Re: [PATCH] pv32: Fix bogus cr2 on fault in emulation gate

[Teddy Astie <teddy.astie@xxxxxxxxxx>]

Le 20/05/2026 à 18:27, Andrew Cooper a écrit :
> On 20/05/2026 4:51 pm, Teddy Astie wrote:
> > __{put,get}_guest returns -EFAULT on access faults which causes
> > the injected cr2 to be off by 14 bytes (as EFAULT is 14) which is
> > incorrect.
> >
> > Fix the computation by relying on copy_{from,to}_guest_pv which
> > reports the number of remaining bytes instead of a negative errno,
> > such that we can compute the offset properly.
> >
> > Fixes: 70ad570b2799 ("x86/64: paravirt 32-on-64 call gate support")
> > Signed-off-by: Teddy Astie <teddy.astie@xxxxxxxxxx>
> Given it was __*_guest() before, I think we can use the
> __copy_*_guest_pv() variants.
>
> I can fix on commit if you're happy?  Reviewed-by: Andrew Cooper
> <andrew.cooper3@xxxxxxxxxx>
Looks good to me.

> Jan, this wants committing ahead of your MISRA change, as it needs
> backporting.
>
> Strictly speaking, gate emulation is all PV right now.  It does want
> moving behind CONFIG_PV32.
This file is already gated behind CONFIG_PV32, so I think it's already 
the case.
> Also, with these two callers dropped, all remaining users of
> __{get,put}_guest() only care about success/failure rather than the
> precise value, so there is probably some cleanup which can be done.
>
> ~Andrew
Teddy

Attachment: OpenPGP_0x660FA9D102CBCFD0.asc
Description: OpenPGP public key

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature