xfs
[Top] [All Lists]

Re: [PATCH v8 00/71] xfs: add reflink and dedupe support

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH v8 00/71] xfs: add reflink and dedupe support
From: "Darrick J. Wong" <darrick.wong@xxxxxxxxxx>
Date: Fri, 26 Aug 2016 11:42:46 -0700
Cc: david@xxxxxxxxxxxxx, linux-xfs@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20160826125636.GA1548@xxxxxxxxxxxxx>
References: <147216791538.867.12413509832420924168.stgit@xxxxxxxxxxxxxxxx> <20160826125636.GA1548@xxxxxxxxxxxxx>
User-agent: Mutt/1.5.24 (2015-08-30)
On Fri, Aug 26, 2016 at 05:56:36AM -0700, Christoph Hellwig wrote:
> FTI, with your djwong-devel tree I see sporadic slab corruption when
> running xfstests with reflinks force enabled in mkfs.
> 
> And while the trace looks like from the rmap code it doesn't happen
> without the reflink patches.
> 
> Example during xfs/313 below:

Ok, so this was caused by a bug in the deferred ops abort code where we'd
unconditionally free a log intent item even if a log done item has already been
attached.  That in turn led to the intent item being freed three times (once
when the CUI was committed, again when the deferred op aborted, and again when
the CUD was committed during transaction abort), hence the slab corruption.
Oops.

Will post patch shortly.

--D

> 
> [14751.774252] Slab corruption (Tainted: G        W      ): xfs_cui_item 
> start=ffff88013afb3748, len=408
> [14751.775170] Redzone: 0x9f911029d74e35b/0x9f911029d74e35b
> [14751.775731] Last user: [<ffffffff815ce4cc>](xfs_cui_item_free+0x1c/0x30)
> [14751.776397] 070: 6b 6b 6b 6b 6b 6b 6b 6b 6a 6b 6b 6b 6b 6b 6b 6b 
> kkkkkkkkjkkkkkkk
> [14751.777157] Single bit error detected. Probably bad RAM.
> [14751.777591] Run memtest86+ or a similar memory test tool.
> [14751.778015] Prev obj: start=ffff88013afb3598, len=408
> [14751.778468] Redzone: 0x9f911029d74e35b/0x9f911029d74e35b
> [14751.778940] Last user:
> [<ffffffff815ce4cc>](xfs_cui_item_free+0x1c/0x30)
> [14751.779621] 000: 6b 6b 6b 6b 6b 00 01 02 03 6b 6b 6b 6b 6b 6b 6b 
> kkkkk....kkkkkkk
> [14751.780378] 010: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 
> kkkkkkkkkkkkkkkk
> [14751.781140] Next obj: start=ffff88013afb38f8, len=408
> [14751.781639] Redzone: 0x9f911029d74e35b/0x9f911029d74e35b
> [14751.782216] Last user:
> [<ffffffff815ce4cc>](xfs_cui_item_free+0x1c/0x30)
> [14751.782885] 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 
> kkkkkkkkkkkkkkkk
> [14751.783608] 010: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 
> kkkkkkkkkkkkkkkk
> [14751.784287] Slab corruption (Tainted: G        W      ): xfs_bui_item 
> start=ffff880131b67048, len=184
> [14751.785126] Redzone: 0x9f911029d74e35b/0x9f911029d74e35b
> [14751.785979] Last user: [<ffffffff815c9584>](xfs_bui_release+0x44/0x70)0
> [14751.786636] 070: 6b 6b 6b 6b 6b 6b 6b 6b 6a 6b 6b 6b 6b 6b 6b 6b 
> kkkkkkkkjkkkkkkk
> [14751.787414] Single bit error detected. Probably bad RAM.
> [14751.787948] Run memtest86+ or a similar memory test tool.
> [14751.788477] Next obj: start=ffff880131b67118, len=184
> [14751.789002] Redzone: 0x9f911029d74e35b/0x9f911029d74e35b
> [14751.789559] Last user: [<ffffffff815c9584>](xfs_bui_release+0x44/0x70)
> [14751.790259] 000: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 
> kkkkkkkkkkkkkkkk
> [14751.791055] 010: 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 
> kkkkkkkkkkkkkkkk
> 

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