time/sched_clock: Use raw_read_seqcount_latch()
authorAhmed S. Darwish <a.darwish@linutronix.de>
Thu, 16 Jul 2020 05:11:25 +0000 (13:11 +0800)
committerWill Deacon <will@kernel.org>
Mon, 20 Jul 2020 10:50:47 +0000 (11:50 +0100)
commitaadd6e5caaacd6feca9691ba30536e7de5a7d152
treef787a5427d498335af8c1abbb1b1ff3091f18624
parent1b86abc1c645ad5c9c7bf70910cb3ce73939d2d7
time/sched_clock: Use raw_read_seqcount_latch()

sched_clock uses seqcount_t latching to switch between two storage
places protected by the sequence counter. This allows it to have
interruptible, NMI-safe, seqcount_t write side critical sections.

Since 7fc26327b756 ("seqlock: Introduce raw_read_seqcount_latch()"),
raw_read_seqcount_latch() became the standardized way for seqcount_t
latch read paths. Due to the dependent load, it also has one read
memory barrier less than the currently used raw_read_seqcount() API.

Use raw_read_seqcount_latch() for the seqcount_t latch read path.

Signed-off-by: Ahmed S. Darwish <a.darwish@linutronix.de>
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Link: https://lkml.kernel.org/r/20200625085745.GD117543@hirez.programming.kicks-ass.net
Link: https://lkml.kernel.org/r/20200715092345.GA231464@debian-buster-darwi.lab.linutronix.de
Link: https://lore.kernel.org/r/20200716051130.4359-3-leo.yan@linaro.org
References: 1809bfa44e10 ("timers, sched/clock: Avoid deadlock during read from NMI")
Signed-off-by: Will Deacon <will@kernel.org>
kernel/time/sched_clock.c