[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Marvel controller gives errors Squeeze



On Tue, 03 Jul 2012 02:32:44 +0200, Mark Panen wrote:

> On 02/07/2012 16:26, Camaleón wrote:
>> Is ata2 your hdd? Run "dmesg | grep -i ata2" to find out.
> ata8 is my HDD

Uh? I had a dejavù... What happened with ata2? :-)

> [    0.888232] ata8: SATA max UDMA/133 abar m2048@0xf7fff800 port 0xf7fff980 irq 36
> [    1.371432] ata8: SATA link up 6.0 Gbps (SStatus 133 SControl 370) 
> [    1.375914] ata8.00: ATA-8: WDC WD2002FAEX-007BA0, 05.01D05, max UDMA/133 
                                     ^^^^^^^^^^^^^^^^^

> [    1.375919] ata8.00: 3907029168 sectors, multi 0: LBA48 NCQ (depth 31/32), AA
> [    1.380941] ata8.00: configured for UDMA/133 

That makes more sense: hard disk is proplery listed and sata speed 
matches with hdd and board specs.

> [   36.917447] ata8.00: exception Emask 0x0 SAct 0xf SErr 0x0 action 0x6 frozen
> [   36.917470] ata8.00: failed command: READ FPDMA QUEUED

These entries are logged in a later stage so it means that all of 
the sata ports are throwing "exception" errors. To list them run:

dmesg | grep -i ata[1-12]

> On boot i receive a message that my Marvel controller on my mobo can
> only handle 2.5Gb/s so ASUS has given the incorrect specs for this mobo
> selling it as 6Gb/s but dmesg seems to give 6Gb/s, so i don't know
> what's what.

Upload the full dmesg at www.pastebin.com so we can review.

> It's a ASUS P7H55-M/USB3 mobo.

According to the specifications, it should have:

6xSATA 3.0 Gb/s ports (intel chipset)
2xSATA 6.0 Gb/s ports (marvel chipset)

What I would worry about is for the "exception" error and "failed command".

Googling around I've found this:

http://iwtf.net/2011/05/19/western-digital-sata-errors-under-linux/

You can try what it suggests but instead making the change permanent, you 
can first append "libata.force=8.00:noncq" to the kernel line in GRUB's 
boot menu) and check if the errors are gone. If yes, congratulations, 
you are hitting either a kernel or WD bug ;-)

Greetings,

-- 
Camaleón


Reply to: