| Message ID | c34f0dcd116683c05c7c1836c381a09f@imap.linux.ibm.com (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
Can you please resend this patch as well? There's something off with the formatting. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 2014-09-17 21:51, Hefty, Sean wrote: > Can you please resend this patch as well? There's something off with > the formatting. Hi Sean, I have resent the patch in v5. Also attached the patch file to message for convenience. Thank You. - Sreedhar -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/src/rsocket.c b/src/rsocket.c index ffea0ca..d201e0a 100644 --- a/src/rsocket.c +++ b/src/rsocket.c @@ -384,6 +384,7 @@ struct rsocket { dlist_entry iomap_list; dlist_entry iomap_queue; int iomap_pending; + int unack_cqe; }; #define DS_UDP_TAG 0x55555555 @@ -638,6 +639,7 @@ static struct rsocket *rs_alloc(struct rsocket *inherited_rs, int type) rs->target_iomap_size = def_iomap_size; } } + rs->unack_cqe = 0; fastlock_init(&rs->slock); fastlock_init(&rs->rlock);
Author: Sreedhar Kodali <srkodali@linux.vnet.ibm.com> Change note from v2 to v3: eliminated setting through option as suggested by Sean Date: Mon Sep 8 15:38:54 2014 +0530 Perform completion event acknowledgments in batches instead of individually to minimze locking overheads. Size of the completion queue decides the size of a batch. Signed-off-by: Sreedhar Kodali <srkodali@linux.vnet.ibm.com> --- fastlock_init(&rs->cq_lock); @@ -1044,8 +1046,11 @@ static void rs_free(struct rsocket *rs) if (rs->cm_id) { rs_free_iomappings(rs); - if (rs->cm_id->qp) + if (rs->cm_id->qp) { + if (rs->unack_cqe > 0) + ibv_ack_cq_events(rs->cm_id->recv_cq, rs->unack_cqe); rdma_destroy_qp(rs->cm_id); + } rdma_destroy_id(rs->cm_id); } @@ -2026,7 +2031,11 @@ static int rs_get_cq_event(struct rsocket *rs) resume_get_cq_event: ret = ibv_get_cq_event(rs->cm_id->recv_cq_channel, &cq, &context); if (!ret) { - ibv_ack_cq_events(rs->cm_id->recv_cq, 1); + rs->unack_cqe += 1; + if (rs->unack_cqe == rs->sq_size + rs->rq_size) { + ibv_ack_cq_events(rs->cm_id->recv_cq, rs->unack_cqe); + rs->unack_cqe = 0; + } rs->cq_armed = 0; } else if (restart_onintr == 1 && errno == EINTR) { errno = 0; -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html