xfs
[Top] [All Lists]

Re: 2.6.22-rc3 hibernate(?) fails totally - regression (xfs on raid6)

To: David Greaves <david@xxxxxxxxxxxx>
Subject: Re: 2.6.22-rc3 hibernate(?) fails totally - regression (xfs on raid6)
From: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Date: Tue, 12 Jun 2007 11:43:37 -0700 (PDT)
Cc: David Chinner <dgc@xxxxxxx>, Tejun Heo <htejun@xxxxxxxxx>, "Rafael J. Wysocki" <rjw@xxxxxxx>, xfs@xxxxxxxxxxx, "'linux-kernel@xxxxxxxxxxxxxxx'" <linux-kernel@xxxxxxxxxxxxxxx>, linux-pm <linux-pm@xxxxxxxxxxxxxx>, Neil Brown <neilb@xxxxxxx>, Jeff Garzik <jgarzik@xxxxxxxxx>
In-reply-to: <4669A965.20403@dgreaves.com>
References: <200706020122.49989.rjw@sisk.pl> <4661EFBB.5010406@dgreaves.com> <alpine.LFD.0.98.0706021538360.23741@woody.linux-foundation.org> <4662D852.4000005@dgreaves.com> <46667160.80905@gmail.com> <46668EE0.2030509@dgreaves.com> <46679D56.7040001@gmail.com> <4667DE2D.6050903@dgreaves.com> <20070607110708.GS86004887@sgi.com> <46680F5E.6070806@dgreaves.com> <20070607222813.GG85884050@sgi.com> <4669A965.20403@dgreaves.com>
Sender: xfs-bounce@xxxxxxxxxxx

On Fri, 8 Jun 2007, David Greaves wrote:
> 
> positive: I can now get sysrq-t :)

Ok, so color me confused, and maybe I have missed some of the emails or 
skimmed over them too fast (there's been too many of them ;), but

 - I haven't actually seen any traces for this (netconsole apparently 
   doesn't work for you, and I'm not surprised: it never really worked 
   well for me over suspend/resume either, but I think I saw a mention of 
   serial console?)

 - You apparently bisected it down to the range

        0a3fd051c7036ef71b58863f8e5da7c3dabd9d3f <- works
        1d30c33d8d07868199560b24f10ed6280e78a89c <- breaks

   but some of the intermediates in that range didn't compile. Correct?

Can you try to bisect down a bit more, despite the compile error? Just do

        git bisect start
        git bisect good 0a3fd051c7036ef71b58863f8e5da7c3dabd9d3f
        git bisect bad 1d30c33d8d07868199560b24f10ed6280e78a89c

and it should pick 

        f4d6d004: libata: ignore EH scheduling during initialization

for you to test. It will apparently break on the fact that "sata_via.c" 
wants "ata_scsi_device_resume/suspend" for the initialization of the
resume/suspend things in the scsi_host_template, but you should just 
remove those lines, and the compile hopefully completes cleanly after 
that.

IOW, it *should* be easy enough to pinpoint this from 9 changes down to 
just one.

Jeff added to the Cc, since he may not have noticed that one of the most 
long-running issues is apparently sata-related.

(Jeff: David Greaves _also_ had issues with -rc4 due to the SETFXSR 
change, but that should hopefully be resolved and is presumably an 
independent bug. Apart from the fact that "sata_via.c" seems problematic)

                Linus


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