From: Peter Zijlstra <peterz@infradead.org>
To: Wanpeng Li <kernellwp@gmail.com>
Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
Paolo Bonzini <pbonzini@redhat.com>,
Radim Kr??m???? <rkrcmar@redhat.com>,
Wanpeng Li <wanpeng.li@hotmail.com>
Subject: Re: [PATCH v3 2/4] KVM: Add paravirt remote TLB flush
Date: Fri, 10 Nov 2017 11:06:42 +0100 [thread overview]
Message-ID: <20171110100642.hmcapsvggzzmbrsb@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <1510307387-14812-3-git-send-email-wanpeng.li@hotmail.com>
On Fri, Nov 10, 2017 at 01:49:45AM -0800, Wanpeng Li wrote:
> +static DEFINE_PER_CPU(cpumask_var_t, __pv_tlb_mask);
> +
> +static void kvm_flush_tlb_others(const struct cpumask *cpumask,
> + const struct flush_tlb_info *info)
> +{
> + u8 state;
> + int cpu;
> + struct kvm_steal_time *src;
> + struct cpumask *flushmask = this_cpu_cpumask_var_ptr(__pv_tlb_mask);
> +
> + if (unlikely(!flushmask))
> + return;
> +
> + cpumask_copy(flushmask, cpumask);
> + /*
> + * We have to call flush only on online vCPUs. And
> + * queue flush_on_enter for pre-empted vCPUs
> + */
> + for_each_cpu(cpu, cpumask) {
> + src = &per_cpu(steal_time, cpu);
> + state = src->preempted;
I think that wants to be: state = READ_ONCE(src->preempted);
Because without that its possible for state to get re-loaded between the
check here:
> + if ((state & KVM_VCPU_PREEMPTED)) {
and its use here.
> + if (cmpxchg(&src->preempted, state, state |
> + KVM_VCPU_SHOULD_FLUSH) == state)
You can actually write that like:
if (try_cmpxchg(&src->preempted, state,
state | KVM_VCPU_SHOULD_FLUSH))
Which should generate ever so slightly better code (it uses the cmpxchg
ZF instead of doing a superfluous compare).
> + __cpumask_clear_cpu(cpu, flushmask);
> + }
> + }
> +
> + native_flush_tlb_others(flushmask, info);
> +}
next prev parent reply other threads:[~2017-11-10 10:06 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-11-10 9:49 [PATCH v3 0/4] KVM: Paravirt " Wanpeng Li
2017-11-10 9:49 ` [PATCH v3 1/4] KVM: Add vCPU running/preempted state Wanpeng Li
2017-11-10 9:49 ` [PATCH v3 2/4] KVM: Add paravirt remote TLB flush Wanpeng Li
2017-11-10 10:06 ` Peter Zijlstra [this message]
2017-11-10 9:49 ` [PATCH v3 3/4] KVM: X86: introduce invalidate_gpa argument to tlb flush Wanpeng Li
2017-11-10 9:49 ` [PATCH v3 4/4] KVM: Add flush_on_enter before guest enter Wanpeng Li
2017-11-10 10:07 ` Peter Zijlstra
2017-11-13 0:38 ` Wanpeng Li
2017-11-10 10:08 ` Peter Zijlstra
2017-11-10 10:15 ` Paolo Bonzini
2017-11-10 10:31 ` Peter Zijlstra
2017-11-10 16:49 ` Paolo Bonzini
2017-11-10 12:43 ` [PATCH v3 0/4] KVM: Paravirt remote TLB flush David Hildenbrand
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20171110100642.hmcapsvggzzmbrsb@hirez.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=kernellwp@gmail.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=rkrcmar@redhat.com \
--cc=wanpeng.li@hotmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®