I sent this to the autofs4 maintainer a while ago, but never heard back.
autofs4 doesn't return a size for the symlinks it creates, which is
inconsistent with other filesystems. (The Almquist shell uses the sizes
of path components to allocate buffers during a walk, so it can't traverse
autofs4-linked paths.)
This patch applies against 2.4.9 (and probably earlier) through 2.4.13.
--- fs/autofs4/inode.c.orig Fri Feb 9 14:29:44 2001
+++ fs/autofs4/inode.c Thu Aug 23 16:01:59 2001
@@ -315,8 +315,10 @@
inode->i_nlink = 2;
inode->i_op = &autofs4_dir_inode_operations;
inode->i_fop = &autofs4_dir_operations;
- } else if (S_ISLNK(inf->mode))
+ } else if (S_ISLNK(inf->mode)) {
+ inode->i_size = inf->size;
inode->i_op = &autofs4_symlink_inode_operations;
+ }
return inode;
}
Followup to: <[email protected]>
By author: [email protected]
In newsgroup: linux.dev.kernel
>
> I sent this to the autofs4 maintainer a while ago, but never heard back.
> autofs4 doesn't return a size for the symlinks it creates, which is
> inconsistent with other filesystems. (The Almquist shell uses the sizes
> of path components to allocate buffers during a walk, so it can't traverse
> autofs4-linked paths.)
>
That's IMNSHO a bug in the Almquist shell, at least if autofs4 returns
zero, but it's also a bug in autofs4.
-hpa
--
<[email protected]> at work, <[email protected]> in private!
"Unix gives you enough rope to shoot yourself in the foot."
http://www.zytor.com/~hpa/puzzle.txt <[email protected]>
On Wed, Oct 24, 2001 at 01:24:10PM -0700, H. Peter Anvin wrote:
> >
> > I sent this to the autofs4 maintainer a while ago, but never heard back.
> > autofs4 doesn't return a size for the symlinks it creates, which is
> > inconsistent with other filesystems. (The Almquist shell uses the sizes
> > of path components to allocate buffers during a walk, so it can't traverse
> > autofs4-linked paths.)
>
> That's IMNSHO a bug in the Almquist shell, at least if autofs4 returns
> zero, but it's also a bug in autofs4.
True, ash could do things differently, but I don't think expecting the
filesystem to return correct metadata is a bug. Should we distrust stat(2)
and count the bytes in a file ourselves, just to be sure?