected = TRUE;
hierName = (HierName *) node->efnode_name->efnn_hier;
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
nsn = nodeSpiceName(hierName, NULL);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
12 x warning: suggest parentheses around '&&' within '||' The '&' has higher precedence, so the expression of the '&' side receive extra parentheses. ResMakeRes.c:671:37: warning: suggest parentheses around '&&' within '||' ResMakeRes.c:942:32: warning: suggest parentheses around '&&' within '||' ResSimple.c:201:70: warning: suggest parentheses around '&&' within '||' ResRex.c:1036:28: warning: suggest parentheses around '&&' within '||' ResRex.c:1038:50: warning: suggest parentheses around '&&' within '||' ext2sim.c:1341:47: warning: suggest parentheses around '&&' within '||' ext2spice.c:172:44: warning: suggest parentheses around '&&' within '||' ext2spice.c:173:49: warning: suggest parentheses around '&&' within '||' ext2spice.c:3655:50: warning: suggest parentheses around '&&' within '||' ext2hier.c:1412:50: warning: suggest parentheses around '&&' within '||' gcrDebug.c:638:55: warning: suggest parentheses around '&&' within '||' selEnum.c:323:28: warning: suggest parentheses around '&&' within '||' GCC14 -Wall cleanup series [-Wparentheses]
2024-10-04 18:24:47 +02:00
if (esFormat == SPICE2 || (esFormat == HSPICE && strncmp(nsn, "z@", 2)==0 )) {
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
static char ntmp[MAX_STR_SIZE];
EFHNSprintf(ntmp, hierName);
Changed the end-of-line comment character from ';' to '$' when formatting a SPICE netlist for NGSPICE. This is the documented character as described in the NGSPICE User Manual, section 2.2.4 End-of-line comments
2020-06-24 21:22:04 +02:00
if (esFormat == NGSPICE) fprintf(esSpiceF, "$ ");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, "** %s == %s\n", ntmp, nsn);
}
cap = cap / 1000;
if (cap > EFCapThreshold)
{
Changed the behavior of capacitance value output in ext2spice and ext2sim to make use of the new esSIvalue() routine, so that it no longer depends on a preformatted string. Corrected an issue where the esSIvalue() routine would output "a" for "atto" which is not supported by ngspice.
2023-03-13 18:21:34 +01:00
fprintf(esSpiceF, "C%d %s %s ", esCapNum++, nsn, esSpiceCapNode);
esSIvalue(esSpiceF, 1.0E-15 * cap);
if (!isConnected)
{
if (esFormat == NGSPICE) fprintf(esSpiceF, " $");
fprintf(esSpiceF, " **FLOATING");
}
fprintf(esSpiceF, "\n");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
if (node->efnode_attrs && !esNoAttrs)
{
Changed the end-of-line comment character from ';' to '$' when formatting a SPICE netlist for NGSPICE. This is the documented character as described in the NGSPICE User Manual, section 2.2.4 End-of-line comments
2020-06-24 21:22:04 +02:00
if (esFormat == NGSPICE) fprintf(esSpiceF, " $ ");
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
fprintf(esSpiceF, "**nodeattr %s :",nsn );
for (fmt = " %s", ap = node->efnode_attrs; ap; ap = ap->efa_next)
{
fprintf(esSpiceF, fmt, ap->efa_text);
fmt = ",%s";
}
putc('\n', esSpiceF);
}
return 0;
}
/* a debugging procedure */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitnodes_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
nodeVisitDebug(
EFNode *node,
int res,
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
double cap,
ClientData cdata) /* unused */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
HierName *hierName;
extflat: nodeSpiceName() consumer
2025-07-23 22:06:12 +02:00
const char *nsn;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
hierName = (HierName *) node->efnode_name->efnn_hier;
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
nsn = nodeSpiceName(hierName, NULL);
ext2spice/ext2spice.c: %x => (intmax_t) %lx Wrong type of arguments to formatting function Looks like SPICE comment, change maintains hex without 0x prefix in portable way. Copilot Autofix rejected: TxError("** %s (%p)\n", nsn, node); CodeQL: https://github.com/dlmiles/magic/security/code-scanning/70
2024-09-30 00:00:00 +02:00
TxError("** %s (%lx)\n", nsn, (intmax_t) node);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
printf("\t client.name=%s, client.m_w=%p\n",
((nodeClient *)node->efnode_client)->spiceNodeName,
gcc11 -Wall -Wpedantic cleanup
2025-01-06 17:12:11 +01:00
(void *)((nodeClient *)node->efnode_client)->m_w.widths);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
/*
* ----------------------------------------------------------------------------
*
* nodeSpiceName --
* Find the real spice name for the node with hierarchical name hname.
* SPICE2 ==> numeric
* SPICE3 ==> full magic path
* HSPICE ==> less than 15 characters long
*
* Results:
* Returns the spice node name.
*
* Side effects:
* Allocates nodeClients for the node.
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
* Returns the node in the "rnode" pointer, if non-NULL.
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
* ----------------------------------------------------------------------------
*/
static char esTempName[MAX_STR_SIZE];
extflat: nodeSpiceName() constify API
2025-07-21 23:46:03 +02:00
const char *
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
nodeSpiceName(
extflat: nodeSpiceName() constify API
2025-07-21 23:46:03 +02:00
const HierName *hname,
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
EFNode **rnode)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
EFNodeName *nn;
HashEntry *he;
EFNode *node;
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
if (rnode) *rnode = (EFNode *)NULL;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
he = EFHNLook(hname, (char *) NULL, "nodeName");
if ( he == NULL )
return "errGnd!";
nn = (EFNodeName *) HashGetValue(he);
node = nn->efnn_node;
Made some changes to the way that ext2spice generates subcircuit ports, to avoid creating ports for node names that are redundant. It would probably be better to avoid creating the redundant node names in the first place; however, I am less certain why these are generated. The incorrect additional ports all have hierarchical names in the cell, which is a sign that they are incorrect, as the cell itself should not have any parents. The level of certainty about this fix is definitely not 100%, but it was tested on a hierarchical analog design, and setting levels of parasitic caps caused new nodes to appear in subcircuits and in no cases did information appear to be lost.
2019-10-29 14:44:28 +01:00
if (rnode) *rnode = node;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if ( (nodeClient *) (node->efnode_client) == NULL ) {
initNodeClient(node);
goto makeName;
} else if ( ((nodeClient *) (node->efnode_client))->spiceNodeName == NULL)
goto makeName;
else goto retName;
makeName:
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if ( esFormat == SPICE2 )
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
sprintf(esTempName, "%d", esNodeNum++);
else {
EFHNSprintf(esTempName, node->efnode_name->efnn_hier);
if ( esFormat == HSPICE ) /* more processing */
nodeHspiceName(esTempName);
}
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
((nodeClient *) (node->efnode_client))->spiceNodeName =
StrDup(NULL, esTempName);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
retName:
return ((nodeClient *) (node->efnode_client))->spiceNodeName;
}
/*
* ----------------------------------------------------------------------------
*
* EFHNSprintf --
*
* Create a hierarchical node name.
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
* The flags in EFOutputFlags control whether global (!) or local (#)
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* suffixes are to be trimmed. Also substitutes \. with \@ if the
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
* format is hspice.
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
* Results:
* None.
*
* Side effects:
* Changes the area pointed to by str
*
* ----------------------------------------------------------------------------
*/
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
EFHNSprintf(
char *str,
HierName *hierName)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Added substitution for commas in SPICE output, because commas in instance names and node names are basically fatal to running ngspice.
2019-10-15 03:46:07 +02:00
bool trimGlob, trimLocal, convertComma, convertEqual, convertBrackets;
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
char *s, *cp, c;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
char *efHNSprintfPrefix(HierName *, char *);
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
s = str;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (hierName->hn_parent) str = efHNSprintfPrefix(hierName->hn_parent, str);
Correction to prior commit to remove collision between defined flag bits for name trimming and short handling method in the ext2spice options.
2021-07-11 03:27:19 +02:00
if (EFOutputFlags & EF_TRIM_MASK)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
cp = hierName->hn_name;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
trimGlob = (EFOutputFlags & EF_TRIMGLOB);
trimLocal = (EFOutputFlags & EF_TRIMLOCAL);
convertComma = (EFOutputFlags & EF_CONVERTCOMMA);
convertEqual = (EFOutputFlags & EF_CONVERTEQUAL);
convertBrackets = (EFOutputFlags & EF_CONVERTBRACKETS);
35 x warning: suggest parentheses around assignment used as truth value PlowRules1.c:439:9: warning: suggest parentheses around assignment used as truth value PlowTech.c:645:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:652:17: warning: suggest parentheses around assignment used as truth value PlowTech.c:1019:17: warning: suggest parentheses around assignment used as truth value ResReadSim.c:270:13: warning: suggest parentheses around assignment used as truth value ResReadSim.c:871:9: warning: suggest parentheses around assignment used as truth value ResRex.c:1840:17: warning: suggest parentheses around assignment used as truth value getrect.c:72:9: warning: suggest parentheses around assignment used as truth value getrect.c:79:9: warning: suggest parentheses around assignment used as truth value getrect.c:86:9: warning: suggest parentheses around assignment used as truth value getrect.c:93:9: warning: suggest parentheses around assignment used as truth value hash.c:732:16: warning: suggest parentheses around assignment used as truth value heap.c:328:17: warning: suggest parentheses around assignment used as truth value heap.c:344:17: warning: suggest parentheses around assignment used as truth value netlist.c:323:17: warning: suggest parentheses around assignment used as truth value niceabort.c:121:9: warning: suggest parentheses around assignment used as truth value path.c:1102:12: warning: suggest parentheses around assignment used as truth value pathvisit.c:245:13: warning: suggest parentheses around assignment used as truth value pathvisit.c:295:17: warning: suggest parentheses around assignment used as truth value tech.c:656:17: warning: suggest parentheses around assignment used as truth value ext2spice.c:1591:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1622:16: warning: suggest parentheses around assignment used as truth value ext2spice.c:1813:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:1862:12: warning: suggest parentheses around assignment used as truth value ext2spice.c:3808:16: warning: suggest parentheses around assignment used as truth value CalmaRdio.c:437:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:396:9: warning: suggest parentheses around assignment used as truth value CalmaWrite.c:1772:29: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:372:9: warning: suggest parentheses around assignment used as truth value CalmaWriteZ.c:1608:29: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:209:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:214:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:220:9: warning: suggest parentheses around assignment used as truth value CIFrdtech.c:226:9: warning: suggest parentheses around assignment used as truth value CIFrdutils.c:1258:12: warning: suggest parentheses around assignment used as truth value GCC14 -Wall cleanup series [-Wparentheses]
2024-10-04 18:21:15 +02:00
while ((c = *cp++))
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
switch (c)
{
case '!': if (!trimGlob) *str++ = c; break;
case '.': *str++ = (esFormat == HSPICE)?'@':'.'; break;
case '=': if (convertEqual) *str++ = ':'; break;
case ',': if (convertComma) *str++ = '|'; break;
case '[': *str++ = (convertBrackets) ? '_' : '['; break;
case ']': *str++ = (convertBrackets) ? '_' : ']'; break;
case '#': if (trimLocal) break; // else fall through
default: *str++ = c; break;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
*str++ = '\0';
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
Revert "Modified the .ext file reading and the .spice file writing so that" This reverts commit 46baae0ce63fa3d140762955ef975f963713d3be. Reverting the last commit, as it does not work completely the way it is supposed to, and will most likely have to be done in a different way.
2021-02-24 18:35:06 +01:00
else strcpy(str, hierName->hn_name);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
return 0;
}
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
char *
efHNSprintfPrefix(
HierName *hierName,
char *str)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
char *cp, c;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
bool convertEqual = (EFOutputFlags & EF_CONVERTEQUAL) ? TRUE : FALSE;
bool convertComma = (EFOutputFlags & EF_CONVERTCOMMA) ? TRUE : FALSE;
bool convertBrackets = (EFOutputFlags & EF_CONVERTBRACKETS) ? TRUE : FALSE;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (hierName->hn_parent)
str = efHNSprintfPrefix(hierName->hn_parent, str);
cp = hierName->hn_name;
while (1) {
if (convertEqual && (*cp == '='))
*str = ':';
Added substitution for commas in SPICE output, because commas in instance names and node names are basically fatal to running ngspice.
2019-10-15 03:46:07 +02:00
else if (convertBrackets && ((*cp == '[') || (*cp == ']')))
*str = '_';
else if (*cp == ',')
{
if (convertComma) *str = '|';
else str--;
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
*str = *cp;
if (!(*str)) break;
str++;
cp++;
}
*str = '/';
return ++str;
}
/*
* ----------------------------------------------------------------------------
*
* nodeHspiceName --
*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* Convert the hierarchical node name used in Berkeley spice
Corrected hierarchical SPICE extraction to reset the subcircuit name-to-number mapping used for the HSPICE format between subcircuits. Otherwise, subcircuits with the same instance ID remain in the table and may cause nodes to be output with a name that collides with other names in the same subcircuit. This only affects output in HSPICE format.
2018-12-12 23:23:35 +01:00
* to a name understood by hspice and hopefully by the user.
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
* Results:
Corrected hierarchical SPICE extraction to reset the subcircuit name-to-number mapping used for the HSPICE format between subcircuits. Otherwise, subcircuits with the same instance ID remain in the table and may cause nodes to be output with a name that collides with other names in the same subcircuit. This only affects output in HSPICE format.
2018-12-12 23:23:35 +01:00
* A somewhat meaningful node name
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*
* Side effects:
* Mucks with the hash table above.
*
* ----------------------------------------------------------------------------
*/
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
int
nodeHspiceName(
char *s)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
char *p, *sf;
int l, snum = -1;
HashEntry *he;
static char map[MAX_STR_SIZE];
/*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* find the suffix
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*/
l = strlen(s);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
for (p = s + l; (p > s) && *p != '/'; p--);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (p == s)
{
strcpy(map, s);
goto topLevel;
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
/*
* break it into prefix '/0' suffix
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*/
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (*p == '/')
*p = 0;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
sf = p + 1;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
/*
* look up prefix in the hash table and create it if doesn't exist
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
*/
if ((he = HashLookOnly(&subcktNameTable, s)) == NULL)
{
snum = esSbckNum++;
he = HashFind(&subcktNameTable, s);
HashSetValue(he, (ClientData)(pointertype) snum);
#ifndef UNSORTED_SUBCKT
DQPushRear(&subcktNameQueue, he);
#endif
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
snum = (spointertype) HashGetValue(he);
sprintf(map, "x%d/%s", snum, sf);
topLevel:
strcpy(s, map);
if (strlen(s) > 15)
{
/* still hspice does not get it */
sprintf(s, "z@%d", esNodeNum++);
if (strlen(s) > 15)
{
/* screw it: hspice will not work */
TxError("Error: too many nodes in this circuit to be "
"output as names\n");
TxError(" use spice2 format or call and complain "
"to Meta software about their stupid parser\n");
#ifdef MAGIC_WRAPPER
return TCL_ERROR;
#else
exit(1);
#endif
}
}
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* printSubcktDict --
*
* Print out the hspice subcircuit dictionary. Ideally this should go to a
* pa0 file but uncfortunately hspice crashes if the node names contain
* dots so we just append it at the end of the spice file
*
* Results:
* None.
*
* Side effects:
* Writes to the output file
*
* ----------------------------------------------------------------------------
*/
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
int
printSubcktDict(void)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
HashEntry *he;
fprintf(esSpiceF,"\n** hspice subcircuit dictionary\n");
#ifndef UNSORTED_SUBCKT
while ((he = (HashEntry *)DQPopFront(&subcktNameQueue)) != NULL)
#else
HashStartSearch(&hs);
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
while ((he = HashNext(&subcktNameTable, &hs)) != NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
#endif
fprintf(esSpiceF,"* x%"DLONG_PREFIX"d\t%s\n", (dlong) HashGetValue(he), he->h_key.h_name);
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* mkDevMerge --
* Create a new devMerge structure.
*
* Results:
* Obvious
*
* Side effects:
* Allocates memory and sets the fields of the structure.
*
* ----------------------------------------------------------------------------
*/
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
devMerge *
mkDevMerge(
float l,
float w,
EFNode *g,
EFNode *s,
EFNode *d,
EFNode *b,
HierName *hn,
Dev *dev)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
devMerge *fp;
fp = (devMerge *) mallocMagic((unsigned) (sizeof(devMerge)));
fp->l = l; fp->w = w;
fp->g = g; fp->s = s;
fp->d = d; fp->b = b;
fp->dev = dev;
fp->esFMIndex = esFMIndex;
fp->hierName = hn;
fp->next = NULL;
addDevMult(1.0);
return fp;
}
/*
* ----------------------------------------------------------------------------
*
* parallelDevs --
*
* Determine if two devs are in parallel
*
* Results:
* NOT_PARALLEL if not in parallel
* PARALLEL if parallel and an exact match
* ANTIPARALLEL if parallel but reversed source<->drain nodes
*
* Side effects:
* None.
*
* ----------------------------------------------------------------------------
*/
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
parallelDevs(
ext2spice: parallelDevs constify API (const devMerge*)
2025-07-24 15:36:22 +02:00
const devMerge *f1,
const devMerge *f2)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
/* If the devices are not in the same class, then */
/* they cannot be parallel. */
if (f1->dev->dev_class != f2->dev->dev_class)
return NOT_PARALLEL;
/* Can't merge devices with different models */
if (f1->dev->dev_type != f2->dev->dev_type)
return NOT_PARALLEL;
/* Class-dependent action */
switch(f1->dev->dev_class)
{
case DEV_MSUBCKT:
case DEV_MOSFET:
case DEV_FET:
if (f1->b != f2->b) return NOT_PARALLEL;
if ((f1->g == f2->g) && (f1->l == f2->l)
&& (esMergeDevsA || (f1->w == f2->w)))
{
if ((f1->d == f2->d) && (f1->s == f2->s))
return PARALLEL;
else if ((f1->s == f2->d) && (f1->d == f2->s))
return ANTIPARALLEL;
}
break;
case DEV_ASYMMETRIC:
if (f1->b != f2->b) return NOT_PARALLEL;
if ((f1->g == f2->g) && (f1->d == f2->d) && (f1->s == f2->s)
&& (f1->l == f2->l) && (esMergeDevsA || (f1->w == f2->w)))
{
return PARALLEL;
}
break;
/* Capacitors match if top ("gate") and bottom ("source") are */
/* the same. Do not attempt to swap top and bottom, as we do */
/* not know when it is safe to do so. */
case DEV_CAP:
case DEV_CAPREV:
if ((f1->g != f2->g) || (f1->s != f2->s))
return NOT_PARALLEL;
else if (f1->dev->dev_type == esNoModelType)
{
/* Unmodeled capacitor */
if (esMergeDevsA || (f1->dev->dev_cap == f2->dev->dev_cap))
return PARALLEL;
}
else if (esMergeDevsA || ((f1->l == f2->l) && (f1->w == f2->w)))
return PARALLEL;
break;
/* We can't merge resistors because we accumulate capacitance */
/* on the central ("gate") node. Merging the devices would */
/* cause nodes to disappear. */
case DEV_RES:
break;
/* For the remaining devices, it might be possible to merge */
/* if we know that the device model level accepts length and */
/* width parameters. However, at this time, no such */
/* information is passed to the SPICE deck, so we don't try to */
/* merge these devices. */
case DEV_BJT:
case DEV_DIODE:
case DEV_NDIODE:
case DEV_PDIODE:
break;
/* There is no way to merge subcircuit devices */
case DEV_SUBCKT:
Updated the version to go along with the merge of pull request version 8.3.521 that, due to an argument size mismatch, causes device parameters in netlist output to be printed as zero. Also added a small extension to the list of extraction types to include "device veriloga", which has the same syntax as "device subcircuit" but generates a component type "N" (Verilog-A component) in the netlist output.
2025-08-07 17:54:49 +02:00
case DEV_VERILOGA:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_RSUBCKT:
Added extraction device type "csubcircuit", which should have been added a long time ago, since capacitors use a different method for calculating width and length than either transistors or resistors, so subcircuits need a special class designator or else the extraction may calculate the wrong dimensions for device width by totalling the perimeter between the device and terminal types, as it would for a MOSFET.
2018-10-30 21:19:20 +01:00
case DEV_CSUBCKT:
Added a new device type "dsubcircuit" which basically behaves like "csubcircuit" but swaps the first two pins (with the device identifier layer becoming the 2nd pin and the other terminal the first), which is needed for n-type diodes modeled as subcircuits where the subcircuit pin order matches the order of pins for a SPICE n-type diode component. Previously "msubcircuit" was used for this purpose, but will calculate the wrong L and W. While use of L and W for diodes is rare, this device type also works for reversed capacitors (where the bottom or non-identifying layer terminal is in the first pin position of the subcircuit).
2025-10-07 22:45:22 +02:00
case DEV_DSUBCKT:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
Implemented a command option "ext2spice short voltage|resistor|none" that implements a method for handling ports in a subcircuit that have different port names and indexes but are shorted together. "none" is the default and backwards-compatible behavior that merges ports together, which will often cause one of the ports to be optimized out of the netlist. "resistor" will separate the port names with a 0-ohm ideal resistor. "voltage" will separate the port names with a 0-volt voltage source. This should work well for simulation and potentially for LVS, although its impact on LVS has not been fully investigated.
2021-07-11 03:13:24 +02:00
case DEV_VOLT:
break;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
return NOT_PARALLEL;
}
/*
* ----------------------------------------------------------------------------
*
* mergeAttr --
*
* merge two attribute strings
*
* Results:
* The merged strings
*
* Side effects:
* Might allocate and free memory.
*
* ----------------------------------------------------------------------------
*/
void
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
mergeAttr(
char **a1,
char **a2)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
if (*a1 == NULL)
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
*a1 = *a2; /* TODO check strdup() is appropiate due to lifetime ambiguity this creates */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
{
ext2spice: mergeAttr() fix, leak removal and restore original intention
2025-02-24 09:31:42 +01:00
size_t l1 = strlen(*a1);
size_t l2 = strlen(*a2);
char *t = (char *) mallocMagic(l1 + l2 + 1);
strcpy(t, *a1); /* strcpy_advance() */
strcat(t, *a2);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
freeMagic(*a1);
*a1 = t;
}
}
/*
* ----------------------------------------------------------------------------
*
* devMergeVisit --
* Visits each dev throu EFVisitDevs and finds if it is in parallel with
* any previously visited dev.
*
* Results:
* 0 always to keep the caller going.
*
* Side effects:
* Numerous.
*
* ----------------------------------------------------------------------------
*/
extflat: integrate cb_extflat_visitdevs_t for EFVisitDevs()
2025-07-20 01:03:14 +02:00
/* ARGSUSED */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitdevs_t (UNUSED) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
devMergeVisit(
Dev *dev, /* Dev to examine */
HierContext *hc, /* Hierarchical context down to this dev */
float scale, /* Scale transform */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
Transform *trans, /* (unused) */
ClientData cdata) /* (unused) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
DevTerm *gate, *source, *drain;
EFNode *subnode, *snode, *dnode, *gnode;
int pmode, l, w;
bool hS, hD, chS, chD;
extraction: constify struct devMerge.next, devMergeList Feel free to undo variables as necessary in the future as requirements change.
2025-07-24 15:36:22 +02:00
devMerge *fp;
const devMerge *cfp;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
float m;
Additional modification to the devVisit routine in extflat to change from passing the HierName to passing the HierContext (of which HierName is a part) so that more information from HierContext (such as the cell use being visited) can be passed to the callback procedure (largely for diagnostic purposes).
2020-06-01 23:14:22 +02:00
HierName *hierName = hc->hc_hierName;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (esDistrJunct)
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
devDistJunctVisit(dev, hc, scale, trans, cdata);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (dev->dev_nterm < 2)
{
TxError("outPremature\n");
return 0;
}
gate = &dev->dev_terms[0];
if (dev->dev_nterm >= 2)
source = drain = &dev->dev_terms[1];
if (dev->dev_nterm >= 3)
drain = &dev->dev_terms[2];
gnode = SpiceGetNode(hierName, gate->dterm_node->efnode_name->efnn_hier);
if (dev->dev_nterm >= 2)
{
snode = SpiceGetNode(hierName, source->dterm_node->efnode_name->efnn_hier);
dnode = SpiceGetNode(hierName, drain->dterm_node->efnode_name->efnn_hier);
}
if (dev->dev_subsnode)
subnode = spcdevSubstrate(hierName,
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
dev->dev_subsnode->efnode_name->efnn_hier,
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
dev->dev_type, NULL);
else
subnode = NULL;
/* Get length and width of the device */
EFGetLengthAndWidth(dev, &l, &w);
fp = mkDevMerge((float)((float)l * scale), (float)((float)w * scale),
gnode, snode, dnode, subnode, hierName, dev);
hS = extHierSDAttr(source);
hD = extHierSDAttr(drain);
/*
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* run the list of devs. compare the current one with
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* each one in the list. if they fullfill the matching requirements
* merge them only if:
* 1) they have both apf S, D attributes
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
* or
* 2) one of them has aph S, D attributes and they have the same
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
* hierarchical prefix
* If one of them has apf and the other aph print a warning.
*/
for (cfp = devMergeList; cfp != NULL; cfp = cfp->next)
{
if ((pmode = parallelDevs(fp, cfp)) != NOT_PARALLEL)
{
ext2spice.c: cleanup unused localvars -Wunused-variable -Wunused-but-set-variable
2025-07-24 15:36:22 +02:00
#if 0 /* -Wunused-but-set-variable */
Dev *cf = cfp->dev;
DevTerm *cg = &cfp->dev->dev_terms[0];
#endif
DevTerm *cs = &cfp->dev->dev_terms[1];
DevTerm *cd = cs;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (cfp->dev->dev_nterm >= 3)
{
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (pmode == PARALLEL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
cd = &cfp->dev->dev_terms[2];
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
else if (pmode == ANTIPARALLEL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
cs = &cfp->dev->dev_terms[2];
}
chS = extHierSDAttr(cs); chD = extHierSDAttr(cd);
if (!(chS || chD || hS || hD)) /* all flat S, D */
goto mergeThem;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (cfp->hierName != hierName &&
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
((chS && !hS) || (chD && !hD) ||
(!chS && hS) || (!chD && hD)))
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
{
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
efHNSprintfPrefix((cfp->hierName)?cfp->hierName:hierName,
esTempName);
TxError("Warning: conflicting SD attributes of parallel"
" devs in cell: %s\n", esTempName);
break;
}
else if (cfp->hierName == hierName)
{
if (hS && !chS)
{
mergeAttr(&cs->dterm_attrs, &source->dterm_attrs);
}
if (hD && !chD)
{
mergeAttr(&cd->dterm_attrs, &drain->dterm_attrs);
}
}
else /* cfp->hierName != hierName */
break;
mergeThem:
Corrected an issue in which the device merging routine can generate NaN results for devices which are not FETs (specifically, devices that are declared using "msubckt" but are not FETs, although there may have been a related issue with non-FET devices not getting the correct M count), due to the device having zero measured width or length. NOTE: This may need more investigation. If a subcircuit device's method of merging cannot be understood, then such devices should never extract with "merge aggressive", and should always merge conservatively if and only if all parameters match.
2023-05-21 18:13:01 +02:00
/* Default case is to add counts */
m = esFMult[cfp->esFMIndex] + esFMult[fp->esFMIndex];
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
switch(dev->dev_class)
{
case DEV_MSUBCKT:
case DEV_MOSFET:
case DEV_ASYMMETRIC:
case DEV_FET:
Corrected an issue in which the device merging routine can generate NaN results for devices which are not FETs (specifically, devices that are declared using "msubckt" but are not FETs, although there may have been a related issue with non-FET devices not getting the correct M count), due to the device having zero measured width or length. NOTE: This may need more investigation. If a subcircuit device's method of merging cannot be understood, then such devices should never extract with "merge aggressive", and should always merge conservatively if and only if all parameters match.
2023-05-21 18:13:01 +02:00
if (cfp->w > 0)
m = esFMult[cfp->esFMIndex] + (fp->w / cfp->w);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
case DEV_RSUBCKT:
case DEV_RES:
if (fp->dev->dev_type == esNoModelType)
Corrected an issue in which the device merging routine can generate NaN results for devices which are not FETs (specifically, devices that are declared using "msubckt" but are not FETs, although there may have been a related issue with non-FET devices not getting the correct M count), due to the device having zero measured width or length. NOTE: This may need more investigation. If a subcircuit device's method of merging cannot be understood, then such devices should never extract with "merge aggressive", and should always merge conservatively if and only if all parameters match.
2023-05-21 18:13:01 +02:00
{
if (cfp->dev->dev_res > 0)
m = esFMult[cfp->esFMIndex] + (fp->dev->dev_res
/ cfp->dev->dev_res);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
Corrected an issue in which the device merging routine can generate NaN results for devices which are not FETs (specifically, devices that are declared using "msubckt" but are not FETs, although there may have been a related issue with non-FET devices not getting the correct M count), due to the device having zero measured width or length. NOTE: This may need more investigation. If a subcircuit device's method of merging cannot be understood, then such devices should never extract with "merge aggressive", and should always merge conservatively if and only if all parameters match.
2023-05-21 18:13:01 +02:00
{
if (cfp->l > 0)
m = esFMult[cfp->esFMIndex] + (fp->l / cfp->l);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
Added extraction device type "csubcircuit", which should have been added a long time ago, since capacitors use a different method for calculating width and length than either transistors or resistors, so subcircuits need a special class designator or else the extraction may calculate the wrong dimensions for device width by totalling the perimeter between the device and terminal types, as it would for a MOSFET.
2018-10-30 21:19:20 +01:00
case DEV_CSUBCKT:
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
case DEV_CAP:
case DEV_CAPREV:
if (fp->dev->dev_type == esNoModelType)
Corrected an issue in which the device merging routine can generate NaN results for devices which are not FETs (specifically, devices that are declared using "msubckt" but are not FETs, although there may have been a related issue with non-FET devices not getting the correct M count), due to the device having zero measured width or length. NOTE: This may need more investigation. If a subcircuit device's method of merging cannot be understood, then such devices should never extract with "merge aggressive", and should always merge conservatively if and only if all parameters match.
2023-05-21 18:13:01 +02:00
{
if (cfp->dev->dev_cap > 0)
m = esFMult[cfp->esFMIndex] + (fp->dev->dev_cap
/ cfp->dev->dev_cap);
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
else
Corrected an issue in which the device merging routine can generate NaN results for devices which are not FETs (specifically, devices that are declared using "msubckt" but are not FETs, although there may have been a related issue with non-FET devices not getting the correct M count), due to the device having zero measured width or length. NOTE: This may need more investigation. If a subcircuit device's method of merging cannot be understood, then such devices should never extract with "merge aggressive", and should always merge conservatively if and only if all parameters match.
2023-05-21 18:13:01 +02:00
{
if ((cfp->l > 0) && (cfp->w > 0))
m = esFMult[cfp->esFMIndex] +
((fp->l * fp->w) / (cfp->l * cfp->w));
}
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
break;
}
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
setDevMult(fp->esFMIndex, DEV_KILLED);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
setDevMult(cfp->esFMIndex, m);
esSpiceDevsMerged++;
/* Need to do attribute stuff here */
freeMagic(fp);
return 0;
}
}
/* No parallel devs to it yet */
fp->next = devMergeList;
devMergeList = fp;
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* update_w --
* Updates the width client of node n with the current dev width
*
* Results:
* N/A
*
* Side effects:
* might allocate node client and widths
*
* ----------------------------------------------------------------------------
*/
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
update_w(
short resClass,
int w,
EFNode *n)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
nodeClient *nc;
int i;
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
if (n->efnode_client == (ClientData)NULL)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
initNodeClient(n);
nc = (nodeClient *) n->efnode_client;
if (nc->m_w.widths == NULL)
{
(nc->m_w.widths) = (float *)mallocMagic((unsigned)sizeof(float)
Got to the bottom of why area and perimeter values are no longer output for transistors. The problem came from a change made to fix an issue with capacitors marked as floating nodes because some nodes are not output as source or drain. But those nodes are output before the parameters, so when generating parameter output, all nodes appear to have already been output. Solution: Specify an additional bit in the "visited" mask for the node having been output that is separate from the mask for resist classes used by the code that writes parameter values, and use that bit as a test for whether the node is connected to some device (not necessarily a FET source or drain).
2021-12-20 19:12:49 +01:00
* (efNumResistClasses + 1));
for (i = 0; i <= efNumResistClasses; i++)
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
nc->m_w.widths[i] = 0.0;
}
nc->m_w.widths[resClass] += (float)w;
return 0;
}
/*
* ----------------------------------------------------------------------------
*
* devDistJunctVisit --
* Called for every dev and updates the nodeclients of its terminals
*
* Results:
* 0 to keep the calling procedure going
*
* Side effects:
* calls update_w which might allocate stuff
*
* ----------------------------------------------------------------------------
*/
extflat: integrate cb_extflat_visitdevs_t for EFVisitDevs()
2025-07-20 01:03:14 +02:00
/* ARGSUSED */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
/* @typedef cb_extflat_visitdevs_t (UNUSED) */
Updates from running the fix-ending-whitespace script.
2020-05-23 23:13:14 +02:00
int
K&R ext2spice/*.c: bulk function implementation conversion int spcnAP() follows K&R param ordering int spcnAPHier() follows K&R param ordering TODO mergeAttr() sharing references (does this setup a future potential use-after-free concern?)
2025-07-18 15:11:04 +02:00
devDistJunctVisit(
Dev *dev, /* Dev to examine */
HierContext *hc, /* Hierarchical path down to this dev */
float scale, /* Scale transform */
ext2spice: use strict callback prototypes
2025-07-18 23:02:48 +02:00
Transform *trans, /* (unused) */
ClientData cdata) /* (unused) */
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
{
EFNode *n;
int i;
int l, w;
Additional modification to the devVisit routine in extflat to change from passing the HierName to passing the HierContext (of which HierName is a part) so that more information from HierContext (such as the cell use being visited) can be passed to the callback procedure (largely for diagnostic purposes).
2020-06-01 23:14:22 +02:00
HierName *hierName = hc->hc_hierName;
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
if (dev->dev_nterm < 2)
{
TxError("outPremature\n");
return 0;
}
EFGetLengthAndWidth(dev, &l, &w);
Cleanup of a number of uninitialized variable issues flagged by the compiler. Some are obscure functions (plot verstatec hasn't been used in years) but others (like SPICE distributed junctions) are potentially significant sources of unexpected crashes on systems that don't zero uninitialized memory.
2019-06-07 20:13:50 +02:00
w = (int)((float)w * scale);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
for (i = 1; i<dev->dev_nterm; i++)
{
n = SpiceGetNode(hierName,
dev->dev_terms[i].dterm_node->efnode_name->efnn_hier);
Separated drain and source records for the esFetInfo array, to support asymmetric FETs and other devices like bipolars that have three distinct terminals. This does not go as far as it should to make the array independent of the number of declared terminals of the device. However, it suffices to make, e.g., parameter "a2=area" work for a bipolar device, and to generate the right drain and source areas and perimeters for asymmetric (e.g., extended-drain) devices.
2020-09-11 23:29:12 +02:00
if (i == 1)
update_w(esFetInfo[dev->dev_type].resClassSource, w, n);
else
update_w(esFetInfo[dev->dev_type].resClassDrain, w, n);
Initial commit at Tue Apr 25 08:41:48 EDT 2017 by tim on stravinsky
2017-04-25 14:41:48 +02:00
}
return 0;
}