From: Linus Torvalds Date: Sat, 9 Sep 2017 19:49:01 +0000 (-0700) Subject: Merge branch 'for-4.14/block-postmerge' of git://git.kernel.dk/linux-block X-Git-Tag: v4.14-rc1~81 X-Git-Url: https://asedeno.scripts.mit.edu/gitweb/?a=commitdiff_plain;h=126e76ffbf78d9e948b641aadb265d16c57f5a3d;p=linux.git Merge branch 'for-4.14/block-postmerge' of git://git.kernel.dk/linux-block Pull followup block layer updates from Jens Axboe: "I ended up splitting the main pull request for this series into two, mainly because of clashes between NVMe fixes that went into 4.13 after the for-4.14 branches were split off. This pull request is mostly NVMe, but not exclusively. In detail, it contains: - Two pull request for NVMe changes from Christoph. Nothing new on the feature front, basically just fixes all over the map for the core bits, transport, rdma, etc. - Series from Bart, cleaning up various bits in the BFQ scheduler. - Series of bcache fixes, which has been lingering for a release or two. Coly sent this in, but patches from various people in this area. - Set of patches for BFQ from Paolo himself, updating both documentation and fixing some corner cases in performance. - Series from Omar, attempting to now get the 4k loop support correct. Our confidence level is higher this time. - Series from Shaohua for loop as well, improving O_DIRECT performance and fixing a use-after-free" * 'for-4.14/block-postmerge' of git://git.kernel.dk/linux-block: (74 commits) bcache: initialize dirty stripes in flash_dev_run() loop: set physical block size to logical block size bcache: fix bch_hprint crash and improve output bcache: Update continue_at() documentation bcache: silence static checker warning bcache: fix for gc and write-back race bcache: increase the number of open buckets bcache: Correct return value for sysfs attach errors bcache: correct cache_dirty_target in __update_writeback_rate() bcache: gc does not work when triggering by manual command bcache: Don't reinvent the wheel but use existing llist API bcache: do not subtract sectors_to_gc for bypassed IO bcache: fix sequential large write IO bypass bcache: Fix leak of bdev reference block/loop: remove unused field block/loop: fix use after free bfq: Use icq_to_bic() consistently bfq: Suppress compiler warnings about comparisons bfq: Check kstrtoul() return value bfq: Declare local functions static ... --- 126e76ffbf78d9e948b641aadb265d16c57f5a3d diff --cc drivers/nvme/host/rdma.c index bf42d31484d4,6a7682620d87..58983000964b --- a/drivers/nvme/host/rdma.c +++ b/drivers/nvme/host/rdma.c @@@ -605,10 -629,9 +626,10 @@@ out_stop_queues return ret; } - static int nvme_rdma_init_io_queues(struct nvme_rdma_ctrl *ctrl) + static int nvme_rdma_alloc_io_queues(struct nvme_rdma_ctrl *ctrl) { struct nvmf_ctrl_options *opts = ctrl->ctrl.opts; + struct ib_device *ibdev = ctrl->device->dev; unsigned int nr_io_queues; int i, ret; @@@ -656,10 -734,144 +741,144 @@@ static void nvme_rdma_destroy_admin_que { nvme_rdma_free_qe(ctrl->queues[0].device->dev, &ctrl->async_event_sqe, sizeof(struct nvme_command), DMA_TO_DEVICE); - nvme_rdma_stop_and_free_queue(&ctrl->queues[0]); - blk_cleanup_queue(ctrl->ctrl.admin_q); - blk_mq_free_tag_set(&ctrl->admin_tag_set); - nvme_rdma_dev_put(ctrl->device); + nvme_rdma_stop_queue(&ctrl->queues[0]); + if (remove) { + blk_cleanup_queue(ctrl->ctrl.admin_q); + nvme_rdma_free_tagset(&ctrl->ctrl, true); + } + nvme_rdma_free_queue(&ctrl->queues[0]); + } + + static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl, + bool new) + { + int error; + + error = nvme_rdma_alloc_queue(ctrl, 0, NVME_AQ_DEPTH); + if (error) + return error; + + ctrl->device = ctrl->queues[0].device; + + ctrl->max_fr_pages = min_t(u32, NVME_RDMA_MAX_SEGMENTS, + ctrl->device->dev->attrs.max_fast_reg_page_list_len); + + if (new) { + ctrl->ctrl.admin_tagset = nvme_rdma_alloc_tagset(&ctrl->ctrl, true); + if (IS_ERR(ctrl->ctrl.admin_tagset)) + goto out_free_queue; + + ctrl->ctrl.admin_q = blk_mq_init_queue(&ctrl->admin_tag_set); + if (IS_ERR(ctrl->ctrl.admin_q)) { + error = PTR_ERR(ctrl->ctrl.admin_q); + goto out_free_tagset; + } + } else { + error = blk_mq_reinit_tagset(&ctrl->admin_tag_set, + nvme_rdma_reinit_request); + if (error) + goto out_free_queue; + } + + error = nvme_rdma_start_queue(ctrl, 0); + if (error) + goto out_cleanup_queue; + + error = ctrl->ctrl.ops->reg_read64(&ctrl->ctrl, NVME_REG_CAP, + &ctrl->ctrl.cap); + if (error) { + dev_err(ctrl->ctrl.device, + "prop_get NVME_REG_CAP failed\n"); + goto out_cleanup_queue; + } + + ctrl->ctrl.sqsize = + min_t(int, NVME_CAP_MQES(ctrl->ctrl.cap), ctrl->ctrl.sqsize); + + error = nvme_enable_ctrl(&ctrl->ctrl, ctrl->ctrl.cap); + if (error) + goto out_cleanup_queue; + + ctrl->ctrl.max_hw_sectors = - (ctrl->max_fr_pages - 1) << (PAGE_SHIFT - 9); ++ (ctrl->max_fr_pages - 1) << (ilog2(SZ_4K) - 9); + + error = nvme_init_identify(&ctrl->ctrl); + if (error) + goto out_cleanup_queue; + + error = nvme_rdma_alloc_qe(ctrl->queues[0].device->dev, + &ctrl->async_event_sqe, sizeof(struct nvme_command), + DMA_TO_DEVICE); + if (error) + goto out_cleanup_queue; + + return 0; + + out_cleanup_queue: + if (new) + blk_cleanup_queue(ctrl->ctrl.admin_q); + out_free_tagset: + if (new) + nvme_rdma_free_tagset(&ctrl->ctrl, true); + out_free_queue: + nvme_rdma_free_queue(&ctrl->queues[0]); + return error; + } + + static void nvme_rdma_destroy_io_queues(struct nvme_rdma_ctrl *ctrl, + bool remove) + { + nvme_rdma_stop_io_queues(ctrl); + if (remove) { + blk_cleanup_queue(ctrl->ctrl.connect_q); + nvme_rdma_free_tagset(&ctrl->ctrl, false); + } + nvme_rdma_free_io_queues(ctrl); + } + + static int nvme_rdma_configure_io_queues(struct nvme_rdma_ctrl *ctrl, bool new) + { + int ret; + + ret = nvme_rdma_alloc_io_queues(ctrl); + if (ret) + return ret; + + if (new) { + ctrl->ctrl.tagset = nvme_rdma_alloc_tagset(&ctrl->ctrl, false); + if (IS_ERR(ctrl->ctrl.tagset)) + goto out_free_io_queues; + + ctrl->ctrl.connect_q = blk_mq_init_queue(&ctrl->tag_set); + if (IS_ERR(ctrl->ctrl.connect_q)) { + ret = PTR_ERR(ctrl->ctrl.connect_q); + goto out_free_tag_set; + } + } else { + ret = blk_mq_reinit_tagset(&ctrl->tag_set, + nvme_rdma_reinit_request); + if (ret) + goto out_free_io_queues; + + blk_mq_update_nr_hw_queues(&ctrl->tag_set, + ctrl->ctrl.queue_count - 1); + } + + ret = nvme_rdma_start_io_queues(ctrl); + if (ret) + goto out_cleanup_connect_q; + + return 0; + + out_cleanup_connect_q: + if (new) + blk_cleanup_queue(ctrl->ctrl.connect_q); + out_free_tag_set: + if (new) + nvme_rdma_free_tagset(&ctrl->ctrl, false); + out_free_io_queues: + nvme_rdma_free_io_queues(ctrl); + return ret; } static void nvme_rdma_free_ctrl(struct nvme_ctrl *nctrl) @@@ -929,12 -1111,8 +1118,12 @@@ static int nvme_rdma_map_sg_fr(struct n struct nvme_keyed_sgl_desc *sg = &c->common.dptr.ksgl; int nr; - nr = ib_map_mr_sg(req->mr, req->sg_table.sgl, count, NULL, PAGE_SIZE); + /* + * Align the MR to a 4K page size to match the ctrl page size and + * the block virtual boundary. + */ + nr = ib_map_mr_sg(req->mr, req->sg_table.sgl, count, NULL, SZ_4K); - if (nr < count) { + if (unlikely(nr < count)) { if (nr < 0) return nr; return -EINVAL;