On 7/7/07, Theodore Tso <[EMAIL PROTECTED]> wrote:
I found the problem. It was actually a big blkid_set_tag() which
failed to clear dev->bid_type when it should have. Thanks for the bug
report; the details you gave were very helpful in figuring out what
was going on.
Glad I could help. Some
tags 432052 +pending
thanks
I found the problem. It was actually a big blkid_set_tag() which
failed to clear dev->bid_type when it should have. Thanks for the bug
report; the details you gave were very helpful in figuring out what
was going on.
This is a pretty serious bug, so I'll be doing an
I can reproduce this bug with the disk image at:
http://olpc.download.redhat.com/olpc/streams/development/build488/devel_ext3/olpc-redhat-stream-development-build-488-20070705_0237-devel_ext3.img.bz2
(Unzipped and with the offset parameter given in the bug report, since
this is a whole-disk image
On Fri, Jul 06, 2007 at 09:49:16PM -0400, C. Scott Ananian wrote:
> So, 'dev->bid_type' is garbage when blkid_verify is called, and that's
> causing an infinite loop. It is supposed to be NULL if we don't know
> what the requested block device type is (the original command was
> 'sudo mount -o loo
Oh, can you also send me your current /etc/blkid.tab file?
Thanks!!
- Ted
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Can you send me a copy of the xo-initrd-test.img file?
Thanks,
- Ted
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
FWIW, here's some more interesting gdb bits:
Breakpoint 1, 0xb7e7dd68 in strcmp () from /lib/i686/cmov/libc.so.6
1: x/i $pc 0xb7e7dd68 : mov(%ecx),%al
(gdb) print (char*)$ecx
$5 = 0xb7f5dd2a "ocfs"
(gdb) print (char*)$edx
$6 = 0x805e338 "`ã\005\b¸â\005\bÀâ\005\b!"
(gdb) finish
Run till
7 matches
Mail list logo
