4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Marta Rybczynska <mrybczyn@xxxxxxxxx>
commit 0544f5494a03b8846db74e02be5685d1f32b06c9 upstream.
In the case of small NVMe-oF queue size (<32) we may enter a deadlock
caused by the fact that the IB completions aren't sent waiting for 32
and the send queue will fill up.
The error is seen as (using mlx5):
[ 2048.693355] mlx5_0:mlx5_ib_post_send:3765:(pid 7273):
[ 2048.693360] nvme nvme1: nvme_rdma_post_send failed with error code -12
This patch changes the way the signaling is done so that it depends on
the queue depth now. The magic define has been removed completely.
Signed-off-by: Marta Rybczynska <marta.rybczynska@xxxxxxxxx>
Signed-off-by: Samuel Jones <sjones@xxxxxxxxx>
Acked-by: Sagi Grimberg <sagi@xxxxxxxxxxx>
Signed-off-by: Christoph Hellwig <hch@xxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/nvme/host/rdma.c | 18 ++++++++++++++----
1 file changed, 14 insertions(+), 4 deletions(-)
--- a/drivers/nvme/host/rdma.c
+++ b/drivers/nvme/host/rdma.c
@@ -1011,6 +1011,19 @@ static void nvme_rdma_send_done(struct i
nvme_rdma_wr_error(cq, wc, "SEND");
}
+static inline int nvme_rdma_queue_sig_limit(struct nvme_rdma_queue *queue)
+{
+ int sig_limit;
+
+ /*
+ * We signal completion every queue depth/2 and also handle the
+ * degenerated case of a device with queue_depth=1, where we
+ * would need to signal every message.
+ */
+ sig_limit = max(queue->queue_size / 2, 1);
+ return (++queue->sig_count % sig_limit) == 0;
+}
+
static int nvme_rdma_post_send(struct nvme_rdma_queue *queue,
struct nvme_rdma_qe *qe, struct ib_sge *sge, u32 num_sge,
struct ib_send_wr *first, bool flush)
@@ -1038,9 +1051,6 @@ static int nvme_rdma_post_send(struct nv
* Would have been way to obvious to handle this in hardware or
* at least the RDMA stack..
*
- * This messy and racy code sniplet is copy and pasted from the iSER
- * initiator, and the magic '32' comes from there as well.
- *
* Always signal the flushes. The magic request used for the flush
* sequencer is not allocated in our driver's tagset and it's
* triggered to be freed by blk_cleanup_queue(). So we need to
@@ -1048,7 +1058,7 @@ static int nvme_rdma_post_send(struct nv
* embeded in request's payload, is not freed when __ib_process_cq()
* calls wr_cqe->done().
*/
- if ((++queue->sig_count % 32) == 0 || flush)
+ if (nvme_rdma_queue_sig_limit(queue) || flush)
wr.send_flags |= IB_SEND_SIGNALED;
if (first)