Re: [tip:x86/fpu] x86/fpu: Don't export __kernel_fpu_{begin,end}()

From: Borislav Petkov
Date: Tue Dec 04 2018 - 07:34:07 EST


On Tue, Dec 04, 2018 at 01:15:10PM +0100, Sebastian Andrzej Siewior wrote:
> Do you refer to
> | * - by IRQ context code to potentially use the FPU
> | * if it's unused.
>
> ? It is possible to use the FPU in IRQ context.

I mean interrupted_user_mode() where we apparently can use the FPU when
handling an IRQ from user mode.

> The FPU could be used in user-context surrounded by kernel_fpu_begin().

Right, that.

> This only disables preemption so an IRQ could interrupt it. This IRQ
> could then use the FPU or raise a SoftIRQ which would use it.
> Therefore on x86 it is required to check with irq_fpu_usable() if the

Yes, and the check that thing does is:

return !in_interrupt() ||
interrupted_user_mode() || ...

so you're either *not* in interrupt, or you've gotten the IRQ while in
user mode.

> FPU can be used. If the FPU can not be used, you have to implement
> fallback code.
>
> With the "restore FPU on return to userland" series we need to modify
> the FPU in a few places. The softirq and preemption is disabled. I
> didn't find any in-IRQ users.
> Going forward I would like to remove the in-IRQ part and
> irq_fpu_usable() and disable softirq as part of kernel_fpu_begin().

Right, and we should document all those new conditions prominently so
that people are aware.

Thx.

--
Regards/Gruss,
Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.