Remove cached RW mapping when the corresponding RX one is released #82841
+55
−32
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
When a RX mapping in the ExecutableAllocation is released and there is a RW mapping cached for it, it incorrectly stays in the cache. So when the same virtual address range that was just released gets reserved again and a request to get RW mapping comes in, the cached RW mapping is used. But it is likely that the new RX mapping has a different offset in the underlying file mapping and thus the RW mapping is unrelated to the new RX mapping. Using this RW mapping results either in an overwrite of a different block of memory or just a silent dropping of what's written. This was discovered when investigating a GC reliability framework crash. It turned out it was also the culprit behind the dotnet#75244 issue that I was unable to reproduce and it kept occuring in the CI on x86 only for quite some time. In addition to the fix, I've found that there was an off by one condition in the ExecutableAllocator::FindRWBlock so I've fixed that. And I've also noticed that in UnlockedLoaderHeap::UnlockedReservePages, if the allocation of LoaderHeapBlock failed, we would leave the data block incorrectly on the m_pRangeList. I've fixed that too. Close dotnet#75244
janvorli
force-pushed
the
fix-caching-after-release
branch
from
c847589 to
ba4592c
Compare
|
cc: @Maoni0 this is the fix for the GC reliability framework crash. |
|
nice! thank you @janvorli! |
|
/azp run runtime-coreclr outerloop |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
/backport to release/7.0 |
|
Started backporting to release/7.0: https://github.com/dotnet/runtime/actions/runs/4324991438 |
|
@janvorli backporting to release/7.0 failed, the patch most likely resulted in conflicts: $ git am --3way --ignore-whitespace --keep-non-patch changes.patch
Applying: Remove cached RW mapping when the corresponding RX one is released
Using index info to reconstruct a base tree...
M src/coreclr/utilcode/executableallocator.cpp
M src/coreclr/utilcode/loaderheap.cpp
Falling back to patching base and 3-way merge...
Auto-merging src/coreclr/utilcode/loaderheap.cpp
CONFLICT (content): Merge conflict in src/coreclr/utilcode/loaderheap.cpp
Auto-merging src/coreclr/utilcode/executableallocator.cpp
CONFLICT (content): Merge conflict in src/coreclr/utilcode/executableallocator.cpp
error: Failed to merge in the changes.
hint: Use 'git am --show-current-patch=diff' to see the failed patch
Patch failed at 0001 Remove cached RW mapping when the corresponding RX one is released
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".
Error: The process '/usr/bin/git' failed with exit code 128Please backport manually! |
|
@janvorli an error occurred while backporting to release/7.0, please check the run log for details! Error: git am failed, most likely due to a merge conflict. |
ghost
locked as resolved and limited conversation to collaborators
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
When a RX mapping in the ExecutableAllocation is released and there is a RW mapping cached for it, it incorrectly stays in the cache. So when the same virtual address range that was just released gets reserved again and a request to get RW mapping comes in, the cached RW mapping is used. But it is likely that the new RX mapping has a different offset in the underlying file mapping and thus the RW mapping is unrelated to the new RX mapping. Using this RW mapping results either in an overwrite of a different block of memory or just a silent dropping of what's written.
This was discovered when investigating a GC reliability framework crash. It turned out it was also the culprit behind the #75244 issue that I was unable to reproduce and it kept occuring in the CI on x86 only for quite some time.
In addition to the fix, while investigating the original issue, I've noticed that in
UnlockedLoaderHeap::UnlockedReservePages,if the allocation of
LoaderHeapBlockfailed, we would leave the data block incorrectly on them_pRangeList. I've fixed that too.Close #75244