* [PATCH] perf/x86: Use local64_try_cmpxchg
@ 2023-07-06 14:16 Uros Bizjak
2023-07-10 8:37 ` [tip: perf/core] " tip-bot2 for Uros Bizjak
0 siblings, 1 reply; 2+ messages in thread
From: Uros Bizjak @ 2023-07-06 14:16 UTC (permalink / raw)
To: x86, linux-perf-users, linux-kernel
Cc: Uros Bizjak, Peter Zijlstra, Ingo Molnar,
Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin,
Jiri Olsa, Namhyung Kim, Ian Rogers, Adrian Hunter,
Thomas Gleixner, Borislav Petkov, Dave Hansen
Use local64_try_cmpxchg instead of local64_cmpxchg (*ptr, old, new) == old.
x86 CMPXCHG instruction returns success in ZF flag, so this change saves a
compare after cmpxchg (and related move instruction in front of cmpxchg).
Also, try_cmpxchg implicitly assigns old *ptr value to "old" when cmpxchg
fails. There is no need to re-read the value in the loop.
No functional change intended.
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Ian Rogers <irogers@google.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc. "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
---
arch/x86/events/amd/ibs.c | 4 ++--
arch/x86/events/core.c | 10 ++++------
arch/x86/events/intel/cstate.c | 10 ++++------
arch/x86/events/msr.c | 8 +++-----
4 files changed, 13 insertions(+), 19 deletions(-)
diff --git a/arch/x86/events/amd/ibs.c b/arch/x86/events/amd/ibs.c
index 371014802191..b54db2138c16 100644
--- a/arch/x86/events/amd/ibs.c
+++ b/arch/x86/events/amd/ibs.c
@@ -156,8 +156,8 @@ perf_event_try_update(struct perf_event *event, u64 new_raw_count, int width)
* count to the generic event atomically:
*/
prev_raw_count = local64_read(&hwc->prev_count);
- if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
- new_raw_count) != prev_raw_count)
+ if (!local64_try_cmpxchg(&hwc->prev_count,
+ &prev_raw_count, new_raw_count))
return 0;
/*
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
index 9d248703cbdd..23c96420dea1 100644
--- a/arch/x86/events/core.c
+++ b/arch/x86/events/core.c
@@ -129,13 +129,11 @@ u64 x86_perf_event_update(struct perf_event *event)
* exchange a new raw count - then add that new-prev delta
* count to the generic event atomically:
*/
-again:
prev_raw_count = local64_read(&hwc->prev_count);
- rdpmcl(hwc->event_base_rdpmc, new_raw_count);
-
- if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
- new_raw_count) != prev_raw_count)
- goto again;
+ do {
+ rdpmcl(hwc->event_base_rdpmc, new_raw_count);
+ } while (!local64_try_cmpxchg(&hwc->prev_count,
+ &prev_raw_count, new_raw_count));
/*
* Now we have the new raw value and have updated the prev
diff --git a/arch/x86/events/intel/cstate.c b/arch/x86/events/intel/cstate.c
index 835862c548cc..97035db68df2 100644
--- a/arch/x86/events/intel/cstate.c
+++ b/arch/x86/events/intel/cstate.c
@@ -365,13 +365,11 @@ static void cstate_pmu_event_update(struct perf_event *event)
struct hw_perf_event *hwc = &event->hw;
u64 prev_raw_count, new_raw_count;
-again:
prev_raw_count = local64_read(&hwc->prev_count);
- new_raw_count = cstate_pmu_read_counter(event);
-
- if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
- new_raw_count) != prev_raw_count)
- goto again;
+ do {
+ new_raw_count = cstate_pmu_read_counter(event);
+ } while (!local64_try_cmpxchg(&hwc->prev_count,
+ &prev_raw_count, new_raw_count));
local64_add(new_raw_count - prev_raw_count, &event->count);
}
diff --git a/arch/x86/events/msr.c b/arch/x86/events/msr.c
index 0feaaa571303..61c59e0627d4 100644
--- a/arch/x86/events/msr.c
+++ b/arch/x86/events/msr.c
@@ -244,12 +244,10 @@ static void msr_event_update(struct perf_event *event)
s64 delta;
/* Careful, an NMI might modify the previous event value: */
-again:
prev = local64_read(&event->hw.prev_count);
- now = msr_read_counter(event);
-
- if (local64_cmpxchg(&event->hw.prev_count, prev, now) != prev)
- goto again;
+ do {
+ now = msr_read_counter(event);
+ } while (!local64_try_cmpxchg(&event->hw.prev_count, &prev, now));
delta = now - prev;
if (unlikely(event->hw.event_base == MSR_SMI_COUNT)) {
--
2.41.0
^ permalink raw reply [flat|nested] 2+ messages in thread
* [tip: perf/core] perf/x86: Use local64_try_cmpxchg
2023-07-06 14:16 [PATCH] perf/x86: Use local64_try_cmpxchg Uros Bizjak
@ 2023-07-10 8:37 ` tip-bot2 for Uros Bizjak
0 siblings, 0 replies; 2+ messages in thread
From: tip-bot2 for Uros Bizjak @ 2023-07-10 8:37 UTC (permalink / raw)
To: linux-tip-commits; +Cc: Uros Bizjak, Peter Zijlstra (Intel), x86, linux-kernel
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 4c1c9dea2089374cb58505a0df6136969ff3c8b9
Gitweb: https://git.kernel.org/tip/4c1c9dea2089374cb58505a0df6136969ff3c8b9
Author: Uros Bizjak <ubizjak@gmail.com>
AuthorDate: Thu, 06 Jul 2023 16:16:48 +02:00
Committer: Peter Zijlstra <peterz@infradead.org>
CommitterDate: Mon, 10 Jul 2023 09:52:35 +02:00
perf/x86: Use local64_try_cmpxchg
Use local64_try_cmpxchg instead of local64_cmpxchg (*ptr, old, new) == old.
x86 CMPXCHG instruction returns success in ZF flag, so this change saves a
compare after cmpxchg (and related move instruction in front of cmpxchg).
Also, try_cmpxchg implicitly assigns old *ptr value to "old" when cmpxchg
fails. There is no need to re-read the value in the loop.
No functional change intended.
Cc. "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20230706141720.2672-1-ubizjak@gmail.com
---
arch/x86/events/amd/ibs.c | 4 ++--
arch/x86/events/core.c | 10 ++++------
arch/x86/events/intel/cstate.c | 10 ++++------
arch/x86/events/msr.c | 8 +++-----
4 files changed, 13 insertions(+), 19 deletions(-)
diff --git a/arch/x86/events/amd/ibs.c b/arch/x86/events/amd/ibs.c
index 74e6642..7d29be0 100644
--- a/arch/x86/events/amd/ibs.c
+++ b/arch/x86/events/amd/ibs.c
@@ -156,8 +156,8 @@ perf_event_try_update(struct perf_event *event, u64 new_raw_count, int width)
* count to the generic event atomically:
*/
prev_raw_count = local64_read(&hwc->prev_count);
- if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
- new_raw_count) != prev_raw_count)
+ if (!local64_try_cmpxchg(&hwc->prev_count,
+ &prev_raw_count, new_raw_count))
return 0;
/*
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c
index 9d24870..23c9642 100644
--- a/arch/x86/events/core.c
+++ b/arch/x86/events/core.c
@@ -129,13 +129,11 @@ u64 x86_perf_event_update(struct perf_event *event)
* exchange a new raw count - then add that new-prev delta
* count to the generic event atomically:
*/
-again:
prev_raw_count = local64_read(&hwc->prev_count);
- rdpmcl(hwc->event_base_rdpmc, new_raw_count);
-
- if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
- new_raw_count) != prev_raw_count)
- goto again;
+ do {
+ rdpmcl(hwc->event_base_rdpmc, new_raw_count);
+ } while (!local64_try_cmpxchg(&hwc->prev_count,
+ &prev_raw_count, new_raw_count));
/*
* Now we have the new raw value and have updated the prev
diff --git a/arch/x86/events/intel/cstate.c b/arch/x86/events/intel/cstate.c
index 835862c..97035db 100644
--- a/arch/x86/events/intel/cstate.c
+++ b/arch/x86/events/intel/cstate.c
@@ -365,13 +365,11 @@ static void cstate_pmu_event_update(struct perf_event *event)
struct hw_perf_event *hwc = &event->hw;
u64 prev_raw_count, new_raw_count;
-again:
prev_raw_count = local64_read(&hwc->prev_count);
- new_raw_count = cstate_pmu_read_counter(event);
-
- if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
- new_raw_count) != prev_raw_count)
- goto again;
+ do {
+ new_raw_count = cstate_pmu_read_counter(event);
+ } while (!local64_try_cmpxchg(&hwc->prev_count,
+ &prev_raw_count, new_raw_count));
local64_add(new_raw_count - prev_raw_count, &event->count);
}
diff --git a/arch/x86/events/msr.c b/arch/x86/events/msr.c
index 0feaaa5..61c59e0 100644
--- a/arch/x86/events/msr.c
+++ b/arch/x86/events/msr.c
@@ -244,12 +244,10 @@ static void msr_event_update(struct perf_event *event)
s64 delta;
/* Careful, an NMI might modify the previous event value: */
-again:
prev = local64_read(&event->hw.prev_count);
- now = msr_read_counter(event);
-
- if (local64_cmpxchg(&event->hw.prev_count, prev, now) != prev)
- goto again;
+ do {
+ now = msr_read_counter(event);
+ } while (!local64_try_cmpxchg(&event->hw.prev_count, &prev, now));
delta = now - prev;
if (unlikely(event->hw.event_base == MSR_SMI_COUNT)) {
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2023-07-10 8:38 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-07-06 14:16 [PATCH] perf/x86: Use local64_try_cmpxchg Uros Bizjak
2023-07-10 8:37 ` [tip: perf/core] " tip-bot2 for Uros Bizjak
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®