Verilator

08/18/2019

01:09 AM Issue #1369: Raise error / warning on continous assignment to reg

In order to raise the warning for port, I've changed like the following:
In 'V3ParseGrammar.cpp'... Kris Jeon

08/14/2019

08:03 PM Issue #1489 (Feature): Python support for Verilated designs

I've improved the patchset by Patrick Stewart (proposed in bug1663) to add Python support for Verilated designs.
It ... Maarten De Braekeleer

08/12/2019

08:17 PM Issue #1487 (Resolved): New WIDTH warnings on genvars

Squashed, pushed and CI is green.
Verilator_ext_tests is working too. That CI environment is still a WIP, but is ... Todd Strader

07:26 PM Issue #1487: New WIDTH warnings on genvars

Great, tighter patch than I was expecting. Please squash and push.
& if have chance (if didn't already), check exten... Wilson Snyder

01:14 PM Issue #1487: New WIDTH warnings on genvars

Here is my proposed fix:
https://github.com/toddstrader/verilator-dev/tree/genvar-width
It appears the specific i... Todd Strader

08/10/2019

05:07 AM Issue #1487: New WIDTH warnings on genvars

Ok, the new test was written without looking at the old warning rules. My thinking was any const assignment to a unra... Wilson Snyder

08/09/2019

10:06 PM Issue #1487: New WIDTH warnings on genvars

Re: t_lint_width_genfor, I'm thinking the WIDTH warning in the procedural block "ri = i;" may not really be new. I w... Todd Strader

10:10 AM Issue #1487: New WIDTH warnings on genvars

Added two new tests in trunk, t_lint_width_genfor_bad which correctly fails
(though when this bug is fixed might hav... Wilson Snyder

08/08/2019

10:31 PM Issue #1276: vcd trace splits packed data type when it comes through a typedef

I'll try to pull the latest version and see what happens. It's been a while since I updated verilator. I'll update ... Christopher Russell

09:51 PM Issue #1468: Add Continuous Integration

P.S. Note that github just announced free open source CI via Actions. From what I can tell it's two active actions ea... Wilson Snyder

08:44 AM Issue #1486 (NoFixNeeded): Issue Linking on Mac OS

Glad you solved it.
Verilator uses the make-standard AR variable, so "make AR=foo ..."
Wilson Snyder

06:36 AM Issue #1486: Issue Linking on Mac OS

Wilson Snyder wrote:
> The _All files include some of the other files you already have listed, look at the contents ... Joseph Shaker

08/07/2019

09:23 PM Issue #1486: Issue Linking on Mac OS

The _All files include some of the other files you already have listed, look at the contents to determine which to in... Wilson Snyder

02:47 AM Issue #1486: Issue Linking on Mac OS

Wilson Snyder wrote:
> Maybe your "ar" is broken somehow? Try making a single "g++" command, where the file you spe... Joseph Shaker

05:18 AM Issue #1488: A strange code generated from a parametric module.

Thank you very much, Wilson! I was not aware about --pins-sc-uint switch! Slava B

08/06/2019

01:35 PM Issue #1488 (NoFixNeeded): A strange code generated from a parametric module.

uints are much faster so are the default. See the Connecting to SystemC part of the manual and the --pins-sc-uint/--... Wilson Snyder

11:51 AM Issue #1488 (NoFixNeeded): A strange code generated from a parametric module.

I've seen a strange code generated from a parametric module. I minimized this scenario to a few lines:... Slava B

08/05/2019

09:15 PM Issue #1363: CMake support

Ok, I'll hold off merging. If you could please merge from master into cmake, then merge cmake into your python branch... Wilson Snyder

06:37 PM Issue #1363: CMake support

I have applied your patch and think I have fixed the problem with t_flag_make_cmake_sc.pl
I would like to delay me... Maarten De Braekeleer

12:50 AM Issue #1363: CMake support

The sc example now works, thanks.
1. The attached has misc whitespace and other fixups you can please check over a... Wilson Snyder

02:35 AM Issue #1429 (Resolved): Feature request: elaboration tasks

Fixed in git towards 4.018.
Wilson Snyder

08/04/2019

10:37 PM Issue #1363: CMake support

Hello,
I've gone for option 3:
First look in SYSTEMC_INCLUDE/SYSTEMC_LIBDIR, then try SYSTEMC_ROOT, then try SY... Maarten De Braekeleer

05:49 PM Issue #1276: vcd trace splits packed data type when it comes through a typedef

Specifically I get:
$var wire 16 % in [15:0] $end
$var wire 16 & out [15:0] $end
Wilson Snyder

05:48 PM Issue #1276: vcd trace splits packed data type when it comes through a typedef

I don't see this. Are you sure you're using the most recent version? If so, please submit a complete test case in te... Wilson Snyder

03:32 PM Issue #1276: vcd trace splits packed data type when it comes through a typedef

Sorry for the extremely delayed response. This is still affecting VCD output. I think I've distilled it to a much s... Christopher Russell

10:10 AM Issue #1475: V3Hashed.cpp Called isIdentical on non-hashed nodes, from Gate dedupe()

Thanks for fixing this. This must have been an obscure edge case, I only ran into it that once. Given the Verilator t... Øyvind Harboe

02:01 AM Issue #1475 (Resolved): V3Hashed.cpp Called isIdentical on non-hashed nodes, from Gate dedupe()

Fixed in git towards 4.017. Thanks for the test case & patience.
Wilson Snyder

02:16 AM Issue #1483: Make verilator_ext_tests head-to-head

bug1487 captures the failures due to new WIDTH warnings.
Wilson Snyder

02:15 AM Issue #1487 (Closed): New WIDTH warnings on genvars

The following in git (unreleased) creates a warning where previously it did not, due to fixing bug1442. While technic... Wilson Snyder

08/03/2019

12:22 PM Issue #1486: Issue Linking on Mac OS

Maybe your "ar" is broken somehow? Try making a single "g++" command, where the file you specify has an #include of ... Wilson Snyder

02:10 AM Issue #1363: CMake support

Thought we were done... but not quite. Can you please change to use these variables:
SYSTEMC_INCLUDE=/example/p... Wilson Snyder

08/02/2019

02:22 PM Issue #1363: CMake support

... Maarten De Braekeleer

08/01/2019

11:52 PM Issue #1486: Issue Linking on Mac OS

Here are the files in my directory. All the intermediate object files seem to be built for the correct archetecture .... Joseph Shaker

11:50 PM Issue #1486: Issue Linking on Mac OS

Joseph Shaker wrote:
> Wilson Snyder wrote:
> > You're not using SystemC so VM_SC_* won't matter.
> >
> > Look a... Joseph Shaker

11:48 PM Issue #1486: Issue Linking on Mac OS

Wilson Snyder wrote:
> You're not using SystemC so VM_SC_* won't matter.
>
> Look at what commands make prints wh... Joseph Shaker

05:16 PM Issue #1486: Issue Linking on Mac OS

You're not using SystemC so VM_SC_* won't matter.
Look at what commands make prints when you build the archive the... Wilson Snyder

05:04 PM Issue #1486: Issue Linking on Mac OS

Wilson Snyder wrote:
> Looks like you used a GCC or set of GCC -march flags to make the .a (this might have come vis... Joseph Shaker

05:02 PM Issue #1486: Issue Linking on Mac OS

BTW it might help to rely on Verilator to do it all....
verilator {all-flags-used-previously} --exe audio_tes... Wilson Snyder

04:56 PM Issue #1486 (AskedReporter): Issue Linking on Mac OS

Looks like you used a GCC or set of GCC -march flags to make the .a (this might have come vis the GCC you used to con... Wilson Snyder

02:26 PM Issue #1486 (NoFixNeeded): Issue Linking on Mac OS

Hi all, I;ve used verilator for a while on Linux and have had no problems, however I wanted to start running it on my... Joseph Shaker

07/31/2019

10:58 AM Issue #1363: CMake support

Yes, the getenv stuff was added to reduce some questions I was getting about why SystemC would later break. I'll rem... Wilson Snyder

12:36 AM Issue #1363: CMake support

I agree to the `make`/`make install`. I reverted that change.... Maarten De Braekeleer

07/30/2019

10:29 PM Issue #1363: CMake support

I agree users should do a "make" before "make install" however I think it's a feature if someone changes code to make... Wilson Snyder

01:08 PM Issue #1363: CMake support

* I did remove the dependency of `make install` on `all_nomsg` because if some verilator source is changed between th... Maarten De Braekeleer

01:03 AM Issue #1363: CMake support

Also please test that "make test" is clean with VERILATOR_AUTHOR_SITE=1 in your environment before you configure.
... Wilson Snyder

12:57 AM Issue #1363: CMake support

Thanks, getting close.
I still can't "make examples" out of the box.
make examples
...
-- CMake ----... Wilson Snyder

07:36 PM Issue #1484 (Duplicate): Parameter of an array of enumeration values breaks conditionals inside a...

Thanks for the report & test. This is a duplicate of bug1315, on which work is currently in progress.
Wilson Snyder

07:21 PM Issue #1484 (Duplicate): Parameter of an array of enumeration values breaks conditionals inside a...

This was tested using the Verilator 4.016 release. When using a parameter that is an array of enumeration values insi... John Martin

07/29/2019

03:58 PM Issue #1363: CMake support

I've added a systemc example. I've tested this using Accellera's systemc.
The test will only execute if systemc ha... Maarten De Braekeleer

07/26/2019

11:31 PM Issue #1482 (Feature): Conditional event controls ("iff")

The event parsing needs some rewriting to handle this, but is relatively straight forward.
Added a disabled t_iff.... Wilson Snyder

12:14 AM Issue #1482 (Feature): Conditional event controls ("iff")

Section 9.4.2.3 of IEEE 1800-2017 allows "iff" qualifiers on @ event controls. The example code in 9.4.2.3 is fairly... Paul Donahue

10:59 PM Issue #1483 (Assigned): Make verilator_ext_tests head-to-head

I agree this is a good approach. Note Verilator, along with SwerRV is part of Chips Alliance, so if you want you can... Wilson Snyder

10:16 PM Issue #1483 (Closed): Make verilator_ext_tests head-to-head

This is still a WIP, but I think it warrants some discussion before I trudge ahead much further:
https://github.com/... Todd Strader

06:51 PM Issue #1305 (Resolved): Error messages do not contain hierarchical information

Wilson Snyder

05:30 PM Issue #1305: Error messages do not contain hierarchical information

Sold. This has been squashed, pushed and is green. Todd Strader

04:30 PM Issue #1305: Error messages do not contain hierarchical information

Long term I think we need a nicer way to handle error messages but a project better left for later. So I agree #2 is... Wilson Snyder

03:51 PM Issue #1305: Error messages do not contain hierarchical information

I updated the branch with all these changes:
https://github.com/toddstrader/verilator-dev/tree/mod_stack_trace_10
... Todd Strader

06:14 PM Issue #1363: CMake support

I've rewritten `configure.ac` to use a foreach loop.
I'll look into adding systemc support for the cmake script. Maarten De Braekeleer

04:59 PM Issue #1363: CMake support

I pushed the contributors patch and typos patch - threw some other typo fixes in after searching for the same typo. T... Wilson Snyder

02:22 PM Issue #1363: CMake support

Hello folks
@Patrick Stewart
* I have modified the `make install` to copy the cmake scripts to /usr/share/verilat... Maarten De Braekeleer

11:20 AM Issue #1363: CMake support

I'm using cmake 3.10.2. I'm running "make examples" inside the parent git repo checkout (not installed).
rm -... Wilson Snyder

10:55 AM Issue #1363: CMake support

If you're putting it into share/verilator/cmake, you need to set the parent directory of ${CMAKE_CURRENT_LIST_DIR} on... Patrick Stewart

10:24 AM Issue #1481 (Resolved): Add rr support to the perl wrappers

Also, please add the bug # to the Changes, did this, e.g.
*** Add --rr, bug1481. [Todd Strader]
Thanks fo... Wilson Snyder

10:23 AM Issue #1481: Add rr support to the perl wrappers

Number of stars is the importance. Usually big things 2, features are 3, bug fixes 4, but not a science.
Wilson Snyder

09:22 AM Issue #1481: Add rr support to the perl wrappers

Squashed, pushed and green.
Can you explain the number of stars for each entry in Changes? It appears that they a... Todd Strader

07/25/2019

11:45 PM Issue #1481 (Feature): Add rr support to the perl wrappers

"gdb!" => \$opt_gdb,
+ "rr!" => \$opt_rr,
"gdbbt!" => \... Wilson Snyder

11:28 PM Issue #1481 (Closed): Add rr support to the perl wrappers

As discussed in the forums:
https://www.veripool.org/boards/3/topics/3028
This allows users to record verilation ... Todd Strader

10:50 PM Issue #1363: CMake support

Thanks for all the cleanups.
If you want just point me to a github branch and you can skip posting patch files.
... Wilson Snyder

03:52 PM Issue #1363: CMake support

I've created a new pull request in which I add myself to docs/CONTRIBUTORS to certify that this and future contributi... Maarten De Braekeleer

03:34 PM Issue #1363: CMake support

Hello,
I think I have incorporated the comments in #5.
They can still be found at https://github.com/madebr/ver... Maarten De Braekeleer

03:33 PM Issue #1363: CMake support

Hello Patrick,
Thanks for the comments!
I think I have incorporated your suggestions with some changes.
Maarten De Braekeleer

10:34 AM Issue #1363: CMake support

Thanks for taking this on Maarten, I haven't had time for a while.
I've just a had a quick go with it, when I instal... Patrick Stewart

02:16 AM Issue #1363: CMake support

Nice job.
1. Please sumbit a one-time independent patch to edit docs/CONTRIBNUTORS to
add your name and certify t... Wilson Snyder

10:39 PM Issue #1305: Error messages do not contain hierarchical information

Multiple cell visiting in AstParam is expected as it does the original cell, then the one it clones to de-parameteriz... Wilson Snyder

05:07 PM Issue #1305: Error messages do not contain hierarchical information

OK, none of the .out files are looking for internal faults now. And CI is green:
https://github.com/toddstrader/ver... Todd Strader

03:27 PM Issue #1305: Error messages do not contain hierarchical information

Nope. Apparently during my git juggling of the _bad.out files, I missed that I committed two test which are now toss... Todd Strader

02:44 PM Issue #1305: Error messages do not contain hierarchical information

After the refactor in bug1315, I think we're probably/hopefully there:
https://github.com/toddstrader/verilator-dev/... Todd Strader

07/24/2019

07:27 PM Issue #1363: CMake support

Hey
I've rebased the cmake patch and updated it a bit.
It would be nice if someone would review them.
The code... Maarten De Braekeleer

07/23/2019

07:53 PM Issue #1315: Using an array in a function called from a parameter

Squashed, pushed and passing CI. Todd Strader

05:53 PM Issue #1315: Using an array in a function called from a parameter

>Do you think I should leave it as-is so that external users of the fetcher methods get a V3Number*? Or should I chan... Wilson Snyder

03:27 PM Issue #1315: Using an array in a function called from a parameter

> I would think that using an AstConst is ok also, as previously we would have fetched an V3Number with similar lifet... Todd Strader

07/22/2019

10:15 PM Issue #1479: Feature Request: VerilatedVcd callback on rolloverMB

Good work, in general it seems fine.
Please also have the patch add your name to docs/CONTRIBUTORS to acknowledge ... Wilson Snyder

08:09 PM Issue #1479: Feature Request: VerilatedVcd callback on rolloverMB

I have rebased the initial patch on top of master, sorry for posting a second patch.
This patch adds a filename() ... Marc Jessome

12:47 PM Issue #1475 (Assigned): V3Hashed.cpp Called isIdentical on non-hashed nodes, from Gate dedupe()

Got a test case which shows a replacement is going into a flop which itself is then subject to replacement. This is ... Wilson Snyder

02:34 AM Issue #1315: Using an array in a function called from a parameter

>I made variants of fetch* that return AstConst* for use inside SimulateVisitor and
>kept V3Number* methods for use ... Wilson Snyder