linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Allison Henderson <allison.henderson@oracle.com>
To: "djwong@kernel.org" <djwong@kernel.org>
Cc: "linux-xfs@vger.kernel.org" <linux-xfs@vger.kernel.org>
Subject: Re: [PATCH v3 18/26] xfs: Add parent pointers to xfs_cross_rename
Date: Mon, 26 Sep 2022 21:50:26 +0000	[thread overview]
Message-ID: <ab6597de111eeaf3898404abef8ba02b0c091e39.camel@oracle.com> (raw)
In-Reply-To: <Yy4qi1rUB6lj542c@magnolia>

On Fri, 2022-09-23 at 14:52 -0700, Darrick J. Wong wrote:
> On Wed, Sep 21, 2022 at 10:44:50PM -0700,
> allison.henderson@oracle.com wrote:
> > From: Allison Henderson <allison.henderson@oracle.com>
> > 
> > Cross renames are handled separately from standard renames, and
> > need different handling to update the parent attributes correctly.
> > 
> > Signed-off-by: Allison Henderson <allison.henderson@oracle.com>
> > ---
> >  fs/xfs/xfs_inode.c | 85 ++++++++++++++++++++++++++++++++++++++----
> > ----
> >  1 file changed, 70 insertions(+), 15 deletions(-)
> > 
> > diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
> > index 1a35dc972d4d..51724af22bf9 100644
> > --- a/fs/xfs/xfs_inode.c
> > +++ b/fs/xfs/xfs_inode.c
> > @@ -2682,27 +2682,49 @@ xfs_finish_rename(
> >   */
> >  STATIC int
> >  xfs_cross_rename(
> > -       struct xfs_trans        *tp,
> > -       struct xfs_inode        *dp1,
> > -       struct xfs_name         *name1,
> > -       struct xfs_inode        *ip1,
> > -       struct xfs_inode        *dp2,
> > -       struct xfs_name         *name2,
> > -       struct xfs_inode        *ip2,
> > -       int                     spaceres)
> > +       struct xfs_trans                *tp,
> > +       struct xfs_inode                *dp1,
> > +       struct xfs_name                 *name1,
> > +       struct xfs_inode                *ip1,
> > +       struct xfs_inode                *dp2,
> > +       struct xfs_name                 *name2,
> > +       struct xfs_inode                *ip2,
> > +       int                             spaceres)
> >  {
> > -       int             error = 0;
> > -       int             ip1_flags = 0;
> > -       int             ip2_flags = 0;
> > -       int             dp2_flags = 0;
> > +       struct xfs_mount                *mp = dp1->i_mount;
> > +       int                             error = 0;
> > +       int                             ip1_flags = 0;
> > +       int                             ip2_flags = 0;
> > +       int                             dp2_flags = 0;
> > +       int                             new_diroffset,
> > old_diroffset;
> > +       struct xfs_parent_defer         *old_parent_ptr = NULL;
> > +       struct xfs_parent_defer         *new_parent_ptr = NULL;
> > +       struct xfs_parent_defer         *old_parent_ptr2 = NULL;
> > +       struct xfs_parent_defer         *new_parent_ptr2 = NULL;
> > +
> > +
> > +       if (xfs_has_parent(mp)) {
> > +               error = xfs_parent_init(mp, &old_parent_ptr);
> > +               if (error)
> > +                       goto out_trans_abort;
> > +               error = xfs_parent_init(mp, &new_parent_ptr);
> > +               if (error)
> > +                       goto out_trans_abort;
> > +               error = xfs_parent_init(mp, &old_parent_ptr2);
> > +               if (error)
> > +                       goto out_trans_abort;
> > +               error = xfs_parent_init(mp, &new_parent_ptr2);
> > +               if (error)
> > +                       goto out_trans_abort;
> 
> Four?  I think this means that xfs_calc_rename_reservation needs to
> do:
> 
>         overhead += 4 * (sizeof(struct xfs_attri_log_item) +
>                                 whatever_else);
> 
> Right?

Ok, I will bump up the overhead in xfs_calc_rename_reservation and see
if I can plumb another param through xfs_trans_alloc to add in the
additional name/value sizes

Allison 


> The rest of the patch looks fine, though...
> 
> > +       }
> >  
> >         /* Swap inode number for dirent in first parent */
> > -       error = xfs_dir_replace(tp, dp1, name1, ip2->i_ino,
> > spaceres, NULL);
> > +       error = xfs_dir_replace(tp, dp1, name1, ip2->i_ino,
> > spaceres, &old_diroffset);
> >         if (error)
> >                 goto out_trans_abort;
> >  
> >         /* Swap inode number for dirent in second parent */
> > -       error = xfs_dir_replace(tp, dp2, name2, ip1->i_ino,
> > spaceres, NULL);
> > +       error = xfs_dir_replace(tp, dp2, name2, ip1->i_ino,
> > spaceres, &new_diroffset);
> >         if (error)
> >                 goto out_trans_abort;
> >  
> > @@ -2763,6 +2785,28 @@ xfs_cross_rename(
> >                 }
> >         }
> >  
> > +       if (xfs_has_parent(mp)) {
> > +               error = xfs_parent_defer_replace(tp, dp1,
> > +                                                 old_parent_ptr,
> > +                                                 old_diroffset,
> > +                                                 name2,
> > +                                                 dp2,
> > +                                                 new_parent_ptr,
> > +                                                 new_diroffset,
> > ip1);
> 
> ...if you reflowed this to do two-tab indenting instead of kernel-
> style
> indenting I wouldn't be sad. ;)
> 
> --D
> 
> > +               if (error)
> > +                       goto out_trans_abort;
> > +
> > +               error = xfs_parent_defer_replace(tp, dp2,
> > +                                                 new_parent_ptr2,
> > +                                                 new_diroffset,
> > +                                                 name1,
> > +                                                 dp1,
> > +                                                 old_parent_ptr2,
> > +                                                 old_diroffset,
> > ip2);
> > +               if (error)
> > +                       goto out_trans_abort;
> > +       }
> > +
> >         if (ip1_flags) {
> >                 xfs_trans_ichgtime(tp, ip1, ip1_flags);
> >                 xfs_trans_log_inode(tp, ip1, XFS_ILOG_CORE);
> > @@ -2777,10 +2821,21 @@ xfs_cross_rename(
> >         }
> >         xfs_trans_ichgtime(tp, dp1, XFS_ICHGTIME_MOD |
> > XFS_ICHGTIME_CHG);
> >         xfs_trans_log_inode(tp, dp1, XFS_ILOG_CORE);
> > -       return xfs_finish_rename(tp);
> >  
> > +       error = xfs_finish_rename(tp);
> > +       goto out;
> >  out_trans_abort:
> >         xfs_trans_cancel(tp);
> > +out:
> > +       if (new_parent_ptr)
> > +               xfs_parent_cancel(mp, new_parent_ptr);
> > +       if (old_parent_ptr)
> > +               xfs_parent_cancel(mp, old_parent_ptr);
> > +       if (new_parent_ptr2)
> > +               xfs_parent_cancel(mp, new_parent_ptr2);
> > +       if (old_parent_ptr2)
> > +               xfs_parent_cancel(mp, old_parent_ptr2);
> > +
> >         return error;
> >  }
> >  
> > -- 
> > 2.25.1
> > 


  reply	other threads:[~2022-09-26 21:50 UTC|newest]

Thread overview: 74+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-22  5:44 [PATCH v3 00/26] Parent Pointers allison.henderson
2022-09-22  5:44 ` [PATCH v3 01/26] xfs: Add new name to attri/d allison.henderson
2022-09-23 18:53   ` Darrick J. Wong
2022-09-23 20:43     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 02/26] xfs: Increase XFS_DEFER_OPS_NR_INODES to 5 allison.henderson
2022-09-23 19:02   ` Darrick J. Wong
2022-09-23 20:45     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 03/26] xfs: Hold inode locks in xfs_ialloc allison.henderson
2022-09-22  5:44 ` [PATCH v3 04/26] xfs: Hold inode locks in xfs_trans_alloc_dir allison.henderson
2022-09-23 19:04   ` Darrick J. Wong
2022-09-23 20:44     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 05/26] xfs: Hold inode locks in xfs_rename allison.henderson
2022-09-23 19:21   ` Darrick J. Wong
2022-09-23 20:44     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 06/26] xfs: Expose init_xattrs in xfs_create_tmpfile allison.henderson
2022-09-23 19:25   ` Darrick J. Wong
2022-09-23 20:45     ` Allison Henderson
2022-09-23 21:18       ` Darrick J. Wong
2022-09-22  5:44 ` [PATCH v3 07/26] xfs: get directory offset when adding directory name allison.henderson
2022-09-22  5:44 ` [PATCH v3 08/26] xfs: get directory offset when removing " allison.henderson
2022-09-22  5:44 ` [PATCH v3 09/26] xfs: get directory offset when replacing a " allison.henderson
2022-09-22  5:44 ` [PATCH v3 10/26] xfs: add parent pointer support to attribute code allison.henderson
2022-09-22  5:44 ` [PATCH v3 11/26] xfs: define parent pointer xattr format allison.henderson
2022-09-22  5:44 ` [PATCH v3 12/26] xfs: Add xfs_verify_pptr allison.henderson
2022-09-22  5:44 ` [PATCH v3 13/26] xfs: extend transaction reservations for parent attributes allison.henderson
2022-09-23 20:17   ` Darrick J. Wong
2022-09-23 23:53     ` Allison Henderson
2022-09-26 23:53       ` Darrick J. Wong
2022-09-27 20:04         ` Allison Henderson
2022-09-27 20:44           ` Darrick J. Wong
2022-09-22  5:44 ` [PATCH v3 14/26] xfs: parent pointer attribute creation allison.henderson
2022-09-23 21:11   ` Darrick J. Wong
2022-09-26 21:48     ` Allison Henderson
2022-09-26 23:54       ` Darrick J. Wong
2022-09-27 20:10         ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 15/26] xfs: add parent attributes to link allison.henderson
2022-09-23 20:31   ` Darrick J. Wong
2022-09-26 21:49     ` Allison Henderson
2022-09-26 23:55       ` Darrick J. Wong
2022-09-22  5:44 ` [PATCH v3 16/26] xfs: add parent attributes to symlink allison.henderson
2022-09-23 21:16   ` Darrick J. Wong
2022-09-26 21:48     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 17/26] xfs: remove parent pointers in unlink allison.henderson
2022-09-23 21:22   ` Darrick J. Wong
2022-09-26 21:49     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 18/26] xfs: Add parent pointers to xfs_cross_rename allison.henderson
2022-09-23 21:52   ` Darrick J. Wong
2022-09-26 21:50     ` Allison Henderson [this message]
2022-09-22  5:44 ` [PATCH v3 19/26] xfs: Indent xfs_rename allison.henderson
2022-09-23 21:22   ` Darrick J. Wong
2022-09-26 21:49     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 20/26] xfs: Add parent pointers to rename allison.henderson
2022-09-23 22:08   ` Darrick J. Wong
2022-09-26 21:50     ` Allison Henderson
2022-09-22  5:44 ` [PATCH v3 21/26] xfs: Add the parent pointer support to the superblock version 5 allison.henderson
2022-09-22  5:44 ` [PATCH v3 22/26] xfs: Add helper function xfs_attr_list_context_init allison.henderson
2022-09-22  5:44 ` [PATCH v3 23/26] xfs: Filter XFS_ATTR_PARENT for getfattr allison.henderson
2022-09-22 16:55   ` Allison Henderson
2022-09-23 21:45   ` Darrick J. Wong
2022-09-26 21:49     ` Allison Henderson
2022-09-27 18:32       ` Darrick J. Wong
2022-09-28 18:22         ` Allison Henderson
2022-09-28  1:13   ` [xfs] b73248c4ee: xfstests.xfs.269.fail kernel test robot
2022-09-22  5:44 ` [PATCH v3 24/26] xfs: Add parent pointer ioctl allison.henderson
2022-09-24  0:30   ` Darrick J. Wong
2022-09-26 21:50     ` Allison Henderson
2022-09-27 18:34       ` Darrick J. Wong
2022-09-22  5:44 ` [PATCH v3 25/26] xfs: fix unit conversion error in xfs_log_calc_max_attrsetm_res allison.henderson
2022-09-23 21:47   ` Darrick J. Wong
2022-09-26 21:50     ` Allison Henderson
2022-09-27  0:02       ` Darrick J. Wong
2022-09-22  5:44 ` [PATCH v3 26/26] xfs: drop compatibility minimum log size computations for reflink allison.henderson
2022-09-23 21:48   ` Darrick J. Wong
2022-09-26 21:50     ` Allison Henderson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ab6597de111eeaf3898404abef8ba02b0c091e39.camel@oracle.com \
    --to=allison.henderson@oracle.com \
    --cc=djwong@kernel.org \
    --cc=linux-xfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).