Hello,

On Mon, Aug 29, 2011 at 04:49:29PM +0100, Bruce Stenning wrote:

I should mention that we are now working on kernel 2.6.39.4 (and are choosing
not to go with 3.x to limit changes as we progress towards final testing.)

I don't think it would make any difference.  The core EH hasn't
changed for some time now.

[...]
ata4: EH complete
ata_scsi_error: EXIT
__ata_port_freeze: ata4 port frozen
ata_port_schedule_eh: port EH scheduled
ata_scsiop_noop: ENTER
ata_scsiop_read_cap: ENTER
ata_scsiop_mode_sense: ENTER
ata_scsiop_mode_sense: ENTER
ata_scsiop_mode_sense: ENTER

That's weird.  Another EH is scheduled after one completes but EH
doesn't kick in.  Hmmm... can't see how that can happen.  How easily
can you reproduce the problem?

Thanks.

-- 
tejun