[Deepin-Kernel-SIG] [linux 6.18.y] [Backport] sched/fair: Revert 6d71a9c ("sched/fair: Fix EEVDF entity placement b… …ug causing scheduling lag")#1909
Open
opsiff wants to merge 3 commits into
Conversation
This reverts commit 4ed4cc0. Signed-off-by: Wentao Guan <guanwentao@uniontech.com>
…ug causing scheduling lag") [ Upstream commit 101f3498b4bdfef97152a444847948de1543f692 ] Zicheng Qu reported that, because avg_vruntime() always includes cfs_rq->curr, when ->on_rq, place_entity() doesn't work right. Specifically, the lag scaling in place_entity() relies on avg_vruntime() being the state *before* placement of the new entity. However in this case avg_vruntime() will actually already include the entity, which breaks things. Also, Zicheng Qu argues that avg_vruntime should be invariant under reweight. IOW commit 6d71a9c ("sched/fair: Fix EEVDF entity placement bug causing scheduling lag") was wrong! The issue reported in 6d71a9c could possibly be explained by rounding artifacts -- notably the extreme weight '2' is outside of the range of avg_vruntime/sum_w_vruntime, since that uses scale_load_down(). By scaling vruntime by the real weight, but accounting it in vruntime with a factor 1024 more, the average moves significantly. However, that is now cured. Tested by reverting 66951e4 ("sched/fair: Fix update_cfs_group() vs DELAY_DEQUEUE") and tracing vruntime and vlag figures again. Reported-by: Zicheng Qu <quzicheng@huawei.com> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org> Tested-by: K Prateek Nayak <kprateek.nayak@amd.com> Tested-by: Shubhang Kaushik <shubhang@os.amperecomputing.com> Link: https://patch.msgid.link/20260219080625.066102672%40infradead.org (cherry picked from commit 101f3498b4bdfef97152a444847948de1543f692) [jstultz: Resolved minor collision in the revert against 6.18-stable] Signed-off-by: John Stultz <jstultz@google.com>
Signed-off-by: Wentao Guan <guanwentao@uniontech.com>
Reviewer's GuideBackports an upstream revert of a previous EEVDF placement fix and reintroduces updated BORE handling by refactoring lag/deadline computation around avg_vruntime, adding a rescale helper for reweighting, and adjusting how vruntime, vlag, deadlines, and vprot are managed during reweight, placement, and delayed requeue. Flow diagram for updated reweight_entity lag/deadline handlingflowchart TD
A[reweight_entity] --> B{se.on_rq?}
B -- no --> H[dequeue_load_avg]
B -- yes --> C[update_curr]
C --> D[avruntime = avg_vruntime]
D --> E[se.vlag = entity_lag cfs_rq se avruntime]
E --> F[se.deadline -= avruntime]
F --> G{curr && protect_slice se?}
G -- yes --> G1[se.vprot -= avruntime; rel_vprot = true]
G -- no --> H
G1 --> H[dequeue_load_avg]
H --> I[rescale_entity se weight rel_vprot]
I --> J[update_load_set se.load weight]
J --> K[recompute se.avg.load_avg]
K --> L[enqueue_load_avg]
L --> M{se.on_rq?}
M -- no --> Z[end]
M -- yes --> N{rel_vprot?}
N -- yes --> O[se.vprot += avruntime]
N -- no --> P
O --> P[se.deadline += avruntime]
P --> Q[se.rel_deadline = 0]
Q --> R[se.vruntime = avruntime - se.vlag]
R --> S[update_load_add cfs_rq.load se.load.weight]
S --> T{curr?}
T -- no --> U[__enqueue_entity]
T -- yes --> Z[end]
U --> Z[end]
Flow diagram for DELAY_ZERO requeue_delayed_entity with BORE handlingflowchart TD
A[requeue_delayed_entity] --> B{DELAY_ZERO feature?}
B -- no --> Z[end]
B -- yes --> C[update_entity_lag]
C --> D{se.vlag > 0?}
D -- no --> Z
D -- yes --> E[cfs_rq.nr_queued--]
E --> F{se != cfs_rq.curr?}
F -- yes --> G[__dequeue_entity]
F -- no --> H
G --> H[se.vlag = 0]
H --> I{CONFIG_SCHED_BORE && sched_bore_key && DELAY_ZERO?}
I -- yes --> J[flags or= ENQUEUE_WAKEUP]
I -- no --> K[flags = 0]
J --> L[place_entity cfs_rq se flags]
K --> L[place_entity cfs_rq se flags]
L --> M{se != cfs_rq.curr?}
M -- yes --> N[__enqueue_entity]
M -- no --> O
N --> O[cfs_rq.nr_queued++]
O --> Z[end]
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Pull request overview
Backports an upstream revert in the CFS fair scheduler (EEVDF) to correct entity placement/lag behavior by making lag/deadline adjustments reference the cfs_rq average vruntime state, and by tightening when relative-deadline placement is applied.
Changes:
- Add
entity_lag()helper to compute lag from a supplied average vruntime and reuse it for lag updates. - Rework
reweight_entity()to adjustvruntime/vlag/deadline/vprotrelative toavg_vruntime(cfs_rq)instead ofse->vruntime. - Gate relative-deadline placement in
place_entity()behindPLACE_REL_DEADLINEand simplifyDELAY_ZEROdelayed requeueing logic.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
3997
to
4006
| if (se->on_rq) { | ||
| /* commit outstanding execution time */ | ||
| update_curr(cfs_rq); | ||
| update_entity_lag(cfs_rq, se); | ||
| #ifdef CONFIG_SCHED_BORE | ||
| vlag_unscaled = se->vlag; | ||
| #endif /* !CONFIG_SCHED_BORE */ | ||
| se->deadline -= se->vruntime; | ||
| avruntime = avg_vruntime(cfs_rq); | ||
| se->vlag = entity_lag(cfs_rq, se, avruntime); | ||
| se->deadline -= avruntime; | ||
| se->rel_deadline = 1; | ||
| if (curr && protect_slice(se)) { | ||
| vprot = se->vprot - se->vruntime; | ||
| se->vprot -= avruntime; | ||
| rel_vprot = true; |
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
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.
[ Upstream commit 101f3498b4bdfef97152a444847948de1543f692 ]
Zicheng Qu reported that, because avg_vruntime() always includes
cfs_rq->curr, when ->on_rq, place_entity() doesn't work right.
Specifically, the lag scaling in place_entity() relies on
avg_vruntime() being the state before placement of the new entity.
However in this case avg_vruntime() will actually already include the
entity, which breaks things.
Also, Zicheng Qu argues that avg_vruntime should be invariant under
reweight. IOW commit 6d71a9c ("sched/fair: Fix EEVDF entity
placement bug causing scheduling lag") was wrong!
The issue reported in 6d71a9c could possibly be explained by
rounding artifacts -- notably the extreme weight '2' is outside of the
range of avg_vruntime/sum_w_vruntime, since that uses
scale_load_down(). By scaling vruntime by the real weight, but
accounting it in vruntime with a factor 1024 more, the average moves
significantly. However, that is now cured.
Tested by reverting 66951e4 ("sched/fair: Fix update_cfs_group()
vs DELAY_DEQUEUE") and tracing vruntime and vlag figures again.
Reported-by: Zicheng Qu quzicheng@huawei.com
Signed-off-by: Peter Zijlstra (Intel) peterz@infradead.org
Reviewed-by: Vincent Guittot vincent.guittot@linaro.org
Tested-by: K Prateek Nayak kprateek.nayak@amd.com
Tested-by: Shubhang Kaushik shubhang@os.amperecomputing.com
Link: https://patch.msgid.link/20260219080625.066102672%40infradead.org
(cherry picked from commit 101f3498b4bdfef97152a444847948de1543f692)
[jstultz: Resolved minor collision in the revert against 6.18-stable]
Signed-off-by: John Stultz jstultz@google.com
and revert and apply new bore for this version!
Summary by Sourcery
Revert a previous EEVDF entity placement fix in the fair scheduler and adjust vruntime, lag, deadline and protection handling around entity reweighting for the 6.18.y kernel backport.
Enhancements: