From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Allison Collins <allison.henderson@oracle.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH v4 03/17] xfs: Embed struct xfs_name in xfs_da_args
Date: Fri, 8 Nov 2019 13:47:19 -0800 [thread overview]
Message-ID: <20191108214719.GI6219@magnolia> (raw)
In-Reply-To: <3e1bb508-a78f-a3c0-2f17-d981b7b7705b@oracle.com>
On Fri, Nov 08, 2019 at 09:11:54AM -0700, Allison Collins wrote:
> On 11/7/19 6:25 PM, Darrick J. Wong wrote:
> > On Wed, Nov 06, 2019 at 06:27:47PM -0700, Allison Collins wrote:
> > > This patch embeds an xfs_name in xfs_da_args, replacing the name,
> > > namelen, and flags members. This helps to clean up the xfs_da_args
> > > structure and make it more uniform with the new xfs_name parameter
> > > being passed around.
> > >
> > > Signed-off-by: Allison Collins <allison.henderson@oracle.com>
> > > Reviewed-by: Brian Foster <bfoster@redhat.com>
> > > ---
> > > fs/xfs/libxfs/xfs_attr.c | 36 +++++++-------
> > > fs/xfs/libxfs/xfs_attr_leaf.c | 106 +++++++++++++++++++++-------------------
> > > fs/xfs/libxfs/xfs_attr_remote.c | 2 +-
> > > fs/xfs/libxfs/xfs_da_btree.c | 5 +-
> > > fs/xfs/libxfs/xfs_da_btree.h | 4 +-
> > > fs/xfs/libxfs/xfs_dir2.c | 18 +++----
> > > fs/xfs/libxfs/xfs_dir2_block.c | 6 +--
> > > fs/xfs/libxfs/xfs_dir2_leaf.c | 6 +--
> > > fs/xfs/libxfs/xfs_dir2_node.c | 8 +--
> > > fs/xfs/libxfs/xfs_dir2_sf.c | 30 ++++++------
> > > fs/xfs/scrub/attr.c | 12 ++---
> > > fs/xfs/xfs_trace.h | 20 ++++----
> > > 12 files changed, 128 insertions(+), 125 deletions(-)
> > >
> > > diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
> > > index 5a9624a..b77b985 100644
> > > --- a/fs/xfs/libxfs/xfs_attr.c
> > > +++ b/fs/xfs/libxfs/xfs_attr.c
> > > @@ -72,13 +72,12 @@ xfs_attr_args_init(
> > > args->geo = dp->i_mount->m_attr_geo;
> > > args->whichfork = XFS_ATTR_FORK;
> > > args->dp = dp;
> > > - args->flags = flags;
> > > - args->name = name->name;
> > > - args->namelen = name->len;
> > > - if (args->namelen >= MAXNAMELEN)
> > > + name->type = flags;
> >
> > Is there a purpose for modifying the caller's @name, instead of setting
> > args.name.type = flags after the memcpy?
> Not really other than to just make them consistent so that the mem copy
> would cover all three members. I think initially they were set
> individually, but people preferred the memcpy, and then later we decided to
> break the flags variable out of the name struct. I can explicitly set
> args.name.type if folks prefer.
>
> A subtle pitfall I noticed about using the name struct like this now:
> callers need to take care to abandon the original name struct or update the
> original pointer to this struct. It's kind of easy to continue using the
> original struct like it's a convenience pointer, but it not the same memory
> so if args gets passed into some subroutine that modifies it, now you're out
> of sync. If it doesnt get passed around or modified, it doesnt matter, but
> we've created a sort of usage rule about how to handle the name param which
> may not be entirely obvious to other developers.
Hm, maybe the da_args should keep a pointer to the caller's name, if
you're worried about coherency between different xfs_names?
Though, the only state that changes is the type field, right? Maybe we
don't care and that should stay internal to the da_args anyway...
> >
> > > + memcpy(&args->name, name, sizeof(struct xfs_name));
> > > + if (args->name.len >= MAXNAMELEN)
> > > return -EFAULT; /* match IRIX behaviour */
> > > - args->hashval = xfs_da_hashname(args->name, args->namelen);
> > > + args->hashval = xfs_da_hashname(args->name.name, args->name.len);
> > > return 0;
> > > }
> > > @@ -236,7 +235,7 @@ xfs_attr_try_sf_addname(
> > > * Commit the shortform mods, and we're done.
> > > * NOTE: this is also the error path (EEXIST, etc).
> > > */
> > > - if (!error && (args->flags & ATTR_KERNOTIME) == 0)
> > > + if (!error && (args->name.type & ATTR_KERNOTIME) == 0)
> > > xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);
> > > if (mp->m_flags & XFS_MOUNT_WSYNC)
> > > @@ -357,6 +356,9 @@ xfs_attr_set(
> > > if (error)
> > > return error;
> > > + /* Use name now stored in args */
> > > + name = &args.name;
> >
> > You could probably set this as part of the variable declaration, e.g.
> >
> > struct xfs_name *name = &args.name;
> No, because name is a parameter here, so you would loose the contents at the
> start of the function if we did this. And then args init would memcopy and
> empty struct onto itself. And then confusion would ensue.... :-(
>
> >
> > > +
> > > args.value = value;
> > > args.valuelen = valuelen;
> > > args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
> > > @@ -372,7 +374,7 @@ xfs_attr_set(
> > > */
> > > if (XFS_IFORK_Q(dp) == 0) {
> > > int sf_size = sizeof(xfs_attr_sf_hdr_t) +
> > > - XFS_ATTR_SF_ENTSIZE_BYNAME(args.namelen, valuelen);
> > > + XFS_ATTR_SF_ENTSIZE_BYNAME(args.name.len, valuelen);
> >
> > If you're going to keep the convenience variable @name, then please use
> > it throughout the function.
> I think this is the only place it's used for this function at this time. I
> use it later in patch 15 though. I think I stumbled across the name sync
> bug then, and corrected it with the parameter pointer assignment above just
> because it seemed like the right place to fix it.
>
> Do you want the function to uniformly use an explicit reference, or the
> param? Seems like a bit of a wart unfortunately :-(
One or the other, just not both mixed in the same function.
--D
> >
> > > error = xfs_bmap_add_attrfork(dp, sf_size, rsvd);
> > > if (error)
> >
> > <snip>
> >
> > > diff --git a/fs/xfs/libxfs/xfs_da_btree.c b/fs/xfs/libxfs/xfs_da_btree.c
> > > index 4fd1223..129ec09 100644
> > > --- a/fs/xfs/libxfs/xfs_da_btree.c
> > > +++ b/fs/xfs/libxfs/xfs_da_btree.c
> > > @@ -2040,8 +2040,9 @@ xfs_da_compname(
> > > const unsigned char *name,
> > > int len)
> > > {
> > > - return (args->namelen == len && memcmp(args->name, name, len) == 0) ?
> > > - XFS_CMP_EXACT : XFS_CMP_DIFFERENT;
> >
> > Wow that's gross. :)
> >
> > if (args->name.len == len && !memcmp(args->name.name, name, len))
> > return XFS_CMP_EXACT;
> >
> > return XFS_CMP_DIFFERENT;
> >
> > Hmm, is that better?
> Sure, will fix.
>
> Allison
>
> >
> > The rest looks reasonable...
> >
> > --D
> >
next prev parent reply other threads:[~2019-11-08 21:47 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-07 1:27 [PATCH v4 00/17] xfs: Delay Ready Attributes Allison Collins
2019-11-07 1:27 ` [PATCH v4 01/17] xfs: Remove all strlen in all xfs_attr_* functions for attr names Allison Collins
2019-11-11 17:47 ` Christoph Hellwig
2019-11-11 23:35 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 02/17] xfs: Replace attribute parameters with struct xfs_name Allison Collins
2019-11-08 1:13 ` Darrick J. Wong
2019-11-08 17:16 ` Allison Collins
2019-11-11 17:49 ` Christoph Hellwig
2019-11-11 20:07 ` Allison Collins
2019-11-13 15:12 ` Allison Collins
2019-11-20 18:20 ` Christoph Hellwig
2019-11-07 1:27 ` [PATCH v4 03/17] xfs: Embed struct xfs_name in xfs_da_args Allison Collins
2019-11-08 1:25 ` Darrick J. Wong
2019-11-08 16:11 ` Allison Collins
2019-11-08 21:47 ` Darrick J. Wong [this message]
2019-11-07 1:27 ` [PATCH v4 04/17] xfs: Add xfs_dabuf defines Allison Collins
2019-11-08 19:19 ` Darrick J. Wong
2019-11-09 17:32 ` Allison Collins
2019-11-09 20:11 ` Darrick J. Wong
2019-11-09 22:06 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 05/17] xfs: Add xfs_has_attr and subroutines Allison Collins
2019-11-08 19:32 ` Darrick J. Wong
2019-11-08 19:51 ` Allison Collins
2019-11-11 17:40 ` Brian Foster
2019-11-11 23:34 ` Allison Collins
2019-11-11 17:53 ` Christoph Hellwig
2019-11-11 23:36 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 06/17] xfs: Factor out new helper functions xfs_attr_rmtval_set Allison Collins
2019-11-08 19:34 ` Darrick J. Wong
2019-11-08 19:51 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 07/17] xfs: Factor up trans handling in xfs_attr3_leaf_flipflags Allison Collins
2019-11-08 19:35 ` Darrick J. Wong
2019-11-08 19:52 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 08/17] xfs: Factor out xfs_attr_leaf_addname helper Allison Collins
2019-11-08 20:57 ` Darrick J. Wong
2019-11-09 21:41 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 09/17] xfs: Factor up commit from xfs_attr_try_sf_addname Allison Collins
2019-11-08 21:04 ` Darrick J. Wong
2019-11-08 23:13 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 10/17] xfs: Factor up trans roll from xfs_attr3_leaf_setflag Allison Collins
2019-11-07 1:27 ` [PATCH v4 11/17] xfs: Add xfs_attr3_leaf helper functions Allison Collins
2019-11-08 21:17 ` Darrick J. Wong
2019-11-09 0:09 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 12/17] xfs: Factor out xfs_attr_rmtval_invalidate Allison Collins
2019-11-08 21:19 ` Darrick J. Wong
2019-11-09 0:10 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 13/17] xfs: Factor up trans roll in xfs_attr3_leaf_clearflag Allison Collins
2019-11-08 21:19 ` Darrick J. Wong
2019-11-09 0:11 ` Allison Collins
2019-11-11 18:23 ` Brian Foster
2019-11-11 23:37 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 14/17] xfs: Add delay context to xfs_da_args Allison Collins
2019-11-08 21:22 ` Darrick J. Wong
2019-11-09 0:23 ` Allison Collins
2019-11-11 18:23 ` Brian Foster
2019-11-11 23:42 ` Allison Collins
2019-11-07 1:27 ` [PATCH v4 15/17] xfs: Check for -ENOATTR or -EEXIST Allison Collins
2019-11-08 21:28 ` Darrick J. Wong
2019-11-08 21:42 ` Allison Collins
2019-11-08 21:51 ` Darrick J. Wong
2019-11-11 18:24 ` Brian Foster
2019-11-12 0:33 ` Allison Collins
2019-11-07 1:28 ` [PATCH v4 16/17] xfs: Add delay ready attr remove routines Allison Collins
2019-11-08 21:37 ` Darrick J. Wong
2019-11-09 0:25 ` Allison Collins
2019-11-12 13:37 ` Brian Foster
2019-11-13 0:43 ` Allison Collins
2019-11-13 11:54 ` Brian Foster
2019-11-13 23:39 ` Allison Collins
2019-11-14 12:48 ` Brian Foster
2019-11-14 17:58 ` Allison Collins
2019-11-07 1:28 ` [PATCH v4 17/17] xfs: Add delay ready attr set routines Allison Collins
2019-11-08 21:42 ` Darrick J. Wong
2019-11-08 21:52 ` Allison Collins
2019-11-09 4:07 ` Allison Collins
2019-11-12 13:37 ` Brian Foster
2019-11-13 4:57 ` Allison Collins
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=20191108214719.GI6219@magnolia \
--to=darrick.wong@oracle.com \
--cc=allison.henderson@oracle.com \
--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).