-
Notifications
You must be signed in to change notification settings - Fork 4.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
SPMI: Hacky workaround for coredistools unsupported instructions #91668
SPMI: Hacky workaround for coredistools unsupported instructions #91668
Conversation
We currently are unable to get SPMI diffs for methods that use ldapur instructions as our current coredistools does not support this instruction. While we wait for coredistools to be updated (this was blocked on some C++ lib issues) here is a hacky temporary workaround for the problem. It rewrites ldapur* instructions into the corresponding ldur* instruction before passing it to the near differ.
Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch Issue DetailsWe currently are unable to get SPMI diffs for methods that use ldapur
|
/azp run runtime-coreclr superpmi-replay, runtime-coreclr superpmi-diffs |
Azure Pipelines successfully started running 2 pipeline(s). |
/azp run runtime-coreclr superpmi-replay, runtime-coreclr superpmi-diffs |
Azure Pipelines successfully started running 2 pipeline(s). |
cc @dotnet/jit-contrib PTAL @BruceForstall No diffs -- fixes the "0 sized diffs" on arm64 we have been seeing recently. Only to be used as a temporary hack while coredistools is outdated. |
This reverts commit 593fb3a.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That's quite a hack :-)
It would be useful to add a comment about why this code is there, and what it is trying to do. Maybe link to an issue (or two) indicating how it should be ultimately fixed (and the issue can point out to remove this code after coredistools is updated).
@jakobbotsch Are you planning to merge this? I wonder if it's contributing to the checked/release asm diffs failures, e.g. https://dev.azure.com/dnceng-public/public/_build/results?buildId=401020&view=logs&j=83516c17-6666-5250-abde-63983ce72a49
|
Done. Yeah, I think the AV could be related. I would expect it to be fixed either by this or #91783. I kicked off https://dev.azure.com/dnceng-public/public/_build/results?buildId=404846&view=results. |
Looks like the x64 run passed, but the x86 run timed out. I think I forgot to increase the imtout of superpmi-asmdiffs-checked-release in #91637. |
The console logs are odd, though: they finish at the end of the download phase, after less than 10 minutes, with no evidence the replay/run phase even started. |
|
ah, ok. so the timeouts do need to be increased. fwiw, I ran the failing cases locally and got no diffs. |
We currently are unable to get SPMI diffs for methods that use ldapur
instructions as our current coredistools does not support this
instruction. While we wait for coredistools to be updated (this was
blocked on some C++ lib issues) here is a hacky temporary workaround for
the problem. It rewrites ldapur* instructions into the corresponding
ldur* instruction before passing it to the near differ.