sched/core: Make core-sched flips wait for in-flight selections
Linux
Affected versions
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
Linux
| Product | Affected | Fixed in |
|---|---|---|
| Linux Linux |
|
≥ 539f65125d20aacab54d02d77f10a839f45b09dc · < f3629c63a4af3e491381780bc6c123cb498c4c40| No fix yet |
| Linux Linux |
|
|
In the Linux kernel, the following vulnerability has been resolved:
sched/core: Make core-sched flips wait for in-flight selections
Core scheduling's pick_next_task() operates on all sibling rqs under one acquisition of the shared core-wide lock. A ->pick_task() that releases the rq lock leaves every sibling __lock momentarily free, letting __sched_core_flip(false) complete mid-selection and rebind rq_lockp() under it. The selection resumes on the split locks, touching sibling state it no longer protects, and __schedule() finally releases a lock that was never taken while leaking the one that was.
Count in-flight core-wide selections in the leader's rq->core_pick_in_flight and make __sched_core_flip() wait for the count to drain. The count only changes under the shared lock, which the flip holds while sampling, so no other ordering is needed. The wait can repeat while selections overlap, but the flip backs off between samples and flips are rare cookie-lifetime events.
sched_core_cpu_deactivate() moves the count to the new leader - a stale copy left behind would bias it forever if that CPU later returns as its own leader.
From the CVE record · Linux
A race in Linux kernel core scheduling can let a core-scheduling flip occur during an in-flight task selection. The selection may then access sibling state without the expected lock protection and release the wrong lock. The input does not specify an attacker path, authentication requirements, or user interaction.
Engineers running the Linux kernel should check their versions against the vendor advisory; affected versions are not specified in the input.