Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 18 May 2021 15:29:10 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Like Xu <like.xu@linux.intel.com>
Cc:     Kan Liang <kan.liang@linux.intel.com>,
        Ingo Molnar <mingo@redhat.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Jiri Olsa <jolsa@redhat.com>,
        Namhyung Kim <namhyung@kernel.org>,
        linux-kernel@vger.kernel.org, Borislav Petkov <bp@alien8.de>,
        seanjc@google.com, x86@kernel.org
Subject: Re: [PATCH v2 2/2] perf/x86/lbr: Move cpuc->lbr_xsave allocation out
 of sleeping region
Message-ID: <YKPBJoK1b/2PIQfk@hirez.programming.kicks-ass.net>
References: <20210430052247.3079672-1-like.xu@linux.intel.com>
 <20210430052247.3079672-2-like.xu@linux.intel.com>
 <YKOQqT+Ktc4C2kQp@hirez.programming.kicks-ass.net>
 <a9032de6-e55d-bfff-e6b5-4c79dcab30f3@linux.intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <a9032de6-e55d-bfff-e6b5-4c79dcab30f3@linux.intel.com>
Precedence: bulk

On Tue, May 18, 2021 at 08:34:33PM +0800, Like Xu wrote:
> > > +	if (!static_cpu_has(X86_FEATURE_ARCH_LBR))
> > > +		return;
> > > +
> > > +	for_each_possible_cpu(cpu) {
> > > +		cpuc = per_cpu_ptr(&cpu_hw_events, cpu);
> > > +		kmem_cache = x86_get_pmu(cpu)->task_ctx_cache;
> > > +		if (kmem_cache && !cpuc->lbr_xsave && !event->attr.precise_ip)
> > > +			cpuc->lbr_xsave =
> > > +				kmem_cache_alloc_node(kmem_cache, GFP_KERNEL,
> > > +						      cpu_to_node(cpu));
> > 
> > (coding style fail)
> > 
> > But then I looked at this function, and srlsy that !precise_ip is the
> > only thing you need @event for? Why do we care?
> 
> Kan once commented, we only need to alloc the buffer for the non-PEBS
> event. It seems the check "(cpuc->lbr_users != cpuc->lbr_pebs_users)"
> is implicitly removed.
> 
> I think we still need to check !precision_ip, right ?

It's just a memory allocation; who cares if we allocate but not use it
ever?