Re: [PATCH 4/5] x86/HVM: drop tsc_scaling.setup() hook

From: "Roger Pau Monné" <roger.pau@citrix.com>
To: Jan Beulich <jbeulich@suse.com>
Cc: "xen-devel@lists.xenproject.org" <xen-devel@lists.xenproject.org>,
	Andrew Cooper <andrew.cooper3@citrix.com>, Wei Liu <wl@xen.org>,
	Kevin Tian <kevin.tian@intel.com>,
	Jun Nakajima <jun.nakajima@intel.com>
Subject: Re: [PATCH 4/5] x86/HVM: drop tsc_scaling.setup() hook
Date: Wed, 22 Nov 2023 10:06:37 +0100	[thread overview]
Message-ID: <ZV3EnVFk7DtlLyXS@macbook.local> (raw)
In-Reply-To: <d5b7124f-b7cd-4a3a-b12f-e8e315e9f89d@suse.com>

On Thu, Nov 16, 2023 at 02:32:47PM +0100, Jan Beulich wrote:
> This was used by VMX only, and the intended VMCS write can as well
> happen from vmx_set_tsc_offset(), invoked (directly or indirectly)
> almost immediately after the present call sites of the hook.
> vmx_set_tsc_offset() isn't invoked frequently elsewhere, so the extra
> VMCS write shouldn't raise performance concerns.
> 
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Acked-by: Roger Pau Monné <roger.pau@citrix.com>

I would be lying if I told you I understand guest TSC management in
Xen, but patch does keep the previous call sites with adding others
that are indeed not hot paths.

I do wonder whether it would be possible to set this before the domain
is started, as AFAICT the TSC scaling ratio is only set before the
domain is started (either by domain create or other toolstack
hypercalls).

Could we maybe further limit the write to the case where the vCPU is
not running?

Thanks, Roger.