[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20251001114702.06e2b1d8@gandalf.local.home>
Date: Wed, 1 Oct 2025 11:47:02 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Peter Zijlstra <peterz@...radead.org>
Cc: jpoimboe@...nel.org, rostedt@...nel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 09/12] unwind: Make unwind_task_info::unwind_mask
consistent
On Wed, 24 Sep 2025 09:59:57 +0200
Peter Zijlstra <peterz@...radead.org> wrote:
> @@ -324,7 +324,8 @@ void unwind_deferred_cancel(struct unwin
> guard(rcu)();
> /* Clear this bit from all threads */
> for_each_process_thread(g, t) {
> - clear_bit(bit, &t->unwind_info.unwind_mask);
> + atomic_long_andnot(UNWIND_USED,
> + &t->unwind_info.unwind_mask);
Shouldn't this be:
atomic_long_andnot(BIT(bit), &t->unwind_info.unwind_mask);
?
As BIT(bit) != UNWIND_USED.
-- Steve
> if (t->unwind_info.cache)
> clear_bit(bit, &t->unwind_info.cache->unwind_completed);
> }
Powered by blists - more mailing lists