From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Robert Richter <robert.richter@amd.com>
Cc: Paul Mackerras <paulus@samba.org>, Ingo Molnar <mingo@elte.hu>,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 04/29] x86/perfcounters: rework pmc_amd_save_disable_all() and pmc_amd_restore_all()
Date: Wed, 29 Apr 2009 13:07:40 +0200 [thread overview]
Message-ID: <1241003260.8021.236.camel@laptop> (raw)
In-Reply-To: <1241002046-8832-5-git-send-email-robert.richter@amd.com>
On Wed, 2009-04-29 at 12:47 +0200, Robert Richter wrote:
> MSR reads and writes are expensive. This patch adds checks to avoid
> its usage where possible.
save_disable_all()
enable(1)
restore_all()
would not correctly enable 1 with the below modification as we do not
write the configuration into the msr, on which restore relies, as it
only toggles the _ENABLE bit.
That said, I'm not sure if that's really an issue, but its why the does
does as it does.
A better abstraction could perhaps avoid this issue all-together.
> Signed-off-by: Robert Richter <robert.richter@amd.com>
> ---
> arch/x86/kernel/cpu/perf_counter.c | 24 ++++++++++++++----------
> 1 files changed, 14 insertions(+), 10 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c
> index d6d6529..75a0903 100644
> --- a/arch/x86/kernel/cpu/perf_counter.c
> +++ b/arch/x86/kernel/cpu/perf_counter.c
> @@ -334,11 +334,13 @@ static u64 pmc_amd_save_disable_all(void)
> for (idx = 0; idx < nr_counters_generic; idx++) {
> u64 val;
>
> + if (!test_bit(idx, cpuc->active_mask))
> + continue;
> rdmsrl(MSR_K7_EVNTSEL0 + idx, val);
> - if (val & ARCH_PERFMON_EVENTSEL0_ENABLE) {
> - val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE;
> - wrmsrl(MSR_K7_EVNTSEL0 + idx, val);
> - }
> + if (!(val & ARCH_PERFMON_EVENTSEL0_ENABLE))
> + continue;
> + val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE;
> + wrmsrl(MSR_K7_EVNTSEL0 + idx, val);
> }
>
> return enabled;
> @@ -372,13 +374,15 @@ static void pmc_amd_restore_all(u64 ctrl)
> return;
>
> for (idx = 0; idx < nr_counters_generic; idx++) {
> - if (test_bit(idx, cpuc->active_mask)) {
> - u64 val;
> + u64 val;
>
> - rdmsrl(MSR_K7_EVNTSEL0 + idx, val);
> - val |= ARCH_PERFMON_EVENTSEL0_ENABLE;
> - wrmsrl(MSR_K7_EVNTSEL0 + idx, val);
> - }
> + if (!test_bit(idx, cpuc->active_mask))
> + continue;
> + rdmsrl(MSR_K7_EVNTSEL0 + idx, val);
> + if (val & ARCH_PERFMON_EVENTSEL0_ENABLE)
> + continue;
> + val |= ARCH_PERFMON_EVENTSEL0_ENABLE;
> + wrmsrl(MSR_K7_EVNTSEL0 + idx, val);
> }
> }
>
next prev parent reply other threads:[~2009-04-29 11:08 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-04-29 10:46 [PATCH 0/29] x86/perfcounters: x86 and AMD cpu updates Robert Richter
2009-04-29 10:46 ` [PATCH 01/29] x86: perfcounter: remove X86_FEATURE_ARCH_PERFMON flag for AMD cpus Robert Richter
2009-04-29 11:21 ` Ingo Molnar
2009-04-29 13:03 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-05-10 5:02 ` Jaswinder Singh Rajput
2009-05-10 5:53 ` Jaswinder Singh Rajput
2009-05-10 6:30 ` Ingo Molnar
2009-05-10 6:51 ` Jaswinder Singh Rajput
2009-05-11 10:02 ` Ingo Molnar
2009-05-11 10:27 ` Jaswinder Singh Rajput
2009-05-11 12:05 ` Ingo Molnar
2009-05-11 12:13 ` Jaswinder Singh Rajput
2009-05-11 12:39 ` Robert Richter
2009-05-11 13:07 ` Jaswinder Singh Rajput
2009-05-11 13:15 ` Robert Richter
2009-04-29 10:46 ` [PATCH 02/29] perfcounter: declare perf_max_counters only for CONFIG_PERF_COUNTERS Robert Richter
2009-04-29 13:03 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 03/29] x86/perfcounters: add default path to cpu detection Robert Richter
2009-04-29 13:03 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 04/29] x86/perfcounters: rework pmc_amd_save_disable_all() and pmc_amd_restore_all() Robert Richter
2009-04-29 11:07 ` Peter Zijlstra [this message]
2009-04-29 11:11 ` Ingo Molnar
2009-04-29 11:16 ` Peter Zijlstra
2009-04-29 11:19 ` Ingo Molnar
2009-04-29 11:27 ` Peter Zijlstra
2009-04-29 12:33 ` Paul Mackerras
2009-04-29 12:55 ` Ingo Molnar
2009-04-29 13:04 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 05/29] x86/perfcounters: protect per-cpu variables with compile barriers only Robert Richter
2009-04-29 13:04 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 06/29] perfcounters: rename struct hw_perf_counter_ops into struct pmu Robert Richter
2009-04-29 13:04 ` [tip:perfcounters/core] " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 07/29] x86/perfcounters: rename struct pmc_x86_ops into struct x86_pmu Robert Richter
2009-04-29 13:04 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 08/29] x86/perfcounters: make interrupt handler model specific Robert Richter
2009-04-29 13:04 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 09/29] x86/perfcounters: remove get_status() from struct x86_pmu Robert Richter
2009-04-29 13:05 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 10/29] x86/perfcounters: remove ack_status() " Robert Richter
2009-04-29 13:05 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 11/29] x86/perfcounters: rename __hw_perf_counter_set_period into x86_perf_counter_set_period Robert Richter
2009-04-29 13:05 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 12/29] x86/perfcounters: rename intel only functions Robert Richter
2009-04-29 13:05 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 13/29] x86/perfcounters: modify initialization of struct x86_pmu Robert Richter
2009-04-29 13:05 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 14/29] x86/perfcounters: make x86_pmu data a static struct Robert Richter
2009-04-29 13:05 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 15/29] x86/perfcounters: move counter parameters to struct x86_pmu Robert Richter
2009-04-29 13:06 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 16/29] x86/perfcounters: make pmu version generic Robert Richter
2009-04-29 13:06 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 17/29] x86/perfcounters: make x86_pmu_read() static inline Robert Richter
2009-04-29 13:06 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 18/29] x86/perfcounters: rename cpuc->active_mask Robert Richter
2009-04-29 11:24 ` Ingo Molnar
2009-04-29 13:06 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 13:10 ` Ingo Molnar
2009-04-29 14:55 ` [PATCH] perf_counter, x86: rename bitmasks to ->used_mask and ->active_mask Robert Richter
2009-04-29 20:21 ` [tip:perfcounters/core] " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 19/29] x86/perfcounters: generic use of cpuc->active Robert Richter
2009-04-29 13:06 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 20/29] x86/perfcounters: consistent use of type int for counter index Robert Richter
2009-04-29 13:07 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 21/29] x86/perfcounters: rework counter enable functions Robert Richter
2009-04-29 13:07 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 22/29] x86/perfcounters: rework counter disable functions Robert Richter
2009-04-29 13:07 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 23/29] x86/perfcounters: change and remove pmu initialization checks Robert Richter
2009-04-29 13:07 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 24/29] x86/perfcounters: implement the interrupt handler for AMD cpus Robert Richter
2009-04-29 13:07 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 25/29] x86/perfcounters: return raw count with x86_perf_counter_update() Robert Richter
2009-04-29 13:08 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 26/29] x86/perfcounters: introduce max_period variable Robert Richter
2009-04-29 11:14 ` Ingo Molnar
2009-04-29 13:08 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 27/29] x86/perfcounters: remove vendor check in fixed_mode_idx() Robert Richter
2009-04-29 13:08 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-05-04 18:24 ` [tip:perfcounters/core] perf_counter: fix fixed-purpose counter support on v2 Intel-PERFMON tip-bot for Ingo Molnar
2009-04-29 10:47 ` [PATCH 28/29] x86/perfcounters: remove unused function argument in intel_pmu_get_status() Robert Richter
2009-04-29 13:08 ` [tip:perfcounters/core] perf_counter, x86: " tip-bot for Robert Richter
2009-04-29 10:47 ` [PATCH 29/29] perfcounters: updating kerneltop documentation Robert Richter
2009-04-29 11:12 ` Ingo Molnar
2009-04-29 13:08 ` [tip:perfcounters/core] perf_counter: update 'perf top' documentation tip-bot for Robert Richter
2009-04-29 11:09 ` [PATCH 0/29] x86/perfcounters: x86 and AMD cpu updates Ingo Molnar
2009-04-29 11:22 ` Peter Zijlstra
2009-04-29 11:28 ` Ingo Molnar
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=1241003260.8021.236.camel@laptop \
--to=a.p.zijlstra@chello.nl \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=paulus@samba.org \
--cc=robert.richter@amd.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