On Wed, 2016-06-01 at 08:29 -0700, Bart Van Assche wrote:

On 06/01/2016 07:36 AM, Kevin Groeneveld wrote:

quoted

quoted

Subject: [PATCH v2 1/2] scsi: fix race between simultaneous
decrements of ->host_failed

I wonder if this could be related to 
http://www.spinics.net/lists/linux-scsi/msg86808.html?

I never did get to the bottom of that.  If I have time I hope to
retest my scsi hang issue with this patch.

I've never seen that hang with the ib_srp driver (which is a SCSI 
LLD). That probably means that the SATA code uses the SCSI error 
handler in another way than other SCSI LLDs.

It's the ata_sas error handler, which is split between libsas and
libata.  It does one thread of execution per ata port when the strategy
handler is invoked.  This behaviour is pretty much unique at the
moment.

James