Hi Tejun,
On Wed, Aug 19, 2026 at 10:55:59AM -1000, Tejun Heo wrote:
Hello,
On Tue, Aug 18, 2026 at 04:59:17PM -0700, Ziyang Men wrote:
> > The bpf_cgroup_base_stat() takes an rstat spinlock_t, which can sleep on
> > PREEMPT_RT.
>
> Is this actually required? This doesn't really make sense to me. Shouldn't
> what SLEEPABLE mean change on RT kernels instead?
Oh sorry, I didn't notice that.
I might be wrong: this function calls the cputime_adjust(), which in turn
acquires raw_spin_lock_irqsave(), so there would be NMI deadlock in the
perf_event program. The __css_rstat_lock() take the spin_lock_irq() as well. So
maybe a SLEEPABLE tag is still necessary?
I think this is for BPF folks to answer. I don't think SLEEPABLE is needed
because of RT but yeah it wouldn't be safe to be called from nmi context.
Ok I see, let's remove the SLEEPABLE for now.
Thanks,
Ziyang
Thanks.
--
tejun
