625606 - closing Thunderbird during 'looking up' stage leaves thunderbird-bin eating 100% CPU - shutdown hang

Status: RESOLVED INCOMPLETE

(Thunderbird :: General, defect)

Description

[nobody10331033]

User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.13) Gecko/20110108 Gentoo Firefox/3.6.13
Build Identifier: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.13) Gecko/20110113 Lightning/1.0b3pre Thunderbird/3.1.7

If I close the Thunderbird window during the 'looking up' stage of connection, the window closes but thunderbird-bin seems to busy-loop, eating 100% CPU until I kill it. I also must kill the process before being able to open Thunderbird again.

Reproducible: Always

Steps to Reproduce:
1. open thunderbird
2. close the window when status bar says 'Looking up' (easier if you have no internet connection)
3. observe thunderbird-bin eating 100% CPU
4. try to open Thunderbird again
5. observe the message saying that the process is running but not responding
6. kill thunderbird-bin
7. open thunderbird
Actual Results:  
Closing the window during DNS lookup leaves thunderbird-bin eating 100% CPU.

Expected Results:  
Closing the window during DNS lookup should quit the process gracefully.

Comment 1

[Wayne Mery (:wsmwk)]

please see https://wiki.mozilla.org/Thunderbird:Testing:Shutdown_Hang

a stacktrace would probably be great

Comment 2

[nobody10331033]

Describe the environment
========================
1. happens in safe mode? yes
2. global indexing enabled? yes
3. account types? one mail account, imap only
4. sleep/hibernate? never
5. offline mode? never
6. change networks? never

7. versions?
thunderbird ua: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.13) Gecko/20110113 Lightning/1.0b3pre Thunderbird/3.1.7
`uname -a`: Linux delan2 2.6.37-gentoo #2 SMP Sat Jan 8 18:16:38 WST 2011 x86_64 Intel(R) Core(TM)2 Quad CPU Q9650 @ 3.00GHz GenuineIntel GNU/Linux

8. frequency? always, when reproduced correctly
9. CPU usage? 100% of one core during hang

10. connections? `netstat -Wtp | grep -i thunderbird`:
tcp        0      0 delan2.local:50853      agaacqmaoc.c03.gridserver.com:imap ESTABLISHED 3009/thunderbird-bi

11. imap:5 log will be attached if you need it

I will attach the imap:5 log and a stacktrace shortly.

Comment 3

[nobody10331033]

Attachment: imap:5 log

Comment 4

[nobody10331033]

How should I obtain a stacktrace in gdb? I have a couple of questions:

1. if the program doesn't crash and doesn't actually quit, how should I get a stack trace?
2. what should I run if I can't run thunderbird/thunderbird because it's a shell script, but running thunderbird/thunderbird-bin without the wrapper fails to load libraries?

Comment 5

[Ludovic Hirlimann [:Usul]]

https://wiki.mozilla.org/Thunderbird:Testing:Get_A_Debug_Thunderbird_Hang_Stack is the way to go.

Comment 6

[nobody10331033]

Done. I've recompiled thunderbird with FEATURES=nostrip before doing this trace, so that all the symbols that are needed are hopefully there.

If you can, please modify the wiki page from

./thunderbird-bin

to

LD_LIBRARY_PATH=. ./thunderbird-bin

so that the thunderbird libraries can be loaded correctly and actually start.

Comment 7

[nobody10331033]

Attachment: hang stacktrace

Comment 8

[Wayne Mery (:wsmwk)]

unfortunately, the stack lacks line numbers

Comment 9

[Ludovic Hirlimann [:Usul]]

Hum looks like a very low level hang. something at the system level :(

But I don't like the last line from the stack at all.

Comment 10

[Wayne Mery (:wsmwk)]

(In reply to comment #9)
> Hum looks like a very low level hang. something at the system level :(
> 
> But I don't like the last line from the stack at all.
"Backtrace stopped: previous frame identical to this frame (corrupt stack?)"

Delan, can you post another stacktrace?

Comment 11

[Sebastian Krämer]

I can't say for sure if it's during lookup but the 100% cpu hang is still there in 3.1.9 release.

Comment 12

[Wayne Mery (:wsmwk)]

Delan,  are you seeing this with v5 or v6 beta or trunk?


Sebastian,   were you using the steps of comment 0? And, do you also still see this when using newer version?

Comment 13

[Joshua Cranmer [:jcranmer]]

RESOLVED INCOMPLETE due to lack of response to previous question. If you feel this change was made in error, please respond to this bug with your reasons why.