Re: [PATCH] x86/cet: Clear IST supervisor token busy bits on S3 resume

[Jan Beulich <jbeulich@xxxxxxxx>]

On 14.03.2022 12:00, Andrew Cooper wrote:
> Stacks are not freed across S3.  Execution just stops, leaving supervisor
> token busy bits active.  Fixing this for the primary shadow stack was done
> previously, but there is a (rare) risk that an IST token is left busy too.
> This will manifest as #DF next time the IST vector gets used.

Under what (rare) condition would this happen? The only scenario I could
come up with (which wouldn't result in a crash anyway) is the NMI watchdog
hitting after a CPU was already taken offline, and the handler not
managing to complete before power is cut. I think it would help to mention
one such specific case.

> --- /dev/null
> +++ b/xen/arch/x86/include/asm/shstk.h
> @@ -0,0 +1,46 @@
> +/******************************************************************************
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation; either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; If not, see <http://www.gnu.org/licenses/>.
> + *
> + * Copyright (c) 2022 Citrix Systems Ltd.
> + */
> +#ifndef XEN_ASM_SHSTK_H
> +#define XEN_ASM_SHSTK_H
> +
> +/*
> + * RDSSP is a nop when shadow stacks are active.

I guess there's a "not" missing here, supported by ...

>  Also, SSP has a minimum
> + * alignment of 4 which enforced by hardware.
> + *
> + * We load 1 into a register, then RDSSP.  If shadow stacks are not active,
> + * RDSSP is a nop, and the 1 is preserved.

... this. As an alternative I wouldn't mind if you removed the redundancy.
Then
Reviewed-by: Jan Beulich <jbeulich@xxxxxxxx>

Jan