Re: [PATCH] mm/mmu_notifier: Remove non_block_start/end() from notifier invocation
Jason Gunthorpe <[email protected]>
| Newsgroups | dev.linux.lists.linux-rt-devel,org.kernel.vger.kvm,org.kernel.vger.linux-kernel,org.kvack.linux-mm |
|---|---|
| Message-ID | <[email protected]> |
On Wed, Aug 12, 2026 at 06:04:47PM +0200, Paolo Bonzini wrote: > > + * Atomic-flavor readers run with preemption disabled, so every > > + * read-side critical section is bounded and running (not blocked) on > > + * some CPU, which is what makes spinning here sane: the wait is > > + * bounded by the longest such section. No index flip and no > > + * grace-period sequence update occur, so concurrent call_srcu(), > > + * synchronize_srcu() and srcu_barrier() are entirely unaffected. > > Making synchronize_srcu_atomic() spin in CONFIG_PREEMPT_RT is a > slightly tall request - it's basically putting it on the same level as > raw_spinlock. But if the MMU notifier guys really want to make OOM > notifiers atomic, this would be basically the only way to do so. IMHO it is fine if PREEMPT_RT does something else, just that the normal cases should follow the current arrangment where it is atomic > If instead it's okay to remove nonblock_start/end under > CONFIG_PREEMPT_RT, for RT kernels I'd rather avoid This seems well agreed at least Jason