[<prev] [next>] [day] [month] [year] [list]
Message-ID: <14a3278e-0299-8138-ff73-654d0208bdfa@redhat.com>
Date: Mon, 27 Jun 2022 09:25:54 -0400
From: Waiman Long <longman@...hat.com>
To: Guo Hui <guohui@...ontech.com>, peterz@...radead.org
Cc: jgross@...e.com, srivatsa@...il.mit.edu, amakhalov@...are.com,
pv-drivers@...are.com, tglx@...utronix.de, mingo@...hat.com,
bp@...en8.de, dave.hansen@...ux.intel.com, x86@...nel.org,
hpa@...or.com, will@...nel.org, boqun.feng@...il.com,
virtualization@...ts.linux-foundation.org,
wangxiaohua@...ontech.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] x86/paravirt: useless assignment instructions cause
Unixbench full core performance degradation
On 6/27/22 01:54, Guo Hui wrote:
> Thank you very much Longman, my patch is as you said, only disable
> node_cpu on X86, enable node_cpu on arm64, powerpc, s390 architectures;
> the code is in file arch/x86/kernel/paravirt-spinlocks.c:
> DECLARE_STATIC_KEY_FALSE(preemted_key);
> static_branch_enable(&preemted_key);
>
> the default value of preemted_key is false and the if conditional
> statement is reversed,
> the code is in file kernel/locking/osq_lock.c:
> DEFINE_STATIC_KEY_FALSE(preemted_key);
>
> static inline int node_cpu(struct optimistic_spin_node *node)
> {
> int cpu = 0;
>
> if (!static_branch_unlikely(&preemted_key))
> cpu = node->cpu - 1;
>
> return cpu;
> }
>
> In this way, only one nop instruction is added to architectures arm64,
> powerpc and s390, including virtual machines, without any other changes.
You are right. I am probably too tired last night to read the patch more
carefully.
Cheers,
Longman
Powered by blists - more mailing lists