[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] SVM: avoid VMSAVE in ctxt-switch-to

  • To: Jan Beulich <jbeulich@xxxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • From: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
  • Date: Mon, 19 Oct 2020 15:10:14 +0100
  • Authentication-results: esa2.hc3370-68.iphmx.com; dkim=none (message not signed) header.i=none
  • Delivery-date: Mon, 19 Oct 2020 14:11:16 +0000
  • Ironport-sdr: vLVREriFjx0v8ejZNAN8fPnm3PHtWYqAF3XvpoLdTRi5TGNpr+ziIhwt6+vG/2vUT6njaNnT1u PhGLPX2r0Fv1yAYQgS+jEzOQilZlHLok2setgqlN5xFsrk/x+V3DQtV3K+IKoHKH9GpXboLLHe T+L16yN4gbG8ZeONtd4tAdbDPdc45WF5/YPwDuXK9V927gNJJ0UjZDuGxhwIJEzC3G9NEGZzPA NQBZ3HMjmDJQqOhWtR8dvGyGU5YgCGOqgWbwBi2xKtq0dZttWH2zb8ftC/GnqYX1gT9pHxFI3O iyk=
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 19/10/2020 14:40, Jan Beulich wrote:
> Of the state saved by the insn and reloaded by the corresponding VMLOAD
> - TR, syscall, and sysenter state are invariant while having Xen's state
>   loaded,
> - FS, GS, and LDTR are not used by Xen and get suitably set in PV
>   context switch code.

I think it would be helpful to state that Xen's state is suitably cached
in _svm_cpu_up(), as this does now introduce an ordering dependency
during boot.

Is it possibly worth putting an assert checking the TR selector?  That
ought to be good enough to catch stray init reordering problems.

> Suggested-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>

Either way, Acked-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.