* [PATCH] x86 / CPU: Avoid unnecessary IPIs in arch_freq_get_on_cpu()
@ 2017-11-13 1:15 Rafael J. Wysocki
2017-11-13 3:00 ` WANG Chao
0 siblings, 1 reply; 2+ messages in thread
From: Rafael J. Wysocki @ 2017-11-13 1:15 UTC (permalink / raw)
To: x86
Cc: Thomas Gleixner, Ingo Molnar, H. Peter Anvin, Linus Torvalds,
LKML, Linux PM, Len Brown, WANG Chao
From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Even though aperfmperf_snapshot_khz() caches the samples.khz value
to return if called again in a sufficiently short time, its caller,
arch_freq_get_on_cpu(), still uses smp_call_function_single() to run
it which may allow user space to trigger an IPI storm by reading from
the scaling_cur_freq cpufreq sysfs file in a tight loop.
To avoid that, move the decision on whether or not to return the
cached samples.khz value to arch_freq_get_on_cpu().
Fixes: 4815d3c56d1e (cpufreq: x86: Make scaling_cur_freq behave more as expected)
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
This change was part of commit 941f5f0f6ef5 (x86: CPU: Fix up "cpu MHz" in
/proc/cpuinfo), but it was not the reason for the revert and it remains
applicable.
Thanks,
Rafael
---
arch/x86/kernel/cpu/Makefile | 2 +-
arch/x86/kernel/cpu/aperfmperf.c | 11 +++++++----
arch/x86/kernel/cpu/proc.c | 4 +++-
3 files changed, 11 insertions(+), 6 deletions(-)
Index: linux-pm/arch/x86/kernel/cpu/aperfmperf.c
===================================================================
--- linux-pm.orig/arch/x86/kernel/cpu/aperfmperf.c
+++ linux-pm/arch/x86/kernel/cpu/aperfmperf.c
@@ -42,10 +42,6 @@ static void aperfmperf_snapshot_khz(void
s64 time_delta = ktime_ms_delta(now, s->time);
unsigned long flags;
- /* Don't bother re-computing within the cache threshold time. */
- if (time_delta < APERFMPERF_CACHE_THRESHOLD_MS)
- return;
-
local_irq_save(flags);
rdmsrl(MSR_IA32_APERF, aperf);
rdmsrl(MSR_IA32_MPERF, mperf);
@@ -74,6 +70,7 @@ static void aperfmperf_snapshot_khz(void
unsigned int arch_freq_get_on_cpu(int cpu)
{
+ s64 time_delta;
unsigned int khz;
if (!cpu_khz)
@@ -82,6 +79,12 @@ unsigned int arch_freq_get_on_cpu(int cp
if (!static_cpu_has(X86_FEATURE_APERFMPERF))
return 0;
+ /* Don't bother re-computing within the cache threshold time. */
+ time_delta = ktime_ms_delta(ktime_get(), per_cpu(samples.time, cpu));
+ khz = per_cpu(samples.khz, cpu);
+ if (khz && time_delta < APERFMPERF_CACHE_THRESHOLD_MS)
+ return khz;
+
smp_call_function_single(cpu, aperfmperf_snapshot_khz, NULL, 1);
khz = per_cpu(samples.khz, cpu);
if (khz)
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] x86 / CPU: Avoid unnecessary IPIs in arch_freq_get_on_cpu()
2017-11-13 1:15 [PATCH] x86 / CPU: Avoid unnecessary IPIs in arch_freq_get_on_cpu() Rafael J. Wysocki
@ 2017-11-13 3:00 ` WANG Chao
0 siblings, 0 replies; 2+ messages in thread
From: WANG Chao @ 2017-11-13 3:00 UTC (permalink / raw)
To: Rafael J. Wysocki
Cc: x86, Thomas Gleixner, Ingo Molnar, H. Peter Anvin,
Linus Torvalds, LKML, Linux PM, Len Brown
On 11/13/17 at 02:15P, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>
> Even though aperfmperf_snapshot_khz() caches the samples.khz value
> to return if called again in a sufficiently short time, its caller,
> arch_freq_get_on_cpu(), still uses smp_call_function_single() to run
> it which may allow user space to trigger an IPI storm by reading from
> the scaling_cur_freq cpufreq sysfs file in a tight loop.
>
> To avoid that, move the decision on whether or not to return the
> cached samples.khz value to arch_freq_get_on_cpu().
>
> Fixes: 4815d3c56d1e (cpufreq: x86: Make scaling_cur_freq behave more as expected)
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Looks good to me.
Reviewed-by: WANG Chao <chao.wang@ucloud.cn>
> ---
>
> This change was part of commit 941f5f0f6ef5 (x86: CPU: Fix up "cpu MHz" in
> /proc/cpuinfo), but it was not the reason for the revert and it remains
> applicable.
>
> Thanks,
> Rafael
>
> ---
> arch/x86/kernel/cpu/Makefile | 2 +-
> arch/x86/kernel/cpu/aperfmperf.c | 11 +++++++----
> arch/x86/kernel/cpu/proc.c | 4 +++-
> 3 files changed, 11 insertions(+), 6 deletions(-)
>
> Index: linux-pm/arch/x86/kernel/cpu/aperfmperf.c
> ===================================================================
> --- linux-pm.orig/arch/x86/kernel/cpu/aperfmperf.c
> +++ linux-pm/arch/x86/kernel/cpu/aperfmperf.c
> @@ -42,10 +42,6 @@ static void aperfmperf_snapshot_khz(void
> s64 time_delta = ktime_ms_delta(now, s->time);
> unsigned long flags;
>
> - /* Don't bother re-computing within the cache threshold time. */
> - if (time_delta < APERFMPERF_CACHE_THRESHOLD_MS)
> - return;
> -
> local_irq_save(flags);
> rdmsrl(MSR_IA32_APERF, aperf);
> rdmsrl(MSR_IA32_MPERF, mperf);
> @@ -74,6 +70,7 @@ static void aperfmperf_snapshot_khz(void
>
> unsigned int arch_freq_get_on_cpu(int cpu)
> {
> + s64 time_delta;
> unsigned int khz;
>
> if (!cpu_khz)
> @@ -82,6 +79,12 @@ unsigned int arch_freq_get_on_cpu(int cp
> if (!static_cpu_has(X86_FEATURE_APERFMPERF))
> return 0;
>
> + /* Don't bother re-computing within the cache threshold time. */
> + time_delta = ktime_ms_delta(ktime_get(), per_cpu(samples.time, cpu));
> + khz = per_cpu(samples.khz, cpu);
> + if (khz && time_delta < APERFMPERF_CACHE_THRESHOLD_MS)
> + return khz;
> +
> smp_call_function_single(cpu, aperfmperf_snapshot_khz, NULL, 1);
> khz = per_cpu(samples.khz, cpu);
> if (khz)
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2017-11-13 3:01 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-11-13 1:15 [PATCH] x86 / CPU: Avoid unnecessary IPIs in arch_freq_get_on_cpu() Rafael J. Wysocki
2017-11-13 3:00 ` WANG Chao
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.