sata_mv hotplug flaky?

sata_mv hotplug flaky?

[Orion Poplawski]

Not sure what the latest status of sata_mv hotplug should be, but it 
seems close.  I'm currently running 2.6.24-0.81.rc4.git7.fc9 with a 
MV88SX5081.  Pulled a couple drives and re-added.  One device got 
re-added, but the other did not.  It seems like I got the system to 
probe and re-add the first drive by doing a "mount -a", but haven't been 
able (no idea how) to re-add the second.


Dec 11 16:51:10 saga kernel: ata8: exception Emask 0x10 SAct 0x0 SErr 
0x180002 action 0xe frozen
Dec 11 16:51:10 saga kernel: ata8: edma_err 0x00000108, dev disconnect, 
EDMA self-disable
Dec 11 16:51:10 saga kernel: ata8: SError: { RecovComm 10B8B Dispar }
Dec 11 16:51:10 saga kernel: ata8: hard resetting link
Dec 11 16:51:10 saga kernel: ata8: SATA link down (SStatus 0 SControl 300)
Dec 11 16:51:10 saga kernel: ata8: failed to recover some devices, 
retrying in 5 secs
Dec 11 16:51:15 saga kernel: ata8: hard resetting link
Dec 11 16:51:15 saga kernel: ata8: SATA link down (SStatus 0 SControl 300)
Dec 11 16:51:15 saga kernel: ata8: failed to recover some devices, 
retrying in 5 secs
Dec 11 16:51:20 saga kernel: ata8: hard resetting link
Dec 11 16:51:20 saga kernel: ata8: SATA link down (SStatus 0 SControl 300)
Dec 11 16:51:20 saga kernel: ata8.00: disabled
Dec 11 16:51:20 saga kernel: ata8: EH complete
Dec 11 16:51:20 saga kernel: ata8.00: detaching (SCSI 7:0:0:0)
Dec 11 16:51:20 saga kernel: sd 7:0:0:0: [sdh] Synchronizing SCSI cache
Dec 11 16:51:20 saga kernel: sd 7:0:0:0: [sdh] Result: 
hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK,SUGGEST_OK
Dec 11 16:51:20 saga kernel: sd 7:0:0:0: [sdh] Stopping disk
Dec 11 16:51:20 saga kernel: sd 7:0:0:0: [sdh] START_STOP FAILED
Dec 11 16:51:20 saga kernel: sd 7:0:0:0: [sdh] Result: 
hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK,SUGGEST_OK

Dec 11 16:51:54 saga kernel: ata5: exception Emask 0x10 SAct 0x0 SErr 
0x180002 action 0xe frozen
Dec 11 16:51:54 saga kernel: ata5: edma_err 0x00000108, dev disconnect, 
EDMA self-disable
Dec 11 16:51:54 saga kernel: ata5: SError: { RecovComm 10B8B Dispar }
Dec 11 16:51:54 saga kernel: ata5: hard resetting link
Dec 11 16:51:54 saga kernel: ata5: SATA link down (SStatus 0 SControl 300)
Dec 11 16:51:54 saga kernel: ata5: failed to recover some devices, 
retrying in 5 secs
Dec 11 16:51:59 saga kernel: ata5: hard resetting link
Dec 11 16:51:59 saga kernel: ata5: SATA link down (SStatus 0 SControl 300)
Dec 11 16:51:59 saga kernel: ata5: failed to recover some devices, 
retrying in 5 secs
Dec 11 16:52:04 saga kernel: ata5: hard resetting link
Dec 11 16:52:04 saga kernel: ata5: SATA link down (SStatus 0 SControl 300)
Dec 11 16:52:04 saga kernel: ata5.00: disabled
Dec 11 16:52:04 saga kernel: ata5: EH complete
Dec 11 16:52:04 saga kernel: ata5.00: detaching (SCSI 4:0:0:0)
Dec 11 16:52:04 saga kernel: sd 4:0:0:0: [sde] Synchronizing SCSI cache
Dec 11 16:52:04 saga kernel: sd 4:0:0:0: [sde] Result: 
hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK,SUGGEST_OK
Dec 11 16:52:04 saga kernel: sd 4:0:0:0: [sde] Stopping disk
Dec 11 16:52:04 saga kernel: sd 4:0:0:0: [sde] START_STOP FAILED
Dec 11 16:52:04 saga kernel: sd 4:0:0:0: [sde] Result: 
hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK,SUGGEST_OK

Dec 11 16:52:59 saga kernel: ata8: exception Emask 0x10 SAct 0x0 SErr 
0x0 action 0xe frozen
Dec 11 16:52:59 saga kernel: ata8: edma_err 0x00000010, dev connect
Dec 11 16:52:59 saga kernel: ata8: hard resetting link
Dec 11 16:53:00 saga kernel: ata8: SATA link up 1.5 Gbps (SStatus 113 
SControl 300)
Dec 11 16:53:00 saga kernel: ata8.00: ATA-6: WDC WD2500PD-75FZB0, 
02.05D02, max UDMA/100
Dec 11 16:53:00 saga kernel: ata8.00: 488281250 sectors, multi 0: LBA48
Dec 11 16:53:00 saga kernel: ata8.00: applying bridge limits
Dec 11 16:53:00 saga kernel: ata8.00: configured for UDMA/100
Dec 11 16:53:00 saga kernel: ata8: exception Emask 0x10 SAct 0x0 SErr 
0x0 action 0xf t4
Dec 11 16:53:00 saga kernel: ata8: edma_err 0x00000018, dev disconnect
Dec 11 16:53:00 saga kernel: ata8: hard resetting link

Dec 11 16:53:00 saga kernel: ata8: SATA link up 1.5 Gbps (SStatus 113 
SControl 300)
Dec 11 16:53:00 saga kernel: ata8.00: configured for UDMA/100
Dec 11 16:53:00 saga kernel: ata8: exception Emask 0x10 SAct 0x0 SErr 
0x0 action 0xf t3
Dec 11 16:53:00 saga kernel: ata8: edma_err 0x00000018, dev disconnect
Dec 11 16:53:00 saga kernel: ata8: hard resetting link
Dec 11 16:53:01 saga kernel: ata8: SATA link up 1.5 Gbps (SStatus 113 
SControl 300)
Dec 11 16:53:01 saga kernel: ata8.00: configured for UDMA/100
Dec 11 16:53:01 saga kernel: ata8: EH complete
Dec 11 16:53:01 saga kernel: scsi 7:0:0:0: Direct-Access     ATA 
WDC WD2500PD-75F 02.0 PQ: 0 ANSI: 5
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: [sde] 488281250 512-byte 
hardware sectors (250000 MB)
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: [sde] Write Protect is off
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: [sde] Write cache: enabled, 
read cache: enabled, doesn't support DPO or FUA
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: [sde] 488281250 512-byte 
hardware sectors (250000 MB)
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: [sde] Write Protect is off
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: [sde] Write cache: enabled, 
read cache: enabled, doesn't support DPO or FUA
Dec 11 16:53:01 saga kernel:  sde: sde1 sde2
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: [sde] Attached SCSI disk
Dec 11 16:53:01 saga kernel: sd 7:0:0:0: Attached scsi generic sg5 type 0



-- 
Orion Poplawski
Technical Manager                     303-415-9701 x222
NWRA/CoRA Division                    FAX: 303-415-9702
3380 Mitchell Lane                  orion@cora.nwra.com
Boulder, CO 80301              http://www.cora.nwra.com

Re: sata_mv hotplug flaky?

[Jeff Garzik]

Orion Poplawski wrote:
> Not sure what the latest status of sata_mv hotplug should be, but it 
> seems close.  I'm currently running 2.6.24-0.81.rc4.git7.fc9 with a 
> MV88SX5081.  Pulled a couple drives and re-added.  One device got 
> re-added, but the other did not.  It seems like I got the system to 
> probe and re-add the first drive by doing a "mount -a", but haven't been 
> able (no idea how) to re-add the second.
> 

hotplug made great strides with the introduction of the new-EH code, but 
it still needs a bit of work.  Mark Lord was looking into that, so you 
can sure your report has been noted...

	Jeff

Re: sata_mv hotplug flaky?

[Mark Lord]

Jeff Garzik wrote:
> Orion Poplawski wrote:
>> Not sure what the latest status of sata_mv hotplug should be, but it 
>> seems close.  I'm currently running 2.6.24-0.81.rc4.git7.fc9 with a 
>> MV88SX5081.  Pulled a couple drives and re-added.  One device got 
>> re-added, but the other did not.  It seems like I got the system to 
>> probe and re-add the first drive by doing a "mount -a", but haven't 
>> been able (no idea how) to re-add the second.
>>
> 
> hotplug made great strides with the introduction of the new-EH code, but 
> it still needs a bit of work.  Mark Lord was looking into that, so you 
> can sure your report has been noted...
..

I am not yet looking into sata_mv hotplug, but it is on the list
for early 2008.  There are some peculiarities and errata that need
to be addressed on some of the Marvell chips for this to work reliably.

Target kernel for it is now 2.6.26, with development/testing happening
during the 2.6.25 timeframe (but too late for inclusion there).

If the required tweaks are small and safe enough, then we might be
able to push a few of those out for 2.6.25.  We'll see.

The 2.6.26 kernel should also get full NCQ and PMP support for sata_mv,
and a smattering of other enhancements as well (eg. ATAPI support).

Cheers