nvme-rdma: rework queue maps handling
Sagi Grimberg
maintainerPATCH 1/1 nvme-rdma: rework queue maps handling
If the device supports less queues than provided (if the device has less completion vectors), we might hit a bug due to the fact that we ignore that in nvme_rdma_map_queues (we override the maps nr_queues with user opts).
Instead, keep track of how many default/read/poll queues we actually allocated (rather than asked by the user) and use that to assign our queue mappings.
--- drivers/nvme/host/rdma.c | 37 ++++++++++++++++++++++++------------- 1 file changed, 24 insertions(+), 13 deletions(-)
drivers/nvme/host/rdma.c +24 -13
@@ -119,6 +119,7 @@ struct nvme_rdma_ctrl {struct nvme_ctrl ctrl;bool use_inline_data; +u32 io_queues[HCTX_MAX_TYPES ]; };static inline struct nvme_rdma_ctrl *to_rdma_ctrl (struct nvme_ctrl *ctrl)
@@ -165,8 +166,8 @@ static inline int nvme_rdma_queue_idx(struct nvme_rdma_queue *queue)static bool nvme_rdma_poll_queue (struct nvme_rdma_queue *queue) {return nvme_rdma_queue_idx (queue) > - queue->ctrl->ctrl.opts->nr_io_queues + - queue->ctrl->ctrl.opts->nr_write_queues; + queue->ctrl->io_queues[HCTX_TYPE_DEFAULT ] + + queue->ctrl->io_queues[HCTX_TYPE_READ ]; }static inline size_t nvme_rdma_inline_data_size (struct nvme_rdma_queue *queue)
@@ -661,8 +662,20 @@ static int nvme_rdma_alloc_io_queues(struct nvme_rdma_ctrl *ctrl) nr_io_queues = min_t(unsigned int, nr_io_queues, ibdev->num_comp_vectors); - nr_io_queues +=min (opts->nr_write_queues,num_online_cpus ()); - nr_io_queues +=min (opts->nr_poll_queues,num_online_cpus ()); + ctrl->io_queues[HCTX_TYPE_READ ] = nr_io_queues; +if (opts->nr_write_queues) { + ctrl->io_queues[HCTX_TYPE_DEFAULT ] = +min (opts->nr_write_queues, nr_io_queues); + nr_io_queues += ctrl->io_queues[HCTX_TYPE_DEFAULT ]; + }else { + ctrl->io_queues[HCTX_TYPE_DEFAULT ] = nr_io_queues; + } + +if (opts->nr_poll_queues) { + ctrl->io_queues[HCTX_TYPE_POLL ] = +min (opts->nr_poll_queues,num_online_cpus ()); + nr_io_queues += ctrl->io_queues[HCTX_TYPE_POLL ]; + } ret =nvme_set_queue_count (&ctrl->ctrl, &nr_io_queues);if (ret)
@@ -1781,17 +1794,15 @@ static int nvme_rdma_map_queues(struct blk_mq_tag_set *set)struct nvme_rdma_ctrl *ctrl = set->driver_data; set->map[HCTX_TYPE_DEFAULT ].queue_offset =0 ; - set->map[HCTX_TYPE_READ ].nr_queues = ctrl->ctrl.opts->nr_io_queues; + set->map[HCTX_TYPE_DEFAULT ].nr_queues = + ctrl->io_queues[HCTX_TYPE_DEFAULT ]; + set->map[HCTX_TYPE_READ ].nr_queues = ctrl->io_queues[HCTX_TYPE_READ ];if (ctrl->ctrl.opts->nr_write_queues) {/* separate read/write queues */ - set->map[HCTX_TYPE_DEFAULT ].nr_queues = - ctrl->ctrl.opts->nr_write_queues; set->map[HCTX_TYPE_READ ].queue_offset = - ctrl->ctrl.opts->nr_write_queues; + ctrl->io_queues[HCTX_TYPE_DEFAULT ]; }else {/* mixed read/write queues */ - set->map[HCTX_TYPE_DEFAULT ].nr_queues = - ctrl->ctrl.opts->nr_io_queues; set->map[HCTX_TYPE_READ ].queue_offset =0 ; } blk_mq_rdma_map_queues(&set->map[HCTX_TYPE_DEFAULT],
@@ -1801,12 +1812,12 @@ static int nvme_rdma_map_queues(struct blk_mq_tag_set *set)if (ctrl->ctrl.opts->nr_poll_queues) { set->map[HCTX_TYPE_POLL ].nr_queues = - ctrl->ctrl.opts->nr_poll_queues; + ctrl->io_queues[HCTX_TYPE_POLL ]; set->map[HCTX_TYPE_POLL ].queue_offset = - ctrl->ctrl.opts->nr_io_queues; + ctrl->io_queues[HCTX_TYPE_DEFAULT ];if (ctrl->ctrl.opts->nr_write_queues) set->map[HCTX_TYPE_POLL ].queue_offset += - ctrl->ctrl.opts->nr_write_queues; + ctrl->io_queues[HCTX_TYPE_READ ];blk_mq_map_queues (&set->map[HCTX_TYPE_POLL ]); }return 0 ;
Signature · 1 line
2.17.1
Christoph Hellwig
maintainerPATCH 1/1 nvme-rdma: rework queue maps handling
Quoted 43 lines · Sagi Grimberg
If the device supports less queues than provided (if the device has less completion vectors), we might hit a bug due to the fact that we ignore that in
nvme_rdma_map_queues(we override the mapsnr_queueswith user opts).Instead, keep track of how many default/read/poll queues we actually allocated (rather than asked by the user) and use that to assign our queue mappings.
Fixes b65bb777ef22 (" nvme-rdma: support separate queue maps for read and write")Reported-by Saleem, Shiraz <shiraz.saleem at intel.com>Signed-off-by Sagi Grimberg <sagi at grimberg.me>--- drivers/nvme/host/rdma.c | 37 ++++++++++++++++++++++++------------- 1 file changed, 24 insertions(+), 13 deletions(-)drivers/nvme/host/rdma.c +3
-2@@ -119,6 +119,7 @@ struct nvme_rdma_ctrl {struct nvme_ctrl ctrl;bool use_inline_data; +u32 io_queues[HCTX_MAX_TYPES ]; };static inline struct nvme_rdma_ctrl *to_rdma_ctrl (struct nvme_ctrl *ctrl)@@ -165,8 +166,8 @@ static inline int nvme_rdma_queue_idx(struct nvme_rdma_queue *queue)static bool nvme_rdma_poll_queue (struct nvme_rdma_queue *queue) {return nvme_rdma_queue_idx (queue) > - queue->ctrl->ctrl.opts->nr_io_queues + - queue->ctrl->ctrl.opts->nr_write_queues; + queue->ctrl->io_queues[HCTX_TYPE_DEFAULT ] + + queue->ctrl->io_queues[HCTX_TYPE_READ ]; }static inline size_t nvme_rdma_inline_data_size (struct nvme_rdma_queue *queue)@@ -661,8 +662,20 @@ static int nvme_rdma_alloc_io_queues(struct nvme_rdma_ctrl *ctrl) nr_io_queues = min_t(unsigned int, nr_io_queues, ibdev->num_comp_vectors);
drivers/nvme/host/rdma.c static int nvme_rdma_alloc_io_queues(struct nvme_rdma_ctrl *ctrl) · line 672
Quoted 10 lines
drivers/nvme/host/rdma.c +8
-2- nr_io_queues +=min (opts->nr_write_queues,num_online_cpus ()); - nr_io_queues +=min (opts->nr_poll_queues,num_online_cpus ()); + ctrl->io_queues[HCTX_TYPE_READ ] = nr_io_queues; +if (opts->nr_write_queues) { + ctrl->io_queues[HCTX_TYPE_DEFAULT ] = +min (opts->nr_write_queues, nr_io_queues); + nr_io_queues += ctrl->io_queues[HCTX_TYPE_DEFAULT ]; + }else { + ctrl->io_queues[HCTX_TYPE_DEFAULT ] = nr_io_queues; + }
Nipick: I'd find this easier to read of the HCTX_TYPE_READ line was after the default one (I know, this is purely cosmetics).
Otherwise looks fine:
PATCH 1/1 nvme-rdma: rework queue maps handling
Quoted 7 lines
Nipick: I'd find this easier to read of the
HCTX_TYPE_READline was after the default one (I know, this is purely cosmetics).Otherwise looks fine:
Reviewed-by Christoph Hellwig <hch at lst.de>
Applied to nvme-5.0 with the cosmetic change. Thanks!
PATCH 1/1 nvme-rdma: rework queue maps handling
In reply to Sagi GrimbergQuoted 5 lines · Sagi Grimberg
If the device supports less queues than provided (if the device has less completion vectors), we might hit a bug due to the fact that we ignore that in
nvme_rdma_map_queues(we override the mapsnr_queueswith user opts).
Quoted 8 lines
Instead, keep track of how many default/read/poll queues we actually allocated (rather than asked by the user) and use that to assign our queue mappings.
Fixes b65bb777ef22 (" nvme-rdma: support separate queue maps for read and write")Reported-by Saleem, Shiraz <shiraz.saleem at intel.com>Signed-off-by Sagi Grimberg <sagi at grimberg.me>---
This patch fixes the reported problem on i40iw. Thanks!
Shiraz