Re: 2.6.18-rc2 Intermittent failures to detect sata disks

From: Keith Owens
Date: Tue Jul 25 2006 - 02:48:18 EST


Keith Owens (on Tue, 25 Jul 2006 16:27:34 +1000) wrote:
>Jeff Garzik (on Tue, 25 Jul 2006 01:57:08 -0400) wrote:
>>Keith Owens wrote:
>>> Keith Owens (on Fri, 21 Jul 2006 16:18:47 +1000) wrote:
>>>> I am seeing an intermittent failures to detect sata disks on
>>>> 2.6.18-rc2. Dell SC1425, PIIX chipset, gcc 4.1.0 (opensuse 10.1).
>>>> Sometimes it will detect both disks, sometimes only one, sometimes none
>>>> at all. AFAICT it only occurs after a soft reboot, and possibly only
>>>> after an emergency reboot. Alas the problem is so intermittent that it
>>>> is hard to tell what conditions will trigger it.
>>>
>>> I applied the debug patch below, turn on prink timing and set
>>> initdefault to 6 so the machine was in a continual soft reboot cycle.
>>> After multiple cycles I got this trace. piix_sata_prereset() reads a
>>> zero config byte for almost 15 seconds then it changes to 0x11,
>>> followed by a hang. Why is the config byte initially zero, and what
>>> makes it change? The normal value for pcs is 0x33.
>>
>>Can you try 2.6.18-rc2-git3?
>>
>> Jeff
>
>Running now, with the trivial bug fix below plus my debug patch. I
>will leave it running overnight, this problem is very intermittent.

Failed again on 2.6.18-rc2-git3. 80+ seconds of this before I killed
the power. No sign of it getting any data from the sata PCI config.

[ 8.689136] piix_sata_prereset: ata1: ENTER, pcs=0x0 base=0
[ 8.755861] piix_sata_prereset: ata1: LEAVE, pcs=0x0 present=0x0
[ 8.827741] piix_sata_prereset: ata1: ENTER, pcs=0x0 base=0
[ 8.894446] piix_sata_prereset: ata1: LEAVE, pcs=0x0 present=0x0

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/