linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Allison Collins <allison.henderson@oracle.com>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH v9 24/43] xfsprogs: Split apart xfs_attr_leaf_addname
Date: Mon, 4 May 2020 15:33:08 -0700	[thread overview]
Message-ID: <42a034b5-36bd-604a-db83-1225ded7fa1a@oracle.com> (raw)
In-Reply-To: <20200504173234.GB13783@magnolia>



On 5/4/20 10:32 AM, Darrick J. Wong wrote:
> On Thu, Apr 30, 2020 at 03:46:41PM -0700, Allison Collins wrote:
>> Split out new helper function xfs_attr_leaf_try_add from
>> xfs_attr_leaf_addname. Because new delayed attribute routines cannot
>> roll transactions, we split off the parts of xfs_attr_leaf_addname that
>> we can use, and move the commit into the calling function.
>>
>> Signed-off-by: Allison Collins <allison.henderson@oracle.com>
>> Reviewed-by: Brian Foster <bfoster@redhat.com>
>> Reviewed-by: Chandan Rajendra <chandanrlinux@gmail.com>
>> ---
>>   libxfs/xfs_attr.c | 94 +++++++++++++++++++++++++++++++++++--------------------
>>   1 file changed, 60 insertions(+), 34 deletions(-)
>>
>> diff --git a/libxfs/xfs_attr.c b/libxfs/xfs_attr.c
>> index 1cdebec..5f622c8 100644
>> --- a/libxfs/xfs_attr.c
>> +++ b/libxfs/xfs_attr.c
>> @@ -257,10 +257,30 @@ xfs_attr_set_args(
>>   		}
>>   	}
>>   
>> -	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
>> +	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
>>   		error = xfs_attr_leaf_addname(args);
>> -	else
>> -		error = xfs_attr_node_addname(args);
>> +		if (error != -ENOSPC)
>> +			return error;
>> +
>> +		/*
>> +		 * Commit that transaction so that the node_addname()
>> +		 * call can manage its own transactions.
>> +		 */
>> +		error = xfs_defer_finish(&args->trans);
>> +		if (error)
>> +			return error;
>> +
>> +		/*
>> +		 * Commit the current trans (including the inode) and
>> +		 * start a new one.
>> +		 */
>> +		error = xfs_trans_roll_inode(&args->trans, dp);
>> +		if (error)
>> +			return error;
> 
> I was just about to debate myself (again) on why it's still necessary to
> call _defer_finish and then _trans_roll_inode, but then I remembered the
> actual reason I unearthed for that, like 5 revisions ago.
> 
> xfs_defer_trans_roll looks for all buffer and inode items attached to a
> transaction.  If it finds bhold'ed buffers or inodes that were ijoined
> with lock_flags==0, it will log and rejoin those items to the new
> transaction after rolling.  Therefore, the xfs_trans_roll_inode here
> makes sure that any buffers that might have been bhold'ed have been
> released.  The goal here is to start each step of an xattr operation
> with the transaction in the same state (clean transaction, inode locked
> and joined) no matter how we got to that step.  That's what "manage its
> own transactions" means, though that's apparently too vague for me to
> pick up on.
> 
> With that in mind, could you change the comments for both to state why
> we're doing this more explicitly? e.g.
> 
> 	/*
> 	 * Finish any deferred work items and roll the transaction once
> 	 * more.  The goal here is to call node_addname with the inode
> 	 * and transaction in the same state (inode locked and joined,
> 	 * transaction clean) no matter how we got to this step.
> 	 */
> 	error = xfs_defer_finish(&args->trans);
> 	if (error)
> 		return error;
> 	error = xfs_trans_roll_inode(&args->trans, dp);
> 	if (error)
> 		return error;
> 
>> +
> 
> Unnecessary blank line here.
> 
Sure, will fix these.  Thanks for the reviews!

Allison

> --D
> 
>> +	}
>> +
>> +	error = xfs_attr_node_addname(args);
>>   	return error;
>>   }
>>   
>> @@ -508,20 +528,21 @@ xfs_attr_shortform_addname(xfs_da_args_t *args)
>>    *========================================================================*/
>>   
>>   /*
>> - * Add a name to the leaf attribute list structure
>> + * Tries to add an attribute to an inode in leaf form
>>    *
>> - * This leaf block cannot have a "remote" value, we only call this routine
>> - * if bmap_one_block() says there is only one block (ie: no remote blks).
>> + * This function is meant to execute as part of a delayed operation and leaves
>> + * the transaction handling to the caller.  On success the attribute is added
>> + * and the inode and transaction are left dirty.  If there is not enough space,
>> + * the attr data is converted to node format and -ENOSPC is returned. Caller is
>> + * responsible for handling the dirty inode and transaction or adding the attr
>> + * in node format.
>>    */
>>   STATIC int
>> -xfs_attr_leaf_addname(
>> -	struct xfs_da_args	*args)
>> +xfs_attr_leaf_try_add(
>> +	struct xfs_da_args	*args,
>> +	struct xfs_buf		*bp)
>>   {
>> -	struct xfs_buf		*bp;
>> -	int			retval, error, forkoff;
>> -	struct xfs_inode	*dp = args->dp;
>> -
>> -	trace_xfs_attr_leaf_addname(args);
>> +	int			retval, error;
>>   
>>   	/*
>>   	 * Look up the given attribute in the leaf block.  Figure out if
>> @@ -563,31 +584,39 @@ xfs_attr_leaf_addname(
>>   	retval = xfs_attr3_leaf_add(bp, args);
>>   	if (retval == -ENOSPC) {
>>   		/*
>> -		 * Promote the attribute list to the Btree format, then
>> -		 * Commit that transaction so that the node_addname() call
>> -		 * can manage its own transactions.
>> +		 * Promote the attribute list to the Btree format. Unless an
>> +		 * error occurs, retain the -ENOSPC retval
>>   		 */
>>   		error = xfs_attr3_leaf_to_node(args);
>>   		if (error)
>>   			return error;
>> -		error = xfs_defer_finish(&args->trans);
>> -		if (error)
>> -			return error;
>> +	}
>> +	return retval;
>> +out_brelse:
>> +	xfs_trans_brelse(args->trans, bp);
>> +	return retval;
>> +}
>>   
>> -		/*
>> -		 * Commit the current trans (including the inode) and start
>> -		 * a new one.
>> -		 */
>> -		error = xfs_trans_roll_inode(&args->trans, dp);
>> -		if (error)
>> -			return error;
>>   
>> -		/*
>> -		 * Fob the whole rest of the problem off on the Btree code.
>> -		 */
>> -		error = xfs_attr_node_addname(args);
>> +/*
>> + * Add a name to the leaf attribute list structure
>> + *
>> + * This leaf block cannot have a "remote" value, we only call this routine
>> + * if bmap_one_block() says there is only one block (ie: no remote blks).
>> + */
>> +STATIC int
>> +xfs_attr_leaf_addname(
>> +	struct xfs_da_args	*args)
>> +{
>> +	int			error, forkoff;
>> +	struct xfs_buf		*bp = NULL;
>> +	struct xfs_inode	*dp = args->dp;
>> +
>> +	trace_xfs_attr_leaf_addname(args);
>> +
>> +	error = xfs_attr_leaf_try_add(args, bp);
>> +	if (error)
>>   		return error;
>> -	}
>>   
>>   	/*
>>   	 * Commit the transaction that added the attr name so that
>> @@ -682,9 +711,6 @@ xfs_attr_leaf_addname(
>>   		error = xfs_attr3_leaf_clearflag(args);
>>   	}
>>   	return error;
>> -out_brelse:
>> -	xfs_trans_brelse(args->trans, bp);
>> -	return retval;
>>   }
>>   
>>   /*
>> -- 
>> 2.7.4
>>

  reply	other threads:[~2020-05-04 22:35 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-30 22:46 [PATCH v9 00/43] xfsprogs: Delay Ready Attributes Allison Collins
2020-04-30 22:46 ` [PATCH v9 01/43] xfsprogs: remove the ATTR_INCOMPLETE flag Allison Collins
2020-04-30 22:46 ` [PATCH v9 02/43] xfsprogs: merge xfs_attr_remove into xfs_attr_set Allison Collins
2020-04-30 22:46 ` [PATCH v9 03/43] xfsprogs: remove the name == NULL check from xfs_attr_args_init Allison Collins
2020-04-30 22:46 ` [PATCH v9 04/43] xfsprogs: remove the MAXNAMELEN " Allison Collins
2020-04-30 22:46 ` [PATCH v9 05/43] xfsprogs: turn xfs_da_args.value into a void pointer Allison Collins
2020-04-30 22:46 ` [PATCH v9 06/43] xfsprogs: pass an initialized xfs_da_args structure to xfs_attr_set Allison Collins
2020-04-30 22:46 ` [PATCH v9 07/43] xfsprogs: pass an initialized xfs_da_args to xfs_attr_get Allison Collins
2020-04-30 22:46 ` [PATCH v9 08/43] xfsprogs: remove the xfs_inode argument to xfs_attr_get_ilocked Allison Collins
2020-04-30 22:46 ` [PATCH v9 09/43] xfsprogs: remove ATTR_KERNOVAL Allison Collins
2020-04-30 22:46 ` [PATCH v9 10/43] xfsprogs: remove ATTR_ALLOC and XFS_DA_OP_ALLOCVAL Allison Collins
2020-04-30 22:46 ` [PATCH v9 11/43] xfsprogs: replace ATTR_KERNOTIME with XFS_DA_OP_NOTIME Allison Collins
2020-04-30 22:46 ` [PATCH v9 12/43] xfsprogs: factor out a xfs_attr_match helper Allison Collins
2020-04-30 22:46 ` [PATCH v9 13/43] xfsprogs: cleanup struct xfs_attr_list_context Allison Collins
2020-04-30 22:46 ` [PATCH v9 14/43] xfsprogs: remove the unused ATTR_ENTRY macro Allison Collins
2020-04-30 22:46 ` [PATCH v9 15/43] xfsprogs: move the legacy xfs_attr_list to xfs_ioctl.c Allison Collins
2020-04-30 22:46 ` [PATCH v9 16/43] xfsprogs: rename xfs_attr_list_int to xfs_attr_list Allison Collins
2020-04-30 22:46 ` [PATCH v9 17/43] xfsprogs: clean up the ATTR_REPLACE checks Allison Collins
2020-04-30 22:46 ` [PATCH v9 18/43] xfsprogs: clean up the attr flag confusion Allison Collins
2020-04-30 22:46 ` [PATCH v9 19/43] xfsprogs: embedded the attrlist cursor into struct xfs_attr_list_context Allison Collins
2020-04-30 22:46 ` [PATCH v9 20/43] xfsprogs: Add xfs_has_attr and subroutines Allison Collins
2020-04-30 22:46 ` [PATCH v9 21/43] xfsprogs: Check for -ENOATTR or -EEXIST Allison Collins
2020-04-30 22:46 ` [PATCH v9 22/43] xfsprogs: Factor out new helper functions xfs_attr_rmtval_set Allison Collins
2020-04-30 22:46 ` [PATCH v9 23/43] xfsprogs: Pull up trans handling in xfs_attr3_leaf_flipflags Allison Collins
2020-04-30 22:46 ` [PATCH v9 24/43] xfsprogs: Split apart xfs_attr_leaf_addname Allison Collins
2020-05-04 17:32   ` Darrick J. Wong
2020-05-04 22:33     ` Allison Collins [this message]
2020-04-30 22:46 ` [PATCH v9 25/43] xfsprogs: Refactor xfs_attr_try_sf_addname Allison Collins
2020-04-30 22:46 ` [PATCH v9 26/43] xfsprogs: Pull up trans roll from xfs_attr3_leaf_setflag Allison Collins
2020-04-30 22:46 ` [PATCH v9 27/43] xfsprogs: Factor out xfs_attr_rmtval_invalidate Allison Collins
2020-04-30 22:46 ` [PATCH v9 28/43] xfsprogs: Pull up trans roll in xfs_attr3_leaf_clearflag Allison Collins
2020-04-30 22:46 ` [PATCH v9 29/43] xfsprogs: Add helper function __xfs_attr_rmtval_remove Allison Collins
2020-04-30 22:46 ` [PATCH v9 30/43] xfsprogs: Pull up xfs_attr_rmtval_invalidate Allison Collins
2020-04-30 22:46 ` [PATCH v9 31/43] xfsprogs: Add helper function xfs_attr_node_shrink Allison Collins
2020-04-30 22:46 ` [PATCH v9 32/43] xfsprogs: Remove unneeded xfs_trans_roll_inode calls Allison Collins
2020-04-30 22:46 ` [PATCH v9 33/43] xfsprogs: Add helpers xfs_attr_is_shortform and xfs_attr_set_shortform Allison Collins
2020-04-30 22:46 ` [PATCH v9 34/43] xfsprogs: Add helper function xfs_attr_leaf_mark_incomplete Allison Collins
2020-04-30 22:46 ` [PATCH v9 35/43] xfsprogs: Add remote block helper functions Allison Collins
2020-04-30 22:46 ` [PATCH v9 36/43] xfsprogs: Add helper function xfs_attr_node_removename_setup Allison Collins
2020-04-30 22:46 ` [PATCH v9 37/43] xfsprogs: Add helper function xfs_attr_node_removename_rmt Allison Collins
2020-04-30 22:46 ` [PATCH v9 38/43] xfsprogs: Simplify xfs_attr_leaf_addname Allison Collins
2020-04-30 22:46 ` [PATCH v9 39/43] xfsprogs: Simplify xfs_attr_node_addname Allison Collins
2020-04-30 22:46 ` [PATCH v9 40/43] xfsprogs: Lift -ENOSPC handler from xfs_attr_leaf_addname Allison Collins
2020-04-30 22:46 ` [PATCH v9 41/43] xfsprogs: Add delay ready attr remove routines Allison Collins
2020-04-30 22:46 ` [PATCH v9 42/43] xfsprogs: Add delay ready attr set routines Allison Collins
2020-04-30 22:47 ` [PATCH v9 43/43] xfsprogs: Rename __xfs_attr_rmtval_remove 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=42a034b5-36bd-604a-db83-1225ded7fa1a@oracle.com \
    --to=allison.henderson@oracle.com \
    --cc=darrick.wong@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).