From: Jeff Garzik <jeff@garzik.org>
To: Volker Lanz <vl@fidra.de>
Cc: linux-ide@vger.kernel.org
Subject: Re: AHCI errors on ICH10 board
Date: Mon, 16 Mar 2009 16:40:03 -0400 [thread overview]
Message-ID: <49BEB923.6070001@garzik.org> (raw)
In-Reply-To: <200903161857.19245.vl@fidra.de>
Volker Lanz wrote:
> On Montag, 16. März 2009 16:06:46 Jeff Garzik wrote:
>> Your hardware is telling you that it is seeing CRC errors on the ATA bus
>> -- which means to say, a hardware problem that causes data transfer over
>> the SATA cable to fail.
>>
>> Possible sources of problems: bad SATA cable, bad power supply, bad
>> SATA port on device or mainboard, bad mainboard, ...
>
> Faulty hardware is of course always possible. It's hard to imagine something
> is wrong with the hardware in this case, though:
>
> * Both drives are affected, a Samsung and a Seagate. We can rule out the
> drives.
> * Both drives worked in AHCI mode with the same cables and power supply with
> the ICH9 based Gigabyte board: We can rule out the drives themselves again and
> also the cables.
> * Different SATA cables did not help. We can rule out the cables again.
> * Unplugging everything in the machine except the drive with the root
> partition and the video card did not help: We can rule out the power supply (a
> high quality Tagan 400W model, not some OEM junk) with a near 100%
> probability, I suppose.
>
> So, all that's left is the mainboard. But on this same machine, AHCI in
> Microsoft Windows XP works without any problem whatsoever. I know the old
I would double-check that you are using an AHCI driver in XP, since that
is uncommon. Most XP drivers for AHCI-capable hardware program the SATA
device in legacy IDE mode.
> Linux-stresses-your-hardware-more-than-Windows tale, but I haven't seen proof
> of that for 10 years now. This finally seems to rule out the mainboard and at
> least to me it appears the software side is all that is left.
> What can I do to diagnose this further short of buying an additional power
> supply and mainboard?
Well, the root cause of these errors are CRC errors during transmission
over the SATA cable. You can extrapolate from there... it could be
heat, poor cable shielding, dirty A/C power, bad RAM, who knows.
Your AHCI chip is telling Linux "hardware CRC error, that I could not
recover from" and Linux is dutifully reporting that back to you.
Software is unlikely as the cause, considering the volume of AHCI chips
in users hands versus the volume of bug reports like yours.
Every problem is different, of course, but the standard recommendations are
- trying different SATA ports on the mainboard. SATA ports are easy to
break or "fuzz" into uselessness
- replacing cables
- replacing the power supply
- running memtest86 or replacing RAM, etc.
- disabling AHCI mode
Mainboards can definitely go bad, but the above tends to be of more help.
Jeff
prev parent reply other threads:[~2009-03-16 20:40 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-03-16 13:06 AHCI errors on ICH10 board Volker Lanz
2009-03-16 15:06 ` Jeff Garzik
2009-03-16 17:57 ` Volker Lanz
2009-03-16 20:40 ` Jeff Garzik [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=49BEB923.6070001@garzik.org \
--to=jeff@garzik.org \
--cc=linux-ide@vger.kernel.org \
--cc=vl@fidra.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).