Re: [GIT PULL] Backing device changes for 3.20

Re: [GIT PULL] Backing device changes for 3.20

[Linus Torvalds]

On Thu, Feb 12, 2015 at 12:38 PM, Jens Axboe <axboe@fb.com> wrote:
>
> This pull request contains a cleanup of how the backing device is
> handled, in preparation for a rework of the life time rules. In this
> part, the most important change is to split the unrelated nommu mmap
> flags from it, but also removing a backing_dev_info pointer from the
> address_space (and inode), and a cleanup of other various minor bits.

Ugh, so this has a semantic conflict with the NFS client, that has
this particular code:

        if (!cinfo->dreq) {
                inc_zone_page_state(req->wb_page, NR_UNSTABLE_NFS);
-               inc_bdi_stat(page_file_mapping(req->wb_page)->backing_dev_info,
+
inc_bdi_stat(inode_to_bdi(page_file_mapping(req->wb_page)->host),
                             BDI_RECLAIMABLE);
                __mark_inode_dirty(req->wb_context->dentry->d_inode,
                                   I_DIRTY_DATASYNC);
        }

duplicated several times, and now one more time in the new
fs/nfs/flexfilelayout/flexfilelayout.c file.

I fixed it the same way it was fixed everywhere else, but while fixing
and looking at the cases, I *really* feel like the nfs code needs some
cleaning up.

That insane complicated and unexplained code exists three times: in
filelayout/filelayout.c, flexfilelayout/flexfilelayout.c and in
write.c.

The "reverse" case (which does the decrements, and doesn't mark the
inode dirty) exists a few more times.

Could we make that a helper function, with a few comments. For
example, looking at it, I wonder if

 - page_file_mapping(req->wb_page)->host
 - req->wb_context->dentry->d_inode

are the same inode?

Hmm?

                          Linus

Re: [GIT PULL] Backing device changes for 3.20

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1457 bytes --]

Hi All,

On Thu, 12 Feb 2015 14:05:00 -0800 Linus Torvalds <torvalds@linux-foundation.org> wrote:
>
> On Thu, Feb 12, 2015 at 12:38 PM, Jens Axboe <axboe@fb.com> wrote:
> >
> > This pull request contains a cleanup of how the backing device is
> > handled, in preparation for a rework of the life time rules. In this
> > part, the most important change is to split the unrelated nommu mmap
> > flags from it, but also removing a backing_dev_info pointer from the
> > address_space (and inode), and a cleanup of other various minor bits.
> 
> Ugh, so this has a semantic conflict with the NFS client, that has
> this particular code:
> 
>         if (!cinfo->dreq) {
>                 inc_zone_page_state(req->wb_page, NR_UNSTABLE_NFS);
> -               inc_bdi_stat(page_file_mapping(req->wb_page)->backing_dev_info,
> +
> inc_bdi_stat(inode_to_bdi(page_file_mapping(req->wb_page)->host),
>                              BDI_RECLAIMABLE);
>                 __mark_inode_dirty(req->wb_context->dentry->d_inode,
>                                    I_DIRTY_DATASYNC);
>         }
> 
> duplicated several times, and now one more time in the new
> fs/nfs/flexfilelayout/flexfilelayout.c file.

Sometime I wonder why I bother pointing these things out in
linux-next :-(  Ah, well.

(Trond did mention it in his pull request, but it wasn't a problem
then.)

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [GIT PULL] Backing device changes for 3.20

[Tom Haynes]

On Thu, Feb 12, 2015 at 02:05:00PM -0800, Linus Torvalds wrote:
> On Thu, Feb 12, 2015 at 12:38 PM, Jens Axboe <axboe@fb.com> wrote:
> >
> > This pull request contains a cleanup of how the backing device is
> > handled, in preparation for a rework of the life time rules. In this
> > part, the most important change is to split the unrelated nommu mmap
> > flags from it, but also removing a backing_dev_info pointer from the
> > address_space (and inode), and a cleanup of other various minor bits.
> 
> Ugh, so this has a semantic conflict with the NFS client, that has
> this particular code:
> 
>         if (!cinfo->dreq) {
>                 inc_zone_page_state(req->wb_page, NR_UNSTABLE_NFS);
> -               inc_bdi_stat(page_file_mapping(req->wb_page)->backing_dev_info,
> +
> inc_bdi_stat(inode_to_bdi(page_file_mapping(req->wb_page)->host),
>                              BDI_RECLAIMABLE);
>                 __mark_inode_dirty(req->wb_context->dentry->d_inode,
>                                    I_DIRTY_DATASYNC);
>         }
> 
> duplicated several times, and now one more time in the new
> fs/nfs/flexfilelayout/flexfilelayout.c file.
> 
> I fixed it the same way it was fixed everywhere else, but while fixing
> and looking at the cases, I *really* feel like the nfs code needs some
> cleaning up.
> 
> That insane complicated and unexplained code exists three times: in
> filelayout/filelayout.c, flexfilelayout/flexfilelayout.c and in
> write.c.
> 
> The "reverse" case (which does the decrements, and doesn't mark the
> inode dirty) exists a few more times.
> 
> Could we make that a helper function, with a few comments.

Yes, I'll make a helper function.

And for the "reverse" case, the second instance can actually call
nfs_clear_page_commit().

> For example, looking at it, I wonder if
> 
>  - page_file_mapping(req->wb_page)->host
>  - req->wb_context->dentry->d_inode
> 
> are the same inode?

Will resolve...

> 
> Hmm?
> 
>                           Linus
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html