xfs
[Top] [All Lists]

Re: [PATCH v3] xfs: probe data buffer from page cache for unwritten exte

To: jeff.liu@xxxxxxxxxx
Subject: Re: [PATCH v3] xfs: probe data buffer from page cache for unwritten extents
From: Mark Tinguely <tinguely@xxxxxxx>
Date: Mon, 09 Jul 2012 11:41:00 -0500
Cc: xfs@xxxxxxxxxxx
In-reply-to: <4FF29C2C.2000407@xxxxxxxxxx>
References: <4FF29C2C.2000407@xxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:9.0) Gecko/20120122 Thunderbird/9.0
On 07/03/12 02:15, Jeff Liu wrote:
Hello,

This is v3 of the patch.

We can trigger BUG() in xfs_seek_data() if met two unwritten without data or 
hole extents at last version.
So making the extents map reading in loop could solve it.

Sorry, Am not yet try the repeated holes scenario according to Dave's comments 
as lack of X64 test env, still
waiting for it ready.  But this patch is already too long delayed, I have 
worked it out one weeks ago.
So I'd like to post it because of it could handle repeated hole/unwritten 
extents well in a loop, and I also improved
xfstests:286 with those cases for the verification, will post it soon.

v2->v3:
Tested by Mark, hit BUG() for continuous unwritten extents without data wrote.
* xfs_seek_data(), remove BUG() and having extents map search in loop.

v1->v2:
suggested by Mark.
* xfs_has_unwritten_buffer(), use the input offset instead of bmap->br_startoff 
to
  calculate page index for data buffer probing.

Thanks,
-Jeff


Signed-off-by: Jie Liu<jeff.liu@xxxxxxxxxx>

---

Nice. Up to refining xfs_seek_hole()?


Reviewed-by: Mark Tinguely <tinguely@xxxxxxx>

<Prev in Thread] Current Thread [Next in Thread>