Not sure if related, but my Samsung Chromebox can no longer recognize MTP after the latest Beta update. Used to be fine before.


8172.25.0 (Official Build) beta-channel stumpy

#CBC-RS/TC-watchlist

This seems to be impacting USB connection of many devices to Chromebooks. We are seeing scattered sports on CBC.

Typo (autocorrect!) - "reports", not "sports". :-)

I can reproduce this using Nexus 5X and Pixel 1.
I did bisection for test images and the result is:

- R51-8100.0.0 (chrome:51.0.2687.0/platform:8100.0.0) => Good
- R51-8101.0.0 (chrome:51.0.2687.0/platform:8101.0.0) => Bad

 Issue 610949  has been merged into this issue.

Also happens on Nexus 5X, 6P, and Pixel C on my Chell system.

Here's my feedback report : https://feedback.corp.google.com/#/Report/8507940863

Here's another one reported from another user : 
https://feedback.corp.google.com/#/Report/8495999343

abodehna@, could you help me triage this bug?

From platform 8101, Chrome can't receive signal "MTPStorageAttached" from mtpd.
mtpd didn't change, and in the change list I could not find a suspicious change.
https://crosland.corp.google.com/log/8100.0.0..8101.0.0
I'd like to ask someone who should know more about it.

--------------------------------
What I conformed is:

* Platform 8101 is the first bad version.

- R51-8100.0.0 (chrome:51.0.2687.0/platform:8100.0.0) => The last good version
- R51-8101.0.0 (chrome:51.0.2687.0/platform:8101.0.0) => The first bad version
Changes between 8100 and 8101 are listed here.
https://crosland.corp.google.com/log/8100.0.0..8101.0.0

* In the bad version, Chrome doesn't receive dbus call from mtpd. 

MediaTransferProtocolDaemonClientImpl::MediaTransferProtocolDaemonClientImpl::OnMTPStorageSignal is not called.
https://code.google.com/p/chromium/codesearch#chromium/src/device/media_transfer_protocol/media_transfer_protocol_daemon_client.cc&sq=package:chromium&type=cs&l=479&q=OnMTPStorageSignal
We wait the signal whose interface is "org.chromium.Mtpd" and signal name is "MTPStorageAttached".

Any progress on this issue? We are still hearing reports of this with Chromebooks and all Android devices.

gwendal@ any suggestion of someone to take a closer look?

Here's another report : https://feedback.corp.google.com/#/Report/8893209417

Here's yet another report : https://feedback.corp.google.com/#/Report/8977384241

Any update on this one?

This is marked as ReleaseBlock-Stable and we are looking to get a stable RC build tomorrow, is this something we are going to fix in the next week or should this be punted to R52?

Guys, is there any plan forward to dig deeper into this and find a fix? Fukino managed to bisect the builds in comment #12, and something definitely changed in R51. We shouldn't ship 51 with a known regression to MTP.

Investigated this issue and found that this is caused because mtpd fails to start a thread at the following line. That call does not return.

https://chromium.googlesource.com/chromiumos/platform/mtpd/+/master/device_manager.cc#751

When I comment out the line, file manager becomes to show the device.

abodenha@: do you have any idea why the call becomes not to return? or could you re-route this issue to someone who might know well about this?

thestig@: Added you since you might know something about why this happens. Please let us know if you have any idea here. Thanks.

yawano: Can you try not starting mtpd automatically, and instead, run it manually outside of minijail? If that works, then it's likely a sandboxing issue.

 Issue 613815  has been merged into this issue.

thestig@: it seems that setpriority becomes necessary to start new thread after https://crrev.com/1660273004. Uploaded a CL to add setpriority in seccomp of mtpd at https://chromium-review.googlesource.com/#/c/347913/

yawano@ it sounds like your CL should fix this. Right?

Same behavior with the Nexus 5X and the Toshiba CB 30 running beta channel (Version 51.0.2704.55 beta (64-bit). Cannot mount the device on the Chromebook.

Re:#24
Yes, yawano@'s CL should fix this issue. (he is OOO until Wednesday)

We are going for a stable RC Wednesday night and since this is a blocker, is there any chance we can land Yuki's change today on master so we have a build on ToT before we merge back?

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/mtpd/+/6a4c1e8f38acfaa0ad2f7054eac6f0c4bfcbad3b

commit 6a4c1e8f38acfaa0ad2f7054eac6f0c4bfcbad3b
Author: Yuki Awano <yawano@chromium.org>
Date: Fri May 27 05:49:55 2016

mtpd: add setpriority to seccomp

BUG= chromium:600780 
TEST=manually tested as described in the issue

Change-Id: Ifef59fce726af2f72ba5f1e146c0caa69e6ea45a
Reviewed-on: https://chromium-review.googlesource.com/347913
Commit-Ready: Naoki Fukino <fukino@chromium.org>
Tested-by: Naoki Fukino <fukino@chromium.org>
Reviewed-by: David Reveman <reveman@chromium.org>
Reviewed-by: Jorge Lucangeli Obes <jorgelo@chromium.org>
Reviewed-by: Bernie Thompson <bhthompson@chromium.org>

[modify] https://crrev.com/6a4c1e8f38acfaa0ad2f7054eac6f0c4bfcbad3b/mtpd-seccomp-x86.policy
[modify] https://crrev.com/6a4c1e8f38acfaa0ad2f7054eac6f0c4bfcbad3b/mtpd-seccomp-arm.policy
[modify] https://crrev.com/6a4c1e8f38acfaa0ad2f7054eac6f0c4bfcbad3b/mtpd-seccomp-amd64.policy

The patch has landed on master branch. (I hope tomorrow's ToT gets it)

It looks like this is on 8397, if someone can verify that works real quick we can merge this back to R51 and R52. 

Was just able to verify this seems to work on 8397, putting this in for the R51 beta build tonight (stable RC). 

Assuming fixed. 

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/mtpd/+/a99bce98af8ad3c06df91bf81698b0a8c7572dd7

commit a99bce98af8ad3c06df91bf81698b0a8c7572dd7
Author: Yuki Awano <yawano@chromium.org>
Date: Fri May 27 05:49:55 2016

mtpd: add setpriority to seccomp

BUG= chromium:600780 
TEST=manually tested as described in the issue

Change-Id: Ifef59fce726af2f72ba5f1e146c0caa69e6ea45a
Reviewed-on: https://chromium-review.googlesource.com/347913
Commit-Ready: Naoki Fukino <fukino@chromium.org>
Tested-by: Naoki Fukino <fukino@chromium.org>
Reviewed-by: David Reveman <reveman@chromium.org>
Reviewed-by: Jorge Lucangeli Obes <jorgelo@chromium.org>
Reviewed-by: Bernie Thompson <bhthompson@chromium.org>
(cherry picked from commit 6a4c1e8f38acfaa0ad2f7054eac6f0c4bfcbad3b)
Reviewed-on: https://chromium-review.googlesource.com/348821
Commit-Queue: Bernie Thompson <bhthompson@chromium.org>
Tested-by: Bernie Thompson <bhthompson@chromium.org>

[modify] https://crrev.com/a99bce98af8ad3c06df91bf81698b0a8c7572dd7/mtpd-seccomp-x86.policy
[modify] https://crrev.com/a99bce98af8ad3c06df91bf81698b0a8c7572dd7/mtpd-seccomp-arm.policy
[modify] https://crrev.com/a99bce98af8ad3c06df91bf81698b0a8c7572dd7/mtpd-seccomp-amd64.policy

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/mtpd/+/d955d08ec42ad58d22dec88c6fac13800220ce32

commit d955d08ec42ad58d22dec88c6fac13800220ce32
Author: Yuki Awano <yawano@chromium.org>
Date: Fri May 27 05:49:55 2016

mtpd: add setpriority to seccomp

BUG= chromium:600780 
TEST=manually tested as described in the issue

Change-Id: Ifef59fce726af2f72ba5f1e146c0caa69e6ea45a
Reviewed-on: https://chromium-review.googlesource.com/347913
Commit-Ready: Naoki Fukino <fukino@chromium.org>
Tested-by: Naoki Fukino <fukino@chromium.org>
Reviewed-by: David Reveman <reveman@chromium.org>
Reviewed-by: Jorge Lucangeli Obes <jorgelo@chromium.org>
Reviewed-by: Bernie Thompson <bhthompson@chromium.org>
(cherry picked from commit 6a4c1e8f38acfaa0ad2f7054eac6f0c4bfcbad3b)
Reviewed-on: https://chromium-review.googlesource.com/348820
Commit-Queue: Bernie Thompson <bhthompson@chromium.org>
Tested-by: Bernie Thompson <bhthompson@chromium.org>

[modify] https://crrev.com/d955d08ec42ad58d22dec88c6fac13800220ce32/mtpd-seccomp-x86.policy
[modify] https://crrev.com/d955d08ec42ad58d22dec88c6fac13800220ce32/mtpd-seccomp-arm.policy
[modify] https://crrev.com/d955d08ec42ad58d22dec88c6fac13800220ce32/mtpd-seccomp-amd64.policy

Thank you for doing them!

Fixed in today's Canary! Thank you :)

Please I beg you, do not email me anylonger

Hit the the star button to follow/unfollow in the top right corner for future reference.

Verified on ChromeOs:8172.47.0/Chrome:51.0.2704.79

I'm still not seeing this fixed!

Version 52.0.2743.19 dev (64-bit)
Platform 8350.14.0 (Official Build) dev-channel samus
Firmware Google_Samus.6300.174.0

Do we need another merge for M52?

akishore@ and thestig@ :

No need for another merge : https://chromium-review.googlesource.com/#/c/348820/2 

Crosland reports that this Landed in 8350.15.0

Please try again with the next dev-channel release!

verified on Lars with 8172.56.0/51.0.2704.103 and nexus phone.

Using 2 different HP Chromebooks and Galaxy S7 Edge. Both on official firmware and still have this issue. Have had it for over a year and still happening now. Can anyone confirm (to a noob) if the fix for this was ever released into the stable channel? I have tried the developer / beta builds (but not canary) and found the issue to persist. Not sure if i'm doing something wrong or if it still a known issue. Also can anyone point me to a thread for the HDMI output issue that still persists as i have struggled to find one.

Many thanks.