[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <tip-07cde313b2d21f728cec2836db7cdb55476f7a26@git.kernel.org>
Date: Tue, 27 Mar 2018 03:10:05 -0700
From: tip-bot for Eric Dumazet <tipbot@...or.com>
To: linux-tip-commits@...r.kernel.org
Cc: mingo@...nel.org, bp@...en8.de, eric.dumazet@...il.com,
linux-kernel@...r.kernel.org, hpa@...or.com, tglx@...utronix.de,
edumazet@...gle.com, hughd@...gle.com
Subject: [tip:x86/cleanups] x86/msr: Allow rdmsr_safe_on_cpu() to schedule
Commit-ID: 07cde313b2d21f728cec2836db7cdb55476f7a26
Gitweb: https://git.kernel.org/tip/07cde313b2d21f728cec2836db7cdb55476f7a26
Author: Eric Dumazet <edumazet@...gle.com>
AuthorDate: Fri, 23 Mar 2018 14:58:17 -0700
Committer: Thomas Gleixner <tglx@...utronix.de>
CommitDate: Tue, 27 Mar 2018 12:01:47 +0200
x86/msr: Allow rdmsr_safe_on_cpu() to schedule
High latencies can be observed caused by a daemon periodically reading
various MSR on all cpus. On KASAN enabled kernels ~10ms latencies can be
observed simply reading one MSR. Even without KASAN, sending an IPI to a
CPU, which is in a deep sleep state or in a long hard IRQ disabled section,
waiting for the answer can consume hundreds of microseconds.
All usage sites are in preemptible context, convert rdmsr_safe_on_cpu() to
use a completion instead of busy polling.
Overall daemon cpu usage was reduced by 35 %, and latencies caused by
msr_read() disappeared.
Signed-off-by: Eric Dumazet <edumazet@...gle.com>
Signed-off-by: Thomas Gleixner <tglx@...utronix.de>
Acked-by: Ingo Molnar <mingo@...nel.org>
Cc: Hugh Dickins <hughd@...gle.com>
Cc: Borislav Petkov <bp@...en8.de>
Cc: Eric Dumazet <eric.dumazet@...il.com>
Link: https://lkml.kernel.org/r/20180323215818.127774-1-edumazet@google.com
---
arch/x86/lib/msr-smp.c | 32 ++++++++++++++++++++++++--------
1 file changed, 24 insertions(+), 8 deletions(-)
diff --git a/arch/x86/lib/msr-smp.c b/arch/x86/lib/msr-smp.c
index 693cce0be82d..761ba062afda 100644
--- a/arch/x86/lib/msr-smp.c
+++ b/arch/x86/lib/msr-smp.c
@@ -2,6 +2,7 @@
#include <linux/export.h>
#include <linux/preempt.h>
#include <linux/smp.h>
+#include <linux/completion.h>
#include <asm/msr.h>
static void __rdmsr_on_cpu(void *info)
@@ -143,13 +144,19 @@ void wrmsr_on_cpus(const struct cpumask *mask, u32 msr_no, struct msr *msrs)
}
EXPORT_SYMBOL(wrmsr_on_cpus);
+struct msr_info_completion {
+ struct msr_info msr;
+ struct completion done;
+};
+
/* These "safe" variants are slower and should be used when the target MSR
may not actually exist. */
static void __rdmsr_safe_on_cpu(void *info)
{
- struct msr_info *rv = info;
+ struct msr_info_completion *rv = info;
- rv->err = rdmsr_safe(rv->msr_no, &rv->reg.l, &rv->reg.h);
+ rv->msr.err = rdmsr_safe(rv->msr.msr_no, &rv->msr.reg.l, &rv->msr.reg.h);
+ complete(&rv->done);
}
static void __wrmsr_safe_on_cpu(void *info)
@@ -161,17 +168,26 @@ static void __wrmsr_safe_on_cpu(void *info)
int rdmsr_safe_on_cpu(unsigned int cpu, u32 msr_no, u32 *l, u32 *h)
{
+ struct msr_info_completion rv;
+ call_single_data_t csd = {
+ .func = __rdmsr_safe_on_cpu,
+ .info = &rv,
+ };
int err;
- struct msr_info rv;
memset(&rv, 0, sizeof(rv));
+ init_completion(&rv.done);
+ rv.msr.msr_no = msr_no;
- rv.msr_no = msr_no;
- err = smp_call_function_single(cpu, __rdmsr_safe_on_cpu, &rv, 1);
- *l = rv.reg.l;
- *h = rv.reg.h;
+ err = smp_call_function_single_async(cpu, &csd);
+ if (!err) {
+ wait_for_completion(&rv.done);
+ err = rv.msr.err;
+ }
+ *l = rv.msr.reg.l;
+ *h = rv.msr.reg.h;
- return err ? err : rv.err;
+ return err;
}
EXPORT_SYMBOL(rdmsr_safe_on_cpu);
Powered by blists - more mailing lists