mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: Robert Richter <rric@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>, Josh Boyer <jwboyer@redhat.com>,
	Ingo Molnar <mingo@redhat.com>,
	Arnaldo Carvalho de Melo <acme@ghostprotocols.net>,
	x86@kernel.org, linux-kernel@vger.kernel.org, gleb@redhat.com,
	Jacob Shin <jacob.shin@amd.com>
Subject: Re: Drop WARN on AMD lack of perfctrs
Date: Tue, 21 May 2013 13:05:37 +0200	[thread overview]
Message-ID: <20130521110537.GG26912@twins.programming.kicks-ass.net> (raw)
In-Reply-To: <20130521085630.GA1860@rric.localhost>

On Tue, May 21, 2013 at 10:56:30AM +0200, Robert Richter wrote:
> On 17.05.13 12:57:30, Peter Zijlstra wrote:
> > 
> > So what about something like the below?
> 
> See my comments below, otherwise it looks fine to me.

I've been liberal and read an Ack there, holler if that needs be amended.

> There is the question about core performance counters and its
> constraints on fam16h. Not sure if there are any. Cc'ing Jacob.

Currently the code doesn't work for Fam16h afaict, so I didn't wreck
that did I?

> >  	/*
> >  	 * If core performance counter extensions exists, we must use
> >  	 * MSR_F15H_PERF_CTL/MSR_F15H_PERF_CTR msrs. See also
> 
> ... amd_pmu_addr_offset():
> 
>          * If core performance counter extensions exists, we must use
>          * MSR_F15H_PERF_CTL/MSR_F15H_PERF_CTR msrs. See also
>          * amd_pmu_addr_offset().
> 

Done.

> > @@ -675,8 +670,7 @@ static int setup_perfctr_core(void)
> >  	x86_pmu.perfctr		= MSR_F15H_PERF_CTR;
> >  	x86_pmu.num_counters	= AMD64_NUM_COUNTERS_CORE;
> >  
> > -	printk(KERN_INFO "perf: AMD core performance counters detected\n");
> > -
> > +	pr_cont("core perfctr, ");
> >  	return 0;
> >  }
> >  
> > @@ -688,8 +682,8 @@ __init int amd_pmu_init(void)
> >  
> >  	x86_pmu = amd_pmu;
> >  
> > -	setup_event_constraints();
> > -	setup_perfctr_core();
> > +	if (cpu_has_perfctr_core && amd_core_pmu_init())
> > +		return -ENODEV;
> 
> Better return result of amd_core_pmu_init().

Done.. that was admittedly a tad lazy of me :-)

---
Subject: perf, x86: Rework AMD PMU init code
From: Peter Zijlstra <peterz@infradead.org>
Date: Fri, 17 May 2013 12:57:30 +0200

Josh reported that his QEMU is a bad hardware emulator and trips a
WARN in the AMD PMU init code. He requested the WARN be turned into a
pr_err() or similar.

While there, rework the code a little.

Reported-by: Josh Boyer <jwboyer@redhat.com>
Acked-by: Robert Richter <rric@kernel.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
---
 arch/x86/kernel/cpu/perf_event_amd.c |   34 +++++++++++++++++-----------------
 1 file changed, 17 insertions(+), 17 deletions(-)

--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -648,48 +648,48 @@ static __initconst const struct x86_pmu
 	.cpu_dead		= amd_pmu_cpu_dead,
 };
 
-static int setup_event_constraints(void)
+__init int amd_core_pmu_init(void)
 {
-	if (boot_cpu_data.x86 == 0x15)
+	if (!cpu_has_perfctr_core)
+		return 0;
+
+	switch (boot_cpu_data.x86) {
+	case 0x15:
+		pr_cont("Fam15h ");
 		x86_pmu.get_event_constraints = amd_get_event_constraints_f15h;
-	return 0;
-}
+		break;
 
-static int setup_perfctr_core(void)
-{
-	if (!cpu_has_perfctr_core) {
-		WARN(x86_pmu.get_event_constraints == amd_get_event_constraints_f15h,
-		     KERN_ERR "Odd, counter constraints enabled but no core perfctrs detected!");
+	default:
+		pr_err("core perfctr but no constraints; unknown hardware!\n");
 		return -ENODEV;
 	}
 
-	WARN(x86_pmu.get_event_constraints == amd_get_event_constraints,
-	     KERN_ERR "hw perf events core counters need constraints handler!");
-
 	/*
 	 * If core performance counter extensions exists, we must use
 	 * MSR_F15H_PERF_CTL/MSR_F15H_PERF_CTR msrs. See also
-	 * x86_pmu_addr_offset().
+	 * amd_pmu_addr_offset().
 	 */
 	x86_pmu.eventsel	= MSR_F15H_PERF_CTL;
 	x86_pmu.perfctr		= MSR_F15H_PERF_CTR;
 	x86_pmu.num_counters	= AMD64_NUM_COUNTERS_CORE;
 
-	printk(KERN_INFO "perf: AMD core performance counters detected\n");
-
+	pr_cont("core perfctr, ");
 	return 0;
 }
 
 __init int amd_pmu_init(void)
 {
+	int ret;
+
 	/* Performance-monitoring supported from K7 and later: */
 	if (boot_cpu_data.x86 < 6)
 		return -ENODEV;
 
 	x86_pmu = amd_pmu;
 
-	setup_event_constraints();
-	setup_perfctr_core();
+	ret = amd_core_pmu_init();
+	if (ret)
+		return ret;
 
 	/* Events are common for all AMDs */
 	memcpy(hw_cache_event_ids, amd_hw_cache_event_ids,


  reply	other threads:[~2013-05-21 13:33 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-05-16 15:10 Josh Boyer
2013-05-16 17:51 ` Peter Zijlstra
2013-05-16 17:55   ` Josh Boyer
2013-05-16 18:10     ` Peter Zijlstra
2013-05-16 20:55       ` Robert Richter
2013-05-16 21:34         ` Borislav Petkov
2013-05-17  9:04           ` Peter Zijlstra
2013-05-17  9:16             ` Borislav Petkov
2013-05-17  9:27               ` Peter Zijlstra
2013-05-17  9:45                 ` Borislav Petkov
2013-05-17 10:36                   ` Robert Richter
2013-05-17 10:57                     ` Peter Zijlstra
2013-05-21  8:56                       ` Robert Richter
2013-05-21 11:05                         ` Peter Zijlstra [this message]
2013-05-21 13:58                           ` Robert Richter
2013-05-21 15:20                           ` Jacob Shin
2013-05-28 13:03                           ` [tip:perf/core] perf/x86/amd: Rework AMD PMU init code tip-bot for Peter Zijlstra
2013-05-16 21:54         ` Drop WARN on AMD lack of perfctrs Josh Boyer
2013-05-16 22:33           ` Robert Richter
2013-05-16 19:31   ` Gleb Natapov
2013-05-16 20:00     ` Borislav Petkov

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=20130521110537.GG26912@twins.programming.kicks-ass.net \
    --to=peterz@infradead.org \
    --cc=acme@ghostprotocols.net \
    --cc=bp@alien8.de \
    --cc=gleb@redhat.com \
    --cc=jacob.shin@amd.com \
    --cc=jwboyer@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=rric@kernel.org \
    --cc=x86@kernel.org \
    /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

Powered by JetHome