bug-hurd
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Error with Ogi's ext2fs patch


From: Michael Banck
Subject: Error with Ogi's ext2fs patch
Date: Sat, 14 Aug 2004 14:22:04 +0200
User-agent: Mutt/1.5.6+20040523i

Hello,

I was compiling glibc over night two days ago, and when I woke up, I had
a frozen box and one screenful of error messages.

If I remember correctly, I formatted the (3.x GB) partition under
GNU/Linux with -o hurd, mounted it under GNU/Linux and performed the
Debian-specific commands to extract the source tarball.

I then rebooted into GNU/Hurd and built it.

The error message was as follows (minor errors might apply, I
transcribed this from the screen and back):

ext2fs: /dev/hd0s8: warning: directory entry too long: inode: 2 offset: 0
make: getcwd: : Input/output error
ext2fs: /dev/hd0s8: warning: directory entry too long: inode: 2 offset: 0
make: getcwd: : Input/output error
ext2fs: /dev/hd0s8: warning: directory entry too long: inode: 48896 offset: 7422
ext2fs: /dev/hd0s8: warning: bad directory entry: inode: 48896 offset: 4096
ext2fs: /dev/hd0s8: warning: bad directory entry: inode: 2 offset: 0

Then followed the usual couple of lines with 'make error' and 'leaving
directory foo' and (I think) a shell prompt.

At the bottom of the screen the following error message was printed,
probably at a later point:

ext2fs.static: hurdmalloc.c: 246: more_memory: Unexpected error:
        (os/kern) invalid argument

The first line I quoted above was also the first line of the screen, so
maybe some other error messages slipped by already.

The file system seemed to be totally corrupted afterwards, this is a
snippet from the e2fsck run:

--8<--
Too many illegal blocks in inode 70.
        Clear inode<y>? yes
Inode 210 has INDEX_FL flag set but is not a directory.
        Clear HTree index<y>? yes
Inode 210, i_size is 219556133516591753, should be 0.  Fix<y>? yes
Inode 210, i_blocks is 989855744, should be 0.  Fix<y>? yes
Inode 169 has compression flag set on filesystem without compression
        support.  Clear<y>? yes
Inode 150 has INDEX_FL flag set but is not a directory.
Clear HTree index<y>? yes

And later, the two inodes referred above:

Directory inode 48896, block 1, offset 0: directory corrupted
Salvage? yes
Entry 'elf' in /???/???/build-tree/??? (14898) is a link to directory
        /???/???/build-tree/???/elf (48896).
Clear? yes
Pass 4: Checking reference counts
Inode 2 ref count is 0, should be 4.  Fix? yes

These errors occured with lots of other inodes as well.


Michael




reply via email to

[Prev in Thread] Current Thread [Next in Thread]