New issue
Advanced search Search tips

Issue 878523 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner: ----
Closed: Oct 16
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

run_oci: crash during startup in libcontainer.so

Reported by mfie...@students.op97.org, Aug 28

Issue description

IMPORTANT: Your crash has already been automatically reported to our crash system. Please file this bug only if you can provide more information about it.


Chrome Version: 67.0.3396.99
Operating System: Linux 4.4.127-13890-g6c206f1711d4

URL (if applicable) where crash occurred:

Can you reproduce this crash?

What steps will reproduce this crash? (If it's not reproducible, what were you doing just before the crash?)
1.using google chrome
2.
3.

****DO NOT CHANGE BELOW THIS LINE****
Crash ID: crash/ee03e251dc14bd60

 
Labels: OS-Chrome
As per the above provided crash id in C#0, issue seems to be specific to chrome OS hence adding the same OS to proceed further.

Thanks..!
Components: OS>Systems>Containers
run_oci is the executable
Cc: vapier@chromium.org
Components: -OS>Systems>Containers Platform>Apps>ARC
Labels: -Restrict-View-EditIssue
Status: Fixed (was: Unconfirmed)
Summary: run_oci: crash during startup in libcontainer.so (was: information being stolen)
run_oci is only used by the ARC++ code nowadays.  containers has moved on to only using VM stuff.

from a quick search, i'm not seeing any crashes beyond R68, and we're stabilizing R70 now.  so if people can't find anything newer, just going to icebox this.

https://crash.corp.google.com/browse?q=product_name%3D%27ChromeOS%27+AND+stable_signature%3D%27_dl_init-c040e0b4%27#-propertyselector,-productname:1000,productversion:30,-stablesignature:50
why is _dl_init the only symbolized one?
The failure is happening in libcontainer/minijail and stack trace doesn't symbolize these for some reason?

it does seem wrong that we're missing symbols from so many modules.  my guess is that it's due to symbols expiring (what with 10452.x being R66), so reports no longer get symbolified, and it works out that ld.so, being fairly stable/simple, happens to generate the same code across releases, so we have matching symbols from newer releases.

the fact that R66 is long out of support, we basically don't care about reports coming from such devices anymore.  i wonder if we should change our crash reporter to stop reporting anything after 6 months.  i'll file a bug about it.
Labels: Pri-2 Type-Bug
issue 915381 fwiw discussing that

Sign in to add a comment