Issue metadata
Sign in to add a comment
|
Regression: Recovery installation via USB is failing on Gnawty, Paine, Falco, Butterfly, Samus,Peppy,wolf,Zako,Glimmer and Link |
||||||||||||||||||||||||||||||||
Issue descriptionVersion: 55.0.2869.0/8837.0.0 (Official Build) dev-channel Falco OS: Chrome What steps will reproduce the problem? (1) Try to recover the build 55.0.2869.0/8837.0.0 via USB>> and observe Expected: Should be able to recover the build. Actual: Instead unable to recover the build (Blank screen is seen on recovery and after few minutes it's going to existing build which was there in earlier) This is a Regression issue as it is working fine in M53: 53.0.2785.123/8530.87.0 (Official Build) stable-channel Falco Note: Recovery is fine with the same version M55: 55.0.2869.0/8837.0.0 (Official Build) dev-channel Gnawty, Paine, Spring, Peppy
,
Sep 26 2016
If autoupdate is successful, this should be marked as Stable blocker.
,
Sep 26 2016
successful auto updated to 8838.0.0 on falco.
,
Sep 26 2016
Do you have the logs handy from the failed recovery attempt? We had some other recovery failures recently involving Flashrom, and I am wondering if this might be the same thing, they should be fixed though, the logs can tell us.
,
Sep 26 2016
Recovery stick does not contain recovery logs. After inserting recovery stick: - The device turned black, & then return with dev screen - Recovery process never start This issue is similar to https://code.google.com/p/chrome-os-partner/issues/detail?id=57345
,
Sep 30 2016
Issue is still seen in M55: 55.0.2874.0/8848.0.0 (Official Build) dev-channel Falco Note: Auto update to latest Dev channel is also not working now
,
Oct 10 2016
Still able to reproduce this issue on M55 build 8872.5.0
,
Oct 10 2016
,
Oct 10 2016
Is this related to the recovery app ?
,
Oct 10 2016
Nope. This is pure recovery mechanism. No Recovery tool needed to re-image the stick.
,
Oct 18 2016
Still reproducible in ChromeOS: 8872.15.0/55.0.2883.17 on Falco-li
,
Oct 19 2016
,
Oct 21 2016
Issue is still reproducible with the build M56: 56.0.2895.0/8918.0.0 dev-channel Falco
,
Oct 24 2016
Issue is still seen in 56.0.2890./8928.0.0 dev -channel Wolf, Gnawty, Glimmer, Zako as well. @Can some one from Dev team please look into this issue.
,
Oct 24 2016
+reinauer Stefan, it sounds like per comment 5 we may not be making it past the firmware portion of the recovery process, whom is the right person to dig into this? If we have no logs otherwise, this might require a serial enabled firmware :-/.
,
Oct 24 2016
,
Oct 24 2016
It could be firmware or early kernel. In my experience it's usually the latter... once the firmware turns off the screen there's very little left that could go wrong. Next step would be to run on Servo with both serial-enabled firmware and earlycon kernel. Command line arguments can either be injected with a small firmware patch (just a strcpy() at the end of depthcharge/src/boot/commandline.c) or by reciting the correct futility incantations over the recovery stick's kernel partition.
,
Oct 24 2016
Issue 658649 has been merged into this issue.
,
Oct 24 2016
Shelley is debugging right now with one falco device from FW team.
,
Oct 24 2016
Shelley, assigning this to you as you're looking into it. Please pass along to right owner if need be.
,
Oct 24 2016
Tracking M56 dev blocker version of this bug in crbug/658649
,
Oct 24 2016
,
Oct 25 2016
If this is only in recovery, we can punt to stable for now.
,
Oct 26 2016
THis looks like a duplicate of https://code.google.com/p/chrome-os-partner/issues/detail?id=58976 but that is in the partner bug database, so I cannot duplicate.. We are working on this pretty hard. We need at most a couple of days to find the culprit. According to the labels here, I see this does not block dev releases. So, is it ok to take a couple of days to fix the issue before we have to revert the change for LLVM migration?
,
Oct 26 2016
Issue is still seen in M56: 56.0.2899.0/8934.0.0 dev-channel Gnawty, Paine, Falco, Butterfly, Samus and Link as well
,
Oct 26 2016
bleung@ says we are blocking dogfooding. So, we will revert the LLVM AMD64 roll for now. revert at: https://chromium-review.googlesource.com/#/c/403288/
,
Oct 26 2016
Updating it to ReleaseBlock-Dev as we are unable to work on many devices on ToT. Feel free to update if not appropriate.
,
Oct 26 2016
,
Oct 26 2016
After chatting offline with Kalin, agreed that this is not release block dev for M56 as recovery sticks isn't the usual way to get build.
,
Oct 26 2016
Re #28 - Thanks for pushing the priority high. It seems shchen@ and team are working on it. I want to propose ReleaseBlock-Beta, so to reflect on the dogfooding being blocked, and the variety of devices where recovery is failing.
,
Oct 26 2016
Updating ownership to the proper people.
,
Oct 26 2016
,
Oct 26 2016
note that by #27, we have reverted the change. So, at the moment this is not blocking anything. We are working on identifying the culprit.
,
Oct 26 2016
,
Oct 26 2016
Recovery images after and including 8935.0 should no longer have this issue. Please take a look.
,
Oct 26 2016
,
Oct 26 2016
Note that LLVM was rolled early on Falco on M-55, so this may still be a releaseblock-beta... Busybox is the culprit right now, so inorder to unbreak the falco recovery images on 55, this needs to be picked back : https://chromium-review.googlesource.com/#/c/404148/
,
Oct 26 2016
,
Oct 27 2016
,
Oct 27 2016
The change in comment 38 appears to cherry-pick cleanly, if we are confident this will resolve the problem in 55 let us go ahead with it.
,
Oct 27 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/b97dbb3b367295d7638b40bc9b60e0801a037c77 commit b97dbb3b367295d7638b40bc9b60e0801a037c77 Author: Yunlian Jiang <yunlian@chromium.org> Date: Wed Oct 26 23:04:11 2016 use gcc to build busybox. Building busybox with llvm breaks recovery image. So we use gcc to build it instead. BUG= chromium:650227 TEST=the recovery image works. Change-Id: Ibd37ab79ad315451ed0863fe7f7b2974930c7e7e Reviewed-on: https://chromium-review.googlesource.com/403437 Reviewed-by: Benson Leung <bleung@chromium.org> Commit-Queue: Bernie Thompson <bhthompson@chromium.org> Tested-by: Bernie Thompson <bhthompson@chromium.org> [add] https://crrev.com/b97dbb3b367295d7638b40bc9b60e0801a037c77/chromeos/config/env/sys-apps/busybox
,
Oct 27 2016
thanks Bernie!
,
Oct 27 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/d3c29bcdb5370e0b76eeee20a1040b4605cf300a commit d3c29bcdb5370e0b76eeee20a1040b4605cf300a Author: Yunlian Jiang <yunlian@chromium.org> Date: Wed Oct 26 23:04:11 2016 use gcc to build busybox. Building busybox with llvm breaks recovery image. So we use gcc to build it instead. BUG= chromium:650227 TEST=the recovery image works. Change-Id: Ibd37ab79ad315451ed0863fe7f7b2974930c7e7e Reviewed-on: https://chromium-review.googlesource.com/404148 Reviewed-by: Benson Leung <bleung@chromium.org> Commit-Queue: Yunlian Jiang <yunlian@chromium.org> Tested-by: Yunlian Jiang <yunlian@chromium.org> [add] https://crrev.com/d3c29bcdb5370e0b76eeee20a1040b4605cf300a/chromeos/config/env/sys-apps/busybox
,
Oct 27 2016
Since we landed the change, assuming this no longer blocks 55 beta.
,
Oct 27 2016
Thank you all to look into this on priority, will update it once we have latest M56 ToT and R55-8872.B available with the Fix landed.
,
Oct 27 2016
Able to recover falco-li/samus/link/peppy with USB stick in 8872.27.0/55.0.2883.29
,
Oct 28 2016
Recovery is successful with the below devices 1. M55: 55.0.2883.29/8872.27.0 dev-channel Falco, Gnawty, Butterfly, Peppy, Paine, Link, Glimmer, Samus 2. M56: 56.0.2900.0/8940.0.0 dev-channel auron_paine, Peach-Pit, Jerry, Falco, Samus, Peppy Thanks!!!
,
Oct 30 2016
This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible! If all merges have been completed, please remove any remaining Merge-Approved labels from this issue. Thanks for your time! To disable nags, add the Disable-Nags label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Nov 1 2016
yunlian@ : Could you please update the Status to Fixed as its already verified.
,
Nov 1 2016
,
Nov 2 2016
This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible! If all merges have been completed, please remove any remaining Merge-Approved labels from this issue. Thanks for your time! To disable nags, add the Disable-Nags label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Nov 2 2016
,
Nov 8 2016
#CBC-RS/TC-watchlist
,
Nov 9 2016
,
Mar 15 2017
|
|||||||||||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||||||||||
Comment 1 by pucchakayala@chromium.org
, Sep 26 2016