Fix crash in SuperPMI PGO apis#47305
Conversation
The real fix is a trivial fix in the linked list handling of AllocJitTempBuffer. The rest of the changes are making the code conform more to SuperPMI conventions, adding comments, and improving SPMI dumping.
41c5ca7 to
2978bb4
Compare
|
@kunalspathak @AndyAyersMS PTAL |
kunalspathak
left a comment
There was a problem hiding this comment.
Fix looks good to me. If you agree, can you also update the pipeline schedule to also trigger if there are any chances in src\coreclr\ToolBox\superpmi\*?
I was thinking about this, and I decided not to do that: I don't want to force a re-collection on the potential to just fix a bug that doesn't require a re-collection. So, I'd prefer giving developers more choice and responsibility to "do the right thing" even if that means people might forget to do a collection when it would be required. Hopefully that doesn't bite us too often. |
Sure, so I guess what you are saying is if I just sync your SPMI changes, my existing SPMI collection should work. Makes sense. |
That's not quite true: I changed the SPMI file format. However, I also changed the JIT/EE GUID, so if you pick up these changes, you need to wait until the triggered SPMI collection is published. |
The real fix is a trivial fix in the linked list handling
of AllocJitTempBuffer.
The rest of the changes are making the code conform more to SuperPMI
conventions, adding comments, and improving SPMI dumping.