linux-kernel - Re: [tip: sched/core] sched: Fix build for modules using set_tsk_need

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] [day] [month] [year] [list]

Message-ID: <20260128224843.GA185446@pauld.westford.csb>
Date: Wed, 28 Jan 2026 17:48:43 -0500
From: Phil Auld <pauld@...hat.com>
To: linux-kernel@...r.kernel.org
Cc: linux-tip-commits@...r.kernel.org, Gabriele Monaco <gmonaco@...hat.com>,
	"Peter Zijlstra (Intel)" <peterz@...radead.org>, x86@...nel.org
Subject: Re: [tip: sched/core] sched: Fix build for modules using
 set_tsk_need_resched()

Hi Peter and Gabriele,

On Thu, Jan 15, 2026 at 06:38:03PM -0500 Phil Auld wrote:
> On Thu, Jan 15, 2026 at 09:44:19PM -0000 tip-bot2 for Gabriele Monaco wrote:
> > The following commit has been merged into the sched/core branch of tip:
> > 
> > Commit-ID:     8d737320166bd145af70a3133a9964b00ca81cba
> > Gitweb:        https://git.kernel.org/tip/8d737320166bd145af70a3133a9964b00ca81cba
> > Author:        Gabriele Monaco <gmonaco@...hat.com>
> > AuthorDate:    Mon, 12 Jan 2026 15:04:13 +01:00
> > Committer:     Peter Zijlstra <peterz@...radead.org>
> > CommitterDate: Thu, 15 Jan 2026 22:41:26 +01:00
> >
> 
> Thank you for both of these, Peter!

Well, I think this one and the related one need to lose the _GPL because
they are called from things, like wake_up_process(), which are exported
without that restriction. This essentially adds _GPL to that and
anything else that ends up in resched_curr().

I prefer to add new symbols _GPL too but this case seems different.

Thoughts?

I can spin up a patch against tip:sched/core if you want it.


Cheers,
Phil

> 
> > sched: Fix build for modules using set_tsk_need_resched()
> > 
> > Commit adcc3bfa8806 ("sched: Adapt sched tracepoints for RV task model")
> > added a tracepoint to the need_resched action that can be triggered also
> > by set_tsk_need_resched.
> > This function was previously accessible from out-of-tree modules but
> > it's no longer available because the __trace_set_need_resched() symbol
> > is not exported (together with the tracepoint itself, which was exported
> > in a separate patch) and building such modules fails.
> > 
> > Export __trace_set_need_resched to modules to fix those build issues.
> > 
> > Fixes: adcc3bfa8806 ("sched: Adapt sched tracepoints for RV task model")
> > Signed-off-by: Gabriele Monaco <gmonaco@...hat.com>
> > Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
> > Reviewed-by: Phil Auld <pauld@...hat.com>
> > Link: https://patch.msgid.link/20260112140413.362202-1-gmonaco@redhat.com
> > ---
> >  kernel/sched/core.c | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index b033f97..3cca012 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -1139,6 +1139,7 @@ void __trace_set_need_resched(struct task_struct *curr, int tif)
> >  {
> >  	trace_sched_set_need_resched_tp(curr, smp_processor_id(), tif);
> >  }
> > +EXPORT_SYMBOL_GPL(__trace_set_need_resched);
> >  
> >  void resched_curr(struct rq *rq)
> >  {
> > 
> 
> -- 
> 

--