xfs
[Top] [All Lists]

TAKE 804398 - XFS hits BUG call in pagebuf_read_full_page

To: cattelan@xxxxxxxxxxxxxxxxxxxx
Subject: TAKE 804398 - XFS hits BUG call in pagebuf_read_full_page
From: pv@xxxxxxxxxxxxxxxxxxxxxx (lord@xxxxxxx)
Date: Fri, 13 Oct 2000 09:00:04 -0700 (PDT)
Cc: linux-xfs@xxxxxxxxxxx
Reply-to: sgi.bugs.xfs@xxxxxxxxxxxxxxxxx
Sender: owner-linux-xfs@xxxxxxxxxxx
 Submitter : cattelan                 *Status : closed                      
 Assigned Engineer : cattelan         *Fixed By : lord                      
*Fixed By Domain : sgi.com            *Closed Date : 10/13/00               
 Priority : 2                         *Modified Date : 10/13/00             
*Modified User : lord                 *Modified User Domain : sgi.com       
*Fix Description :
From: steve lord <lord@xxxxxxx> (TAKE)
Date: Oct 13 2000 09:00:03AM
[pvnews version: 1.71]
----------------------------

This was the cause of panics during shutdown for systems with
an XFS root disk.

Date:  Fri Oct 13 08:56:54 PDT 2000
Workarea:  jen.americas.sgi.com:/src/lord/xfs-linux.2.4

The following file(s) were checked into:
  bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs


Modid:  2.4.x-xfs:slinx:76152a
linux/mm/filemap.c - 1.55
        - Do not call convert page on the partial page in truncate_inode_pages,
          this causes us to clear the delalloc bit and never flush the page to 
disk.
Description :
The problem appears to be a page that was orginaly delay alloc
(the underlying disk blocks cam back as zero) but has either
been reused or its state has been cleared.

I don't have all the kbd output... I;ll add to it as
I capture more traces.

[0]kdb> bt
    EBP       EIP         Function(args)
0xc4c13e58 0xc880c914 [pagebuf]pagebuf_read_full_page+0x198 (0xc49eeb60, 
0xc11939ac)

.....


==========================
ADDITIONAL INFORMATION (ADD)
From: russell cattelan <cattelan@xxxxxxxxxxx>
Date: Oct 10 2000 12:25:10PM
[pvnews version: 1.71]
==========================
"ananth@xxxxxxxxxxxx" wrote:

> View Incident: 
> http://co-op.engr.sgi.com/BugWorks/code/bwxquery.cgi?search=Search&wlong=1&view_type=Bug&wi=804398
>
>  Status : open                         Priority : 2
>  Assigned Engineer : cattelan          Submitter : cattelan
> *Modified User : ananth               *Modified User Domain : engr
> *Description :
> The problem appears to be a page that was orginaly delay alloc
> (the underlying disk blocks cam back as zero) but has either
> been reused or its state has been cleared.
>
> I don't have all the kbd output... I;ll add to it as
> I capture more traces.
>
> [0]kdb> bt
>     EBP       EIP         Function(args)
> 0xc4c13e58 0xc880c914 [pagebuf]pagebuf_read_full_page+0x198 (0xc49eeb60, 
> 0xc11939ac)
>
> .....
>
> ==========================
> ADDITIONAL INFORMATION (ADD)
> From: ananth@engr (BugWorks)
> Date: Oct 10 2000 12:11:51PM
> ==========================
>
> Couple of quick suggestions: Can you please
> dump out the page struct (both by using the "page"
> command & as raw memory) and the "map" in
> page_buf_read_full_page ... you can use "pbmap"
> kdb command for the map.

that is what  "I don't have all the kbd output... I;ll add to it as
I capture more traces"  means

I didn't capture all that
kdb crashed and I was forced to reboot.
The points of interest though
page flags: PG_locked

xextlist showed the extent to be delay alloc.

>
>
> thanks,
>
> ananth.

<Prev in Thread] Current Thread [Next in Thread>
  • TAKE 804398 - XFS hits BUG call in pagebuf_read_full_page, lord@xxxxxxx <=