linux-kernel - Re: [PATCH v5] tracing: Guard __DECLARE_TRACE() use of __DO_TRACE

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <13926791.uLZWGnKmhe@7940hx>
Date: Mon, 12 Jan 2026 15:23:13 +0800
From: Menglong Dong <menglong.dong@...ux.dev>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
 Alexei Starovoitov <alexei.starovoitov@...il.com>,
 LKML <linux-kernel@...r.kernel.org>,
 Linux trace kernel <linux-trace-kernel@...r.kernel.org>,
 bpf <bpf@...r.kernel.org>, Masami Hiramatsu <mhiramat@...nel.org>,
 "Paul E. McKenney" <paulmck@...nel.org>,
 Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
 Thomas Gleixner <tglx@...utronix.de>, Peter Zijlstra <peterz@...radead.org>,
 Linus Torvalds <torvalds@...ux-foundation.org>
Subject:
 Re: [PATCH v5] tracing: Guard __DECLARE_TRACE() use of __DO_TRACE_CALL() with
 SRCU-fast

On 2026/1/10 05:02 Steven Rostedt <rostedt@...dmis.org> write:
> On Fri, 9 Jan 2026 15:21:19 -0500
> Mathieu Desnoyers <mathieu.desnoyers@...icios.com> wrote:
> 
> > * preempt disable/enable pair:                                     1.1 ns
> > * srcu-fast lock/unlock:                                           1.5 ns
> > 
> > CONFIG_RCU_REF_SCALE_TEST=y
> > * migrate disable/enable pair:                                     3.0 ns
> > * calls to migrate disable/enable pair within noinline functions: 17.0 ns
> > 
> > CONFIG_RCU_REF_SCALE_TEST=m
> > * migrate disable/enable pair:                                    22.0 ns
> 
> OUCH! So migrate disable/enable has a much larger overhead when executed in
> a module than in the kernel? This means all spin_locks() in modules
> converted to mutexes in PREEMPT_RT are taking this hit!
> 
> It looks like it has to allow access to the rq->nr_pinned. There's a hack to
> expose this part of the rq struct for in-kernel by the following:
> 
> kernel/sched/rq-offsets.c:      DEFINE(RQ_nr_pinned, offsetof(struct rq, nr_pinned));
> 
> Then for the in-kernel code we have:
> 
> #define this_rq_raw() arch_raw_cpu_ptr(&runqueues)
> #else
> #define this_rq_raw() PERCPU_PTR(&runqueues)
> #endif
> #define this_rq_pinned() (*(unsigned int *)((void *)this_rq_raw() + RQ_nr_pinned))
> 
> Looking at the scheduler code, the rq->nr_pinned is referenced by a static
> function with:
> 
> static inline bool rq_has_pinned_tasks(struct rq *rq)
> {
> 	return rq->nr_pinned;
> }
> 
> Which is only referenced in hotplug code and a balance_push() path in load
> balancing. Does this variable really need to be in the runqueue struct?
> 
> Why not just make it a per-cpu variable. Maybe call it cpu_nr_pinned_tasks,
> and export that for all to use?
> 
> It will not only fix the discrepancy between the overhead of
> migrate_disable/enable in modules vs in-kernel. But it also removes the
> hack to expose a portion of the runqueue.

I think it's a good idea to factor out the "nr_pinned" from struct rq.
The current approach that we inline the migrate_disable is a little
obscure. The initial propose of inline migrate_disable is to optimize the
performance of bpf trampoline, so the modules are not considered.

As you said, rq_has_pinned_tasks() is the only place that use the
nr_pinned, except the migrate_disable/migrate_enable. After more
analysis, I think maybe we can do it this way:

DEFINE_PER_CPU_SHARED_ALIGNED(int, cpu_nr_pinned_tasks);

And change rq_has_pinned_tasks() to:
static inline bool rq_has_pinned_tasks(struct rq *rq)
{
	return *per_cpu_ptr(&cpu_nr_pinned_tasks, rq->cpu);
}

The "rq" in rq_has_pinned_tasks() may come from other CPU, so we
can't use "return this_cpu_read(cpu_nr_pinned_tasks)" directly.

Thanks!
Menglong Dong

> 
> -- Steve
> 
>