Spelling fixes

This commit is contained in:
Larry Doolittle 2015-04-13 09:23:43 -07:00 committed by Cary R
parent ccf7c042c9
commit 5dc1396eea
6 changed files with 16 additions and 16 deletions

View File

@ -94,7 +94,7 @@ class netrange_t {
inline netrange_t() : msb_(LONG_MAX), lsb_(LONG_MAX) { }
// Create a properly defined netrange
inline netrange_t(long m, long l) : msb_(m), lsb_(l) { }
// Copy constructure.
// Copy constructor.
inline netrange_t(const netrange_t&that)
: msb_(that.msb_), lsb_(that.lsb_) { }

View File

@ -1153,7 +1153,7 @@ description_list
;
/* This implements the [ : INDENTIFIER ] part of the constructure
/* This implements the [ : IDENTIFIER ] part of the constructor
rule documented in IEEE1800-2005: A.1.8 */
endnew_opt : ':' K_new | ;
@ -1584,7 +1584,7 @@ modport_item
/* The modport_ports_list is a LALR(2) grammar. When the parser sees a
',' it needs to look ahead to the next token to decide whether it is
a continuation of the preceeding modport_ports_declaration, or the
a continuation of the preceding modport_ports_declaration, or the
start of a new modport_ports_declaration. bison only supports LALR(1),
so we have to handcraft a mini parser for this part of the syntax.
last_modport_port holds the state for this mini parser.*/

View File

@ -88,7 +88,7 @@ static bool symbol_search(const LineInfo*li, Design*des, NetScope*scope,
return false;
// The prefix is found to be a scope, so switch to that
// scope, set the hier_path to turn of upwards searches,
// scope, set the hier_path to turn off upwards searches,
// and continue our search for the tail.
if (recurse.is_scope()) {
scope = recurse.scope;

View File

@ -117,7 +117,7 @@ static void draw_ufunc_preamble(ivl_expr_t expr)
function ports. Do this in two passes - evaluate,
then send - this avoids the function input variables
being overwritten if the same (non-automatic) function
is called in one of the exressions. */
is called in one of the expressions. */
assert(ivl_expr_parms(expr) == (ivl_scope_ports(def)-1));
for (idx = 0 ; idx < ivl_expr_parms(expr) ; idx += 1) {

View File

@ -201,7 +201,7 @@ class ComponentInstantiation : public Architecture::Statement {
virtual int emit(ostream&out, Entity*entity, Architecture*arc);
virtual void dump(ostream&out, int indent =0) const;
// Returns the expression that initalizes a generic (or NULL if not found).
// Returns the expression that initializes a generic (or NULL if not found).
Expression*find_generic_map(perm_string by_name) const;
inline perm_string instance_name() const { return iname_; }

View File

@ -199,7 +199,7 @@ event control registers to determine when to perform the assign.
This instruction blends the bits of two vectors into a result in a
manner line the expressions ('bx ? <a> : <b>). The two source vectors
are popped from the vec4 stack (and must have the same width) and the
result poshed in their place. The truth table for each bit is:
result pushed in their place. The truth table for each bit is:
1 1 --> 1
0 0 --> 0
@ -510,7 +510,7 @@ This instruction copies the flag bit from <flag2> to <flag1>.
* %flag_or <flag1>, <flag2>
This instruction calculates the verilog OR of the flag bits in <flag1>
This instruction calculates the Verilog OR of the flag bits in <flag1>
and <flag2>, and leaves the result in <flag1>.
* %flag_set/imm <flag>, <value>
@ -529,11 +529,11 @@ writes the LSB to the selected flag.
* %force/vec4 <label>
* %force/vec4/off <label>, <off>
Perform a "force" assign of a values to the tarvet variable. The value
Perform a "force" assign of a values to the target variable. The value
to be forced is popped from the vec4 stack and forced to the target
variable. The /off variant forces a part of a vector. The width of the
part comes from the width of the popped value, and the <off> is an
index register that contains the cannonical offset where the value
index register that contains the canonical offset where the value
sets written.
The %foce/vec4/off instruction will test the value if flags[4], and if
@ -806,7 +806,7 @@ The supported types are:
* %nor
Perform the bitwise nor of vec4 vectors, and push the result. Eack bit
Perform the bitwise nor of vec4 vectors, and push the result. Each bit
in the source vectors is combined to make a result bit according to the
truth table.
@ -841,7 +841,7 @@ can be used with any class or darray object, so it is not typed.
* %or
Perform the bitwise or of twp vectors. Pop two values from the vec4
Perform the bitwise or of two vectors. Pop two values from the vec4
stack to get the input arguments. Each bit in the result is combined
with the corresponding bit in the input arguments, according to the
truth table:
@ -893,7 +893,7 @@ includes an implicit pop, but sometimes it is necessary to pop
explicitly.
The <skip> is the number of top positions on the stack to keep,
beforing starting to pop. This allows for popping positions other than
before starting to pop. This allows for popping positions other than
the top of the stack.
* %pow <bit-l>, <bit-r>, <wid>
@ -966,7 +966,7 @@ result pushed to the string stack.
* %putc/str/v <functor-label>, <muxr>
Pop a vector byte from the thread vec4 stack and write it to a
character of the string variable at <functtor-label>. This is
character of the string variable at <functor-label>. This is
basically an implementation of <string>.putc(<muxr>, <val>) where
<val> is the 8bit vector popped from the stack.
@ -1045,7 +1045,7 @@ The instruction also checks flag bit 4. If it is true, the result is
replaced with X instead of a shifted result. This is intended to
detect that the index contents were not valid.
For a negative shift, %shiftr will pad the vlaue with 'bx.
For a negative shift, %shiftr will pad the value with 'bx.
* %split/vec4 <wid>
@ -1054,7 +1054,7 @@ parts. Split off <wid> bits from the LSB, then push the remaining bits
of the original (the MSB) back to the stack. Then push the split off
LSB vector.
The <wid> must be less then the width of the original, unsplit vector.
The <wid> must be less than the width of the original, unsplit vector.
* %store/obj <var-label>