Re: 2.6.32.2 SATA link detect failed, 2.6.32.1 works fine

[Bjarke Istrup Pedersen]

2009/12/23 fengxiangjun <fengxiangjun@xxxxxxxxxxx>:
>
> Hi,
>
> with this patch, the problem still exists.
>
> ata1.01: XXX debounce start, SStatus=4
> ata1.01: XXX debounce done, SStatus=4, DET stable for 520 msecs
> ata1.00: SATA link down (SStatus 0 SControl 301)
> ata1.01: SATA link down (SStatus 0 Scontrol 300)
> ...
> VFS; cannot open root device ...
>
> The kernel just panic, and I can't see what's the messege printing before
> that.
>
> So, I tried 'console=ttyS0 console=tty1'ïfull dmesg attachedã
>
> $ hdparm -I /dev/sda:
>
> /dev/sda:
>
> ATA device, with non-removable media
> Â Â Â ÂModel Number: Â Â Â WDC WD800JD-60LSA5
> Â Â Â ÂSerial Number: Â Â ÂWD-WCAM9T882872
> Â Â Â ÂFirmware Revision: Â10.01E03
> Standards:
> Â Â Â ÂSupported: 7 6 5 4
> Â Â Â ÂLikely used: 8
> Configuration:
>    ÂLogical     max   current
>    Âcylinders    16383  16383
>    Âheads      16   Â16
>    Âsectors/track  63   Â63
> Â Â Â Â--
> Â Â Â ÂCHS current addressable sectors: Â 16514064
> Â Â Â ÂLBA Â Âuser addressable sectors: Â156301488
> Â Â Â ÂLBA48 Âuser addressable sectors: Â156301488
> Â Â Â Âdevice size with M = 1024*1024: Â Â Â 76319 MBytes
> Â Â Â Âdevice size with M = 1000*1000: Â Â Â 80026 MBytes (80 GB)
> Â Â Â Âcache/buffer size Â= 8192 KBytes
> Capabilities:
> Â Â Â ÂLBA, IORDY(can be disabled)
> Â Â Â ÂStandby timer values: spec'd by Standard, with device specific
> minimum
> Â Â Â ÂR/W multiple sector transfer: Max = 16 ÂCurrent = 16
> Â Â Â ÂDMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 *udma5
> Â Â Â Â Â Â Cycle time: min=120ns recommended=120ns
> Â Â Â ÂPIO: pio0 pio1 pio2 pio3 pio4
> Â Â Â Â Â Â Cycle time: no flow control=120ns ÂIORDY flow control=120ns
> Commands/features:
> Â Â Â ÂEnabled Supported:
> Â Â Â Â Â * Â ÂSMART feature set
> Â Â Â Â Â Â Â ÂSecurity Mode feature set
> Â Â Â Â Â * Â ÂPower Management feature set
> Â Â Â Â Â * Â ÂWrite cache
> Â Â Â Â Â * Â ÂLook-ahead
> Â Â Â Â Â * Â ÂWRITE_BUFFER command
> Â Â Â Â Â * Â ÂREAD_BUFFER command
> Â Â Â Â Â * Â ÂNOP cmd
> Â Â Â Â Â * Â ÂDOWNLOAD_MICROCODE
> Â Â Â Â Â * Â Â48-bit Address feature set
> Â Â Â Â Â * Â ÂDevice Configuration Overlay feature set
> Â Â Â Â Â * Â ÂMandatory FLUSH_CACHE
> Â Â Â Â Â * Â ÂFLUSH_CACHE_EXT
> Â Â Â Â Â * Â ÂSMART error logging
> Â Â Â Â Â * Â ÂSMART self-test
> Â Â Â Â Â * Â ÂGeneral Purpose Logging feature set
> Â Â Â Â Â * Â ÂSATA-I signaling speed (1.5Gb/s)
> Â Â Â Â Â * Â ÂSATA-II signaling speed (3.0Gb/s)
> Â Â Â Â Â * Â ÂPhy event counters
> Â Â Â Â Â Â Â ÂDevice-initiated interface power management
> Â Â Â Â Â * Â ÂSoftware settings preservation
> Â Â Â Â Â * Â ÂSMART Command Transport (SCT) feature set
> Â Â Â Â Â * Â ÂSCT Long Sector Access (AC1)
> Â Â Â Â Â * Â ÂSCT LBA Segment Access (AC2)
> Â Â Â Â Â * Â ÂSCT Error Recovery Control (AC3)
> Â Â Â Â Â * Â ÂSCT Features Control (AC4)
> Â Â Â Â Â * Â ÂSCT Data Tables (AC5)
> Â Â Â Â Â Â Â Âunknown 206[12] (vendor specific)
> Security:
> Â Â Â ÂMaster password revision code = 65534
> Â Â Â Â Â Â Â Âsupported
>    Ânot   enabled
>    Ânot   locked
> Â Â Â Â Â Â Â Âfrozen
>    Ânot   expired: security count
> Â Â Â Â Â Â Â Âsupported: enhanced erase
> Â Â Â Â24min for SECURITY ERASE UNIT. 24min for ENHANCED SECURITY ERASE
> UNIT.
> Checksum: correct
>
> Thanks.
>
> On Wed, 23 Dec 2009, Tejun Heo wrote:
>
>> Hello,
>>
>> On 12/22/2009 11:02 AM, fengxiangjun wrote:
>> > You are right, it's definitely timing related.
>> >
>> > I just tried capturing the failed booting message through a null modem
>> > cable, but the problem goes away! It seems that I can't reproduce it
>> > with serial console attached.
>>
>> Hmmm... this is the second report of probing failure with SIDPR based
>> hardreset. ÂCan you please apply the attached patch and see whether
>> the problem is reproducible? ÂAnd if so, please attach full boot log.
>> Can you please also attach hdparm -I output of the occassionally
>> failing drive.
>>
>> Thanks.
>>
>> --
>> tejun
>>
> ---------------------------------------------------------------------------------------------------
> Confidentiality Notice: The information contained in this e-mail and any accompanying attachment(s)
> is intended only for the use of the intended recipient and may be confidential and/or privileged of
> Neusoft Corporation, its subsidiaries and/or its affiliates. If any reader of this communication is
> not the intended recipient, unauthorized use, forwarding, printing, Âstoring, disclosure or copying
> is strictly prohibited, and may be unlawful.If you have received this communication in error,please
> immediately notify the sender by return e-mail, and delete the original message and all copies from
> your system. Thank you.
> ---------------------------------------------------------------------------------------------------
>

Hey,

Could it be an idea to do a bisect for the patch that makes it break? :)

/Bjarke
--
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/