Re: [PATCH] struct super_block cleanup - msdos/vfat

[OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>]

Brian Gerst <bgerst@didntduck.org> writes:

> OGAWA Hirofumi wrote:
> > OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> writes:
> >
> >>Hi,
> >>
> >>Brian Gerst <bgerst@didntduck.org> writes:
> >>
> >>
> >>>diff -urN linux-2.5.7-pre1/fs/msdos/namei.c linux/fs/msdos/namei.c
> >>>--- linux-2.5.7-pre1/fs/msdos/namei.c	Thu Mar  7 21:18:32 2002
> >>>+++ linux/fs/msdos/namei.c	Wed Mar 13 08:20:12 2002
> >>>@@ -603,17 +603,14 @@
> >>> int msdos_fill_super(struct super_block *sb,void *data, int silent)
> >>> {
> >>>-	struct super_block *res;
> >>>+	int res;
> >>> -	MSDOS_SB(sb)->options.isvfat = 0;
> >>>-	res = fat_read_super(sb, data, silent, &msdos_dir_inode_operations);
> >>>-	if (IS_ERR(res))
> >>>-		return PTR_ERR(res);
> >>>-	if (res == NULL) {
> >>>+	res = fat_fill_super(sb, data, silent, &msdos_dir_inode_operations, 0);
> >>>+	if (res) {
> >>> 		if (!silent)
> >>> 			printk(KERN_INFO "VFS: Can't find a valid"
> >>> 			       " MSDOS filesystem on dev %s.\n", sb->s_id);
> >>
> >>If the error is I/O error, I think we shouldn't output this message.
> >   ^^^^^^^^^^^^^^^^^^^^^^^^^^
> > If the error is except -EINVAL,
> > Sorry.
> >
> >>What do you think about this?
> >
> 
> Why not?  The statement is true, and other filesystems do complain
> when there is an I/O error.

Umm, almost all filesystems doesn't output this message when the I/O error
occurs, AFAIK.

I think that this message indicate that a device isn't a FAT
filesystem.  And, of course, if error is the I/O error,
fat_full_super() can't detect whether it is FAT filesystem or not.
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>