On Mon, 06 Oct 2008 16:25:23 +0200 Doron Shoham <dorons at voltaire.com> wrote: > Alexander Nezhinsky wrote: > > Backing store based on SCSI Generic driver (sg), v.3. > > allows to route scsi commands to the native scsi devices > > unchanged. It uses Direct I/O (dio) to avoid memcopy > > to the buffer cache. > > > > Using sg allows asynchronous execution, commands are > > written to a sg-backed char device, completions are > > read from the same device, which can be polled > > asynchronously. > > > > This bs provides significant performance improvement when > > working with native scsi devices. In a setup, where > > the scsi devices are exported by a tgt with bs_null, > > and both links (from initiator to target and from the > > target to the "backing-store" target) are iSER/IB > > sustained bandwidth of 1450 MB/s for READ and > > 1350 MB/s for WRITE is achieved. This to be compared to > > 700-800 MB/s when running with bs_rdwr in the same setup. > > Some improvements are seen with IOPS as well: > > 60 kIOPS for READ, 38 kIOPS for WRITE > > (compared to 31/35KIOPS with bs_rdwr). > > > > Hi, > > I applied this patch and it works, but I noticed that the serial number > reported by the device (using sg_inq) returned 'beaf11' instead of the real physical serial number. > > Does it mean that the inquiry command is not transferred to the device and handled internally? Yeah, this just uses sg for I/Os. You are talking about passthrough. -- To unsubscribe from this list: send the line "unsubscribe stgt" in the body of a message to majordomo at vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html |