Re: [PATCH] ARM: tegra: cpuidle: use CPUIDLE_FLAG_TIMER_STOP flag

[Stephen Warren <swarren-3lzwWm7+Weoh9ZMKESR00Q@public.gmane.org>]

On 07/17/2013 04:15 AM, Joseph Lo wrote:
> On Wed, 2013-07-17 at 03:51 +0800, Stephen Warren wrote:
>> On 07/16/2013 05:17 AM, Joseph Lo wrote:
>>> On Tue, 2013-07-16 at 02:04 +0800, Stephen Warren wrote:
>>>> On 06/25/2013 03:23 AM, Joseph Lo wrote:
>>>>> Use the CPUIDLE_FLAG_TIMER_STOP and let the cpuidle framework
>>>>> to handle the CLOCK_EVT_NOTIFY_BROADCAST_ENTER/EXIT when entering
>>>>> this state.
... [ discussion of issues with Joesph's patches applied]
>
> OK. I did more stress tests last night and today. I found it cause by
> the patch "ARM: tegra: cpuidle: use CPUIDLE_FLAG_TIMER_STOP flag" and
> only impact the Tegra20 platform. The hot plug regression seems due to
> this patch. After dropping this patch on top of v3.11-rc1, the Tegra20
> can back to normal.
> 
> And the hop plug and suspend stress test can pass on Tegra30/114 too.
> 
> Can the other two patch series for Tegra114 to support CPU idle power
> down mode and system suspend still moving forward, not be blocked by
> this patch?
> 
> Looks the CPUIDLE_FLAG_TIMER_STOP flag still cause some other issue for
> hot plug on Tegra20, I will continue to check this. You can just drop
> this patch.

OK, if I drop that patch, then everything on Tegra20 and Tegra30 seems
fine again.

However, I've found some new and exciting issue on Tegra114!

With unmodified v3.11-rc1, I can do the following without issue:

* Unplug/replug CPUs, so that I had all combinations of CPU 1, 2, 3
plugged/unpplugged (with CPU 0 always plugged).

* Unplug/replug CPUs, so that I had all combinations of CPU 0, 1, 2, 3
plugged/unpplugged (with the obvious exception of never having all CPUs
unplugged).

However, if I try this with your Tegra114 cpuidle and suspend patches
applied, I see the following issues:

1) If I boot, unplug CPU 0, then replug CPU 0, the system immediately
hard-hangs.

2) If I run the hotplug test script, leaving CPU 0 always present, I
sometimes see:

> root@localhost:~# for i in `seq 1 50`; do echo ITERATION $i; ./cpuonline.py; done
> ITERATION 1
> echo 0 > /sys/devices/system/cpu/cpu2/online
> [  458.910054] CPU2: shutdown
> echo 0 > /sys/devices/system/cpu/cpu1/online
> [  461.004371] CPU1: shutdown
> echo 0 > /sys/devices/system/cpu/cpu3/online
> [  463.027341] CPU3: shutdown
> echo 1 > /sys/devices/system/cpu/cpu1/online
> [  465.061412] CPU1: Booted secondary processor
> echo 1 > /sys/devices/system/cpu/cpu2/online
> [  467.095313] CPU2: Booted secondary processor
> [  467.113243] ------------[ cut here ]------------
> [  467.117948] WARNING: CPU: 2 PID: 0 at kernel/time/tick-broadcast.c:667 tick_broadcast_oneshot_control+0x19c/0x1c4()
> [  467.128352] Modules linked in:
> [  467.131455] CPU: 2 PID: 0 Comm: swapper/2 Not tainted 3.11.0-rc1-00022-g7487363-dirty #49
> [  467.139678] [<c0015620>] (unwind_backtrace+0x0/0xf8) from [<c001154c>] (show_stack+0x10/0x14)
> [  467.148228] [<c001154c>] (show_stack+0x10/0x14) from [<c05135a8>] (dump_stack+0x80/0xc4)
> [  467.156336] [<c05135a8>] (dump_stack+0x80/0xc4) from [<c0024590>] (warn_slowpath_common+0x64/0x88)
> [  467.165300] [<c0024590>] (warn_slowpath_common+0x64/0x88) from [<c00245d0>] (warn_slowpath_null+0x1c/0x24)
> [  467.174959] [<c00245d0>] (warn_slowpath_null+0x1c/0x24) from [<c00695e4>] (tick_broadcast_oneshot_control+0x19c/0x1c4)
> [  467.185659] [<c00695e4>] (tick_broadcast_oneshot_control+0x19c/0x1c4) from [<c0067cdc>] (clockevents_notify+0x1b0/0x1dc)
> [  467.196538] [<c0067cdc>] (clockevents_notify+0x1b0/0x1dc) from [<c034f348>] (cpuidle_idle_call+0x11c/0x168)
> [  467.206292] [<c034f348>] (cpuidle_idle_call+0x11c/0x168) from [<c000f134>] (arch_cpu_idle+0x8/0x38)
> [  467.215359] [<c000f134>] (arch_cpu_idle+0x8/0x38) from [<c0061038>] (cpu_startup_entry+0x60/0x134)
> [  467.224325] [<c0061038>] (cpu_startup_entry+0x60/0x134) from [<800083d8>] (0x800083d8)
> [  467.232227] ---[ end trace ea579be22a00e7fb ]---
> echo 0 > /sys/devices/system/cpu/cpu1/online
> [  469.126682] CPU1: shutdown

I have found no solution for (1) (although I didn't look hard!).

(2) can be solved with the following (at least 50 iterations of my test
script worked with this patch applied):

> diff --git a/arch/arm/mach-tegra/cpuidle-tegra114.c b/arch/arm/mach-tegra/cpuidle-tegra114.c
> index 658b205..896408d 100644
> --- a/arch/arm/mach-tegra/cpuidle-tegra114.c
> +++ b/arch/arm/mach-tegra/cpuidle-tegra114.c
> @@ -66,8 +66,7 @@ static struct cpuidle_driver tegra_idle_driver = {
>                         .exit_latency           = 500,
>                         .target_residency       = 1000,
>                         .power_usage            = 0,
> -                       .flags                  = CPUIDLE_FLAG_TIME_VALID |
> -                                                 CPUIDLE_FLAG_TIMER_STOP,
> +                       .flags                  = CPUIDLE_FLAG_TIME_VALID,
>                         .name                   = "powered-down",
>                         .desc                   = "CPU power gated",
>                 },

Here's my test script for reference:

#!/usr/bin/env python

import multiprocessing
import os
import sys
import time

cpus = multiprocessing.cpu_count()
if cpus == 4:
  socf = file('/sys/devices/soc0/soc_id')
  soc = socf.readline().strip()
  socf.close()
  if True: #soc == '48':
    gc = (11, 9, 1, 3, 7, 5, 13, 15)
  else:
    gc = (14, 10, 11, 9, 8, 1, 3, 2, 6, 7, 5, 4, 12, 13, 15)
elif cpus == 2:
  gc = (1, 3)
else:
  raise Exception("Invalid CPU count %d" % cpus)

oldidx = len(gc) - 1
oldmask = gc[oldidx]

for newidx in range(len(gc)):
  newmask = gc[newidx]
  for cpu in range(cpus):
    oldon = oldmask & (1 << cpu)
    newon = newmask & (1 << cpu)
    if oldon != newon:
      if newon:
        newonval = 1
      else:
        newonval = 0
      cmd = "echo %d > /sys/devices/system/cpu/cpu%d/online" \
% (newonval, cpu)
      print cmd
      os.system(cmd)
  time.sleep(2)
  oldidx = newidx
  oldmask = newmask