[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Y35yGltCxrwueOqn@google.com>
Date: Wed, 23 Nov 2022 19:18:50 +0000
From: Sean Christopherson <seanjc@...gle.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: "Li, Xin3" <xin3.li@...el.com>,
Paolo Bonzini <pbonzini@...hat.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"x86@...nel.org" <x86@...nel.org>,
"kvm@...r.kernel.org" <kvm@...r.kernel.org>,
"tglx@...utronix.de" <tglx@...utronix.de>,
"mingo@...hat.com" <mingo@...hat.com>,
"bp@...en8.de" <bp@...en8.de>,
"dave.hansen@...ux.intel.com" <dave.hansen@...ux.intel.com>,
"hpa@...or.com" <hpa@...or.com>,
"Tian, Kevin" <kevin.tian@...el.com>
Subject: Re: [RESEND PATCH 5/6] KVM: x86/VMX: add kvm_vmx_reinject_nmi_irq()
for NMI/IRQ reinjection
On Wed, Nov 23, 2022, Peter Zijlstra wrote:
> On Tue, Nov 22, 2022 at 08:52:35PM +0000, Sean Christopherson wrote:
>
> > Another wart that needs to be addressed is trace_kvm_exit(). IIRC, tracepoints
> > must be outside of noinstr, though maybe I'm misremembering that.
>
> You are not, that is correct. Another point to be careful with is usage
> of jump_label and static_call, both can be used in noinstr *provided*
> they don't actually ever change -- so boot time setup only.
>
> If either of them were to change, text_poke_bp() has a clue in the name.
I think we're mostly ok on that front. kvm_wait_lapic_expire() consumes multiple
static keys that can change at will, but that can be kept outside of the noinstr
section.
Thanks!
Powered by blists - more mailing lists