On Sat, Jan 10, 2026 at 8:14 AM Steven Rostedt <[email protected]> wrote:
>
> On Fri, 9 Jan 2026 16:35:10 -0800
> Alexei Starovoitov <[email protected]> wrote:
>
> > migrate_enable/disable() wasn't inlined for a long time.
> > It bothered us enough, since sleepable bpf is the main user
> > of it besides RT, so we made an effort to inline it.
>
> It did bother us too. it went through lots of iterations to become more
> efficient over the years (it was really bad in the beginning while
> still in the rt-patch), and hopefully that will continue.
>
> >
> > RT, at the same time, doesn't inline rt_spin_lock() itself
> > so inlining migrate_disable() or not is not 10x at all.
> > Benchmark spin_lock on RT in-tree and in-module and I bet
> > there won't be a big difference.
>
> I'll put that on my todo list. But still, having migrate_disable a
> function for modules and 100% inlined for in-kernel code just because
> it needs access to a field in the run queue that doesn't need to be in
> the run queue seems like it should be fixed.

There was plenty of discussion with Peter regarding different
ways to inline migrate_disable. What was landed was the best
option at that point, but feel free to restart the discussion.

>
> As for tracepoints, BPF is the only one that needs migrate disable.
> It's not needed for ftrace or perf (although perf uses preempt
> disable). It should be moved into the BPF callback code as perf has its
> preempt disable in its callback code.
>
> If BPF doesn't care about the extra overhead of migrate_disable() for
> modules, then why should XFS suffer from that too?

The diff has nothing to do with bpf needs and/or bpf internals.
It's really about being a good citizen of PREEMP_RT.
bpf side already does migrate_disable,
rcu_read_lock, srcu_fast/task_trace when necessary.
Most of the time we don't rely on any external preempt state or rcu/srcu.
Removing guard(preempt_notrace)(); from tracepoint invocation
would be just fine for bpf. Simple remove will trigger bug
on cant_sleep(), but that's a trivial fix.

Reply via email to