Hi, On 07/30/2014 02:21 PM, Oliver Neukum wrote: > On Wed, 2014-07-30 at 12:58 +0200, Hans de Goede wrote: >> Hi, >> >> On 07/30/2014 12:00 PM, Oliver Neukum wrote: >>> Hi, >>> >>> looking at this code made me think whether there actually is a limit >>> on the host. It seems to me that we should not have more concurrent >>> commands than the host controller supports streams. >>> What do you think? >> >> AFAIK this is already enforced by the scsi_activate_tcq(sdev, devinfo->qdepth - 2); >> call in uas_slave_configure(). I assume the host can_queue setting is relevant >> for hosts which can have more then one slave, but for us not so much. > > I can see no reason why a UAS device shouldn't have multiple > slaves. A disk enclosure would be the obvious example. Ok, my bad let me try to be more clear here. There can be only one target as the uas protocol has no target-id. There can be multiple slaves by having multi-lun devices. Still this is not a problem, as uas.c uses scsi_init_shared_tag_map so all queues / slaves will share the same tag map and we still cannot get more then one outstanding command per tag. Regards, Hans -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
