Re: [CFT PATCH 0/2] KVM: support XSAVES usage in the host

[Nadav Amit]

> On Nov 24, 2014, at 13:39, Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
> 
> 
> 
> On 23/11/2014 09:16, Nadav Amit wrote:
>> Iâll try to check it tomorrow (I donât have access to the failing machine at the moment).
> 
> Thanks, you'll need to squash this in:
> 
> diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c
> index 4c540c4719d8..0de1fae2bdf0 100644
> --- a/arch/x86/kernel/xsave.c
> +++ b/arch/x86/kernel/xsave.c
> @@ -738,3 +738,4 @@ void *get_xsave_addr(struct xsave_struct *xsave, int xstate)
> 
> 	return (void *)xsave + xstate_comp_offsets[feature];
> }
> +EXPORT_SYMBOL_GPL(get_xsave_addr);

I tested the patches but there are still problems.

Since kvm_load_guest_fpu is called before the guest_fpu is ever stored, there are 2 more problems that currently cause #GP:
1. XCOMP_BV[63] = 0
2. XSTATE_BV sets a bit (including bit 63) that is not set in XCOMP_BV (XCOMP_BV is initialised to zero).

[see SDM 13.11 "OPERATION OF XRSTORSâ]

Once I initialise XCOMP_BV to (1ull << 63) | XSTATE_BV, the guest runs successfully.
I have not checked any other qemu functionality that might be affected by the patch.

Nadav



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/