sched: add 16-HWT check_sanity for V81, fix MAX_HTHREADS table#36
Open
zbelinsk wants to merge 3 commits into
Open
sched: add 16-HWT check_sanity for V81, fix MAX_HTHREADS table#36zbelinsk wants to merge 3 commits into
zbelinsk wants to merge 3 commits into
Conversation
v81opt.S was a symlink to v73opt.S, which only ran vrmaxh over 2 batches (runlist_prios[0..7], 8 threads max). On V81 hardware with up to 16 HW threads the scheduler was silently ignoring threads 8-15 when picking the next thread to run. Fix by: - Adding a real check_sanity.v81opt.S with 4-batch vrmaxh covering runlist_prios[0..15]. - Correcting MAX_HTHREADS: V68/V73=8, V81+=16 (was V60+=16, which was wrong for V68/V73 and masked the missing v81 implementation). - Replacing the hardcoded 0xffff hthreads_mask ANDs in setup.ref.c and hwconfig.ref.c with (1U << MAX_HTHREADS) - 1. - Updating register map comments in v4opt.S, v68opt.S, v81opt.S to reflect actual register usage (old comments were incorrect). - replacing hardcoded offsets for check_sanity.v4opt.S, check_sanity.v68opt.S, check_sanity.v81opt.S. - v68opt.S promoted from symlink to real file (owns the 8-HWT path). - testlist.v81 symlinked to testlist.v73; futex/pi test now passes since the scheduler correctly finds ready threads. Signed-off-by: Zeev Belinsky <zbelinsk@qti.qualcomm.com>
Resolve type-change conflict on scripts/testlist.v81: keep this branch's symlink to testlist.v73. master's edit (enabling time/timer/test_standalone) is preserved because the v81->v73->...->v61 chain already includes it. Signed-off-by: Zeev Belinsky <zbelinsk@qti.qualcomm.com>
…ched Signed-off-by: Zeev Belinsky <zbelinsk@qti.qualcomm.com>
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.
On V81 hardware (up to 16 HW threads),
check_sanity.v81opt.Swas a symlink tov73opt.S,which only scanned
runlist_prios[0..7](8 threads). Threads 8–15 were invisible to thescheduler — silently never preempted or rescheduled.
This PR fixes that root cause and cleans up the surrounding infrastructure:
check_sanity.v81opt.S: real 4-batchvrmaxhimplementation covering all 16 HW threadsMAX_HTHREADStable corrected: V68/V73=8, V81+=16 (previously V60+=16, which was wrong and hid the missing implementation)hthreads_mask &= (1U << MAX_HTHREADS) - 1replaces three hardcoded0xffffinsetup.ref.candhwconfig.ref.cv68opt.Spromoted from symlink to real filetestlist.v81→ symlink totestlist.v73;futex/pitest now passes as the scheduler correctly finds all ready threads