[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [dm-devel] [PATCHSET block#for-2.6.36-post] block: replace barrier with sequenced flush

Hannes Reinecke, on 08/30/2010 01:54 PM wrote:
As I've wrote multiple times, I'm pretty skeptical it will bring much.
Ordered tag mandates draining inside the device just like the original
barrier implementation.  Sure, it's done at a lower layer and command
issue latencies will be reduced thanks to that but ordered-by-waiting
doesn't require _any_ draining at all.  The whole pipeline can be kept
full all the time.  I'm often wrong tho, so please feel free to go
ahead and prove me wrong.  :-)

Actually, I thought about ordered tag writes, too.
But eventually I had to give up on this for a simple reason:
Ordered tag controls the ordering on the SCSI _TARGET_. But for a
meaningful implementation we need to control the ordering all the way
down from ->queuecommand(). Which means we have three areas we need
to cover here:
- driver (ie between ->queuecommand() and passing it off to the firmware)
- firmware
- fabric

Sadly, the latter two are really hard to influence. And, what's more,
with the new/modern CNAs with multiple queues and possible multiple
routes to the target it becomes impossible to guarantee ordering.
So using ordered tags for FibreChannel is not going to work, which
makes implementing it a bit of a pointless exercise for me.

The situation is, actually, much better than you think. An SCSI transport should provide an in-order delivery of commands. In some transports it is required (e.g. iSCSI), in some - optional (e.g. FC). For FC "an application client may determine if a device server supports the precise delivery function by using the MODE SENSE and MODE SELECT commands to examine and set the enable precise delivery checking (EPDC) bit in the Fibre Channel Logical Unit Control page" (Fibre Channel Protocol for SCSI (FCP)). You can find more details in FCP section "Precise delivery of SCSI commands".

Regarding multiple queues, in case of a multipath access to a device SCSI requires either each path be a separate I_T nexus, where order of commands is maintained, or a transport required to maintain in-order commands delivery among multiple paths in a single I_T nexus (session) as it is done in iSCSI's MC/S and, most likely, wide SAS ports.

So, everything is in the specs. We only need to use it properly. How it can be done on the drivers level as well as how errors recovery can be done using ACA and UA_INTLCK facilities I wrote few weeks ago in the "[RFC] relaxed barrier semantics" thread.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]