| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CAJNi4rPaPeVWzBve6Toi8hhrxP6GROAGRug7+c3zg1crDeOt7Q@mail.gmail.com>
Date: Sat, 22 Mar 2025 11:20:36 +0800
From: richard clark <richard.xnu.clark@...il.com>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
Linus Torvalds <torvalds@...ux-foundation.org>
Subject: hrtimer precision issue/question??
With diff below under the 'cyclictest -a 0 -t 1 -m -p99' trigger from
the arm64-based linux box, the interval is 1000us and the arch_timer
in the system is: arch_timer: cp15 timer(s) running at 31.25MHz
(phys). 1tick = 32ns for the arch timer, I am not sure if those
durations less than 1000us are expected?
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index 11c704bfd4fb..691d1fe6e944 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -2036,14 +2036,25 @@ int nanosleep_copyout(struct restart_block
*restart, struct timespec64 *ts)
static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum
hrtimer_mode mode)
{
struct restart_block *restart;
+ ktime_t t0, t1;
+ int c0, c1;
+ bool hit = !strncmp(current->comm, "cyclictest", 10);
do {
+ if (hit) {
+ c0 = smp_processor_id();
+ t0 = ktime_get();
+ }
set_current_state(TASK_INTERRUPTIBLE|TASK_FREEZABLE);
hrtimer_sleeper_start_expires(t, mode);
if (likely(t->task))
schedule();
-
+ if (hit) {
+ t1 = ktime_get();
+ c1 = smp_processor_id();
+ pr_info("[%2d-%2d]t0=%lld,t1=%lld,d=%lld
us\n", c0, c1, t0, t1, ktime_to_us(ktime_sub(t1, t0)));
+ }
hrtimer_cancel(&t->timer);
mode = HRTIMER_MODE_ABS;
Part of the real output:
[ 165.555795] [ 0- 0]t0=165550399226,t1=165551394303,d=995 us
[ 165.556802] [ 0- 0]t0=165551398751,t1=165552400997,d=1002 us
[ 165.557797] [ 0- 0]t0=165552415141,t1=165553396106,d=980 us
[ 165.558794] [ 0- 0]t0=165553402090,t1=165554394160,d=992 us
[ 165.559794] [ 0- 0]t0=165554398832,t1=165555393941,d=995 us
[ 165.559995] [ 6- 5]t0=165545533151,t1=165555594519,d=10061 us
[ 165.560802] [ 0- 0]t0=165555398485,t1=165556400795,d=1002 us
[ 165.561796] [ 0- 0]t0=165556415259,t1=165557395457,d=980 us
[ 165.562794] [ 0- 0]t0=165557401601,t1=165558394086,d=992 us
[ 165.563794] [ 0- 0]t0=165558398694,t1=165559393996,d=995 us
[ 165.564802] [ 0- 0]t0=165559398188,t1=165560400785,d=1002 us
[ 165.565796] [ 0- 0]t0=165560414673,t1=165561395671,d=980 us
[ 165.566794] [ 0- 0]t0=165561402327,t1=165562394108,d=991 us
[ 165.567794] [ 0- 0]t0=165562399068,t1=165563393890,d=994 us
[ 165.568802] [ 0- 0]t0=165563399362,t1=165564401640,d=1002 us
[ 165.569796] [ 0- 0]t0=165564413640,t1=165565395885,d=982 us
[ 165.570104] [ 5- 6]t0=165555639767,t1=165565703087,d=10063 us
[ 165.570795] [ 0- 0]t0=165565401421,t1=165566394739,d=993 us
[ 165.571794] [ 0- 0]t0=165566400531,t1=165567393816,d=993 us
[ 165.572799] [ 0- 0]t0=165567399800,t1=165568398270,d=998 us
[ 165.573798] [ 0- 0]t0=165568412702,t1=165569398115,d=985 us
[ 165.574794] [ 0- 0]t0=165569404451,t1=165570394249,d=989 us
[ 165.575794] [ 0- 0]t0=165570399145,t1=165571393838,d=994 us
[ 165.576801] [ 0- 0]t0=165571398606,t1=165572400276,d=1001 us
[ 165.577796] [ 0- 0]t0=165572414260,t1=165573395417,d=981 us
[ 165.578794] [ 0- 0]t0=165573401561,t1=165574394015,d=992 us
[ 165.579794] [ 0- 0]t0=165574398975,t1=165575393957,d=994 us
[ 165.580208] [ 6- 5]t0=165565749263,t1=165575807879,d=10058 us
[ 165.580801] [ 0- 0]t0=165575398405,t1=165576399978,d=1001 us
Thanks
Powered by blists - more mailing lists