nvidia-tegra-modules.git/drivers/watchdog, branch gpu-paging

watchdog: t21x: Remove error while pinging

2020-08-21T15:39:44+00:00

At time of ping there might be chance of the watchdog timer status is
reset thus removing error print.

Bug 200648045

Signed-off-by: Ankit Patel 
Change-Id: Ic9c1db3efdade6c7a8b88db87856b839868025ec
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvidia/+/2400602
Reviewed-by: automaticguardword 
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bibek Basu 
Reviewed-by: mobile promotions 
GVS: Gerrit_Virtual_Submit
Tested-by: mobile promotions

softdog: stop watchdog on suspend

2020-07-28T17:27:36+00:00

To prevent unexpected watchdog fire, stop it on suspend

Bug 200635557

Change-Id: I9ab209e75a411a1a03b89f9f2a921b26ef14ff34
Signed-off-by: dmitry pervushin 
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvidia/+/2384698
Tested-by: mobile promotions 
Reviewed-by: automaticguardword 
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Phoenix Jung 
Reviewed-by: mobile promotions 
GVS: Gerrit_Virtual_Submit

watchdog: t21x: Remove IRQ handling as a daemon

2020-07-08T17:23:54+00:00

As per Bug 200577867 and Bug 2509728, In T210 during warm boot stress
test watchdog reset gets triggered on reboot. Upon the conclusion of the
bug It's looks like tegra watchdog hardware is buggy.

In effort to resolve it, Need to simplify the tegra legacy driver made
for earlier version of the Linux. Also, need to remove watchdog handling
way in which HW bug can affect more.

To resolve this below changes are made,
   o Remove IRQ daemon to ping the watchdog, Now Linux watchdog
   subsystem daemon will ping the tegra timer register
   o Remove restarting of counter in the ping, Now just ping to counter
   register instead of restart it
   o If watchdog id enabled by bootloader, don't restart the counter
   and just set WDOG_HW_RUNNING so Linux watchdog subsystem keep pinging
   the HW
   o If "enable-boot-in" flag has been passed then start the counter and
   set the WDOG_HW_RUNNING
   o Check the watchdog is active/inactive status by the status bit from
   the timer register rather than "watchdog_active()" function.

Bug 200624822 Bug 200577867

Signed-off-by: Ankit Patel 
Change-Id: I175f04fb9864fdf2d80763c6297ade4ac1295b06
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvidia/+/2370395
Reviewed-by: automaticguardword 
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Bibek Basu 
Reviewed-by: mobile promotions 
Tested-by: mobile promotions 
GVS: Gerrit_Virtual_Submit

watchdog: tegra: bring wdt timeout to 1 sec

2019-12-18T16:12:20+00:00

With the current watchdog driver implementation, minimum timeout can
be set to 3 secs while we can provide till 1 sec in timeout-sec dt
property. This change will bring down timeout to minimum 1 sec
irrespective of number of expiry levels.

Bug 200576640

Change-Id: I1cc795b4c27fb3b3a8b02d368e6501b98c640f0f
Signed-off-by: Shubhi Garg 
Reviewed-on: https://git-master.nvidia.com/r/2261784
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Laxman Dewangan 
Reviewed-by: mobile promotions 
Tested-by: mobile promotions

drivers:wdt: Read WDT refresh interval from DT

2019-09-04T19:00:24+00:00

Add support to read refresh interval from DT. If not found use
default value of 60 secs.

Bug 2565917

Change-Id: Icfe8fc5e66541cf408c6f8d0be7ca5fbd7eadb64
Signed-off-by: Hardik Tushar Shah 
(cherry picked from commit a1f804e565bc885df3ff497d0a9837687aa7da32)
Reviewed-on: https://git-master.nvidia.com/r/2146286
Reviewed-on: https://git-master.nvidia.com/r/2142724
Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit
Reviewed-by: Dmitry Pervushin 
Reviewed-by: Phoenix Jung 
Reviewed-by: mobile promotions 
Tested-by: mobile promotions

watchdog: fix nested sleep

2019-03-11T17:58:41+00:00

The current implementation can cause the following warning:

[    0.266234] do not call blocking ops when !TASK_RUNNING; state=2 set at
               [] prepare_to_wait_event+0x5c/0x124
[    0.266257] ------------[ cut here ]------------
[    0.266264] WARNING: CPU: 0 PID: 28 at kernel/sched/core.c:7946 __might_sleep+0x7c/0x84
[    0.266267] Modules linked in:

[    0.266285] CPU: 0 PID: 28 Comm: tegra-hv-wdt Not tainted 4.9.140-tegra #11
[    0.266288] Hardware name: chiem2_t194i (DT)
[    0.266293] task: ffffffc3953d3700 task.stack: ffffffc394d24000
[    0.266297] PC is at __might_sleep+0x7c/0x84
[    0.266300] LR is at __might_sleep+0x7c/0x84
[    0.266310] pc : [] lr : [] pstate: 80800045
[    0.266315] sp : ffffffc394d27d40
[    0.266318] x29: ffffffc394d27d40 x28: 0000000000000000
[    0.266335] x27: 0000000000000000 x26: 0000000000000000
[    0.266342] x25: 0000000000000000 x24: 0000000000000000
[    0.266348] x23: ffffffc394d1e6d8 x22: ffffff800965f000
[    0.266354] x21: 0000000000000000 x20: 0000000000000061
[    0.266360] x19: ffffff8008f33848 x18: 0000000000000000
[    0.266365] x17: 00000000f24bd100 x16: 00000000add69dcb
[    0.266371] x15: 00000000c1952591 x14: 66663c5b20746120
[    0.266401] x13: 000000000000012a x12: 0000000000000000
[    0.266408] x11: 0000000000000004 x10: 0000000000000002
[    0.266420] x9 : 000000000000012a x8 : 0000000000000001
[    0.266426] x7 : 0000000000000000 x6 : ffffff800965fe50
[    0.266432] x5 : ffffffc3a68ee050 x4 : ffffff8009661670
[    0.266458] x3 : 0000000000000001 x2 : 000000000004098c
[    0.266464] x1 : 0000000000000000 x0 : 0000000000000071

[    0.266475] ---[ end trace 1beec3e153d15e03 ]---
[    0.266479] Call trace:
[    0.266486] [] __might_sleep+0x7c/0x84
[    0.266496] [] mutex_lock+0x24/0x40
[    0.266513] [] tegra_hv_wdt_loop+0x134/0x160
[    0.266522] [] kthread+0xf4/0x108
[    0.266528] [] ret_from_fork+0x10/0x30

In short, the code that verifies the condition in wait_event should
not sleep (wait on mutexes), as explained in [1]. So, fixing this.

[1] https://lwn.net/Articles/628628/

Bug 2473707

Change-Id: I70d497554f97863c68fcaeb662915f4c0c894c46
Reviewed-on: https://git-master.nvidia.com/r/2020042
(cherry picked from commit f53660df33a7c2cff1367f319684a7cb91d86c5b)
Reviewed-on: https://git-master.nvidia.com/r/2023865
Reviewed-by: svc-mobile-coverity 
Reviewed-by: Dmitry Pervushin 
Tested-by: Dmitry Pervushin 
GVS: Gerrit_Virtual_Submit
Reviewed-by: Rahul Jain (SW-TEGRA) 
Reviewed-by: mobile promotions 
Tested-by: mobile promotions

watchdog: soft: add platform driver base softwatchdog driver

2018-08-27T19:23:53+00:00

Add platform driver based soft watchdog driver to control the watchdog
parameter through DT. Also it will be possible to control driver status
which helps on enabling/disabling the driver on given platform.

bug 1916745

Signed-off-by: Laxman Dewangan 
Change-Id: I00f8531d6e317fab1e131b77e554a671ed00f1fc
Reviewed-on: http://git-master/r/1471246
(cherry picked from commit 278af0140d832ab04e29b5f796cb34ce74a45d90)
Signed-off-by: Krishna Yarlagadda 
Reviewed-on: https://git-master.nvidia.com/r/1799281
GVS: Gerrit_Virtual_Submit
Reviewed-by: Sachin Nikam 
Reviewed-by: mobile promotions 
Tested-by: mobile promotions

hv-wdt: Stop pinging upon receiving magic char

2018-06-12T18:44:36+00:00

HV WDT used to defer pinging to the user once the watchdog device
file is opened, but failed to return to automatic pinging afterwards.
With this CL, WDT will only stop pinging when it receives the magic
character.

JIRA VFND-3318
Bug 200264246

Change-Id: I64f595505349b92017b6a458dd0571129a3eb8e8
Signed-off-by: Faisal Zaghloul 
Reviewed-on: https://git-master.nvidia.com/r/1746474
Reviewed-by: svc-mobile-coverity 
GVS: Gerrit_Virtual_Submit
Reviewed-by: Automatic_Commit_Validation_User
Reviewed-by: Vladislav Buzov 
Reviewed-by: mobile promotions 
Tested-by: mobile promotions

tegra: watchdog: update compatible string

2018-04-06T16:23:47+00:00

Update compatible string to "nvidia,tegra-wdt-t21x"
to  match watchdog strings of newer tegra soc.

Bug 200353269

Change-Id: Id602946a8ff89df264750cc63811a62f50a58f33
Signed-off-by: Shreshtha SAHU 
Reviewed-on: https://git-master.nvidia.com/r/1684327
Reviewed-by: Krishna Yarlagadda 
Reviewed-by: Varun Wadekar 
GVS: Gerrit_Virtual_Submit
Reviewed-by: mobile promotions 
Tested-by: mobile promotions

watchdog: t21x: Remove WDT IRQ handler only once

2018-03-17T00:11:44+00:00

WDT IRQ is requested when WDT daemon is started early
during kernel boot. WDT daemon is stopped when user
space opens /dev/watchdog and then WDT IRQ is also freed.

If user space open/close /dev/watchdog multiple times,
then WDT IRQ free is getting called each time, even though
WDT daemon was already disabled first time user space
started watchdog.

To prevent this multiple free, IRQ free is moved to driver
remove. Also when user space is taking over WDT ping, IRQ is
disabled to stop WDT daemon from pinging WDT.

Bug 1976162

Change-Id: Iae2b8f78311f1f012b4d4078100662295472f069
Signed-off-by: Shreshtha SAHU 
Reviewed-on: https://git-master.nvidia.com/r/1629847
(cherry picked from commit c7d7331a7cbaf0194c31a5a7de8cb4961e3df245)
Signed-off-by: Shreshtha SAHU 
Reviewed-on: https://git-master.nvidia.com/r/1675997
GVS: Gerrit_Virtual_Submit
Reviewed-by: mobile promotions 
Tested-by: mobile promotions