Wilson Snyder
c233a39052
Merge from master for release.
2026-04-26 01:57:12 -04:00
Wilson Snyder
d0aa828c21
Version bump
2026-04-26 01:52:06 -04:00
Geza Lore
42ea9fda17
Commentary: Changes update
2026-04-25 17:56:03 +01:00
Wilson Snyder
00211c290c
Commentary
2026-04-25 11:41:30 -04:00
Wilson Snyder
5c989025c7
Commentary: Changes update
2026-04-25 11:26:51 -04:00
github action
d1c0184916
Apply 'make format'
2026-04-25 15:17:45 +00:00
M2kar
07d91fb253
Fix infinite recursion with VERILATOR_BIN ( #7496 ) ( #7497 )
...
Fixes #7496 .
2026-04-25 17:16:50 +02:00
Wilson Snyder
555d7cc043
Tests: Remove old .out files and related cleanups
2026-04-24 12:25:28 -04:00
Wilson Snyder
b8d59ed7a8
Commentary: Changes update
2026-04-24 12:22:26 -04:00
em2machine
b1d3d63e9f
Fix for unction argument expects a class refetence ( #7483 ) ( #7489 )
...
Fixes #7483 .
2026-04-24 10:31:57 +02:00
Wilson Snyder
2cd9b8df66
Fix multiple inclusion of verilated_std.sv ( #7478 ).
2026-04-23 15:32:02 -04:00
Wilson Snyder
64b91023b1
Commentary: Changes update
2026-04-23 12:47:22 -04:00
em2machine
121fd3f613
Fix module parameters not re-evaluated upon instantiation ( #7463 ) ( #7477 )
2026-04-23 12:30:42 -04:00
Geza Lore
ba3937734f
Optimize mix of Concat/Extend assignments ( #7479 )
...
Generalize concatenation splitting/balancing in V3FuncOpt to also handle
AstExtend (which is a Concat with a 0 LHS).
2026-04-23 16:57:43 +01:00
Alex Zhou
7e0f25b41b
Fix access to static variable inside function ( #7474 ) ( #7475 )
...
Fixes #7474 .
2026-04-23 11:17:14 -04:00
Yilou Wang
27e7fdc9eb
Fix std::randomize inside {typedef array} internal error ( #7481 )
...
Fixes #7481 .
2026-04-23 11:16:24 -04:00
Yilou Wang
ae642facf5
Fix function-in-constraint internal error on bare port VarRef return ( #7473 ) ( #7480 )
...
Fixes #7473 .
2026-04-23 10:47:39 -04:00
Geza Lore
3149f43372
Internals: Add special Dfg vertex for replication ( #7471 )
2026-04-23 10:14:27 +01:00
Yilou Wang
cfebe805c9
Support implication operator with constraint_set ( #7300 ) ( #7448 )
...
* Support implication operator with constraint_set
* improve coverage, achieve 100 line cov
* Address review: simplify addNext; tighten disable soft grammar to constraint_primary
* Apply 'make format'
* re-run
---------
Co-authored-by: github action <action@example.com>
2026-04-23 10:50:23 +02:00
github action
46a6884df6
Apply 'make format'
2026-04-23 07:14:49 +00:00
Yogish Sekhar
72bbccb543
Add Fsm coverage to print summary ( #7462 )
2026-04-23 09:13:45 +02:00
Wilson Snyder
c1d1b333ac
With -Wpendantic allow circular assigns
2026-04-23 00:50:12 -04:00
Wilson Snyder
5064a5ee65
Commentary: Changes update
2026-04-23 00:44:50 -04:00
Wilson Snyder
6945f04531
Internals: Use addNext. No functional change intended
2026-04-23 00:34:47 -04:00
Yilou Wang
51eb5346c8
Support sequence within operator ( #7461 )
2026-04-23 05:47:24 +02:00
github action
e4da16caf0
Apply 'make format'
2026-04-22 19:20:00 +00:00
Yogish Sekhar
a680919edc
Support native FSM state and arc coverage ( #7412 )
2026-04-22 15:18:59 -04:00
github action
41ddf4d9b6
Apply 'make format'
2026-04-22 16:13:36 +00:00
Nikolay Puzanov
5cddbd7fda
Fix tracing virtual interface member written from classes ( #5044 ) ( #7465 )
...
Track AstMemberSel writes through virtual interface refs and connect them to matching interface-member VarScopes, so class-driven interface clocks get proper VCD activity updates.
Fixes #5044 .
2026-04-22 12:09:30 -04:00
Geza Lore
496665800d
Internals: Annotate AstNode classes to generate matching DfgVertex ( #7469 )
...
Explicitly annotate those AstNodeExpr subclasses that should have a
corresponding DfgVertex subclass generated by astgen. This avoids having
to tweak things in Dfg when adding new AstNode subclasses, which can
then be handled separately.
2026-04-22 16:35:07 +01:00
em2machine
59fe8066ff
Support multidimensional arrays of interfaces ( #6230 ) ( #7451 )
...
Fixes #6230
2026-04-22 13:42:10 +02:00
Todd Strader
15163d1e39
Tests: Switch VCD/FST compare to wavediff ( #7426 )
2026-04-21 13:53:53 -04:00
Geza Lore
686594b4ab
Optimize temporary insertion in Dfg ( #7459 )
...
Insert temporaries for highest fanout vertices first. Omit them for
cheap comparisons.
2026-04-21 18:34:26 +01:00
Artur Bieniek
a1a8b9624c
Support IEEE-compliant force/release handling ( #7391 )
2026-04-21 11:54:42 -04:00
Yilou Wang
280cff06f3
Fix std::randomize `inside` corrupting class-member queue operand ( #7449 ) ( #7456 )
...
Fixes #7449
2026-04-21 17:37:58 +02:00
github action
c60d8f002f
Apply 'make format'
2026-04-21 12:22:33 +00:00
Yogish Sekhar
bca5839d46
Add printed summary to verilator_coverage ( #7438 )
2026-04-21 14:21:38 +02:00
Geza Lore
2b9d006097
Change Dfg pattern dumps to use --dump-dfg-patterns ( #7455 )
...
Dumping Dfg patterns can take a non-trivial amount of time, so do it
only with --dump-dfg-patterns, instead of with --stats.
Also further improve dumping format.
2026-04-21 12:07:19 +01:00
Wilson Snyder
2ef979a39c
Fix dead removing packages with only DPI.
2026-04-21 00:34:20 -04:00
Wilson Snyder
23ea3d7f11
Commentary: Changes update
2026-04-21 00:33:40 -04:00
Wilson Snyder
94e3f387a7
Internals: With debug, show cwd
2026-04-20 14:06:01 -04:00
Geza Lore
cd30c22d1c
Optimize more Dfg patterns ( #7452 )
2026-04-20 18:28:11 +01:00
dependabot[bot]
81f6cdc32c
CI: Bump actions/upload-pages-artifact from 4 to 5 in the everything group ( #7453 )
2026-04-20 19:16:37 +02:00
em2machine
23ca23b7b5
Fix for ariane/CVA6 false member call on object error ( #7445 ) ( #7450 )
...
Fixes #7445 .
2026-04-20 18:21:59 +02:00
Geza Lore
f3c63d017a
Optimize Sel patterns in DfgPeephole
2026-04-20 09:01:31 +01:00
Geza Lore
ffd2c5c69e
Optimize temporary insertion in DfgPeephole
2026-04-20 09:01:31 +01:00
Geza Lore
93d7d9c417
Optimize DfgDataType::hash()
2026-04-20 09:01:31 +01:00
Yilou Wang
935b2564eb
Use NFA in SVA pass (V3AssertNfa: NFA-based multi-lcycle SVA evaluation engine) ( #7430 )
2026-04-20 07:43:18 +02:00
Geza Lore
e82bd52fa3
Optimize arithmetic right shift (>>>) in DfgBreakCycles ( #7447 )
2026-04-19 20:28:17 +01:00
Geza Lore
60b41c7822
Optimize DfgBreakCycles IndependentBits analysis ordering ( #7446 )
...
Perform the data flow analysis based on reverse post-order numbering of
vertices. This speeds up convergence by ~10x. Partial fix for #7201 .
2026-04-19 19:48:02 +01:00