* kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used
@ 2023-09-11 7:54 kernel test robot
0 siblings, 0 replies; 4+ messages in thread
From: kernel test robot @ 2023-09-11 7:54 UTC (permalink / raw)
To: Sebastian Andrzej Siewior; +Cc: oe-kbuild-all, linux-kernel, Thomas Gleixner
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
head: 0bb80ecc33a8fb5a682236443c1e740d5c917d1d
commit: 73d20564e0dcae003e0d79977f044d5e57496304 hrtimer: Don't dereference the hrtimer pointer after the callback
date: 3 years, 5 months ago
config: openrisc-allnoconfig (https://download.01.org/0day-ci/archive/20230911/202309111514.seVql7NU-lkp@intel.com/config)
compiler: or1k-linux-gcc (GCC) 12.3.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20230911/202309111514.seVql7NU-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202309111514.seVql7NU-lkp@intel.com/
All warnings (new ones prefixed by >>):
kernel/time/hrtimer.c:120:35: warning: initialized field overwritten [-Woverride-init]
120 | [CLOCK_REALTIME] = HRTIMER_BASE_REALTIME,
| ^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:120:35: note: (near initialization for 'hrtimer_clock_to_base_table[0]')
kernel/time/hrtimer.c:121:35: warning: initialized field overwritten [-Woverride-init]
121 | [CLOCK_MONOTONIC] = HRTIMER_BASE_MONOTONIC,
| ^~~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:121:35: note: (near initialization for 'hrtimer_clock_to_base_table[1]')
kernel/time/hrtimer.c:122:35: warning: initialized field overwritten [-Woverride-init]
122 | [CLOCK_BOOTTIME] = HRTIMER_BASE_BOOTTIME,
| ^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:122:35: note: (near initialization for 'hrtimer_clock_to_base_table[7]')
kernel/time/hrtimer.c:123:35: warning: initialized field overwritten [-Woverride-init]
123 | [CLOCK_TAI] = HRTIMER_BASE_TAI,
| ^~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:123:35: note: (near initialization for 'hrtimer_clock_to_base_table[11]')
kernel/time/hrtimer.c: In function '__run_hrtimer':
>> kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used [-Wunused-but-set-variable]
1483 | bool expires_in_hardirq;
| ^~~~~~~~~~~~~~~~~~
vim +/expires_in_hardirq +1483 kernel/time/hrtimer.c
1458
1459 /*
1460 * The write_seqcount_barrier()s in __run_hrtimer() split the thing into 3
1461 * distinct sections:
1462 *
1463 * - queued: the timer is queued
1464 * - callback: the timer is being ran
1465 * - post: the timer is inactive or (re)queued
1466 *
1467 * On the read side we ensure we observe timer->state and cpu_base->running
1468 * from the same section, if anything changed while we looked at it, we retry.
1469 * This includes timer->base changing because sequence numbers alone are
1470 * insufficient for that.
1471 *
1472 * The sequence numbers are required because otherwise we could still observe
1473 * a false negative if the read side got smeared over multiple consequtive
1474 * __run_hrtimer() invocations.
1475 */
1476
1477 static void __run_hrtimer(struct hrtimer_cpu_base *cpu_base,
1478 struct hrtimer_clock_base *base,
1479 struct hrtimer *timer, ktime_t *now,
1480 unsigned long flags) __must_hold(&cpu_base->lock)
1481 {
1482 enum hrtimer_restart (*fn)(struct hrtimer *);
> 1483 bool expires_in_hardirq;
1484 int restart;
1485
1486 lockdep_assert_held(&cpu_base->lock);
1487
1488 debug_deactivate(timer);
1489 base->running = timer;
1490
1491 /*
1492 * Separate the ->running assignment from the ->state assignment.
1493 *
1494 * As with a regular write barrier, this ensures the read side in
1495 * hrtimer_active() cannot observe base->running == NULL &&
1496 * timer->state == INACTIVE.
1497 */
1498 raw_write_seqcount_barrier(&base->seq);
1499
1500 __remove_hrtimer(timer, base, HRTIMER_STATE_INACTIVE, 0);
1501 fn = timer->function;
1502
1503 /*
1504 * Clear the 'is relative' flag for the TIME_LOW_RES case. If the
1505 * timer is restarted with a period then it becomes an absolute
1506 * timer. If its not restarted it does not matter.
1507 */
1508 if (IS_ENABLED(CONFIG_TIME_LOW_RES))
1509 timer->is_rel = false;
1510
1511 /*
1512 * The timer is marked as running in the CPU base, so it is
1513 * protected against migration to a different CPU even if the lock
1514 * is dropped.
1515 */
1516 raw_spin_unlock_irqrestore(&cpu_base->lock, flags);
1517 trace_hrtimer_expire_entry(timer, now);
1518 expires_in_hardirq = lockdep_hrtimer_enter(timer);
1519
1520 restart = fn(timer);
1521
1522 lockdep_hrtimer_exit(expires_in_hardirq);
1523 trace_hrtimer_expire_exit(timer);
1524 raw_spin_lock_irq(&cpu_base->lock);
1525
1526 /*
1527 * Note: We clear the running state after enqueue_hrtimer and
1528 * we do not reprogram the event hardware. Happens either in
1529 * hrtimer_start_range_ns() or in hrtimer_interrupt()
1530 *
1531 * Note: Because we dropped the cpu_base->lock above,
1532 * hrtimer_start_range_ns() can have popped in and enqueued the timer
1533 * for us already.
1534 */
1535 if (restart != HRTIMER_NORESTART &&
1536 !(timer->state & HRTIMER_STATE_ENQUEUED))
1537 enqueue_hrtimer(timer, base, HRTIMER_MODE_ABS);
1538
1539 /*
1540 * Separate the ->running assignment from the ->state assignment.
1541 *
1542 * As with a regular write barrier, this ensures the read side in
1543 * hrtimer_active() cannot observe base->running.timer == NULL &&
1544 * timer->state == INACTIVE.
1545 */
1546 raw_write_seqcount_barrier(&base->seq);
1547
1548 WARN_ON_ONCE(base->running != timer);
1549 base->running = NULL;
1550 }
1551
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 4+ messages in thread
* kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used
@ 2023-11-12 0:52 kernel test robot
2023-11-20 15:17 ` Sebastian Andrzej Siewior
0 siblings, 1 reply; 4+ messages in thread
From: kernel test robot @ 2023-11-12 0:52 UTC (permalink / raw)
To: Sebastian Andrzej Siewior; +Cc: oe-kbuild-all, linux-kernel, Thomas Gleixner
Hi Sebastian,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
head: 3ca112b71f35dd5d99fc4571a56b5fc6f0c15814
commit: 73d20564e0dcae003e0d79977f044d5e57496304 hrtimer: Don't dereference the hrtimer pointer after the callback
date: 3 years, 7 months ago
config: i386-allnoconfig (https://download.01.org/0day-ci/archive/20231112/202311120815.33YfGjzD-lkp@intel.com/config)
compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231112/202311120815.33YfGjzD-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202311120815.33YfGjzD-lkp@intel.com/
All warnings (new ones prefixed by >>):
kernel/time/hrtimer.c:120:35: warning: initialized field overwritten [-Woverride-init]
120 | [CLOCK_REALTIME] = HRTIMER_BASE_REALTIME,
| ^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:120:35: note: (near initialization for 'hrtimer_clock_to_base_table[0]')
kernel/time/hrtimer.c:121:35: warning: initialized field overwritten [-Woverride-init]
121 | [CLOCK_MONOTONIC] = HRTIMER_BASE_MONOTONIC,
| ^~~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:121:35: note: (near initialization for 'hrtimer_clock_to_base_table[1]')
kernel/time/hrtimer.c:122:35: warning: initialized field overwritten [-Woverride-init]
122 | [CLOCK_BOOTTIME] = HRTIMER_BASE_BOOTTIME,
| ^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:122:35: note: (near initialization for 'hrtimer_clock_to_base_table[7]')
kernel/time/hrtimer.c:123:35: warning: initialized field overwritten [-Woverride-init]
123 | [CLOCK_TAI] = HRTIMER_BASE_TAI,
| ^~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:123:35: note: (near initialization for 'hrtimer_clock_to_base_table[11]')
kernel/time/hrtimer.c: In function '__run_hrtimer':
>> kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used [-Wunused-but-set-variable]
1483 | bool expires_in_hardirq;
| ^~~~~~~~~~~~~~~~~~
vim +/expires_in_hardirq +1483 kernel/time/hrtimer.c
1458
1459 /*
1460 * The write_seqcount_barrier()s in __run_hrtimer() split the thing into 3
1461 * distinct sections:
1462 *
1463 * - queued: the timer is queued
1464 * - callback: the timer is being ran
1465 * - post: the timer is inactive or (re)queued
1466 *
1467 * On the read side we ensure we observe timer->state and cpu_base->running
1468 * from the same section, if anything changed while we looked at it, we retry.
1469 * This includes timer->base changing because sequence numbers alone are
1470 * insufficient for that.
1471 *
1472 * The sequence numbers are required because otherwise we could still observe
1473 * a false negative if the read side got smeared over multiple consequtive
1474 * __run_hrtimer() invocations.
1475 */
1476
1477 static void __run_hrtimer(struct hrtimer_cpu_base *cpu_base,
1478 struct hrtimer_clock_base *base,
1479 struct hrtimer *timer, ktime_t *now,
1480 unsigned long flags) __must_hold(&cpu_base->lock)
1481 {
1482 enum hrtimer_restart (*fn)(struct hrtimer *);
> 1483 bool expires_in_hardirq;
1484 int restart;
1485
1486 lockdep_assert_held(&cpu_base->lock);
1487
1488 debug_deactivate(timer);
1489 base->running = timer;
1490
1491 /*
1492 * Separate the ->running assignment from the ->state assignment.
1493 *
1494 * As with a regular write barrier, this ensures the read side in
1495 * hrtimer_active() cannot observe base->running == NULL &&
1496 * timer->state == INACTIVE.
1497 */
1498 raw_write_seqcount_barrier(&base->seq);
1499
1500 __remove_hrtimer(timer, base, HRTIMER_STATE_INACTIVE, 0);
1501 fn = timer->function;
1502
1503 /*
1504 * Clear the 'is relative' flag for the TIME_LOW_RES case. If the
1505 * timer is restarted with a period then it becomes an absolute
1506 * timer. If its not restarted it does not matter.
1507 */
1508 if (IS_ENABLED(CONFIG_TIME_LOW_RES))
1509 timer->is_rel = false;
1510
1511 /*
1512 * The timer is marked as running in the CPU base, so it is
1513 * protected against migration to a different CPU even if the lock
1514 * is dropped.
1515 */
1516 raw_spin_unlock_irqrestore(&cpu_base->lock, flags);
1517 trace_hrtimer_expire_entry(timer, now);
1518 expires_in_hardirq = lockdep_hrtimer_enter(timer);
1519
1520 restart = fn(timer);
1521
1522 lockdep_hrtimer_exit(expires_in_hardirq);
1523 trace_hrtimer_expire_exit(timer);
1524 raw_spin_lock_irq(&cpu_base->lock);
1525
1526 /*
1527 * Note: We clear the running state after enqueue_hrtimer and
1528 * we do not reprogram the event hardware. Happens either in
1529 * hrtimer_start_range_ns() or in hrtimer_interrupt()
1530 *
1531 * Note: Because we dropped the cpu_base->lock above,
1532 * hrtimer_start_range_ns() can have popped in and enqueued the timer
1533 * for us already.
1534 */
1535 if (restart != HRTIMER_NORESTART &&
1536 !(timer->state & HRTIMER_STATE_ENQUEUED))
1537 enqueue_hrtimer(timer, base, HRTIMER_MODE_ABS);
1538
1539 /*
1540 * Separate the ->running assignment from the ->state assignment.
1541 *
1542 * As with a regular write barrier, this ensures the read side in
1543 * hrtimer_active() cannot observe base->running.timer == NULL &&
1544 * timer->state == INACTIVE.
1545 */
1546 raw_write_seqcount_barrier(&base->seq);
1547
1548 WARN_ON_ONCE(base->running != timer);
1549 base->running = NULL;
1550 }
1551
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 4+ messages in thread
* kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used
@ 2023-11-19 4:35 kernel test robot
0 siblings, 0 replies; 4+ messages in thread
From: kernel test robot @ 2023-11-19 4:35 UTC (permalink / raw)
To: Sebastian Andrzej Siewior; +Cc: oe-kbuild-all, linux-kernel, Thomas Gleixner
Hi Sebastian,
FYI, the error/warning still remains.
tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
head: b8f1fa2419c19c81bc386a6b350879ba54a573e1
commit: 73d20564e0dcae003e0d79977f044d5e57496304 hrtimer: Don't dereference the hrtimer pointer after the callback
date: 3 years, 8 months ago
config: i386-allnoconfig (https://download.01.org/0day-ci/archive/20231119/202311191229.55QXHVc6-lkp@intel.com/config)
compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231119/202311191229.55QXHVc6-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202311191229.55QXHVc6-lkp@intel.com/
All warnings (new ones prefixed by >>):
kernel/time/hrtimer.c:120:35: warning: initialized field overwritten [-Woverride-init]
120 | [CLOCK_REALTIME] = HRTIMER_BASE_REALTIME,
| ^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:120:35: note: (near initialization for 'hrtimer_clock_to_base_table[0]')
kernel/time/hrtimer.c:121:35: warning: initialized field overwritten [-Woverride-init]
121 | [CLOCK_MONOTONIC] = HRTIMER_BASE_MONOTONIC,
| ^~~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:121:35: note: (near initialization for 'hrtimer_clock_to_base_table[1]')
kernel/time/hrtimer.c:122:35: warning: initialized field overwritten [-Woverride-init]
122 | [CLOCK_BOOTTIME] = HRTIMER_BASE_BOOTTIME,
| ^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:122:35: note: (near initialization for 'hrtimer_clock_to_base_table[7]')
kernel/time/hrtimer.c:123:35: warning: initialized field overwritten [-Woverride-init]
123 | [CLOCK_TAI] = HRTIMER_BASE_TAI,
| ^~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:123:35: note: (near initialization for 'hrtimer_clock_to_base_table[11]')
kernel/time/hrtimer.c: In function '__run_hrtimer':
>> kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used [-Wunused-but-set-variable]
1483 | bool expires_in_hardirq;
| ^~~~~~~~~~~~~~~~~~
vim +/expires_in_hardirq +1483 kernel/time/hrtimer.c
1458
1459 /*
1460 * The write_seqcount_barrier()s in __run_hrtimer() split the thing into 3
1461 * distinct sections:
1462 *
1463 * - queued: the timer is queued
1464 * - callback: the timer is being ran
1465 * - post: the timer is inactive or (re)queued
1466 *
1467 * On the read side we ensure we observe timer->state and cpu_base->running
1468 * from the same section, if anything changed while we looked at it, we retry.
1469 * This includes timer->base changing because sequence numbers alone are
1470 * insufficient for that.
1471 *
1472 * The sequence numbers are required because otherwise we could still observe
1473 * a false negative if the read side got smeared over multiple consequtive
1474 * __run_hrtimer() invocations.
1475 */
1476
1477 static void __run_hrtimer(struct hrtimer_cpu_base *cpu_base,
1478 struct hrtimer_clock_base *base,
1479 struct hrtimer *timer, ktime_t *now,
1480 unsigned long flags) __must_hold(&cpu_base->lock)
1481 {
1482 enum hrtimer_restart (*fn)(struct hrtimer *);
> 1483 bool expires_in_hardirq;
1484 int restart;
1485
1486 lockdep_assert_held(&cpu_base->lock);
1487
1488 debug_deactivate(timer);
1489 base->running = timer;
1490
1491 /*
1492 * Separate the ->running assignment from the ->state assignment.
1493 *
1494 * As with a regular write barrier, this ensures the read side in
1495 * hrtimer_active() cannot observe base->running == NULL &&
1496 * timer->state == INACTIVE.
1497 */
1498 raw_write_seqcount_barrier(&base->seq);
1499
1500 __remove_hrtimer(timer, base, HRTIMER_STATE_INACTIVE, 0);
1501 fn = timer->function;
1502
1503 /*
1504 * Clear the 'is relative' flag for the TIME_LOW_RES case. If the
1505 * timer is restarted with a period then it becomes an absolute
1506 * timer. If its not restarted it does not matter.
1507 */
1508 if (IS_ENABLED(CONFIG_TIME_LOW_RES))
1509 timer->is_rel = false;
1510
1511 /*
1512 * The timer is marked as running in the CPU base, so it is
1513 * protected against migration to a different CPU even if the lock
1514 * is dropped.
1515 */
1516 raw_spin_unlock_irqrestore(&cpu_base->lock, flags);
1517 trace_hrtimer_expire_entry(timer, now);
1518 expires_in_hardirq = lockdep_hrtimer_enter(timer);
1519
1520 restart = fn(timer);
1521
1522 lockdep_hrtimer_exit(expires_in_hardirq);
1523 trace_hrtimer_expire_exit(timer);
1524 raw_spin_lock_irq(&cpu_base->lock);
1525
1526 /*
1527 * Note: We clear the running state after enqueue_hrtimer and
1528 * we do not reprogram the event hardware. Happens either in
1529 * hrtimer_start_range_ns() or in hrtimer_interrupt()
1530 *
1531 * Note: Because we dropped the cpu_base->lock above,
1532 * hrtimer_start_range_ns() can have popped in and enqueued the timer
1533 * for us already.
1534 */
1535 if (restart != HRTIMER_NORESTART &&
1536 !(timer->state & HRTIMER_STATE_ENQUEUED))
1537 enqueue_hrtimer(timer, base, HRTIMER_MODE_ABS);
1538
1539 /*
1540 * Separate the ->running assignment from the ->state assignment.
1541 *
1542 * As with a regular write barrier, this ensures the read side in
1543 * hrtimer_active() cannot observe base->running.timer == NULL &&
1544 * timer->state == INACTIVE.
1545 */
1546 raw_write_seqcount_barrier(&base->seq);
1547
1548 WARN_ON_ONCE(base->running != timer);
1549 base->running = NULL;
1550 }
1551
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used
2023-11-12 0:52 kernel test robot
@ 2023-11-20 15:17 ` Sebastian Andrzej Siewior
0 siblings, 0 replies; 4+ messages in thread
From: Sebastian Andrzej Siewior @ 2023-11-20 15:17 UTC (permalink / raw)
To: kernel test robot; +Cc: oe-kbuild-all, linux-kernel, Thomas Gleixner
On 2023-11-12 08:52:44 [+0800], kernel test robot wrote:
> Hi Sebastian,
Hi,
> FYI, the error/warning still remains.
>
> kernel/time/hrtimer.c:120:35: warning: initialized field overwritten [-Woverride-init]
> 120 | [CLOCK_REALTIME] = HRTIMER_BASE_REALTIME,
> | ^~~~~~~~~~~~~~~~~~~~~
> kernel/time/hrtimer.c:120:35: note: (near initialization for 'hrtimer_clock_to_base_table[0]')
> kernel/time/hrtimer.c:121:35: warning: initialized field overwritten [-Woverride-init]
> 121 | [CLOCK_MONOTONIC] = HRTIMER_BASE_MONOTONIC,
> | ^~~~~~~~~~~~~~~~~~~~~~
> kernel/time/hrtimer.c:121:35: note: (near initialization for 'hrtimer_clock_to_base_table[1]')
> kernel/time/hrtimer.c:122:35: warning: initialized field overwritten [-Woverride-init]
> 122 | [CLOCK_BOOTTIME] = HRTIMER_BASE_BOOTTIME,
> | ^~~~~~~~~~~~~~~~~~~~~
> kernel/time/hrtimer.c:122:35: note: (near initialization for 'hrtimer_clock_to_base_table[7]')
> kernel/time/hrtimer.c:123:35: warning: initialized field overwritten [-Woverride-init]
> 123 | [CLOCK_TAI] = HRTIMER_BASE_TAI,
> | ^~~~~~~~~~~~~~~~
> kernel/time/hrtimer.c:123:35: note: (near initialization for 'hrtimer_clock_to_base_table[11]')
This is "okay". The array is first initialized completely to a safe
value and then a few of the individual values are overwritten.
So this "initialized field overwritten", that the compiler complains
about, is intended here.
> kernel/time/hrtimer.c: In function '__run_hrtimer':
> >> kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used [-Wunused-but-set-variable]
> 1483 | bool expires_in_hardirq;
> | ^~~~~~~~~~~~~~~~~~
This is because the code using it is disabled at build time. I could
tweak the code to make look it used in the off case so the warning goes
away.
Sebastian
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2023-11-20 15:17 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-09-11 7:54 kernel/time/hrtimer.c:1483:14: warning: variable 'expires_in_hardirq' set but not used kernel test robot
-- strict thread matches above, loose matches on Subject: below --
2023-11-12 0:52 kernel test robot
2023-11-20 15:17 ` Sebastian Andrzej Siewior
2023-11-19 4:35 kernel test robot
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox