Tejun (and all) -

I've attached a log from my machine marked up with a few extra printk's.
I did add the ssleep() you suggested off line - made no difference.

I'll be happy to add more prinks in places that I did not ...

...tom


the working log (visible then):
[4194022.295042] ata6: SATA max UDMA/100 host m128@0x9aa9000 port 
0x9aa4000 irq 31
[4194022.295042] ata7: SATA max UDMA/100 host m128@0x9aa9000 port 
0x9aa6000 irq 31
[4194024.462033] ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[4194024.462033] ata4.15: Port Multiplier 1.1, 0x1095:0x4726 r31, 7 
ports, feat 0x1/0x9
[4194024.463009] ata4.00: hard resetting link
[4194024.770626] ata4.00: SATA link up 3.0 Gbps (SStatus 123 SControl 320)

Now insetad, I see:
ata1: SATA max UDMA/100 host m128@0x9aa9000 port 0x9aa0000 irq 31
ata2: SATA max UDMA/100 host m128@0x9aa9000 port 0x9aa2000 irq 31
ata3: SATA max UDMA/100 host m128@0x9aa9000 port 0x9aa4000 irq 31
ata4: SATA max UDMA/100 host m128@0x9aa9000 port 0x9aa6000 irq 31
ata1: ata_eh_recover - 1
ata1: ata_eh_recover - 2 re-enable link?
ata1.00: ata_eh_recover - 3
ata1.00: ata_eh_recover - 5
ata1.00: ata_eh_recover - 6
ata1.00: ata_eh_recover -> ata_dev_init 7
ata_eh_recover - 8 - retry
ata1: ata_eh_recover - 8a prep for EH
ata1: ata_eh_recover - reset
ata1: ata_eh_recover - ATA_DEV_UNKNOWN
ata_eh_recover - 9 - reset is true
ata_eh_recover -> freezing port -> nr_pmp_links == 0
ata1: ata_eh_recover -> ata_eh_reset


Tejun Heo wrote:
> Thomas Evans wrote:
>   
>> I'm gonna throw a bunch of printfk's around to see what happens as well.
>>     
>
> Great, please don't forget to cc linux-ide when reporting result.
>
>