3.12-stable review patch. If anyone has any objections, please let me know. ------------------ From: Nicholas Bellinger <nab@xxxxxxxxxxxxxxx> commit 4a4caa29f1abcb14377e05d57c0793d338fb945d upstream. This patch addresses an traditional iscsi-target fabric ack starvation issue where iscsit_allocate_cmd() -> percpu_ida_alloc_state() ends up hitting slow path percpu-ida code, because iscsit_ack_from_expstatsn() is expected to free ack'ed tags after tag allocation. This is done to take into account the tags waiting to be acknowledged and released in iscsit_ack_from_expstatsn(), but who's number are not directly limited by the CmdSN Window queue_depth being enforced by the target. So that said, this patch bumps up the pre-allocated number of per session tags to: (max(queue_depth, ISCSIT_MIN_TAGS) * 2) + ISCSIT_EXTRA_TAGS for good measure to avoid the percpu_ida_alloc_state() slow path. Signed-off-by: Nicholas Bellinger <nab@xxxxxxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/target/iscsi/iscsi_target_nego.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) --- a/drivers/target/iscsi/iscsi_target_nego.c +++ b/drivers/target/iscsi/iscsi_target_nego.c @@ -1192,7 +1192,7 @@ get_target: */ alloc_tags: tag_num = max_t(u32, ISCSIT_MIN_TAGS, queue_depth); - tag_num += (tag_num / 2) + ISCSIT_EXTRA_TAGS; + tag_num = (tag_num * 2) + ISCSIT_EXTRA_TAGS; tag_size = sizeof(struct iscsi_cmd) + conn->conn_transport->priv_size; ret = transport_alloc_session_tags(sess->se_sess, tag_num, tag_size); -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html