Re: + ext2-get-rid-of-most-mentions-of-xip-in-ext2.patch added to -mm tree

Re: + ext2-get-rid-of-most-mentions-of-xip-in-ext2.patch added to -mm tree

[Jan Kara]

On Mon 12-01-15 15:11:11, Andrew Morton wrote:
...
> diff -puN fs/ext2/ext2.h~ext2-get-rid-of-most-mentions-of-xip-in-ext2 fs/ext2/ext2.h
> --- a/fs/ext2/ext2.h~ext2-get-rid-of-most-mentions-of-xip-in-ext2
> +++ a/fs/ext2/ext2.h
> @@ -380,14 +380,15 @@ struct ext2_inode {
>  #define EXT2_MOUNT_NO_UID32		0x000200  /* Disable 32-bit UIDs */
>  #define EXT2_MOUNT_XATTR_USER		0x004000  /* Extended user attributes */
>  #define EXT2_MOUNT_POSIX_ACL		0x008000  /* POSIX Access Control Lists */
> -#ifdef CONFIG_FS_DAX
> -#define EXT2_MOUNT_XIP			0x010000  /* Execute in place */
> -#else
> -#define EXT2_MOUNT_XIP			0
> -#endif
> +#define EXT2_MOUNT_XIP			0x010000  /* Obsolete, use DAX */
>  #define EXT2_MOUNT_USRQUOTA		0x020000  /* user quota */
>  #define EXT2_MOUNT_GRPQUOTA		0x040000  /* group quota */
>  #define EXT2_MOUNT_RESERVATION		0x080000  /* Preallocation */
> +#ifdef CONFIG_FS_DAX
> +#define EXT2_MOUNT_DAX			0x100000  /* Direct Access */
> +#else
> +#define EXT2_MOUNT_DAX			0
> +#endif
  Why do you define EXT2_MOUNT_DAX differently depending on CONFIG_FS_DAX?
  That seems unnecessary...

...
> diff -puN fs/ext2/super.c~ext2-get-rid-of-most-mentions-of-xip-in-ext2 fs/ext2/super.c
> --- a/fs/ext2/super.c~ext2-get-rid-of-most-mentions-of-xip-in-ext2
> +++ a/fs/ext2/super.c
> @@ -294,6 +294,8 @@ static int ext2_show_options(struct seq_
>  #ifdef CONFIG_FS_DAX
>  	if (sbi->s_mount_opt & EXT2_MOUNT_XIP)
>  		seq_puts(seq, ",xip");
> +	if (sbi->s_mount_opt & EXT2_MOUNT_DAX)
> +		seq_puts(seq, ",dax");
>  #endif
>  
>  	if (!test_opt(sb, RESERVATION))
  So if someone mounts filesystem with -o xip, /proc/mounts will show both
'xip' and 'dax'. OK, probably makes sense.

								Honza
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR

RE: + ext2-get-rid-of-most-mentions-of-xip-in-ext2.patch added to -mm tree

[Wilcox, Matthew R]

Ugh ... this works better from my linux.intel.com address ...

-----Original Message-----
From: Jan Kara [mailto:jack@suse.cz] 
Sent: Thursday, January 15, 2015 2:31 AM
To: Wilcox, Matthew R
Cc: akpm@linux-foundation.org; Dilger, Andreas; axboe@kernel.dk; boaz@plexistor.com; david@fromorbit.com; hch@lst.de; jack@suse.cz; kirill.shutemov@linux.intel.com; mathieu.desnoyers@efficios.com; rdunlap@infradead.org; ross.zwisler@linux.intel.com; tytso@mit.edu; mm-commits@vger.kernel.org; linux-ext4@vger.kernel.org
Subject: Re: + ext2-get-rid-of-most-mentions-of-xip-in-ext2.patch added to -mm tree

On Mon 12-01-15 15:11:11, Andrew Morton wrote:
...
> diff -puN fs/ext2/ext2.h~ext2-get-rid-of-most-mentions-of-xip-in-ext2 fs/ext2/ext2.h
> --- a/fs/ext2/ext2.h~ext2-get-rid-of-most-mentions-of-xip-in-ext2
> +++ a/fs/ext2/ext2.h
> @@ -380,14 +380,15 @@ struct ext2_inode {
>  #define EXT2_MOUNT_NO_UID32		0x000200  /* Disable 32-bit UIDs */
>  #define EXT2_MOUNT_XATTR_USER		0x004000  /* Extended user attributes */
>  #define EXT2_MOUNT_POSIX_ACL		0x008000  /* POSIX Access Control Lists */
> -#ifdef CONFIG_FS_DAX
> -#define EXT2_MOUNT_XIP			0x010000  /* Execute in place */
> -#else
> -#define EXT2_MOUNT_XIP			0
> -#endif
> +#define EXT2_MOUNT_XIP			0x010000  /* Obsolete, use DAX */
>  #define EXT2_MOUNT_USRQUOTA		0x020000  /* user quota */
>  #define EXT2_MOUNT_GRPQUOTA		0x040000  /* group quota */
>  #define EXT2_MOUNT_RESERVATION		0x080000  /* Preallocation */
> +#ifdef CONFIG_FS_DAX
> +#define EXT2_MOUNT_DAX			0x100000  /* Direct Access */
> +#else
> +#define EXT2_MOUNT_DAX			0
> +#endif
  Why do you define EXT2_MOUNT_DAX differently depending on CONFIG_FS_DAX?
  That seems unnecessary...

That's copied from the original XIP code.  I think the intent is to make the XIP/DAX code disappear from the module, rather than being dead code cluttering up the object file.

...
> diff -puN fs/ext2/super.c~ext2-get-rid-of-most-mentions-of-xip-in-ext2 fs/ext2/super.c
> --- a/fs/ext2/super.c~ext2-get-rid-of-most-mentions-of-xip-in-ext2
> +++ a/fs/ext2/super.c
> @@ -294,6 +294,8 @@ static int ext2_show_options(struct seq_
>  #ifdef CONFIG_FS_DAX
>  	if (sbi->s_mount_opt & EXT2_MOUNT_XIP)
>  		seq_puts(seq, ",xip");
> +	if (sbi->s_mount_opt & EXT2_MOUNT_DAX)
> +		seq_puts(seq, ",dax");
>  #endif
>  
>  	if (!test_opt(sb, RESERVATION))
  So if someone mounts filesystem with -o xip, /proc/mounts will show both
'xip' and 'dax'. OK, probably makes sense.

Yeah, I thought that was the best migration path; somebody might have a script that's checking for 'xip' in /proc/mounts.

Re: + ext2-get-rid-of-most-mentions-of-xip-in-ext2.patch added to -mm tree

[Jan Kara]

On Fri 16-01-15 21:09:03, Wilcox, Matthew R wrote:
> -----Original Message-----
> From: Jan Kara [mailto:jack@suse.cz] 
> Sent: Thursday, January 15, 2015 2:31 AM
> To: Wilcox, Matthew R
> Cc: akpm@linux-foundation.org; Dilger, Andreas; axboe@kernel.dk; boaz@plexistor.com; david@fromorbit.com; hch@lst.de; jack@suse.cz; kirill.shutemov@linux.intel.com; mathieu.desnoyers@efficios.com; rdunlap@infradead.org; ross.zwisler@linux.intel.com; tytso@mit.edu; mm-commits@vger.kernel.org; linux-ext4@vger.kernel.org
> Subject: Re: + ext2-get-rid-of-most-mentions-of-xip-in-ext2.patch added to -mm tree
> 
> On Mon 12-01-15 15:11:11, Andrew Morton wrote:
> ...
> > diff -puN fs/ext2/ext2.h~ext2-get-rid-of-most-mentions-of-xip-in-ext2 fs/ext2/ext2.h
> > --- a/fs/ext2/ext2.h~ext2-get-rid-of-most-mentions-of-xip-in-ext2
> > +++ a/fs/ext2/ext2.h
> > @@ -380,14 +380,15 @@ struct ext2_inode {
> >  #define EXT2_MOUNT_NO_UID32		0x000200  /* Disable 32-bit UIDs */
> >  #define EXT2_MOUNT_XATTR_USER		0x004000  /* Extended user attributes */
> >  #define EXT2_MOUNT_POSIX_ACL		0x008000  /* POSIX Access Control Lists */
> > -#ifdef CONFIG_FS_DAX
> > -#define EXT2_MOUNT_XIP			0x010000  /* Execute in place */
> > -#else
> > -#define EXT2_MOUNT_XIP			0
> > -#endif
> > +#define EXT2_MOUNT_XIP			0x010000  /* Obsolete, use DAX */
> >  #define EXT2_MOUNT_USRQUOTA		0x020000  /* user quota */
> >  #define EXT2_MOUNT_GRPQUOTA		0x040000  /* group quota */
> >  #define EXT2_MOUNT_RESERVATION		0x080000  /* Preallocation */
> > +#ifdef CONFIG_FS_DAX
> > +#define EXT2_MOUNT_DAX			0x100000  /* Direct Access */
> > +#else
> > +#define EXT2_MOUNT_DAX			0
> > +#endif
>   Why do you define EXT2_MOUNT_DAX differently depending on CONFIG_FS_DAX?
>   That seems unnecessary...
> 
> That's copied from the original XIP code.  I think the intent is to make the XIP/DAX code disappear from the module, rather than being dead code cluttering up the object file.
  I see. OK.

								Honza

-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR