Comment 41 for bug 1063474

On 8/15/2013 9:47 PM, Bruce Link wrote:
> On 8/15/2013 12:36 AM, Robert Hancock wrote:
>> On Wed, Aug 14, 2013 at 4:17 PM, Bruce Link <Bruce@1045.ca> wrote:
>>>> These NVIDIA SATA controllers are a pain as they all seem to have a
>>>> different set of bugs related to hardreset, etc. What happens if
>>>> you boot up
>>>> without the drive connected and then plug in the cable after the
>>>> system
>>>> boots up?
>>> Roughly the same error is returned. See below for the dmesg output.
>>>
>>> root@teevee:/sys/devices/pci0000:00/0000:00:06.0/ata1/host0/scsi_host/host0#
>>>
>>> echo "- - -" > /sys/class/scsi_host/host4/scan
>>> root@teevee:/sys/devices/pci0000:00/0000:00:06.0/ata1/host0/scsi_host/host0#
>>>
>>> dmesg|grep ata5
>>> [ 1.030224] ata5: SATA max UDMA/133 cmd 0x9e0 ctl 0xbe0 bmdma
>>> 0xc400 irq
>>> 20
>>> [ 1.354216] ata5: SATA link down (SStatus 0 SControl 300)
>>> [ 944.773288] ata5: exception Emask 0x10 SAct 0x0 SErr 0x50000
>>> action 0xf
>>> [ 944.773304] ata5: SError: { PHYRdyChg CommWake }
>>> [ 944.773322] ata5: hard resetting link
>>> [ 945.652070] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
>>> [ 945.660176] ata5.00: ATAPI: ATAPI iHOS104, WL0F, max UDMA/100
>>> [ 945.676165] ata5.00: configured for UDMA/100
>>> [ 950.676049] ata5.00: qc timeout (cmd 0xa0)
>>> [ 950.676056] ata5.00: TEST_UNIT_READY failed (err_mask=0x5)
>>> [ 950.676064] ata5: hard resetting link
>>> [ 950.676066] ata5: nv: skipping hardreset on occupied port
>>> [ 951.144073] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
>>> [ 951.168168] ata5.00: configured for UDMA/100
>>> [ 956.168049] ata5.00: qc timeout (cmd 0xa0)
>>> [ 956.168057] ata5.00: TEST_UNIT_READY failed (err_mask=0x4)
>>> [ 956.168061] ata5: limiting SATA link speed to 1.5 Gbps
>>> [ 956.168064] ata5.00: limiting speed to UDMA/100:PIO3
>>> [ 956.168070] ata5: hard resetting link
>>> [ 956.168072] ata5: nv: skipping hardreset on occupied port
>>> [ 956.636056] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
>>> [ 956.660169] ata5.00: configured for UDMA/100
>>> [ 961.660036] ata5.00: qc timeout (cmd 0xa0)
>>> [ 961.660044] ata5.00: TEST_UNIT_READY failed (err_mask=0x4)
>>> [ 961.660046] ata5.00: disabled
>>> [ 961.660061] ata5: hard resetting link
>>> [ 962.544045] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
>>> [ 962.544061] ata5: EH complete
>>> root@teevee:/sys/devices/pci0000:00/0000:00:06.0/ata1/host0/scsi_host/host0#
>>>
>>>
>> Hmm, it did a hard reset on the hotplug with seemingly the same
>> effect. Can we narrow down the problem any more: does this drive work
>> on this machine under any Linux version, or in Windows? Can you try it
>> in another Linux machine with a different controller to see if it
>> works there?
> Well, this is embarrassing.
>
> I've always assumed the problem was with the with the motherboard, as
> the DVD will always successfully boot a windows installer, so I
> assumed everything with the drive was OK. On further inspection, it
> appears that the drive falls down when being accessed from the WinPE
> environment, much the same as in linux. I've tested this on another PC
> and the result looks to be the same. I'm going to do some more
> testing, and will send another email if I find out otherwise. But I
> think we can consider this closed.
>
> Sorry to be a bother.
> Bruce
I hooked up the DVD drive to a USB-SATA adapter to a PC and the DVD
drive works great (under both windows and linux). I was able to copy a
large number of files with no performance issues.
The PC that I did the second test on turns out to have a nVidia MCP51
chipset (It's a gateway GT4016 with a KTBC51G-LF motherboard). I should
have checked this beforehand.

So this problem looks to persist across different nVidia chipsets.