Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.os.linux.misc > #836

Re: ata errors on kernel 2.6.35.12

From Douglas Mayne <none@invalid.com>
Newsgroups comp.os.linux.misc
Subject Re: ata errors on kernel 2.6.35.12
Date 2011-04-19 19:57 +0000
Organization XMission http://xmission.com/
Message-ID <iokpec$ca0$2@news.xmission.com> (permalink)
References <ioeuh1$dbu$1@news.xmission.com> <i3npq61nfnki923fqna9emnlp1f8u3i6gj@4ax.com>

Show all headers | View raw


On Tue, 19 Apr 2011 10:59:45 +1000, Grant wrote:

> On Sun, 17 Apr 2011 14:46:57 +0000 (UTC), Douglas Mayne
> <none@invalid.com> wrote:
> 
>>Lately, I have been trying out slackware -current. The linux kernel was
>>a little "bleeding" edge, so I tried rolling back to an earlier version.
>>I chose the 2.6.35.x series, because IIRC it is set for long term
>>maintenance by the kernel team. I have been seeing these errors crop up
>>every couple of days, as copied from the syslog:
>>
>>Please, pardon long lines:
>>Apr 16 16:16:48 darkstar kernel: [159300.229115] ata2.00: qc timeout
>>(cmd 0xa0)
> ...
>>Apr 16 16:38:20 darkstar kernel: [160592.240088] ata2: device not ready
>>(errno=-16), forcing hardreset
>>
>>Some googling reveals that there are a lot of errors like this. My
>>question is whether anyone here is seeing the same thing, or has any
>>advice. This bug is going to keep my "production" computers at 2.6.33.x
>>for the near term.
>>
>>TIA
>>
>>p.s. I have actually "thrown money" at this problem by replacing the
>>hard disk. Now, I am really questioning whether it was a hardware fault
>>at all, and has been a kernel issue all along.
> 
> First thing I suspect is that you have the wrong driver installed for
> the hard drive.  Suggest you get rid of unused drivers, or make sure
> they're not loaded.  Check lspci vs menuconfig to make sure you have
> correct drivers for the hardware.  There is no size fits all in this
> area.
>
At least some of the SATA drivers are "built" into the official Slackware 
kernel. The hardware I am using is using the popular Intel ICHx pci 
chipset. I think the right driver is loaded, but I don't think it is 
doing something right with the hardware.
> 
> Actually, you don't match to the hard drive, you match to the whatsit
> talking to the hard drive.  Sometimes on odd hardware it's difficult to
> see which is the correct driver.
> 
> Also dmesg may help here.  Use the latest libata drivers, don't use the
> old PATA drivers unless you have really old stuff.
> 
> Want examples?  You could check my kernel web page for hardware similar
> to yours.  You know I make a custom kernel for each machine here, but
> there may be something in the hardware + driver selection help you in
> that info?  See:
> 
>   http://bugsplatter.id.au/kernel/boxen/
> 
> You can tell by the dates which are the more supported boxes there, some
> are no longer in use.
> 
> If I'm not specific enough, could you post lspci result and I'll look
> see what driver I'd select -- this would help?
> 
> If the machine is very new you might want o download the latest lspci
> data, I for get from where, it's in the docs somewhere ;)
> 
> Grant.
>
Note: comment inline.

Thanks for the response. I have just upgraded one of my test machines 
(samsung nc10) to the official kernel in slackware -current, "testing":
kernel-generic-smp-2.6.38.3_smp-i686-1

So far, that upgrade is working fine, but it hasn't been booted for a 
full day, yet. Fingers crossed that it keeps working! Note: I upgraded 
the other relevant packages per the official slackware changelog to this 
mark in the changelog: Mon Apr 18 19:13:18 UTC 2011

There were quite a few package upgrades and the changes may have played a 
part in the fix, but the kernel was most likely the biggest change. I 
haven't had a chance to test the other test machine, yet (that is, the 
Dell D620). These continuous crashes have not been kind to that disk 
filesystem (even though I am using a journalling filesystem, xfs). I also 
want to review the kernel changelog from 2.6.36.12 vs 2.6.38.3 for clues 
to the fix, but I am guessing the number of changes are substantial.

BTW, I have just posted a response to my "hibernation" thread in 
alt.os.linux.slackware.

-- 
Douglas Mayne

Back to comp.os.linux.misc | Previous | NextPrevious in thread | Find similar


Thread

ata errors on kernel 2.6.35.12 Douglas Mayne <none@invalid.com> - 2011-04-17 14:46 +0000
  Re: ata errors on kernel 2.6.35.12 The Natural Philosopher <tnp@invalid.invalid> - 2011-04-17 16:59 +0100
  Re: ata errors on kernel 2.6.35.12 "David W. Hodgins" <dwhodgins@nomail.afraid.org> - 2011-04-17 17:05 -0400
    Re: ata errors on kernel 2.6.35.12 Douglas Mayne <none@invalid.com> - 2011-04-17 21:37 +0000
  Re: ata errors on kernel 2.6.35.12 Jens <jens@tuxane.com> - 2011-04-18 10:59 +0200
    Re: ata errors on kernel 2.6.35.12 Douglas Mayne <none@invalid.com> - 2011-04-18 13:19 +0000
  Re: ata errors on kernel 2.6.35.12 Grant <omg@grrr.id.au> - 2011-04-19 10:59 +1000
    Re: ata errors on kernel 2.6.35.12 Douglas Mayne <none@invalid.com> - 2011-04-19 19:57 +0000

csiph-web