The manual for the BLKANDNBLK warning describes that it is safe to
disable that error if the updated ranges are non-overlapping. This
however was not true (see the added t_nba_mixed_update* tests).
In this patch we change V3Delayed to use a new ShadowVarMasked
scheme for variables that have mixed blocking and non-blocking
updates (or the FlagUnique scheme for unpacked variables), which
is in fact safe to use when the updated parts are non-overlapping.
Furthermore, mixed assignments are safe as far as scheduling is
concerned if either:
- They are to independent parts (bits/members/etc) (with this patch)
- Or if the blocking assignment is in clocked (or suspendable) logic.
The risk in scheduling is a race between the Post scheduled NBA
commit, and blocking assignments in combinational logic, which might
order incorrectly.
The second point highlights that we can handle stuff like this safely,
which is sometimes used in testbenches:
```systemverilog
always @(posedge clk) begin
if ($time == 0) a = 0;
end
always @(posedge clk) begin
if ($time > 0) a <= 2;
end
````
The only dangerous case is:
```systemverilog
always @(posedge clk) foo[idx] <= val;
assign foo[0] = bar;
```
Whit this patch, this will still resolve fine at run-time if 'idx' is
never zero, but might resolve incorrectly if 'idx' is zero.
With the above in mind, the BLKANDNBLK warning is now only issued if:
- We can't prove that the assignments are to non-overlapping bits
- And the blocking assignment is in combinational logic
These are the cases that genuinely require user attention to resolve.
With this patch, there are no more BLKANDNBLK warnings in the RTLMeter
designs.
Fixes#6122.
These passes blow up the Ast size on some designs, so delaying running V3Const
until after the whole pass can notably increase peak memory usage. In this
patch we apply V3Const per CFunc within these passes, which saves on memory.
Added -fno-const-eager to disable the intra-pass V3Const application, for
debugging.
This patch adds a heuristic to V3SplitVar, and it attempts to split up
packed variables that are only referenced via constant index,
non-overlapping bit/range selects. This can eliminate some UNOPTFLAT cases.
Add a new pass to split up (recursively):
foo = {l, r};
into the following, with the right indices, iff the concatenation
straddles a wide word boundary.
foo[_:_] = r;
foo[_:_] = l;
This eliminates more wide temporaries.
Another 23% speedup on VeeR EH2 high_perf. Also brings the predicted
stack size from 8M to 40k.
* Refactor V3Delay for extensibility
Introduce the concept of an "NBA Scheme", which is the lowering pattern
we can use for various variables that are the targets of NBAs.
E.g.:
- ShadowVariable (old default scheme)
- FlagShared (old array set flag scheme)
- ValueQueueWhole (recent dynamic commit queue)
We now analyse all AstAssignDly before making any decisions on which
scheme to apply. We then choose a specific scheme for each variable that
is the target of an NBA, and then all NBAs targeting that variable use
the same scheme. This enables easy mix and match of schemes as needed,
while remaining consistent by design after extensions.
Output is perturbed due to node insertion order, but no functional
or performance change is intended.
The necessary options to support C++ coroutines vary greatly between
compilers, its versions, and the standard library being used. This
patch makes the check for coroutine support more robust by adding a
declaration of a coroutine variable, instead of just including the
header. It also makes sure that the HAVE_COROUTINE and
CFG_CXXFLAGS_COROUTINES flags are always set together, and only when
coroutine support is detected.
Documentation states that minimum of all reported coverage of all signals in a line should be taken.
Previous logic would break if there were any signals with zero coverage followed by signals with
nonzero coverage - a minimum from those nonzero toggle count would be taken, disregarding zero
coverage of previous signals.
Internal-tag: [#62193]
Signed-off-by: Krzysztof Obłonczek <koblonczek@antmicro.com>
`$psprintf` is a non-standard system function present in some other
simulators, and has been rejected for standardization by IEEE because
of being basically the same as `$sformatf`.
To encourage users to fix their codebase, a warning is emitted by
default, but it gets otherwise interpreted as `$sformatf` as early as
during lexing.
Signed-off-by: Arkadiusz Kozdra <akozdra@antmicro.com>
* wording/formatting
Signed-off-by: Arkadiusz Kozdra <akozdra@antmicro.com>
---------
Signed-off-by: Arkadiusz Kozdra <akozdra@antmicro.com>
* Support 2D dynamic array initialization (#4700)
- new[] on sub arrays (as per original issue)
- Built-in methods for sub-arrays
- Initialization and literals assignmensts
- Dynamic array as an element for other arrays and queues
libcxx has removed the experimental/coroutine include file in favor of
the C++20-standard coroutine include. If the latter is available we
use it otherwise falling back to the existing experimental version (in
which case we also disable the deprecated-experimental-coroutine warning).
(See also https://reviews.llvm.org/D108697.)
operand order reversed for AstCMethodHard "neq"
interface between C-style arrays and VlUnpacked
overloads added to VlUnpacked::neq(), VlUnpacked::assign()
VlUnpacked::operator=() added
Fixes #5125
For NBAs that might execute a dynamic number of times in a single
evaluation (specifically: those that assign to array elements inside
loops), we introduce a new run-time VlNBACommitQueue data-structure
(currently a vector), which stores all pending updates and the necessary
info to reconstruct the LHS reference of the AstAssignDly at run-time.
All variables needing a commit queue has their corresponding unique
commit queue.
All NBAs to a variable that requires a commit queue go through the
commit queue. This is necessary to preserve update order in sequential
code, e.g.:
a[7] <= 10
for (int i = 1 ; i < 10; ++i) a[i] <= i;
a[2] <= 10
needs to end with array elements 1..9 being 1, 10, 3, 4, 5, 6, 7, 8, 9.
This enables supporting common forms of NBAs to arrays on the left hand
side of <= in non-suspendable/non-fork code. (Suspendable/fork
implementation is unclear to me so I left it unchanged, see #5084).
Any NBA that does not need a commit queue (i.e.: those that were
supported before), use the same scheme as before, and this patch should
have no effect on the generated code for those NBAs.
No functional change. Postpone the conversion of all AstAssignDlys that
use the 'VdlySet' scheme for array LHSs until after the complete
traversal of the netlist. The next patch takes advantage of this by
using some extra information also gathered through the traversal to
change the conversion.
AstAssignDlys inside suspendable or fork are not deferred and are
processed identical to the previous version.
There are some TODOs in this patch that are fixed in the next patch.
Output code perturbed due to variable ordering.
MULTIDRIVEN message ordering perturbed due to processing order change.
Geza Lore
Wilson Snyder
Peter Birch
Todd Strader
Bartłomiej Chmiel
Petr Nohavica
Ryszard Rozak
Joel Bodenmann
Shou-Li Hsu
Zhen Yan
Dominick Grochowina
John Khoo
Zhouyi Shen
Robin Heinemann
Bartłomiej Chmiel
Krzysztof Sychla
Brian Li
Tobias Jensen
Andrew Voznytsa
Gilberto Abram
Mateusz Gancarz
Kamil Rakoczy
Dave Sargeant
Trung Nguyen
Moubarak Jeje
Krzysztof Bieganski
Natan Kreimer
Andrew Nolte
Nick Brereton
Anthony Moore
sumpster
Tom Manner
Greg Davill
Krzysztof Starecki
Tomasz Gorochowik
Eric Müller
Aidan McNay
Furqan Nadir
Nathan Graybeal
Han Qi
Fabian Keßler
Mariusz Glebocki
Chris Bachhuber
Yilou Wang
James Bailey
Luca Colagrande
Ricardo Barbedo
Udaya Raj Subedi
Yangyu Chen
Krzysztof Obłonczek
Ryan Ziegler
Liam Braun
404allen404
Philip Axer
Alex Solomatnikov
Arthur Rosa
Arkadiusz Kozdra
Valentin Atepalikhin
Paul Bowen-Huggett
Pawel Jewstafjew