OpenVZ Forum


Home » Mailing lists » Devel » Re: [PATCH][usercr]: Ghost tasks must be detached
Re: [PATCH][usercr]: Ghost tasks must be detached [message #41627 is a reply to message #41623] Thu, 10 February 2011 22:31 Go to previous message
Sukadev Bhattiprolu is currently offline  Sukadev Bhattiprolu
Messages: 413
Registered: August 2006
Senior Member
Louis Rilling [Louis.Rilling@kerlabs.com] wrote:
| On 10/02/11 9:54 -0800, Sukadev Bhattiprolu wrote:
| > Louis Rilling [Louis.Rilling@kerlabs.com] wrote:

| > If it was a detached task, do_wait_thread() skipped over it. In the C/R
| > kernel we were setting the ->exit_signal to -1 for a "ghost" process.
| > I assumed that the container-init exited without waiting for the "ghost"
| > and we were getting the crash in proc_flush_mnt() when the ghost exited.
|
| The point is that wait_consider_task() skips detached tasks as soon as they are
| not ptraced. So removing the check in do_wait_thread() should not have changed
| the behavior. Am I missing something?

No. I was :-) You are right that it did not change the behavior. I still
need to investigate why the crash does not occur on 2.6.37 even without
Eric's fix.

Sukadev
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containe rs
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: [PATCH] Reduce uidhash lock hold time when lookup succeeds
Next Topic: [PATCH] Don't crash if we are self-checkpointing with a child
Goto Forum:
  


Current Time: Thu Aug 28 08:46:50 GMT 2025

Total time taken to generate the page: 0.12592 seconds