Fix RecyclerVisited related crashes during marking #4290
Merged
Conversation
This file contains hidden or 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
(courtesy of Bo Cupp) Fixing two issues in marking. 1. If there are only precisely traced objects allocated, the markStack used for non-precisely traced objects will be empty and crash when Pop is called. An IsEmpty check was added to prevent popping from an empty stack. 2. A race between allocating and background marking can lead to calling a virtual Trace method on an object that has a null v-table pointer. It isn't possible to call trace until after a flag has been set by the allocator, so prior to that happening a dummy v-table pointer is now installed until control returns to the caller of the recycler's alloc method where a final v-table pointer will overwrite the dummy one. This must be done with a MemoryBarrier on ARM to guarantee the writing of the v-table happens before the attribute updates. We also need to fixup the address in FillCheckPad We were tripping an assert that the v-table doesn't match the padding pattern. Because we install the v-table earlier, there's a new mode for FillCheckPad where the first void* bytes should be skipped. The object is not initialized, but already has a valid v-table that doesn't need to be verified and shouldn't be 0-filled.
curtisman
approved these changes
Nov 22, 2017
chakrabot
pushed a commit
that referenced
this pull request
Nov 22, 2017
…king Merge pull request #4290 from dlibby-:build/dlibby/dummyvtable_fix_1.8 (courtesy of Bo Cupp) Fixing two issues in marking. 1. If there are only precisely traced objects allocated, the markStack used for non-precisely traced objects will be empty and crash when Pop is called. An IsEmpty check was added to prevent popping from an empty stack. 2. A race between allocating and background marking can lead to calling a virtual Trace method on an object that has a null v-table pointer. It isn't possible to call trace until after a flag has been set by the allocator, so prior to that happening a dummy v-table pointer is now installed until control returns to the caller of the recycler's alloc method where a final v-table pointer will overwrite the dummy one. This must be done with a MemoryBarrier on ARM to guarantee the writing of the v-table happens before the attribute updates. We also need to fixup the address in FillCheckPad We were tripping an assert that the v-table doesn't match the padding pattern. Because we install the v-table earlier, there's a new mode for FillCheckPad where the first void* bytes should be skipped. The object is not initialized, but already has a valid v-table that doesn't need to be verified and shouldn't be 0-filled.
chakrabot
pushed a commit
that referenced
this pull request
Nov 22, 2017
…es during marking Merge pull request #4290 from dlibby-:build/dlibby/dummyvtable_fix_1.8 (courtesy of Bo Cupp) Fixing two issues in marking. 1. If there are only precisely traced objects allocated, the markStack used for non-precisely traced objects will be empty and crash when Pop is called. An IsEmpty check was added to prevent popping from an empty stack. 2. A race between allocating and background marking can lead to calling a virtual Trace method on an object that has a null v-table pointer. It isn't possible to call trace until after a flag has been set by the allocator, so prior to that happening a dummy v-table pointer is now installed until control returns to the caller of the recycler's alloc method where a final v-table pointer will overwrite the dummy one. This must be done with a MemoryBarrier on ARM to guarantee the writing of the v-table happens before the attribute updates. We also need to fixup the address in FillCheckPad We were tripping an assert that the v-table doesn't match the padding pattern. Because we install the v-table earlier, there's a new mode for FillCheckPad where the first void* bytes should be skipped. The object is not initialized, but already has a valid v-table that doesn't need to be verified and shouldn't be 0-filled.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
(courtesy of Bo Cupp)
Fixing two issues in marking.
We also need to fixup the address in FillCheckPad
We were tripping an assert that the v-table doesn't match the padding pattern. Because we install
the v-table earlier, there's a new mode for FillCheckPad where the first void* bytes should be skipped.
The object is not initialized, but already has a valid v-table that doesn't need to be verified and shouldn't
be 0-filled.