linux-stable.git/kernel/time, branch linux-4.12.y

timers: Fix excessive granularity of new timers after a nohz idle

2017-08-30T08:26:40+00:00

commit 2fe59f507a65dbd734b990a11ebc7488f6f87a24 upstream.

When a timer base is idle, it is forwarded when a new timer is added
to ensure that granularity does not become excessive. When not idle,
the timer tick is expected to increment the base.

However there are several problems:

- If an existing timer is modified, the base is forwarded only after
  the index is calculated.

- The base is not forwarded by add_timer_on.

- There is a window after a timer is restarted from a nohz idle, after
  it is marked not-idle and before the timer tick on this CPU, where a
  timer may be added but the ancient base does not get forwarded.

These result in excessive granularity (a 1 jiffy timeout can blow out
to 100s of jiffies), which cause the rcu lockup detector to trigger,
among other things.

Fix this by keeping track of whether the timer base has been idle
since it was last run or forwarded, and if so then forward it before
adding a new timer.

There is still a case where mod_timer optimises the case of a pending
timer mod with the same expiry time, where the timer can see excessive
granularity relative to the new, shorter interval. A comment is added,
but it's not changed because it is an important fastpath for
networking.

This has been tested and found to fix the RCU softlockup messages.

Testing was also done with tracing to measure requested versus
achieved wakeup latencies for all non-deferrable timers in an idle
system (with no lockup watchdogs running). Wakeup latency relative to
absolute latency is calculated (note this suffers from round-up skew
at low absolute times) and analysed:

             max     avg      std
upstream   506.0    1.20     4.68
patched      2.0    1.08     0.15

The bug was noticed due to the lockup detector Kconfig changes
dropping it out of people's .configs and resulting in larger base
clk skew When the lockup detectors are enabled, no CPU can go idle for
longer than 4 seconds, which limits the granularity errors.
Sub-optimal timer behaviour is observable on a smaller scale in that
case:

	     max     avg      std
upstream     9.0    1.05     0.19
patched      2.0    1.04     0.11

Fixes: Fixes: a683f390b93f ("timers: Forward the wheel clock whenever possible")
Signed-off-by: Nicholas Piggin 
Signed-off-by: Thomas Gleixner 
Tested-by: Jonathan Cameron 
Tested-by: David Miller 
Cc: dzickus@redhat.com
Cc: sfr@canb.auug.org.au
Cc: mpe@ellerman.id.au
Cc: Stephen Boyd 
Cc: linuxarm@huawei.com
Cc: abdhalee@linux.vnet.ibm.com
Cc: John Stultz 
Cc: akpm@linux-foundation.org
Cc: paulmck@linux.vnet.ibm.com
Cc: torvalds@linux-foundation.org
Link: http://lkml.kernel.org/r/20170822084348.21436-1-npiggin@gmail.com
Signed-off-by: Greg Kroah-Hartman

timers: Fix overflow in get_next_timer_interrupt

2017-08-11T15:33:52+00:00

commit 34f41c0316ed52b0b44542491d89278efdaa70e4 upstream.

For e.g. HZ=100, timer being 430 jiffies in the future, and 32 bit
unsigned int, there is an overflow on unsigned int right-hand side
of the expression which results with wrong values being returned.

Type cast the multiplier to 64bit to avoid that issue.

Fixes: 46c8f0b077a8 ("timers: Fix get_next_timer_interrupt() computation")
Signed-off-by: Matija Glavinic Pecotic 
Signed-off-by: Thomas Gleixner 
Reviewed-by: Alexander Sverdlin 
Cc: khilman@baylibre.com
Cc: akpm@linux-foundation.org
Link: http://lkml.kernel.org/r/a7900f04-2a21-c9fd-67be-ab334d459ee5@nokia.com
Signed-off-by: Greg Kroah-Hartman

alarmtimer: don't rate limit one-shot timers

2017-07-27T22:10:25+00:00

Commit ff86bf0c65f1 ("alarmtimer: Rate limit periodic intervals") sets a
minimum bound on the alarm timer interval.  This minimum bound shouldn't
be applied if the interval is 0.  Otherwise, one-shot timers will be
converted into periodic ones.

This patch is specific to 4.11.y and 4.12.y.  Older -stable trees have a
slightly different patch, and 4.13-rc2 isn't impacted due to a later
refactoring.

Fixes: ff86bf0c65f1 ("alarmtimer: Rate limit periodic intervals")
Reported-by: Ben Fennema 
Signed-off-by: Greg Hackmann 
Cc: John Stultz 
Reviewed-by: Thomas Gleixner 
Signed-off-by: Greg Kroah-Hartman

time: Fix CLOCK_MONOTONIC_RAW sub-nanosecond accounting

2017-06-20T08:41:50+00:00

Due to how the MONOTONIC_RAW accumulation logic was handled,
there is the potential for a 1ns discontinuity when we do
accumulations. This small discontinuity has for the most part
gone un-noticed, but since ARM64 enabled CLOCK_MONOTONIC_RAW
in their vDSO clock_gettime implementation, we've seen failures
with the inconsistency-check test in kselftest.

This patch addresses the issue by using the same sub-ns
accumulation handling that CLOCK_MONOTONIC uses, which avoids
the issue for in-kernel users.

Since the ARM64 vDSO implementation has its own clock_gettime
calculation logic, this patch reduces the frequency of errors,
but failures are still seen. The ARM64 vDSO will need to be
updated to include the sub-nanosecond xtime_nsec values in its
calculation for this issue to be completely fixed.

Signed-off-by: John Stultz 
Tested-by: Daniel Mentz 
Cc: Prarit Bhargava 
Cc: Kevin Brodsky 
Cc: Richard Cochran 
Cc: Stephen Boyd 
Cc: Will Deacon 
Cc: "stable #4 . 8+" 
Cc: Miroslav Lichvar 
Link: http://lkml.kernel.org/r/1496965462-20003-3-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner

time: Fix clock->read(clock) race around clocksource changes

2017-06-20T08:41:50+00:00

In tests, which excercise switching of clocksources, a NULL
pointer dereference can be observed on AMR64 platforms in the
clocksource read() function:

u64 clocksource_mmio_readl_down(struct clocksource *c)
{
	return ~(u64)readl_relaxed(to_mmio_clksrc(c)->reg) & c->mask;
}

This is called from the core timekeeping code via:

	cycle_now = tkr->read(tkr->clock);

tkr->read is the cached tkr->clock->read() function pointer.
When the clocksource is changed then tkr->clock and tkr->read
are updated sequentially. The code above results in a sequential
load operation of tkr->read and tkr->clock as well.

If the store to tkr->clock hits between the loads of tkr->read
and tkr->clock, then the old read() function is called with the
new clock pointer. As a consequence the read() function
dereferences a different data structure and the resulting 'reg'
pointer can point anywhere including NULL.

This problem was introduced when the timekeeping code was
switched over to use struct tk_read_base. Before that, it was
theoretically possible as well when the compiler decided to
reload clock in the code sequence:

     now = tk->clock->read(tk->clock);

Add a helper function which avoids the issue by reading
tk_read_base->clock once into a local variable clk and then issue
the read function via clk->read(clk). This guarantees that the
read() function always gets the proper clocksource pointer handed
in.

Since there is now no use for the tkr.read pointer, this patch
also removes it, and to address stopping the fast timekeeper
during suspend/resume, it introduces a dummy clocksource to use
rather then just a dummy read function.

Signed-off-by: John Stultz 
Acked-by: Ingo Molnar 
Cc: Prarit Bhargava 
Cc: Richard Cochran 
Cc: Stephen Boyd 
Cc: stable 
Cc: Miroslav Lichvar 
Cc: Daniel Mentz 
Link: http://lkml.kernel.org/r/1496965462-20003-2-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner

tick/broadcast: Make tick_broadcast_setup_oneshot() static

2017-06-12T16:56:01+00:00

This function isn't used outside of tick-broadcast.c, so let's
mark it static.

Signed-off-by: Stephen Boyd 
Link: http://lkml.kernel.org/r/20170608063603.13276-1-sboyd@codeaurora.org
Signed-off-by: Thomas Gleixner

alarmtimer: Rate limit periodic intervals

2017-06-04T13:21:18+00:00

The alarmtimer code has another source of potentially rearming itself too
fast. Interval timers with a very samll interval have a similar CPU hog
effect as the previously fixed overflow issue.

The reason is that alarmtimers do not implement the normal protection
against this kind of problem which the other posix timer use:

  timer expires -> queue signal -> deliver signal -> rearm timer

This scheme brings the rearming under scheduler control and prevents
permanently firing timers which hog the CPU.

Bringing this scheme to the alarm timer code is a major overhaul because it
lacks all the necessary mechanisms completely.

So for a quick fix limit the interval to one jiffie. This is not
problematic in practice as alarmtimers are usually backed by an RTC for
suspend which have 1 second resolution. It could be therefor argued that
the resolution of this clock should be set to 1 second in general, but
that's outside the scope of this fix.

Signed-off-by: Thomas Gleixner 
Cc: Peter Zijlstra 
Cc: Kostya Serebryany 
Cc: syzkaller 
Cc: John Stultz 
Cc: Dmitry Vyukov 
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/20170530211655.896767100@linutronix.de

alarmtimer: Prevent overflow of relative timers

2017-06-04T13:21:18+00:00

Andrey reported a alartimer related RCU stall while fuzzing the kernel with
syzkaller.

The reason for this is an overflow in ktime_add() which brings the
resulting time into negative space and causes immediate expiry of the
timer. The following rearm with a small interval does not bring the timer
back into positive space due to the same issue.

This results in a permanent firing alarmtimer which hogs the CPU.

Use ktime_add_safe() instead which detects the overflow and clamps the
result to KTIME_SEC_MAX.

Reported-by: Andrey Konovalov 
Signed-off-by: Thomas Gleixner 
Cc: Peter Zijlstra 
Cc: Kostya Serebryany 
Cc: syzkaller 
Cc: John Stultz 
Cc: Dmitry Vyukov 
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/20170530211655.802921648@linutronix.de

posix-timers: Make signal printks conditional

2017-05-23T21:39:57+00:00

A recent commit added extra printks for CPU/RT limits. This can result in
excessive spam in dmesg.

Make the printks conditional on print_fatal_signals.

Fixes: e7ea7c9806a2 ("rlimits: Print more information when CPU/RT limits are exceeded")
Reported-by: Dave Jones 
Signed-off-by: Thomas Gleixner 
Cc: Arun Raghavan

time: delete current_fs_time()

2017-05-12T22:57:15+00:00

All uses of the current_fs_time() function have been replaced by other
time interfaces.

And, its use cases can be fulfilled by current_time() or ktime_get_*
variants.

Link: http://lkml.kernel.org/r/1491613030-11599-13-git-send-email-deepa.kernel@gmail.com
Signed-off-by: Deepa Dinamani 
Reviewed-by: Arnd Bergmann 
Cc: John Stultz 
Cc: Thomas Gleixner 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds