lightnvm: update bad block table format

The specification was changed to reflect a multi-value bad block table.
Instead of bit-based bad block table, the bad block table now allows
eight bad block categories. Currently four are defined:

 * Factory bad blocks
 * Grown bad blocks
 * Device-side reserved blocks
 * Host-side reserved blocks

The factory and grown bad blocks are the regular bad blocks. The
reserved blocks are either for internal use or external use. In
particular, the device-side reserved blocks allows the host to
bootstrap from a limited number of flash blocks. Reducing the flash
blocks to scan upon super block initialization.

Support for both get bad block table and set bad block table is added.

Signed-off-by: Matias Bjørling <m@bjorling.me>
Signed-off-by: Jens Axboe <axboe@fb.com>

1 files changed, 90 insertions, 23 deletions

diff --git a/drivers/nvme/host/lightnvm.c b/drivers/nvme/host/lightnvm.c
index e0b7b95813bc..2c3546516300 100644
--- a/drivers/nvme/host/lightnvm.c
+++ b/drivers/nvme/host/lightnvm.c
@@ -93,7 +93,7 @@ struct nvme_nvm_l2ptbl {
         __le16                  cdw14[6];
 };
 
-struct nvme_nvm_bbtbl {
+struct nvme_nvm_getbbtbl {
         __u8                    opcode;
         __u8                    flags;
         __u16                   command_id;
@@ -101,10 +101,23 @@ struct nvme_nvm_bbtbl {
         __u64                   rsvd[2];
         __le64                  prp1;
         __le64                  prp2;
-        __le32                  prp1_len;
-        __le32                  prp2_len;
-        __le32                  lbb;
-        __u32                   rsvd11[3];
+        __le64                  spba;
+        __u32                   rsvd4[4];
+};
+
+struct nvme_nvm_setbbtbl {
+        __u8                    opcode;
+        __u8                    flags;
+        __u16                   command_id;
+        __le32                  nsid;
+        __le64                  rsvd[2];
+        __le64                  prp1;
+        __le64                  prp2;
+        __le64                  spba;
+        __le16                  nlb;
+        __u8                    value;
+        __u8                    rsvd3;
+        __u32                   rsvd4[3];
 };
 
 struct nvme_nvm_erase_blk {
@@ -129,8 +142,8 @@ struct nvme_nvm_command {
                 struct nvme_nvm_hb_rw hb_rw;
                 struct nvme_nvm_ph_rw ph_rw;
                 struct nvme_nvm_l2ptbl l2p;
-                struct nvme_nvm_bbtbl get_bb;
-                struct nvme_nvm_bbtbl set_bb;
+                struct nvme_nvm_getbbtbl get_bb;
+                struct nvme_nvm_setbbtbl set_bb;
                 struct nvme_nvm_erase_blk erase;
         };
 };
@@ -187,6 +200,20 @@ struct nvme_nvm_id {
         struct nvme_nvm_id_group groups[4];
 } __packed;
 
+struct nvme_nvm_bb_tbl {
+        __u8    tblid[4];
+        __le16  verid;
+        __le16  revid;
+        __le32  rvsd1;
+        __le32  tblks;
+        __le32  tfact;
+        __le32  tgrown;
+        __le32  tdresv;
+        __le32  thresv;
+        __le32  rsvd2[8];
+        __u8    blk[0];
+};
+
 /*
  * Check we didn't inadvertently grow the command struct
  */
@@ -195,12 +222,14 @@ static inline void _nvme_nvm_check_size(void)
         BUILD_BUG_ON(sizeof(struct nvme_nvm_identity) != 64);
         BUILD_BUG_ON(sizeof(struct nvme_nvm_hb_rw) != 64);
         BUILD_BUG_ON(sizeof(struct nvme_nvm_ph_rw) != 64);
-        BUILD_BUG_ON(sizeof(struct nvme_nvm_bbtbl) != 64);
+        BUILD_BUG_ON(sizeof(struct nvme_nvm_getbbtbl) != 64);
+        BUILD_BUG_ON(sizeof(struct nvme_nvm_setbbtbl) != 64);
         BUILD_BUG_ON(sizeof(struct nvme_nvm_l2ptbl) != 64);
         BUILD_BUG_ON(sizeof(struct nvme_nvm_erase_blk) != 64);
         BUILD_BUG_ON(sizeof(struct nvme_nvm_id_group) != 960);
         BUILD_BUG_ON(sizeof(struct nvme_nvm_addr_format) != 128);
         BUILD_BUG_ON(sizeof(struct nvme_nvm_id) != 4096);
+        BUILD_BUG_ON(sizeof(struct nvme_nvm_bb_tbl) != 512);
 }
 
 static int init_grps(struct nvm_id *nvm_id, struct nvme_nvm_id *nvme_nvm_id)
@@ -322,43 +351,80 @@ out:
         return ret;
 }
 
-static int nvme_nvm_get_bb_tbl(struct request_queue *q, int lunid,
-                                unsigned int nr_blocks,
-                                nvm_bb_update_fn *update_bbtbl, void *priv)
+static int nvme_nvm_get_bb_tbl(struct request_queue *q, struct ppa_addr ppa,
+                                int nr_blocks, nvm_bb_update_fn *update_bbtbl,
+                                void *priv)
 {
         struct nvme_ns *ns = q->queuedata;
         struct nvme_dev *dev = ns->dev;
         struct nvme_nvm_command c = {};
-        void *bb_bitmap;
-        u16 bb_bitmap_size;
+        struct nvme_nvm_bb_tbl *bb_tbl;
+        int tblsz = sizeof(struct nvme_nvm_bb_tbl) + nr_blocks;
         int ret = 0;
 
         c.get_bb.opcode = nvme_nvm_admin_get_bb_tbl;
         c.get_bb.nsid = cpu_to_le32(ns->ns_id);
-        c.get_bb.lbb = cpu_to_le32(lunid);
-        bb_bitmap_size = ((nr_blocks >> 15) + 1) * PAGE_SIZE;
-        bb_bitmap = kmalloc(bb_bitmap_size, GFP_KERNEL);
-        if (!bb_bitmap)
-                return -ENOMEM;
+        c.get_bb.spba = cpu_to_le64(ppa.ppa);
 
-        bitmap_zero(bb_bitmap, nr_blocks);
+        bb_tbl = kzalloc(tblsz, GFP_KERNEL);
+        if (!bb_tbl)
+                return -ENOMEM;
 
-        ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, bb_bitmap,
-                                                                bb_bitmap_size);
+        ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, bb_tbl, tblsz);
         if (ret) {
                 dev_err(dev->dev, "get bad block table failed (%d)\n", ret);
                 ret = -EIO;
                 goto out;
         }
 
-        ret = update_bbtbl(lunid, bb_bitmap, nr_blocks, priv);
+        if (bb_tbl->tblid[0] != 'B' || bb_tbl->tblid[1] != 'B' ||
+                bb_tbl->tblid[2] != 'L' || bb_tbl->tblid[3] != 'T') {
+                dev_err(dev->dev, "bbt format mismatch\n");
+                ret = -EINVAL;
+                goto out;
+        }
+
+        if (le16_to_cpu(bb_tbl->verid) != 1) {
+                ret = -EINVAL;
+                dev_err(dev->dev, "bbt version not supported\n");
+                goto out;
+        }
+
+        if (le32_to_cpu(bb_tbl->tblks) != nr_blocks) {
+                ret = -EINVAL;
+                dev_err(dev->dev, "bbt unsuspected blocks returned (%u!=%u)",
+                                        le32_to_cpu(bb_tbl->tblks), nr_blocks);
+                goto out;
+        }
+
+        ret = update_bbtbl(ppa, nr_blocks, bb_tbl->blk, priv);
         if (ret) {
                 ret = -EINTR;
                 goto out;
         }
 
 out:
-        kfree(bb_bitmap);
+        kfree(bb_tbl);
+        return ret;
+}
+
+static int nvme_nvm_set_bb_tbl(struct request_queue *q, struct nvm_rq *rqd,
+                                                                int type)
+{
+        struct nvme_ns *ns = q->queuedata;
+        struct nvme_dev *dev = ns->dev;
+        struct nvme_nvm_command c = {};
+        int ret = 0;
+
+        c.set_bb.opcode = nvme_nvm_admin_set_bb_tbl;
+        c.set_bb.nsid = cpu_to_le32(ns->ns_id);
+        c.set_bb.spba = cpu_to_le64(rqd->ppa_addr.ppa);
+        c.set_bb.nlb = cpu_to_le16(rqd->nr_pages - 1);
+        c.set_bb.value = type;
+
+        ret = nvme_submit_sync_cmd(q, (struct nvme_command *)&c, NULL, 0);
+        if (ret)
+                dev_err(dev->dev, "set bad block table failed (%d)\n", ret);
         return ret;
 }
 
@@ -474,6 +540,7 @@ static struct nvm_dev_ops nvme_nvm_dev_ops = {
         .get_l2p_tbl            = nvme_nvm_get_l2p_tbl,
 
         .get_bb_tbl             = nvme_nvm_get_bb_tbl,
+        .set_bb_tbl             = nvme_nvm_set_bb_tbl,
 
         .submit_io              = nvme_nvm_submit_io,
         .erase_block            = nvme_nvm_erase_block,