Bug 144625

Summary: hald freezes on hal-hotplug-map callout
Product: [Fedora] Fedora Reporter: Sean Middleditch <sean>
Component: halAssignee: David Zeuthen <davidz>
Status: CLOSED RAWHIDE QA Contact:
Severity: high Docs Contact:
Priority: medium    
Version: rawhideCC: mclasen
Target Milestone: ---   
Target Release: ---   
Hardware: i386   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2005-03-21 17:54:02 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
log of strace hald --daemon=no --verbose=yes
none
log of: hald --daemon=no --verbose=yes
none
log of: strace hald --daemon=no
none
*correct* log of: strace hald --daemon=no none

Description Sean Middleditch 2005-01-09 23:45:34 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.5)
Gecko/20041220 Epiphany/1.4.7

Description of problem:
hald started freezing on me after a hard reboot on the hal-hotplug-map
callout.  I completely removed and reinstalled hal in case something
had become corrupted (yes, using a journalad FS, XFS in particular)
but that changed nothing.

An strace on the executable shows it freezing after forking the
callout and calling write().  The callout finishes and is left in a
zombie state.  Presumably (I don't know the code) HAL is expecting to
write something to the child and is getting stuck in a race condition
when it exits unexpectedly.

Will attach a log of:
strace /usr/bin/hald --daemon=no --verbose=yes

Because HAL never finishes initializing, no HAL-using components work.
 In particular, Nautilus/gnome-vfs-daemon start behaving very badly. 
(Will file a separate bug.)

Version-Release number of selected component (if applicable):
hal-0.4.2.cvs20041210-1 dbus-0.22-12
dbus-0.22-12 

How reproducible:
Always

Steps to Reproduce:
1. start HAL
2.
3.
    

Additional info:

Comment 1 Sean Middleditch 2005-01-09 23:48:39 UTC
Created attachment 109543 [details]
log of strace hald --daemon=no --verbose=yes

Comment 2 David Zeuthen 2005-01-13 19:10:55 UTC
Could you test this against hal-0.4.5; I've made some fixes to the
callout code that I think will fix this. 0.4.5 is available in both
Rawhide and as a FC3 update.

Thanks,
David

Comment 3 Sean Middleditch 2005-01-13 21:25:34 UTC
Created attachment 109739 [details]
log of: hald --daemon=no --verbose=yes

Comment 4 Sean Middleditch 2005-01-13 21:26:01 UTC
Created attachment 109740 [details]
log of: strace hald --daemon=no

Comment 5 Sean Middleditch 2005-01-13 21:27:14 UTC
above two logs are from hal-0.4.5-1, still seems to lock up.

Comment 6 Sean Middleditch 2005-01-13 21:29:06 UTC
Created attachment 109741 [details]
*correct* log of: strace hald --daemon=no

Comment 7 David Zeuthen 2005-01-17 20:45:53 UTC
Is the hald process left in the D state?

Comment 8 David Zeuthen 2005-03-21 17:54:02 UTC
There is no hal-hotplug-map callout in the latest hal in Rawhide so I can now
close this bug :-)

Anyway, now that hal is more careful wrt handling hardware, this problem
shouldn't occur anymore. If it does, feel free to reopen this bug.