BusyBox Bug and Patch Tracking
BusyBox
Main | My View | View Issues | Change Log | Docs
  

Viewing Issue Simple Details [ Jump to Notes ] [ View Advanced ] [ Issue History ] [ Print ]
ID Category Severity Reproducibility Date Submitted Last Update
0000498 [BusyBox] Other major always 10-21-05 20:15 12-14-05 21:11
Reporter smiley73 View Status public  
Assigned To BusyBox
Priority normal Resolution fixed  
Status closed   Product Version
Summary 0000498: mount loop device support broken
Description Starting from some release after 1.0rc1, mounting loop devices (in my case an ISO image) doesn't work right anymore. Right now I use 1.01
Under Kernel 2.4.31 it seems the mount doesn't find a free loop device and you have to specife "mount -o loop=/dev/loop ...."
Under Kernel 2.6.13.2 I couldn't get it to work at all. I tried the latest code from SVN about 3 weeks ago, but it doesn't work either.
Unfortunately I can't give you more information.
Additional Information
Attached Files

- Relationships

- Notes
(0000640)
smiley73
10-24-05 06:44

 I did some further testing and it doesn't seem to be dependend on the Kernel release.
Modifying the mount command to use "-o loop=/dev/loop0" returns the error message, that it's not a valid block device.
Setting up the loopback device via losetup before the mount also doesn't solve the problem.
One test used glibc 2.3.2 (kernel 2.4) and the other glibc 2.3.5 (kernel 2.6).
Both times busybox was statically compiled and used in the InitRD.
 
(0000641)
smiley73
10-24-05 10:28

 Some further testing showed, that busybox 1.00 seems to work fine.
So the error must have been introduced after that.

Heiko
 
(0000656)
landley
11-03-05 18:00

 Does it work in 1.1.0-pre1? This code underwent a major rewrite. The old stuff was really brittle and tended to break a lot.
 
(0000668)
smiley73
11-06-05 09:12

 OK I tested with 1.1.0-pre1, I didn't have any luck.

mount -o loop /mnt/bootcd.iso /cdrom
Returns 'could not set up loop device'

losetup /dev/loop0 /mnt/bootcd.iso
Works the first time, but hangs when it's being run a second time.
And yes, I ran "losetup -d /dev/loop0" bfore.
When running losetup the second time, it doesn't matter if you specify a different loop device, it always hangs.
It also seems that in most cases when this happens, the loop file (the ISO image in my case) actually gets damaged and I have to copy it again.

When first setup the loop device via losetup and then try to mount loop0, then I get the error message: bio too big device loop0 (0 > 2)

Heiko
 
(0000669)
landley
11-06-05 15:14

 Ah sorry, I forgot:

Disable "devfs support". It breaks random stuff. (For example, it means that the loop devices mount is checking for are /dev/loop/# rather than /dev/loop#)

I recently fixed allyesconfig so it wouldn't accidentally enable devfs support and cause this kind of accidental breakage. In 1.2 I hope we can rip it out entirely.

Rob
 
(0000670)
smiley73
11-07-05 17:16

 OK the 1.1.0-pre1 version seems to work a lot better without devfs support.
I have to do more testing, but it takes forever to go through one test cycle...

Heiko
 
(0000775)
landley
12-14-05 21:11

 svn 11912
 

- Issue History
Date Modified Username Field Change
10-21-05 20:15 smiley73 New Issue
10-21-05 20:15 smiley73 Status new => assigned
10-21-05 20:15 smiley73 Assigned To  => BusyBox
10-24-05 06:44 smiley73 Note Added: 0000640
10-24-05 10:28 smiley73 Note Added: 0000641
11-02-05 17:15 pschulz01 Issue Monitored: pschulz01
11-03-05 18:00 landley Note Added: 0000656
11-03-05 18:00 landley Status assigned => feedback
11-06-05 09:12 smiley73 Note Added: 0000668
11-06-05 15:14 landley Note Added: 0000669
11-07-05 17:16 smiley73 Note Added: 0000670
12-14-05 21:11 landley Status feedback => closed
12-14-05 21:11 landley Note Added: 0000775
12-14-05 21:11 landley Resolution open => fixed
12-14-05 21:11 landley Fixed in Version  => svn

Copyright © 2000 - 2006 Mantis Group
Powered by Mantis Bugtracker