On 2025-04-03 19:28, Andrew Cooper wrote:
On 04/04/2025 12:22 am, Alexander M. Merritt wrote:
The new toolchain baseline knows both the XSAVEOPT and CLWB
instructions.
I know that's what I wrote on the ticket, but what I'd forgotten was
that we only use XSAVEOPT for it's operand.
Really what
On 2025-04-03 19:34, Andrew Cooper wrote:
On 04/04/2025 12:28 am, Andrew Cooper wrote:
Also, I forgot to write in the ticket, clflushopt wants similar
treatment, even if there isn't an outward define for it. I think the
following two hunks should do:
~Andrew
diff --git a/xen/arch/x86/flushtlb.
On 04/04/2025 8:21 am, Jan Beulich wrote:
> On 04.04.2025 01:22, Alexander M. Merritt wrote:
>> The new toolchain baseline knows both the XSAVEOPT and CLWB instructions.
>>
>> Resolves: https://gitlab.com/xen-project/xen/-/work_items/205
>> Signed-off-by: Alexander M. Merritt
>> ---
>> xen/arch/x
On 04.04.2025 01:22, Alexander M. Merritt wrote:
> The new toolchain baseline knows both the XSAVEOPT and CLWB instructions.
>
> Resolves: https://gitlab.com/xen-project/xen/-/work_items/205
> Signed-off-by: Alexander M. Merritt
> ---
> xen/arch/x86/arch.mk | 2 --
> xen/arch/x86/f
On 04/04/2025 12:22 am, Alexander M. Merritt wrote:
> The new toolchain baseline knows both the XSAVEOPT and CLWB instructions.
I know that's what I wrote on the ticket, but what I'd forgotten was
that we only use XSAVEOPT for it's operand.
Really what we're doing here is knowing CLWB, and also g
On 04/04/2025 12:28 am, Andrew Cooper wrote:
> On 04/04/2025 12:22 am, Alexander M. Merritt wrote:
>> The new toolchain baseline knows both the XSAVEOPT and CLWB instructions.
> I know that's what I wrote on the ticket, but what I'd forgotten was
> that we only use XSAVEOPT for it's operand.
>
> Re
