CVE-2026-89753

In the Linux kernel, the following vulnerability has been resolved:

mm/vmscan: report RCU-tasks quiescent states in shrink lruvec()

I am seeing some rcu tasks stalls in the Meta fleet during reclaim.

INFO: rcu tasks detected stalls on tasks: 0000000088620d09: .. nvcsw: 6735/6735 holdout: 1 idle cpu: -1/8 task:GlobalCPUThread state:R running task pid:2552016 tgid:2524552 Call Trace: shrink lruvec mem cgroup iter shrink node do try to free pages try to free pages alloc frozen pages noprof alloc pages noprof pte alloc one pte alloc handle mm fault

Nothing promises direct reclaim returns in bounded time, and the scan loop in shrink lruvec() only calls cond resched(), which is a no-op on PREEMPTION kernels. Involuntary preemption is not a Tasks-RCU quiescent state, so the reclaiming task never reports one and becomes a holdout.

Upgrade it to cond resched tasks rcu qs(), which reports a quiescent state even when cond resched() does nothing.

PS: This has been discussed in [1]