On 03.12.19 23:30, Andrew Cooper wrote:
VT-x updates RF before vmexit, so eflags written into the outgoing TSS happens
to be correct. SVM does not update RF before vmexit, and instead provides it
via a bit in exitinfo2.
In practice, needing RF set in the outgoing state occurs when a task gate
> From: Andrew Cooper
> Sent: Wednesday, December 4, 2019 6:31 AM
>
> VT-x updates RF before vmexit, so eflags written into the outgoing TSS
> happens
> to be correct. SVM does not update RF before vmexit, and instead provides
> it
> via a bit in exitinfo2.
>
> In practice, needing RF set in
On 03.12.2019 23:30, Andrew Cooper wrote:
> VT-x updates RF before vmexit, so eflags written into the outgoing TSS happens
> to be correct. SVM does not update RF before vmexit, and instead provides it
> via a bit in exitinfo2.
>
> In practice, needing RF set in the outgoing state occurs when a
VT-x updates RF before vmexit, so eflags written into the outgoing TSS happens
to be correct. SVM does not update RF before vmexit, and instead provides it
via a bit in exitinfo2.
In practice, needing RF set in the outgoing state occurs when a task gate is
used to handle faults.
Extend
