[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Y7VixOou0y+VPIXS@FVFF77S0Q05N>
Date: Wed, 4 Jan 2023 11:28:04 +0000
From: Mark Rutland <mark.rutland@....com>
To: Yann Sionneau <ysionneau@...ray.eu>
Cc: linux-kernel@...r.kernel.org,
Clement Leger <clement.leger@...tlin.com>,
Jules Maselbas <jmaselbas@...ray.eu>,
Julian Vetter <jvetter@...ray.eu>
Subject: Re: [RFC PATCH 13/25] kvx: Add signal handling support
On Tue, Jan 03, 2023 at 05:43:47PM +0100, Yann Sionneau wrote:
> Add sigcontext definition and signal handling support for
> kvx.
[...]
> +asmlinkage void do_work_pending(struct pt_regs *regs,
> + unsigned long thread_flags)
> +{
> + /* We are called with IRQs disabled */
> + trace_hardirqs_off();
> +
> + do {
> + if (thread_flags & _TIF_NEED_RESCHED) {
> + schedule();
> + } else {
> + local_irq_enable();
> + if (thread_flags & (_TIF_SIGPENDING | _TIF_NOTIFY_SIGNAL))
> + do_signal(regs);
> +
> + if (thread_flags & _TIF_NOTIFY_RESUME) {
> + clear_thread_flag(TIF_NOTIFY_RESUME);
> + resume_user_mode_work(regs);
> + }
> + }
> + /* Guarantee task flag atomic read */
> + local_irq_disable();
> + thread_flags = READ_ONCE(current_thread_info()->flags);
The comment here is a bit misleading; disabling IRQs doesn't affect the
atomicity of the read (and flags can be set by remote threads), so I'd
recommend removing it.
Please use read_thread_flags() to read the flags.
> + } while (thread_flags & _TIF_WORK_MASK);
> +}
Thanks,
Mark.
Powered by blists - more mailing lists