diff mbox series

[3/3] ext2: remove dax EXPERIMENTAL warning

Message ID 1631726561-16358-4-git-send-email-sandeen@redhat.com (mailing list archive)
State New, archived
Headers show
Series Remove DAX experimental warnings | expand

Commit Message

Eric Sandeen Sept. 15, 2021, 5:22 p.m. UTC
As there seems to be no significant outstanding concern about
dax on ext2 at this point, remove the scary EXPERIMENTAL
warning when in use.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
---
 fs/ext2/super.c | 2 --
 1 file changed, 2 deletions(-)

Comments

Jan Kara Sept. 17, 2021, 9:47 a.m. UTC | #1
On Wed 15-09-21 12:22:41, Eric Sandeen wrote:
> As there seems to be no significant outstanding concern about
> dax on ext2 at this point, remove the scary EXPERIMENTAL
> warning when in use.
> 
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>

Agreed. Do you want my ack or should I just merge this patch?

								Honza

> ---
>  fs/ext2/super.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/fs/ext2/super.c b/fs/ext2/super.c
> index d8d580b..1915733 100644
> --- a/fs/ext2/super.c
> +++ b/fs/ext2/super.c
> @@ -587,8 +587,6 @@ static int parse_options(char *options, struct super_block *sb,
>  			fallthrough;
>  		case Opt_dax:
>  #ifdef CONFIG_FS_DAX
> -			ext2_msg(sb, KERN_WARNING,
> -		"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
>  			set_opt(opts->s_mount_opt, DAX);
>  #else
>  			ext2_msg(sb, KERN_INFO, "dax option not supported");
> -- 
> 1.8.3.1
>
Christoph Hellwig Sept. 17, 2021, 12:59 p.m. UTC | #2
On Fri, Sep 17, 2021 at 11:47:07AM +0200, Jan Kara wrote:
> On Wed 15-09-21 12:22:41, Eric Sandeen wrote:
> > As there seems to be no significant outstanding concern about
> > dax on ext2 at this point, remove the scary EXPERIMENTAL
> > warning when in use.
> > 
> > Signed-off-by: Eric Sandeen <sandeen@redhat.com>
> 
> Agreed. Do you want my ack or should I just merge this patch?

Please do not merge it.  The whole DAX path is still a mess and should
not be elevated to non-EXPERMINTAL state in this form.
Eric Sandeen Sept. 17, 2021, 2:09 p.m. UTC | #3
On 9/17/21 7:59 AM, Christoph Hellwig wrote:
> On Fri, Sep 17, 2021 at 11:47:07AM +0200, Jan Kara wrote:
>> On Wed 15-09-21 12:22:41, Eric Sandeen wrote:
>>> As there seems to be no significant outstanding concern about
>>> dax on ext2 at this point, remove the scary EXPERIMENTAL
>>> warning when in use.
>>>
>>> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
>>
>> Agreed. Do you want my ack or should I just merge this patch?
> 
> Please do not merge it.  The whole DAX path is still a mess and should
> not be elevated to non-EXPERMINTAL state in this form.

Hi Christoph, "a mess" is tough to work with. What work remains before
we can lift the warning?

-Eric
Darrick J. Wong Sept. 22, 2021, 2:36 a.m. UTC | #4
On Fri, Sep 17, 2021 at 01:59:10PM +0100, Christoph Hellwig wrote:
> On Fri, Sep 17, 2021 at 11:47:07AM +0200, Jan Kara wrote:
> > On Wed 15-09-21 12:22:41, Eric Sandeen wrote:
> > > As there seems to be no significant outstanding concern about
> > > dax on ext2 at this point, remove the scary EXPERIMENTAL
> > > warning when in use.
> > > 
> > > Signed-off-by: Eric Sandeen <sandeen@redhat.com>
> > 
> > Agreed. Do you want my ack or should I just merge this patch?
> 
> Please do not merge it.  The whole DAX path is still a mess and should
> not be elevated to non-EXPERMINTAL state in this form.

Hi Christoph,

'still a mess' isn't all that useful for figuring out what still needs
to be done and splitting up the work.  Do you have items beyond my own
list below?

 - still arguing over what exactly FALLOC_FL_ZERO_REINIT_WHATEVER_PONIES
   should be doing
 - no reflink support, encompassing:
 - hwpoison from mmap regions really ought to tell the fs that bad stuff
   happened
 - mm rmap can't handle more than one owner

--D
Christoph Hellwig Sept. 22, 2021, 5:25 a.m. UTC | #5
On Tue, Sep 21, 2021 at 07:36:22PM -0700, Darrick J. Wong wrote:
> 'still a mess' isn't all that useful for figuring out what still needs
> to be done and splitting up the work.  Do you have items beyond my own
> list below?
> 
>  - still arguing over what exactly FALLOC_FL_ZERO_REINIT_WHATEVER_PONIES
>    should be doing
>  - no reflink support, encompassing:
>  - hwpoison from mmap regions really ought to tell the fs that bad stuff
>    happened
>  - mm rmap can't handle more than one owner

My main really big item is that we're still mounting through a fake
block device, suporting partitions and all that crap.  We need to sort
out the whole story of how pmem/nvdimm is actually treated, because
what we have right now is not sustainable at all.
diff mbox series

Patch

diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index d8d580b..1915733 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -587,8 +587,6 @@  static int parse_options(char *options, struct super_block *sb,
 			fallthrough;
 		case Opt_dax:
 #ifdef CONFIG_FS_DAX
-			ext2_msg(sb, KERN_WARNING,
-		"DAX enabled. Warning: EXPERIMENTAL, use at your own risk");
 			set_opt(opts->s_mount_opt, DAX);
 #else
 			ext2_msg(sb, KERN_INFO, "dax option not supported");