[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7c86c4470906220727l2a50753eg93c44006b8c2c589@mail.gmail.com>
Date: Mon, 22 Jun 2009 16:27:31 +0200
From: stephane eranian <eranian@...glemail.com>
To: Ingo Molnar <mingo@...e.hu>
Cc: LKML <linux-kernel@...r.kernel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Thomas Gleixner <tglx@...utronix.de>,
Robert Richter <robert.richter@....com>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
Paul Mackerras <paulus@...ba.org>,
Andi Kleen <andi@...stfloor.org>,
Maynard Johnson <mpjohn@...ibm.com>,
Carl Love <cel@...ibm.com>,
Corey J Ashford <cjashfor@...ibm.com>,
Philip Mucci <mucci@...s.utk.edu>,
Dan Terpstra <terpstra@...s.utk.edu>,
perfmon2-devel <perfmon2-devel@...ts.sourceforge.net>
Subject: Re: II.1 - Fixed counters on Intel
On Mon, Jun 22, 2009 at 1:57 PM, Ingo Molnar<mingo@...e.hu> wrote:
>> II/ X86 comments
>>
>> 1/ Fixed counters on Intel
>>
>> You cannot simply fall back to generic counters if you cannot find
>> a fixed counter. There are model-specific bugs, for instance
>> UNHALTED_REFERENCE_CYCLES (0x013c), does not measure the same
>> thing on Nehalem when it is used in fixed counter 2 or a generic
>> counter. The same is true on Core.
>
> This could be handled via a model specific quirk, if the erratum is
> serious enough.
Better demonstrated with an actual example on a 2.4GHz Quad: 10s noploop
$ pfmon -v --us-c
-eunhalted_core_cycles,unhalted_reference_cycles,cpu_clk_unhalted:bus
-u noploop 10
[FIXED_CTRL(pmc16)=0xaa0 pmi0=1 en0=0x0 pmi1=1 en1=0x2 pmi2=1 en2=0x2]
UNHALTED_CORE_CYCLES UNHALTED_REFERENCE_CYCLES
[PERFEVTSEL0(pmc0)=0x51013c event_sel=0x3c umask=0x1 os=0 usr=1 en=1
int=1 inv=0 edge=0 cnt_mask=0] CPU_CLK_UNHALTED
noploop for 10 seconds
23,833,577,042 UNHALTED_CORE_CYCLES
23,853,100,716 UNHALTED_REFERENCE_CYCLES
2,650,345,853 CPU_CLK_UNHALTED:BUS
0x013c on fixed counter2 = unhalted_reference_cycles
0x013c on generic counter = cpu_clk_unhalted:bus
Difference is significant, isn't it?
And the question becomes: how do I express that I want the fixed counter version
of the event using the current PCL API, given both have the same encoding?
>
>> You cannot simply look at the event field code to determine
>> whether this is an event supported by a fixed counters. You must
>> look at the other fields such as edge, invert, cnt-mask. If those
>> are present then you have to fall back to using a generic counter
>> as fixed counters only support priv level filtering. As indicated
>> above, though, programming UNHALTED_REFERENCE_CYCLES on a generic
>> counter does not count the same thing, therefore you need to fail
>> if filters other than priv levels are present on this event.
>
> Agreed, we'll fix this.
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists