New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 768495 link

Starred by 2 users

Issue metadata

Status: Fixed
Owner:
Closed: Sep 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

kevin: No touchscreen on chromeos6-row2-rack24-host3

Project Member Reported by mka@chromium.org, Sep 25 2017

Issue description

In recent canary builds (>= R63-9969.0.0) touchscreen related tests often fail because the touchscreen is not detected.

https://wmatrix.googleplex.com/platform/unfiltered?suites=bvt-perbuild&days_back=10&releases=tot&platforms=kevin&builds=R63-99[567].*

An example is http://cautotest-prod.corp.google.com/afe/#tab_id=view_job&object_id=144526184 (R63-9971.0.0):

2017-09-24T20:40:31.739364+00:00 INFO kernel: [    0.425805] rk3x-i2c ff130000.i2c: Initialized RK3xxx I2C bus at ffffff80001f8000
2017-09-24T20:40:31.739369+00:00 WARNING kernel: [    0.426024] 3-004b supply vdd not found, using dummy regulator
2017-09-24T20:40:31.739374+00:00 WARNING kernel: [    0.426078] 3-004b supply avdd not found, using dummy regulator
2017-09-24T20:40:31.739379+00:00 DEBUG kernel: [    0.426116] atmel_mxt_ts 3-004b: GPIO lookup for consumer atmel,reset
2017-09-24T20:40:31.739383+00:00 DEBUG kernel: [    0.426121] atmel_mxt_ts 3-004b: using device tree for GPIO lookup
2017-09-24T20:40:31.739388+00:00 DEBUG kernel: [    0.426128] of_get_named_gpiod_flags: can't parse 'atmel,reset-gpios' property of node '/i2c@ff130000/touchscreen@4b[0]'
2017-09-24T20:40:31.739394+00:00 DEBUG kernel: [    0.426134] of_get_named_gpiod_flags: can't parse 'atmel,reset-gpio' property of node '/i2c@ff130000/touchscreen@4b[0]'
2017-09-24T20:40:31.739407+00:00 DEBUG kernel: [    0.426139] atmel_mxt_ts 3-004b: using lookup tables for GPIO lookup
2017-09-24T20:40:31.739412+00:00 DEBUG kernel: [    0.426145] atmel_mxt_ts 3-004b: lookup for GPIO atmel,reset failed
2017-09-24T20:40:31.739417+00:00 INFO kernel: [    0.426429] rk3x-i2c ff140000.i2c: Initialized RK3xxx I2C bus at ffffff80001fa000
2017-09-24T20:40:31.739421+00:00 WARNING kernel: [    0.426610] 5-004a supply vdd not found, using dummy regulator
2017-09-24T20:40:31.739426+00:00 WARNING kernel: [    0.426678] 5-004a supply avdd not found, using dummy regulator
2017-09-24T20:40:31.739431+00:00 DEBUG kernel: [    0.426711] atmel_mxt_ts 5-004a: GPIO lookup for consumer atmel,reset
2017-09-24T20:40:31.739435+00:00 DEBUG kernel: [    0.426717] atmel_mxt_ts 5-004a: using device tree for GPIO lookup
2017-09-24T20:40:31.739440+00:00 DEBUG kernel: [    0.426726] of_get_named_gpiod_flags: can't parse 'atmel,reset-gpios' property of node '/i2c@ff140000/trackpad@4a[0]'
2017-09-24T20:40:31.739451+00:00 DEBUG kernel: [    0.426732] of_get_named_gpiod_flags: can't parse 'atmel,reset-gpio' property of node '/i2c@ff140000/trackpad@4a[0]'
2017-09-24T20:40:31.739456+00:00 DEBUG kernel: [    0.426737] atmel_mxt_ts 5-004a: using lookup tables for GPIO lookup
2017-09-24T20:40:31.739460+00:00 DEBUG kernel: [    0.426743] atmel_mxt_ts 5-004a: lookup for GPIO atmel,reset failed
2017-09-24T20:40:31.739465+00:00 INFO kernel: [    0.428188] tpm_i2c_infineon 0-0020: 1.2 TPM (device-id 0x1A) [gentle shutdown]
2017-09-24T20:40:31.739469+00:00 INFO kernel: [    0.432518] atmel_mxt_ts 5-004a: Information Block Checksum = 9bd26b
2017-09-24T20:40:31.739474+00:00 INFO kernel: [    0.432530] atmel_mxt_ts 5-004a: No platform data provided

...

2017-09-24T20:40:31.739961+00:00 INFO kernel: [    0.633299] atmel_mxt_ts 5-004a: Family ID: 164 Variant ID: 17 Major.Minor.Build: 2.0.AA
2017-09-24T20:40:31.739965+00:00 INFO kernel: [    0.633308] atmel_mxt_ts 5-004a: Matrix X Size: 24 Matrix Y Size: 14 Object Num: 31
2017-09-24T20:40:31.739970+00:00 INFO kernel: [    0.636799] atmel_mxt_ts 5-004a: T100 Config: XSIZE 1920, YSIZE 1080, XLINE 19, YLINE 11
2017-09-24T20:40:31.739975+00:00 INFO kernel: [    0.637193] atmel_mxt_ts 5-004a: T100 Config: SCRAUX : 0, TCHAUX : 17
2017-09-24T20:40:31.739979+00:00 INFO kernel: [    0.639776] input: Atmel maXTouch Touchpad as /devices/platform/ff140000.i2c/i2c-5/5-004a/input/input2
2017-09-24T20:40:31.739986+00:00 INFO kernel: [    0.647325] atmel_mxt_ts 5-004a: Status: 00 Config Checksum: 0af6ba

Initialization of the touchscreen device (3-004b) starts but is not completed.

So far I didn't have success with reproducing this locally.

I didn't identify any specific CL that might be to blame, the issue could be related with the switch to clang kernel builds with R63-9967.0.0 (CL:673005).
 
Cc: adlr@chromium.org
Hrm.  Something sounds awfully fishy.  Apparently atmel has been behaving strangely on R61 too, but there's not lots of data because it seems like our special atmel logging script isn't working.  See <https://b.corp.google.com/issues/66684735>.

I'm told that R61 stopped behaving badly at .70 though (???).  

---

I believe that some early revs of kevin have problems where the connector can sometimes fail.  Any chance that's what's happening here?  Are all the failures on the same hardware?

Comment 2 by mka@chromium.org, Sep 25 2017

Yes, so far all the failures I've looked into are on chromeos6-row2-rack24-host3.
Cc: alexpau@chromium.org
My suggestion would be to pull that device and hand it off to adlr@ if he'll take it.

For more details about connector problems, b/35775059
...there's also a way to lock the device so the lab won't use it, then ssh over to it and confirm that the touchscreen just won't enumerate on that device...
I've locked it.

It's running this:

CHROMEOS_RELEASE_DESCRIPTION=9974.0.0 (Official Build) dev-channel kevin test

And it's consistently missing both touchscreen and digitizer:

# evtest
No device specified, trying to scan all of /dev/input/event*
Available devices:
/dev/input/event0:	cros_ec
/dev/input/event1:	cros_ec_buttons
/dev/input/event2:	Atmel maXTouch Touchpad
/dev/input/event3:	rk3399-gru-sound Headset Jack
/dev/input/event4:	rk3399-gru-sound DP Jack
/dev/input/event5:	gpio-keys
Select the device event number [0-5]:

Anyone's free to try flashing old images to that DUT (to check for regressions?), or get the lab to pull it.
BTW, locking is done in cautotest:

http://cautotest/afe/#tab_id=view_host&object_id=7052

Comment 7 by mka@chromium.org, Sep 26 2017

Labels: -Pri-1 Pri-2
Summary: kevin: No touchscreen on chromeos6-row2-rack24-host3 (was: kevin: Touchscreen not working)
Thanks,

I tried two older images, 9962.0.0 and 9945.0.0, with both the touchscreen is missing.
Cc: akes...@chromium.org
Owner: xixuan@chromium.org
I _think_ the right thing here is to assign to current Infra Deputy to get the device pulled and replaced with a new Kevin.  Feel free to deliver the bad Kevin to my desk and I'll walk around and poke people to find someone who wants to do a Failure Analysis.

Comment 9 by xixuan@chromium.org, Sep 26 2017

I move this DUT to pool:suites

"""
Removing labels ['pool:bvt'] from host chromeos6-row2-rack24-host3
Adding labels ['pool:suites'] to host chromeos6-row2-rack24-host3
"""

But I don't know how to deliver it since it's supposed in englab.

Any further actions you need? Or I will mark this as fixed or assign back?


> I move this DUT to pool:suites

I personally don't know what that means. Will it get used by the lab if I unlock it? Is it somehow flagged for eventual removal/investigation?
Cc: xixuan@chromium.org
Owner: johndhong@chromium.org
> But I don't know how to deliver it since it's supposed in englab.

I don't know quite what this means.  Is englab not handled by the Infrastructure Deputy?  I guess I must have gotten confused.  Maybe John is the right person?
When a DUT in serious pool (cq, bvt, cts, ...) is broken, deputy's work is to move it to pool:suites, and englab (experts who responsible for maintaining all these hardwares) will fix bad DUTs in this pool (pool:suites) on time.

It may get used if its state is 'Ready'. But tests scheduled on pool:suites can be regarded as 'non-critical' tests, so it doesn't matter there're some bad DUTs in pool:suites that causes tests to fail.
For future reference

Request for DUT Repair/Logs - File a ticket for work on a device in either chromeos2 (Atlantis Lab), chromeos4 (Destiny Lab), or chromeos6 (Prometheus Lab). (go/cros-lab-device-repair)

https://b.corp.google.com/u/0/issues/66950736
Owner: ----
Cc: haoweiw@chromium.org englab-sys-cros@google.com
Owner: xixuan@chromium.org
Status: Started (was: Unconfirmed)
DUT has been repaired and shows touchscreen and digitizer, on multiple reboots. I've unlocked the device.

I guess it should go back in pool:bvt sometime? Or just close this, if we're fine as-is.
Status: Fixed (was: Started)
Currently let's leave it in pool:suites since pool:bvt has enough working DUTs. Next time when there's shortage of DUTs in pool:bvt, it will be moved to pool:bvt then.

Comment 18 by dchan@chromium.org, Jan 22 2018

Status: Archived (was: Fixed)

Comment 19 by dchan@chromium.org, Jan 23 2018

Status: Fixed (was: Archived)

Sign in to add a comment