Fold the partition select into the sized-free tag compare in tcmalloc.cc. - #1000
Draft
copybara-service[bot] wants to merge 1 commit into
Draft
copybara-service[bot] wants to merge 1 commit into
copybara-service[bot] wants to merge 1 commit into
Conversation
….cc. do_free_with_size previously tested (uptr & kNormalOrBadDeallocationMask) != kNormalMask (accepting both kNormalP0 and kNormalP1 via their shared bit) and then had fast_free_with_size re-derive the partition from the pointer with PartitionFromPointerFast (x86: btq $43; jb / aarch64: tbnz). Heap partitioning defaults to off, so essentially every sized free is P0. Widen the mask to kTagOrBadDeallocationMask and compare against the exact kNormalP0 tag value: the not-taken fallthrough now is P0, and the partition is known without a second test. On the taken side, test the exact kNormalP1 value first (a partitioned heap still pays two compares, as before), then nullptr (tag 0 fails both compares), cold, and the sampled/illformed handler. fast_free_with_size takes the partition as a template parameter, which also removes the duplicated GetSizeClass/FreeSmall arms; kNormalPartitions == 1 builds compile the P1 test out. x86 TCMallocInternalDeleteSized drops btq/jb (-2 instructions, -1 cmp/-1 jcc in the fast_path goldens for delete(size) and delete(size,align)); aarch64 drops the tbnz (-1 jcc). No other golden rows change. The cold-hint path gains one compare (cycles unchanged in the benchmark). PiperOrigin-RevId: 983583823
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fold the partition select into the sized-free tag compare in tcmalloc.cc.
do_free_with_size previously tested (uptr & kNormalOrBadDeallocationMask) !=
kNormalMask (accepting both kNormalP0 and kNormalP1 via their shared bit) and
then had fast_free_with_size re-derive the partition from the pointer with
PartitionFromPointerFast (x86: btq $43; jb / aarch64: tbnz).
Heap partitioning defaults to off, so essentially every sized free is P0.
Widen the mask to kTagOrBadDeallocationMask and compare against the exact
kNormalP0 tag value: the not-taken fallthrough now is P0, and the partition is
known without a second test. On the taken side, test the exact kNormalP1 value
first (a partitioned heap still pays two compares, as before), then nullptr
(tag 0 fails both compares), cold, and the sampled/illformed handler.
fast_free_with_size takes the partition as a template parameter, which also
removes the duplicated GetSizeClass/FreeSmall arms; kNormalPartitions == 1
builds compile the P1 test out.
x86 TCMallocInternalDeleteSized drops btq/jb (-2 instructions, -1 cmp/-1 jcc in
the fast_path goldens for delete(size) and delete(size,align)); aarch64 drops
the tbnz (-1 jcc). No other golden rows change. The cold-hint path gains one
compare (cycles unchanged in the benchmark).