NVMe: Retry failed commands with non-fatal errors

For commands returned with failed status, queue these for resubmission
and continue retrying them until success or for a limited amount of
time. The final timeout was arbitrarily chosen so requests can't be
retried indefinitely.

Since these are requeued on the nvmeq that submitted the command, the
callbacks have to take an nvmeq instead of an nvme_dev as a parameter
so that we can use the locked queue to append the iod to retry later.

The nvme_iod conviently can be used to track how long we've been trying
to successfully complete an iod request. The nvme_iod also provides the
nvme prp dma mappings, so I had to move a few things around so we can
keep those mappings.

Signed-off-by: Keith Busch <keith.busch@intel.com>
[fixed checkpatch issue with long line]
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>

1 files changed, 6 insertions, 4 deletions

diff --git a/drivers/block/nvme-scsi.c b/drivers/block/nvme-scsi.c
index 111c920c1574..2c3f5be06da1 100644
--- a/drivers/block/nvme-scsi.c
+++ b/drivers/block/nvme-scsi.c
@@ -1562,13 +1562,14 @@ static int nvme_trans_send_fw_cmd(struct nvme_ns *ns, struct sg_io_hdr *hdr,
                         res = PTR_ERR(iod);
                         goto out;
                 }
-                length = nvme_setup_prps(dev, &c.common, iod, tot_len,
-                                                                GFP_KERNEL);
+                length = nvme_setup_prps(dev, iod, tot_len, GFP_KERNEL);
                 if (length != tot_len) {
                         res = -ENOMEM;
                         goto out_unmap;
                 }
 
+                c.dlfw.prp1 = cpu_to_le64(sg_dma_address(iod->sg));
+                c.dlfw.prp2 = cpu_to_le64(iod->first_dma);
                 c.dlfw.numd = cpu_to_le32((tot_len/BYTES_TO_DWORDS) - 1);
                 c.dlfw.offset = cpu_to_le32(offset/BYTES_TO_DWORDS);
         } else if (opcode == nvme_admin_activate_fw) {
@@ -2092,8 +2093,7 @@ static int nvme_trans_do_nvme_io(struct nvme_ns *ns, struct sg_io_hdr *hdr,
                         res = PTR_ERR(iod);
                         goto out;
                 }
-                retcode = nvme_setup_prps(dev, &c.common, iod, unit_len,
-                                                        GFP_KERNEL);
+                retcode = nvme_setup_prps(dev, iod, unit_len, GFP_KERNEL);
                 if (retcode != unit_len) {
                         nvme_unmap_user_pages(dev,
                                 (is_write) ? DMA_TO_DEVICE : DMA_FROM_DEVICE,
@@ -2102,6 +2102,8 @@ static int nvme_trans_do_nvme_io(struct nvme_ns *ns, struct sg_io_hdr *hdr,
                         res = -ENOMEM;
                         goto out;
                 }
+                c.rw.prp1 = cpu_to_le64(sg_dma_address(iod->sg));
+                c.rw.prp2 = cpu_to_le64(iod->first_dma);
 
                 nvme_offset += unit_num_blocks;