[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CALCETrV2jEG1b7deGeyxwRrXLk30Q2nn6hstzeTpE80i+vu1mQ@mail.gmail.com>
Date: Tue, 16 Oct 2018 16:00:23 -0700
From: Andy Lutomirski <luto@...nel.org>
To: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc: LKML <linux-kernel@...r.kernel.org>, X86 ML <x86@...nel.org>,
Dave Hansen <dave.hansen@...ux.intel.com>,
Andrew Lutomirski <luto@...nel.org>,
Peter Zijlstra <peterz@...radead.org>,
Borislav Petkov <bp@...en8.de>, stable <stable@...r.kernel.org>
Subject: Re: [PATCH 3/3] x86/fpu: Save FPU registers on context switch if
there is a FPU
On Tue, Oct 16, 2018 at 1:25 PM Sebastian Andrzej Siewior
<bigeasy@...utronix.de> wrote:
>
> Booting a 486 with "no387 nofxsr" ends with
> | math_emulate: 0060:c101987d
> | Kernel panic - not syncing: Math emulation needed in kernel
>
> on the first context switch in user land. The reason is that
> copy_fpregs_to_fpstate() tries `fnsave' which does not work. This
> happens since commit f1c8cd0176078 ("x86/fpu: Change fpu->fpregs_active
> users to fpu->fpstate_active").
>
> Add a check for X86_FEATURE_FPU before trying to save FPU registers (we
> have such a check switch_fpu_finish() already).
>
> Fixes: f1c8cd0176078 ("x86/fpu: Change fpu->fpregs_active users to fpu->fpstate_active")
> Cc: stable@...r.kernel.org
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Reviewed-by: Andy Lutomirski <luto@...nel.org>
Powered by blists - more mailing lists