The routine ext2_readdir() will, when reading a directory page
returns an error, try the next page, without reporting the
error to user space. That is bad, and the patch below changes that.
In my case the filesystem was damaged, and ext2_readdir wanted
to read 60000+ pages and wrote as many error messages to syslog
("attempt to access beyond end"), not what one wants.
Andries
[no doubt a similar patch is appropriate for ext3]
diff -uprN -X /linux/dontdiff a/fs/ext2/dir.c b/fs/ext2/dir.c
--- a/fs/ext2/dir.c 2004-10-30 21:44:02.000000000 +0200
+++ b/fs/ext2/dir.c 2004-11-04 00:14:14.000000000 +0100
@@ -275,7 +275,8 @@ ext2_readdir (struct file * filp, void *
"bad page in #%lu",
inode->i_ino);
filp->f_pos += PAGE_CACHE_SIZE - offset;
- continue;
+ ret = -EIO;
+ goto done;
}
kaddr = page_address(page);
if (need_revalidate) {
On Thu, 4 Nov 2004, Andries Brouwer wrote:
>
> [no doubt a similar patch is appropriate for ext3]
ext3 is different here, and uses the old-style buffer cache rather than
page cache. It has the equivalent case for a hole and/or IO error, and
like ext2, it just continues with the next block.
I _think_ that case should be updated to do the same thing you did for
ext2, but I'll leave it up to Andrew, since he is the ext3 master anyway.
Andrew?
Linus
> diff -uprN -X /linux/dontdiff a/fs/ext2/dir.c b/fs/ext2/dir.c
> --- a/fs/ext2/dir.c 2004-10-30 21:44:02.000000000 +0200
> +++ b/fs/ext2/dir.c 2004-11-04 00:14:14.000000000 +0100
> @@ -275,7 +275,8 @@ ext2_readdir (struct file * filp, void *
> "bad page in #%lu",
> inode->i_ino);
> filp->f_pos += PAGE_CACHE_SIZE - offset;
> - continue;
> + ret = -EIO;
> + goto done;
> }
> kaddr = page_address(page);
> if (need_revalidate) {
>