linux.git/drivers/clocksource, branch v4.19.39

clocksource/drivers/arch_timer: Workaround for Allwinner A64 timer instability

2019-03-23T19:09:58+00:00

commit c950ca8c35eeb32224a63adc47e12f9e226da241 upstream.

The Allwinner A64 SoC is known[1] to have an unstable architectural
timer, which manifests itself most obviously in the time jumping forward
a multiple of 95 years[2][3]. This coincides with 2^56 cycles at a
timer frequency of 24 MHz, implying that the time went slightly backward
(and this was interpreted by the kernel as it jumping forward and
wrapping around past the epoch).

Investigation revealed instability in the low bits of CNTVCT at the
point a high bit rolls over. This leads to power-of-two cycle forward
and backward jumps. (Testing shows that forward jumps are about twice as
likely as backward jumps.) Since the counter value returns to normal
after an indeterminate read, each "jump" really consists of both a
forward and backward jump from the software perspective.

Unless the kernel is trapping CNTVCT reads, a userspace program is able
to read the register in a loop faster than it changes. A test program
running on all 4 CPU cores that reported jumps larger than 100 ms was
run for 13.6 hours and reported the following:

 Count | Event
-------+---------------------------
  9940 | jumped backward      699ms
   268 | jumped backward     1398ms
     1 | jumped backward     2097ms
 16020 | jumped forward       175ms
  6443 | jumped forward       699ms
  2976 | jumped forward      1398ms
     9 | jumped forward    356516ms
     9 | jumped forward    357215ms
     4 | jumped forward    714430ms
     1 | jumped forward   3578440ms

This works out to a jump larger than 100 ms about every 5.5 seconds on
each CPU core.

The largest jump (almost an hour!) was the following sequence of reads:
    0x0000007fffffffff → 0x00000093feffffff → 0x0000008000000000

Note that the middle bits don't necessarily all read as all zeroes or
all ones during the anomalous behavior; however the low 10 bits checked
by the function in this patch have never been observed with any other
value.

Also note that smaller jumps are much more common, with backward jumps
of 2048 (2^11) cycles observed over 400 times per second on each core.
(Of course, this is partially explained by lower bits rolling over more
frequently.) Any one of these could have caused the 95 year time skip.

Similar anomalies were observed while reading CNTPCT (after patching the
kernel to allow reads from userspace). However, the CNTPCT jumps are
much less frequent, and only small jumps were observed. The same program
as before (except now reading CNTPCT) observed after 72 hours:

 Count | Event
-------+---------------------------
    17 | jumped backward      699ms
    52 | jumped forward       175ms
  2831 | jumped forward       699ms
     5 | jumped forward      1398ms

Further investigation showed that the instability in CNTPCT/CNTVCT also
affected the respective timer's TVAL register. The following values were
observed immediately after writing CNVT_TVAL to 0x10000000:

 CNTVCT             | CNTV_TVAL  | CNTV_CVAL          | CNTV_TVAL Error
--------------------+------------+--------------------+-----------------
 0x000000d4a2d8bfff | 0x10003fff | 0x000000d4b2d8bfff | +0x00004000
 0x000000d4a2d94000 | 0x0fffffff | 0x000000d4b2d97fff | -0x00004000
 0x000000d4a2d97fff | 0x10003fff | 0x000000d4b2d97fff | +0x00004000
 0x000000d4a2d9c000 | 0x0fffffff | 0x000000d4b2d9ffff | -0x00004000

The pattern of errors in CNTV_TVAL seemed to depend on exactly which
value was written to it. For example, after writing 0x10101010:

 CNTVCT             | CNTV_TVAL  | CNTV_CVAL          | CNTV_TVAL Error
--------------------+------------+--------------------+-----------------
 0x000001ac3effffff | 0x1110100f | 0x000001ac4f10100f | +0x1000000
 0x000001ac40000000 | 0x1010100f | 0x000001ac5110100f | -0x1000000
 0x000001ac58ffffff | 0x1110100f | 0x000001ac6910100f | +0x1000000
 0x000001ac66000000 | 0x1010100f | 0x000001ac7710100f | -0x1000000
 0x000001ac6affffff | 0x1110100f | 0x000001ac7b10100f | +0x1000000
 0x000001ac6e000000 | 0x1010100f | 0x000001ac7f10100f | -0x1000000

I was also twice able to reproduce the issue covered by Allwinner's
workaround[4], that writing to TVAL sometimes fails, and both CVAL and
TVAL are left with entirely bogus values. One was the following values:

 CNTVCT             | CNTV_TVAL  | CNTV_CVAL
--------------------+------------+--------------------------------------
 0x000000d4a2d6014c | 0x8fbd5721 | 0x000000d132935fff (615s in the past)
Reviewed-by: Marc Zyngier 

========================================================================

Because the CPU can read the CNTPCT/CNTVCT registers faster than they
change, performing two reads of the register and comparing the high bits
(like other workarounds) is not a workable solution. And because the
timer can jump both forward and backward, no pair of reads can
distinguish a good value from a bad one. The only way to guarantee a
good value from consecutive reads would be to read _three_ times, and
take the middle value only if the three values are 1) each unique and
2) increasing. This takes at minimum 3 counter cycles (125 ns), or more
if an anomaly is detected.

However, since there is a distinct pattern to the bad values, we can
optimize the common case (1022/1024 of the time) to a single read by
simply ignoring values that match the error pattern. This still takes no
more than 3 cycles in the worst case, and requires much less code. As an
additional safety check, we still limit the loop iteration to the number
of max-frequency (1.2 GHz) CPU cycles in three 24 MHz counter periods.

For the TVAL registers, the simple solution is to not use them. Instead,
read or write the CVAL and calculate the TVAL value in software.

Although the manufacturer is aware of at least part of the erratum[4],
there is no official name for it. For now, use the kernel-internal name
"UNKNOWN1".

[1]: https://github.com/armbian/build/commit/a08cd6fe7ae9
[2]: https://forum.armbian.com/topic/3458-a64-datetime-clock-issue/
[3]: https://irclog.whitequark.org/linux-sunxi/2018-01-26
[4]: https://github.com/Allwinner-Homlet/H6-BSP4.9-linux/blob/master/drivers/clocksource/arm_arch_timer.c#L272

Acked-by: Maxime Ripard 
Tested-by: Andre Przywara 
Signed-off-by: Samuel Holland 
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Lezcano 
Signed-off-by: Greg Kroah-Hartman

clocksource/drivers/exynos_mct: Clear timer interrupt when shutdown

2019-03-23T19:09:58+00:00

commit d2f276c8d3c224d5b493c42b6cf006ae4e64fb1c upstream.

When shutting down the timer, ensure that after we have stopped the
timer any pending interrupts are cleared. This fixes a problem when
suspending, as interrupts are disabled before the timer is stopped,
so the timer interrupt may still be asserted, preventing the system
entering a low power state when the wfi is executed.

Signed-off-by: Stuart Menefy 
Reviewed-by: Krzysztof Kozlowski 
Tested-by: Marek Szyprowski 
Cc:  # v4.3+
Signed-off-by: Daniel Lezcano 
Signed-off-by: Greg Kroah-Hartman

clocksource/drivers/exynos_mct: Move one-shot check from tick clear to ISR

2019-03-23T19:09:58+00:00

commit a5719a40aef956ba704f2aa1c7b977224d60fa96 upstream.

When a timer tick occurs and the clock is in one-shot mode, the timer
needs to be stopped to prevent it triggering subsequent interrupts.
Currently this code is in exynos4_mct_tick_clear(), but as it is
only needed when an ISR occurs move it into exynos4_mct_tick_isr(),
leaving exynos4_mct_tick_clear() just doing what its name suggests it
should.

Signed-off-by: Stuart Menefy 
Reviewed-by: Krzysztof Kozlowski 
Tested-by: Marek Szyprowski 
Cc: stable@vger.kernel.org # v4.3+
Signed-off-by: Daniel Lezcano 
Signed-off-by: Greg Kroah-Hartman

clocksource: timer-ti-dm: Fix pwm dmtimer usage of fck reparenting

2019-03-23T19:09:42+00:00

[ Upstream commit 983a5a43ec254cd5ddf3254db80ca96e8f8bb2a4 ]

Commit 84badc5ec5fc ("ARM: dts: omap4: Move l4 child devices to probe
them with ti-sysc") moved some omap4 timers to probe with ti-sysc
interconnect target module. Turns out this broke pwm-omap-dmtimer
where we now try to reparent the clock to itself with the following:

omap_dm_timer_of_set_source: failed to set parent

With ti-sysc, we can now configure the clock sources in the dts
with assigned-clocks and assigned-clock-parents. So we should be able
to remove omap_dm_timer_of_set_source with clean-up patches later on.
But for now, let's just fix it first by checking if parent and fck
are the same and bail out of so.

Fixes: 84badc5ec5fc ("ARM: dts: omap4: Move l4 child devices to probe them with ti-sysc")
Cc: Bartosz Golaszewski 
Cc: Daniel Lezcano 
Cc: H. Nikolaus Schaller 
Cc: Keerthy 
Cc: Ladislav Michl 
Cc: Pavel Machek 
Cc: Sebastian Reichel 
Cc: Tero Kristo 
Cc: Thierry Reding 
Cc: Thomas Gleixner 
Reported-by: H. Nikolaus Schaller 
Tested-By: Andreas Kemnade 
Tested-By: H. Nikolaus Schaller 
Signed-off-by: Tony Lindgren 
Signed-off-by: Sasha Levin

clocksource/drivers/integrator-ap: Add missing of_node_put()

2019-01-26T08:32:42+00:00

[ Upstream commit 5eb73c831171115d3b4347e1e7124a5a35d8086c ]

The function of_find_node_by_path() acquires a reference to the node
returned by it and that reference needs to be dropped by its caller.

integrator_ap_timer_init_of() doesn't do that.  The pri_node and the
sec_node are used as an identifier to compare against the current
node, so we can directly drop the refcount after getting the node from
the path as it is not used as pointer.

By dropping the refcount right after getting it, a single variable is
needed instead of two.

Fix this by use a single variable and drop the refcount right after
of_find_node_by_path().

Signed-off-by: Yangtao Li 
Signed-off-by: Daniel Lezcano 
Signed-off-by: Sasha Levin

clocksource/drivers/arc_timer: Utilize generic sched_clock

2019-01-09T16:38:42+00:00

commit bf287607c80f24387fedb431a346dc67f25be12c upstream.

It turned out we used to use default implementation of sched_clock()
from kernel/sched/clock.c which was as precise as 1/HZ, i.e.
by default we had 10 msec granularity of time measurement.

Now given ARC built-in timers are clocked with the same frequency as
CPU cores we may get much higher precision of time tracking.

Thus we switch to generic sched_clock which really reads ARC hardware
counters.

This is especially helpful for measuring short events.
That's what we used to have:
------------------------------>8------------------------
$ perf stat /bin/sh -c /root/lmbench-master/bin/arc/hello > /dev/null

 Performance counter stats for '/bin/sh -c /root/lmbench-master/bin/arc/hello':

         10.000000      task-clock (msec)         #    2.832 CPUs utilized
                 1      context-switches          #    0.100 K/sec
                 1      cpu-migrations            #    0.100 K/sec
                63      page-faults               #    0.006 M/sec
           3049480      cycles                    #    0.305 GHz
           1091259      instructions              #    0.36  insn per cycle
            256828      branches                  #   25.683 M/sec
             27026      branch-misses             #   10.52% of all branches

       0.003530687 seconds time elapsed

       0.000000000 seconds user
       0.010000000 seconds sys
------------------------------>8------------------------

And now we'll see:
------------------------------>8------------------------
$ perf stat /bin/sh -c /root/lmbench-master/bin/arc/hello > /dev/null

 Performance counter stats for '/bin/sh -c /root/lmbench-master/bin/arc/hello':

          3.004322      task-clock (msec)         #    0.865 CPUs utilized
                 1      context-switches          #    0.333 K/sec
                 1      cpu-migrations            #    0.333 K/sec
                63      page-faults               #    0.021 M/sec
           2986734      cycles                    #    0.994 GHz
           1087466      instructions              #    0.36  insn per cycle
            255209      branches                  #   84.947 M/sec
             26002      branch-misses             #   10.19% of all branches

       0.003474829 seconds time elapsed

       0.003519000 seconds user
       0.000000000 seconds sys
------------------------------>8------------------------

Note how much more meaningful is the second output - time spent for
execution pretty much matches number of cycles spent (we're runnign
@ 1GHz here).

Signed-off-by: Alexey Brodkin 
Cc: Daniel Lezcano 
Cc: Vineet Gupta 
Cc: Thomas Gleixner 
Cc: stable@vger.kernel.org
Acked-by: Vineet Gupta 
Signed-off-by: Daniel Lezcano 
Signed-off-by: Greg Kroah-Hartman

clockevents/drivers/i8253: Add support for PIT shutdown quirk

2018-11-21T08:19:20+00:00

commit 35b69a420bfb56b7b74cb635ea903db05e357bec upstream.

Add support for platforms where pit_shutdown() doesn't work because of a
quirk in the PIT emulation. On these platforms setting the counter register
to zero causes the PIT to start running again, negating the shutdown.

Provide a global variable that controls whether the counter register is
zero'ed, which platform specific code can override.

Signed-off-by: Michael Kelley 
Signed-off-by: Thomas Gleixner 
Cc: "gregkh@linuxfoundation.org" 
Cc: "devel@linuxdriverproject.org" 
Cc: "daniel.lezcano@linaro.org" 
Cc: "virtualization@lists.linux-foundation.org" 
Cc: "jgross@suse.com" 
Cc: "akataria@vmware.com" 
Cc: "olaf@aepfle.de" 
Cc: "apw@canonical.com" 
Cc: vkuznets 
Cc: "jasowang@redhat.com" 
Cc: "marcelo.cerri@canonical.com" 
Cc: KY Srinivasan 
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/1541303219-11142-2-git-send-email-mikelley@microsoft.com
Signed-off-by: Greg Kroah-Hartman

clocksource/drivers/timer-atmel-pit: Properly handle error cases

2018-09-27T10:01:45+00:00

The smatch utility reports a possible leak:

smatch warnings:
drivers/clocksource/timer-atmel-pit.c:183 at91sam926x_pit_dt_init() warn: possible memory leak of 'data'

Ensure data is freed before exiting with an error.

Reported-by: Dan Carpenter 
Signed-off-by: Alexandre Belloni 
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Lezcano

clocksource/drivers/fttmr010: Fix set_next_event handler

2018-09-24T04:13:31+00:00

Currently, the aspeed MATCH1 register is updated to  in set_next_event handler, with the assumption that COUNT
register value is preserved when the timer is disabled and it continues
decrementing after the timer is enabled. But the assumption is wrong:
RELOAD register is loaded into COUNT register when the aspeed timer is
enabled, which means the next event may be delayed because timer
interrupt won't be generated until <0xFFFFFFFF - current_count +
cycles>.

The problem can be fixed by updating RELOAD register to , and
COUNT register will be re-loaded when the timer is enabled and interrupt
is generated when COUNT register overflows.

The test result on Facebook Backpack-CMM BMC hardware (AST2500) shows
the issue is fixed: without the patch, usleep(100) suspends the process
for several milliseconds (and sometimes even over 40 milliseconds);
after applying the fix, usleep(100) takes averagely 240 microseconds to
return under the same workload level.

Signed-off-by: Tao Ren 
Reviewed-by: Linus Walleij 
Tested-by: Lei YU 
Signed-off-by: Daniel Lezcano

clocksource/drivers/ti-32k: Add CLOCK_SOURCE_SUSPEND_NONSTOP flag for non-am43 SoCs

2018-09-13T13:23:35+00:00

The 32k clocksource is NONSTOP for non-am43 SoCs. Hence
add the flag for all the other SoCs.

Reported-by: Tony Lindgren 
Signed-off-by: Keerthy 
Acked-by: Tony Lindgren 
Signed-off-by: Daniel Lezcano