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

Issue 714236 link

Starred by 5 users

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

Ding-Dong: the HDP IRQ is unreliable

Reported by marc.her...@intel.com, Apr 21 2017

Issue description

UserAgent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36
Platform: DingDong dongle

Steps to reproduce the problem:
From Kevin K Wong at https://chromium-review.googlesource.com/c/481399

With a back-to-back HPD IRQ pulse (500us low width / 2ms high width),
dingdong is not able to send out the same number of PD message that
matches the HPD from a MST monitor.

What is the expected behavior?
MST is reliable

What went wrong?
MST is slower or not reliable

Did this work before? N/A 

Chrome version:   Channel: canary
OS Version: https://chromium.googlesource.com/chromiumos/platform/ec/+log/391056f9eea24c/board/dingdong
Flash Version: N/A
 

Comment 1 by vpalatin@google.com, May 15 2017

Cc: sha...@chromium.org vpalatin@chromium.org tbroch@chromium.org
Components: OS>Firmware>EC
Labels: -Via-Wizard-Other Arch-ARM
Summary: Ding-Dong: the HDP IRQ is unreliable (was: Ding-Dong: the HDP IRQ is dead)
Thanks Todd for sharing the following information with respect to the firmware version of Ding Dong (and others):

https://chromium.googlesource.com/chromiumos/third_party/kernel/+log/chromeos-4.4/drivers/mfd/cros_ec_pd_update.c
https://chromium-review.googlesource.com/c/chromiumos/third_party/kernel/+/276700/2/drivers/mfd/cros_ec_pd_update.c
https://chromium-review.googlesource.com/c/chromiumos/third_party/kernel/+/465354/2/drivers/mfd/cros_ec_pd_update.c

The (secure) hashes in the static const struct cros_ec_pd_firmware_image firmware_images[] array appear in the output of "ectool infopddev 0"

Left out the most useful command:
   dmesg | grep cros_ec_pd

I was working on another bug and noticed few things which I thought would be useful as data point here or if someone wants to reproduce on production device.

I was using APL/Electro and connecting MST Monitor using type-c/dingdong cable.
I tried connecting/reconnecting mst monitor 10 time and all of this time my monitor comes up within 20s. 

NOTE: Kernel v4.4 has by default mst turned off. Enable by doing "echo Y > /sys/module/i915/parameters/enable_dp_mst && restart ui"

Configuration:

MST Monitor: Viewsonic VG2249

Board Info:

 Board: reef
    OS: 9901.0.0 (Official Build) dev-channel reef test
Chrome: Google Chrome 62.0.3201.0 unknown
    FW: Google_Reef.9042.87.0
    EC: reef_v1.1.5909-bd1f0c9 2017-05-10 18:51:15 @build169-m2.golo.chromium.org
   CPU: Intel(R) Celeron(R) CPU N3350 @ 1.10GHz
   Mem: 4G (physical)
   Mem: 3.7G (usable)

Dingdong fw version:   
[ 1133.375525] cros_ec_pd_update GOOG0003:00: Port1 FW is already up-to-date cros-pd/dingdong_v1.7.684-69498dd.bin
Cc: marc...@chromium.org
Owner: tbroch@chromium.org
Status: Assigned (was: Unconfirmed)
Todd can you take a look at this one?
For convenience, linking to MST background information:

- A lot of information and many MST related links in  issue 329267  starting from comment 33

- Similar EC HPD issue fixed (access now restricted?)
 https://bugs.chromium.org/p/chromium/issues/detail?id=711334
 https://chromium-review.googlesource.com/c/chromiumos/platform/ec/+/477414

- https://partnerissuetracker.corp.google.com/issues/63923404
  Enabling MST in chromeos-3.18 breaks Ctrl-Alt-F2 frecon
 (should probably have been filed in this tracker, feel free to transfer)

- issue 698290 screen off is randomly selected when more than 3

- issue 775757 switching resolution reverts other monitor to default resolution (MST or not)
  
- Partner issues (access restricted) https://partnerissuetracker.corp.google.com/issues?q=intel_dp_mst

BTW DingDong doesn't work at all with a Macbook Pro, not even in SST mode. The screen is flashing a couple times and then the Mac stops trying. Reproduced with multiple monitors (another non-Apple Type-C -> DP dongle works fine in the same situations)
> Left out the most useful command:
>   dmesg | grep cros_ec_pd
>   cros_ec_pd_update GOOG0003:00: Port1 FW is already up-to-date cros-pd/dingdong_v1.7.6...

BTW it's not clear when/why this update feature runs versus not. Only on AC? Only once per boot? I suggest trial and error.

Sign in to add a comment