CentOS 7 / VCS fixes #2067
Merged
CentOS 7 / VCS fixes #2067
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Comment on lines
44
to
46
| '-dbname', str(md.dir_cov/'merged.vdb'), | ||
| '-report', str(md.dir_cov/'report'), | ||
| '-log', str(md.dir_cov/'merge.log'), |
There was a problem hiding this comment.
Was this broken before?
There was a problem hiding this comment.
Hmm, looks like it! (Or my grep-fu is not good enough :-))
|
Also, apparently clang format is failing. |
rswarbrick
approved these changes
Aug 3, 2023
| # building, riscv-fesvr isn't strictly required but the DV flow has been | ||
| # observed to see build failures where it isn't present with CentOS 7. | ||
| spike_iss_pc = ['riscv-riscv', 'riscv-disasm', 'riscv-fdt', | ||
| 'riscv-fesvr'] |
There was a problem hiding this comment.
Nit: Maybe worth indenting to line up with the preceding [.
| @@ -18,7 +18,7 @@ | |||
| # As a result, passing -fno-extended-identifiers tells G++ to pretend that | |||
| # everything is ASCII, preventing strange compilation errors. | |||
| - tool: vcs | |||
| env_var: IBEX_ROOT | |||
| env_var: IBEX_ROOT, EXTRA_COSIM_CFLAGS | |||
There was a problem hiding this comment.
Do we also need to tweak some documentation in this commit to tell (some) people to pass this flag?
There was a problem hiding this comment.
We should document it but as it's only really relevant to CentOS 7 and there's other things that need documenting for that also I'd prefer to deal with that in a follow up PR (in particular so we can get the various fixes in this PR in).
Comment on lines
44
to
46
| '-dbname', str(md.dir_cov/'merged.vdb'), | ||
| '-report', str(md.dir_cov/'report'), | ||
| '-log', str(md.dir_cov/'merge.log'), |
There was a problem hiding this comment.
Hmm, looks like it! (Or my grep-fu is not good enough :-))
Fixes two issues observed on CentOS 7 during testbench compile - pkg-config doesn't behave properly when it receives multiple flags - riscv-fesvr library needs to be included in build
When SpikeCosim is getting destructed a SIGSEGV was observed on CentOS 7. The root cause hasn't been identified other than it relates to the deletion of `isa_parser_t`, potentially some kind of use after free error. This is an (optional) hacky workaround that simply never deletes the `isa_parser_t` pointer in SpikeCosim. As in practise this occurs at the end of simulation when the process is terminating the memory leak is of little consequence. Longer term this issue should be investigated and properly fixed.
|
Thanks for the reviews @marnovandermaas and @rswarbrick. Once again apologies I didn't push this over the line when it was first submitted. I've added a bunch of other little fixes mostly related to improving the flow under VCS PTAL. |
Comment on lines
50
to
57
| with LockedMetadata(md.dir_metadata, __file__) as md: | ||
| md.cov_merge_log = md.dir_cov / 'merge.log' | ||
| md.cov_merge_stdout = md.dir_cov / 'merge.log.stdout' | ||
| md.cov_merge_cmds = [cmd] | ||
|
|
||
| with open(md.cov_merge_stdout, 'wb') as fd: | ||
| logging.info("Generating merged coverage directory") | ||
| return run_one(md.verbose, cmd, redirect_stdstreams=fd) |
There was a problem hiding this comment.
No need to keep the Metadata file open while we actually perform the coverage merge.
Suggested change
| with LockedMetadata(md.dir_metadata, __file__) as md: | |
| md.cov_merge_log = md.dir_cov / 'merge.log' | |
| md.cov_merge_stdout = md.dir_cov / 'merge.log.stdout' | |
| md.cov_merge_cmds = [cmd] | |
| with open(md.cov_merge_stdout, 'wb') as fd: | |
| logging.info("Generating merged coverage directory") | |
| return run_one(md.verbose, cmd, redirect_stdstreams=fd) | |
| with LockedMetadata(md.dir_metadata, __file__) as md: | |
| md.cov_merge_log = md.dir_cov / 'merge.log' | |
| md.cov_merge_stdout = md.dir_cov / 'merge.log.stdout' | |
| md.cov_merge_cmds = [cmd] | |
| with open(md.cov_merge_stdout, 'wb') as fd: | |
| logging.info("Generating merged coverage directory") | |
| return run_one(md.verbose, cmd, redirect_stdstreams=fd) |
Additionally update metadata with the appropriate coverage merge log information.
Programs generated from RISC-V were exiting early in configs that don't have PMP
It's possible for a TestRunResult to contain an entry that has a path to a build/run artifact but for that to be None rather than an actual path. This causes the collect_results.py script to fail. With this change such paths will be described as 'MISSING' in the regression log instead.
An instruction stall on a FENCE.I had been mistakenly placed in an illegal bin. A FENCE.I acts much like a branch or jump so can produce instruction stalls just as those instructions can.
Previously some $value$plusargs calls weren't explictly specifying a format for a number to read from the plusarg. Under some simulators this is acceptable under others it generates an error.
Tests targetting security features make use of force to corrupt values. Under VCS this requires building with a particular flag.
The DV flow is expecting log files to be produced with a particular file name, without it the reporting mechanisms do not work correctly. This adds VCS log output to a named rather rather than just capturing stdout.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I've been getting the regression running on CentOS 7 under both VCS and Xcelium. This fixes up a mixture of CentOS 7 and VCS issues.
There will be another PR at some point with some documentation clean-ups and a specific guide for CentOS 7 as setting up the requirements needs some guidance.