[Wip][Draft] Restore SP correctly in functions with dynamic allocas #122676
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
easyonaadit
force-pushed
the
restore-SP
branch
from
deb78e5 to
e07e8fc
Compare
easyonaadit
changed the title
easyonaadit
force-pushed
the
restore-SP
branch
from
e07e8fc to
0c47e9d
Compare
easyonaadit
changed the title
easyonaadit
force-pushed
the
restore-SP
branch
from
0c47e9d to
eb69dcb
Compare
Currently, the AMDGPU backend sets up FP and then increments SP by fixed size, from FP, in the prolog and decrements it by the same amount in the epilog. Prolog: tmp = sp + (alignment - 1) fp &= -alignment sp += frameSize + alignment Epilog: sp -= (frameSize + alignment) In the presence of dynamic alloca, this leads to incorrect restoration of SP. This patch enforces the presence of a base pointer for all functions with dynamic allocas, and SP is restored from the saved BP in the epilog. Prolog: tmp = sp + (alignment - 1) fp &= -alignment bp = sp sp += frameSize + alignment Epilog: sp += bp + frameSize + alignment sp -= (frameSize + alignment) (Note: for dynamic allocas with default alignment, SP can be restored with saved FP as well. However, for the sake of uniformity, presence of BP is enforced) Fixes: SWDEV-408164
easyonaadit
force-pushed
the
restore-SP
branch
from
eb69dcb to
c34194a
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
Currently, the AMDGPU backend sets up FP and then
increments SP by fixed size, from FP, in the prolog
and decrements it by the same amount in the epilog.
Prolog:
tmp = sp + (alignment - 1)fp &= -alignmentsp += frameSize + alignmentEpilog:
sp -= (frameSize + alignment)In the presence of dynamic alloca, this leads to incorrect restoration of SP.
This patch enforces the presence of a base pointer for all functions with dynamic allocas,
and SP is restored from the saved BP in the epilog.
Prolog:
tmp = sp + (alignment - 1)fp &= -alignmentbp = spsp += frameSize + alignmentEpilog:
sp += bp + frameSize + alignmentsp -= (frameSize + alignment)(Note: for dynamic allocas with default alignment, SP can be restored with saved FP as well.
However, for the sake of uniformity, presence of BP is enforced)