[PATCH] xfs: minor inode64 mount option cleanup

[PATCH] xfs: minor inode64 mount option cleanup

[Alex Elder]

Supplying the "inode64" mount option means inode allocations should
not be restricted to "small" inode numbers (within a 32 bit range).
So in xfs_parseargs(), the XFS_MOUNT_SMALL_INUMS flag is cleared if
"inode64" is requested.

If a kernel does not support 64-bit inodes (which may be because it
doesn't support 64-bit block addresses), the "inode64" mount option
is not allowed.  But even though it's treated as an invalid mount
option xfs_parseargs() still clears SMALL_INUMS flag if "inode64" is
requested.

The end result may not actually matter, but I think we should not be
changing any state (clearing the SMALL_INUMS flag) when an invalid
mount option is supplied.

So this just makes sure the SMALL_INUMS flag is only cleared if the
kernel supports big inode numbers.

Signed-off-by: Alex Elder <elder@inktank.com>
---
 fs/xfs/xfs_super.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

Index: b/fs/xfs/xfs_super.c
===================================================================
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -295,12 +295,13 @@ xfs_parseargs(
 			}
 			dswidth = simple_strtoul(value, &eov, 10);
 		} else if (!strcmp(this_char, MNTOPT_64BITINODE)) {
+#if XFS_BIG_INUMS
 			mp->m_flags &= ~XFS_MOUNT_SMALL_INUMS;
-#if !XFS_BIG_INUMS
+#else /* !XFS_BIG_INUMS */
 			xfs_warn(mp, "%s option not allowed on this system",
 				this_char);
 			return EINVAL;
-#endif
+#endif /* !XFS_BIG_INUMS */
 		} else if (!strcmp(this_char, MNTOPT_NOUUID)) {
 			mp->m_flags |= XFS_MOUNT_NOUUID;
 		} else if (!strcmp(this_char, MNTOPT_BARRIER)) {

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH] xfs: minor inode64 mount option cleanup

[Christoph Hellwig]

> @@ -295,12 +295,13 @@ xfs_parseargs(
>  			}
>  			dswidth = simple_strtoul(value, &eov, 10);
>  		} else if (!strcmp(this_char, MNTOPT_64BITINODE)) {
> +#if XFS_BIG_INUMS
>  			mp->m_flags &= ~XFS_MOUNT_SMALL_INUMS;
> -#if !XFS_BIG_INUMS
> +#else /* !XFS_BIG_INUMS */

Looks good for now, but in the long run I'd prefer thast XFS simplify
requires 64-bit block numbers like gfs2 as there's not too much of a
point using it in a super-constrained envrionment and we could kill
all the XFS_BIG_INUMS conditionals, as well as the types that are
different based on the in-memory block number size.

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs