On Wed, 2016-01-13 at 10:33 +0100, Christoph Hellwig wrote: > On Wed, Jan 13, 2016 at 01:22:49AM -0800, Nicholas A. Bellinger wrote: > > Huh..? It's used by iscsi-target, ib_isert, tcm_qla2xxx, vhost-scsi and > > tcm_fc. > > > > That's the majority of the drivers people care about. > > ib_isert never calls transport_init_session / transport_init_session_tags > directly. > Yes. Not to mention, there is a new struct iscsit_transport based driver on the way that assumes percpu-ida tag preallocation using this same method. > We've got 3 callers to transport_init_session_tags vs 6 to > transport_init_session. > In that case, I'll look into converting sbp-target, usb-gadget and xen-scsiback to use percpu-ida like the other consumers here for v4.6 code. > iSCSI already works roughly the way I suggest everything should > work in the future - it uses transport_init_session but manually > allocates the tags. iscsi-target needs it open-coded for discovery purposes, because se_node_acl lookup does not occur. However for the normal session login se_node_acl lookup back, I don't see a reason why iscsi-target can't use target_alloc_session() here too. -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
