mmc: support packed write command for eMMC4.5 devices

This patch supports packed write command of eMMC4.5 devices.  Several
writes can be grouped in packed command and all data of the individual
commands can be sent in a single transfer on the bus. Large amounts of
data in one transfer rather than several data of small size are
effective for eMMC write internally.  As a result, packed command help
write throughput be improved.  The following tables show the results
of packed write.

Type A:
test     none |  packed
iozone   25.8 |  31
tiotest  27.6 |  31.2
lmdd     31.2 |  35.4

Type B:
test     none |  packed
iozone   44.1 |  51.1
tiotest  47.9 |  52.5
lmdd     51.6 |  59.2

Type C:
test     none |  packed
iozone   19.5 |  32
tiotest  19.9 |  34.5
lmdd     22.8 |  40.7

Signed-off-by: Seungwon Jeon <tgih.jun@samsung.com>
Reviewed-by: Maya Erez <merez@codeaurora.org>
Reviewed-by: Namjae Jeon <linkinjeon@gmail.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

4 files changed, 557 insertions, 17 deletions

diff --git a/drivers/mmc/card/block.c b/drivers/mmc/card/block.c
index 1170afe1a59..5bab73b91c2 100644
--- a/drivers/mmc/card/block.c
+++ b/drivers/mmc/card/block.c
@@ -59,6 +59,12 @@ MODULE_ALIAS("mmc:block");
 #define INAND_CMD38_ARG_SECTRIM2 0x88
 #define MMC_BLK_TIMEOUT_MS  (10 * 60 * 1000)        /* 10 minute timeout */
 
+#define mmc_req_rel_wr(req)	(((req->cmd_flags & REQ_FUA) || \
+				  (req->cmd_flags & REQ_META)) && \
+				  (rq_data_dir(req) == WRITE))
+#define PACKED_CMD_VER	0x01
+#define PACKED_CMD_WR	0x02
+
 static DEFINE_MUTEX(block_mutex);
 
 /*
@@ -89,6 +95,7 @@ struct mmc_blk_data {
 	unsigned int	flags;
 #define MMC_BLK_CMD23	(1 << 0)	/* Can do SET_BLOCK_COUNT for multiblock */
 #define MMC_BLK_REL_WR	(1 << 1)	/* MMC Reliable write support */
+#define MMC_BLK_PACKED_CMD	(1 << 2)	/* MMC packed command support */
 
 	unsigned int	usage;
 	unsigned int	read_only;
@@ -113,6 +120,12 @@ struct mmc_blk_data {
 
 static DEFINE_MUTEX(open_lock);
 
+enum {
+	MMC_PACKED_NR_IDX = -1,
+	MMC_PACKED_NR_ZERO,
+	MMC_PACKED_NR_SINGLE,
+};
+
 module_param(perdev_minors, int, 0444);
 MODULE_PARM_DESC(perdev_minors, "Minors numbers to allocate per device");
 
@@ -120,6 +133,19 @@ static inline int mmc_blk_part_switch(struct mmc_card *card,
 				      struct mmc_blk_data *md);
 static int get_card_status(struct mmc_card *card, u32 *status, int retries);
 
+static inline void mmc_blk_clear_packed(struct mmc_queue_req *mqrq)
+{
+	struct mmc_packed *packed = mqrq->packed;
+
+	BUG_ON(!packed);
+
+	mqrq->cmd_type = MMC_PACKED_NONE;
+	packed->nr_entries = MMC_PACKED_NR_ZERO;
+	packed->idx_failure = MMC_PACKED_NR_IDX;
+	packed->retries = 0;
+	packed->blocks = 0;
+}
+
 static struct mmc_blk_data *mmc_blk_get(struct gendisk *disk)
 {
 	struct mmc_blk_data *md;
@@ -1137,12 +1163,78 @@ static int mmc_blk_err_check(struct mmc_card *card,
 	if (!brq->data.bytes_xfered)
 		return MMC_BLK_RETRY;
 
+	if (mmc_packed_cmd(mq_mrq->cmd_type)) {
+		if (unlikely(brq->data.blocks << 9 != brq->data.bytes_xfered))
+			return MMC_BLK_PARTIAL;
+		else
+			return MMC_BLK_SUCCESS;
+	}
+
 	if (blk_rq_bytes(req) != brq->data.bytes_xfered)
 		return MMC_BLK_PARTIAL;
 
 	return MMC_BLK_SUCCESS;
 }
 
+static int mmc_blk_packed_err_check(struct mmc_card *card,
+				    struct mmc_async_req *areq)
+{
+	struct mmc_queue_req *mq_rq = container_of(areq, struct mmc_queue_req,
+			mmc_active);
+	struct request *req = mq_rq->req;
+	struct mmc_packed *packed = mq_rq->packed;
+	int err, check, status;
+	u8 *ext_csd;
+
+	BUG_ON(!packed);
+
+	packed->retries--;
+	check = mmc_blk_err_check(card, areq);
+	err = get_card_status(card, &status, 0);
+	if (err) {
+		pr_err("%s: error %d sending status command\n",
+		       req->rq_disk->disk_name, err);
+		return MMC_BLK_ABORT;
+	}
+
+	if (status & R1_EXCEPTION_EVENT) {
+		ext_csd = kzalloc(512, GFP_KERNEL);
+		if (!ext_csd) {
+			pr_err("%s: unable to allocate buffer for ext_csd\n",
+			       req->rq_disk->disk_name);
+			return -ENOMEM;
+		}
+
+		err = mmc_send_ext_csd(card, ext_csd);
+		if (err) {
+			pr_err("%s: error %d sending ext_csd\n",
+			       req->rq_disk->disk_name, err);
+			check = MMC_BLK_ABORT;
+			goto free;
+		}
+
+		if ((ext_csd[EXT_CSD_EXP_EVENTS_STATUS] &
+		     EXT_CSD_PACKED_FAILURE) &&
+		    (ext_csd[EXT_CSD_PACKED_CMD_STATUS] &
+		     EXT_CSD_PACKED_GENERIC_ERROR)) {
+			if (ext_csd[EXT_CSD_PACKED_CMD_STATUS] &
+			    EXT_CSD_PACKED_INDEXED_ERROR) {
+				packed->idx_failure =
+				  ext_csd[EXT_CSD_PACKED_FAILURE_INDEX] - 1;
+				check = MMC_BLK_PARTIAL;
+			}
+			pr_err("%s: packed cmd failed, nr %u, sectors %u, "
+			       "failure index: %d\n",
+			       req->rq_disk->disk_name, packed->nr_entries,
+			       packed->blocks, packed->idx_failure);
+		}
+free:
+		kfree(ext_csd);
+	}
+
+	return check;
+}
+
 static void mmc_blk_rw_rq_prep(struct mmc_queue_req *mqrq,
 			       struct mmc_card *card,
 			       int disable_multi,
@@ -1297,10 +1389,221 @@ static void mmc_blk_rw_rq_prep(struct mmc_queue_req *mqrq,
 	mmc_queue_bounce_pre(mqrq);
 }
 
+static inline u8 mmc_calc_packed_hdr_segs(struct request_queue *q,
+					  struct mmc_card *card)
+{
+	unsigned int hdr_sz = mmc_large_sector(card) ? 4096 : 512;
+	unsigned int max_seg_sz = queue_max_segment_size(q);
+	unsigned int len, nr_segs = 0;
+
+	do {
+		len = min(hdr_sz, max_seg_sz);
+		hdr_sz -= len;
+		nr_segs++;
+	} while (hdr_sz);
+
+	return nr_segs;
+}
+
+static u8 mmc_blk_prep_packed_list(struct mmc_queue *mq, struct request *req)
+{
+	struct request_queue *q = mq->queue;
+	struct mmc_card *card = mq->card;
+	struct request *cur = req, *next = NULL;
+	struct mmc_blk_data *md = mq->data;
+	struct mmc_queue_req *mqrq = mq->mqrq_cur;
+	bool en_rel_wr = card->ext_csd.rel_param & EXT_CSD_WR_REL_PARAM_EN;
+	unsigned int req_sectors = 0, phys_segments = 0;
+	unsigned int max_blk_count, max_phys_segs;
+	bool put_back = true;
+	u8 max_packed_rw = 0;
+	u8 reqs = 0;
+
+	if (!(md->flags & MMC_BLK_PACKED_CMD))
+		goto no_packed;
+
+	if ((rq_data_dir(cur) == WRITE) &&
+	    mmc_host_packed_wr(card->host))
+		max_packed_rw = card->ext_csd.max_packed_writes;
+
+	if (max_packed_rw == 0)
+		goto no_packed;
+
+	if (mmc_req_rel_wr(cur) &&
+	    (md->flags & MMC_BLK_REL_WR) && !en_rel_wr)
+		goto no_packed;
+
+	if (mmc_large_sector(card) &&
+	    !IS_ALIGNED(blk_rq_sectors(cur), 8))
+		goto no_packed;
+
+	mmc_blk_clear_packed(mqrq);
+
+	max_blk_count = min(card->host->max_blk_count,
+			    card->host->max_req_size >> 9);
+	if (unlikely(max_blk_count > 0xffff))
+		max_blk_count = 0xffff;
+
+	max_phys_segs = queue_max_segments(q);
+	req_sectors += blk_rq_sectors(cur);
+	phys_segments += cur->nr_phys_segments;
+
+	if (rq_data_dir(cur) == WRITE) {
+		req_sectors += mmc_large_sector(card) ? 8 : 1;
+		phys_segments += mmc_calc_packed_hdr_segs(q, card);
+	}
+
+	do {
+		if (reqs >= max_packed_rw - 1) {
+			put_back = false;
+			break;
+		}
+
+		spin_lock_irq(q->queue_lock);
+		next = blk_fetch_request(q);
+		spin_unlock_irq(q->queue_lock);
+		if (!next) {
+			put_back = false;
+			break;
+		}
+
+		if (mmc_large_sector(card) &&
+		    !IS_ALIGNED(blk_rq_sectors(next), 8))
+			break;
+
+		if (next->cmd_flags & REQ_DISCARD ||
+		    next->cmd_flags & REQ_FLUSH)
+			break;
+
+		if (rq_data_dir(cur) != rq_data_dir(next))
+			break;
+
+		if (mmc_req_rel_wr(next) &&
+		    (md->flags & MMC_BLK_REL_WR) && !en_rel_wr)
+			break;
+
+		req_sectors += blk_rq_sectors(next);
+		if (req_sectors > max_blk_count)
+			break;
+
+		phys_segments +=  next->nr_phys_segments;
+		if (phys_segments > max_phys_segs)
+			break;
+
+		list_add_tail(&next->queuelist, &mqrq->packed->list);
+		cur = next;
+		reqs++;
+	} while (1);
+
+	if (put_back) {
+		spin_lock_irq(q->queue_lock);
+		blk_requeue_request(q, next);
+		spin_unlock_irq(q->queue_lock);
+	}
+
+	if (reqs > 0) {
+		list_add(&req->queuelist, &mqrq->packed->list);
+		mqrq->packed->nr_entries = ++reqs;
+		mqrq->packed->retries = reqs;
+		return reqs;
+	}
+
+no_packed:
+	mqrq->cmd_type = MMC_PACKED_NONE;
+	return 0;
+}
+
+static void mmc_blk_packed_hdr_wrq_prep(struct mmc_queue_req *mqrq,
+					struct mmc_card *card,
+					struct mmc_queue *mq)
+{
+	struct mmc_blk_request *brq = &mqrq->brq;
+	struct request *req = mqrq->req;
+	struct request *prq;
+	struct mmc_blk_data *md = mq->data;
+	struct mmc_packed *packed = mqrq->packed;
+	bool do_rel_wr, do_data_tag;
+	u32 *packed_cmd_hdr;
+	u8 hdr_blocks;
+	u8 i = 1;
+
+	BUG_ON(!packed);
+
+	mqrq->cmd_type = MMC_PACKED_WRITE;
+	packed->blocks = 0;
+	packed->idx_failure = MMC_PACKED_NR_IDX;
+
+	packed_cmd_hdr = packed->cmd_hdr;
+	memset(packed_cmd_hdr, 0, sizeof(packed->cmd_hdr));
+	packed_cmd_hdr[0] = (packed->nr_entries << 16) |
+		(PACKED_CMD_WR << 8) | PACKED_CMD_VER;
+	hdr_blocks = mmc_large_sector(card) ? 8 : 1;
+
+	/*
+	 * Argument for each entry of packed group
+	 */
+	list_for_each_entry(prq, &packed->list, queuelist) {
+		do_rel_wr = mmc_req_rel_wr(prq) && (md->flags & MMC_BLK_REL_WR);
+		do_data_tag = (card->ext_csd.data_tag_unit_size) &&
+			(prq->cmd_flags & REQ_META) &&
+			(rq_data_dir(prq) == WRITE) &&
+			((brq->data.blocks * brq->data.blksz) >=
+			 card->ext_csd.data_tag_unit_size);
+		/* Argument of CMD23 */
+		packed_cmd_hdr[(i * 2)] =
+			(do_rel_wr ? MMC_CMD23_ARG_REL_WR : 0) |
+			(do_data_tag ? MMC_CMD23_ARG_TAG_REQ : 0) |
+			blk_rq_sectors(prq);
+		/* Argument of CMD18 or CMD25 */
+		packed_cmd_hdr[((i * 2)) + 1] =
+			mmc_card_blockaddr(card) ?
+			blk_rq_pos(prq) : blk_rq_pos(prq) << 9;
+		packed->blocks += blk_rq_sectors(prq);
+		i++;
+	}
+
+	memset(brq, 0, sizeof(struct mmc_blk_request));
+	brq->mrq.cmd = &brq->cmd;
+	brq->mrq.data = &brq->data;
+	brq->mrq.sbc = &brq->sbc;
+	brq->mrq.stop = &brq->stop;
+
+	brq->sbc.opcode = MMC_SET_BLOCK_COUNT;
+	brq->sbc.arg = MMC_CMD23_ARG_PACKED | (packed->blocks + hdr_blocks);
+	brq->sbc.flags = MMC_RSP_R1 | MMC_CMD_AC;
+
+	brq->cmd.opcode = MMC_WRITE_MULTIPLE_BLOCK;
+	brq->cmd.arg = blk_rq_pos(req);
+	if (!mmc_card_blockaddr(card))
+		brq->cmd.arg <<= 9;
+	brq->cmd.flags = MMC_RSP_SPI_R1 | MMC_RSP_R1 | MMC_CMD_ADTC;
+
+	brq->data.blksz = 512;
+	brq->data.blocks = packed->blocks + hdr_blocks;
+	brq->data.flags |= MMC_DATA_WRITE;
+
+	brq->stop.opcode = MMC_STOP_TRANSMISSION;
+	brq->stop.arg = 0;
+	brq->stop.flags = MMC_RSP_SPI_R1B | MMC_RSP_R1B | MMC_CMD_AC;
+
+	mmc_set_data_timeout(&brq->data, card);
+
+	brq->data.sg = mqrq->sg;
+	brq->data.sg_len = mmc_queue_map_sg(mq, mqrq);
+
+	mqrq->mmc_active.mrq = &brq->mrq;
+	mqrq->mmc_active.err_check = mmc_blk_packed_err_check;
+
+	mmc_queue_bounce_pre(mqrq);
+}
+
 static int mmc_blk_cmd_err(struct mmc_blk_data *md, struct mmc_card *card,
 			   struct mmc_blk_request *brq, struct request *req,
 			   int ret)
 {
+	struct mmc_queue_req *mq_rq;
+	mq_rq = container_of(brq, struct mmc_queue_req, brq);
+
 	/*
 	 * If this is an SD card and we're writing, we can first
 	 * mark the known good sectors as ok.
@@ -1317,11 +1620,84 @@ static int mmc_blk_cmd_err(struct mmc_blk_data *md, struct mmc_card *card,
 			ret = blk_end_request(req, 0, blocks << 9);
 		}
 	} else {
-		ret = blk_end_request(req, 0, brq->data.bytes_xfered);
+		if (!mmc_packed_cmd(mq_rq->cmd_type))
+			ret = blk_end_request(req, 0, brq->data.bytes_xfered);
 	}
 	return ret;
 }
 
+static int mmc_blk_end_packed_req(struct mmc_queue_req *mq_rq)
+{
+	struct request *prq;
+	struct mmc_packed *packed = mq_rq->packed;
+	int idx = packed->idx_failure, i = 0;
+	int ret = 0;
+
+	BUG_ON(!packed);
+
+	while (!list_empty(&packed->list)) {
+		prq = list_entry_rq(packed->list.next);
+		if (idx == i) {
+			/* retry from error index */
+			packed->nr_entries -= idx;
+			mq_rq->req = prq;
+			ret = 1;
+
+			if (packed->nr_entries == MMC_PACKED_NR_SINGLE) {
+				list_del_init(&prq->queuelist);
+				mmc_blk_clear_packed(mq_rq);
+			}
+			return ret;
+		}
+		list_del_init(&prq->queuelist);
+		blk_end_request(prq, 0, blk_rq_bytes(prq));
+		i++;
+	}
+
+	mmc_blk_clear_packed(mq_rq);
+	return ret;
+}
+
+static void mmc_blk_abort_packed_req(struct mmc_queue_req *mq_rq)
+{
+	struct request *prq;
+	struct mmc_packed *packed = mq_rq->packed;
+
+	BUG_ON(!packed);
+
+	while (!list_empty(&packed->list)) {
+		prq = list_entry_rq(packed->list.next);
+		list_del_init(&prq->queuelist);
+		blk_end_request(prq, -EIO, blk_rq_bytes(prq));
+	}
+
+	mmc_blk_clear_packed(mq_rq);
+}
+
+static void mmc_blk_revert_packed_req(struct mmc_queue *mq,
+				      struct mmc_queue_req *mq_rq)
+{
+	struct request *prq;
+	struct request_queue *q = mq->queue;
+	struct mmc_packed *packed = mq_rq->packed;
+
+	BUG_ON(!packed);
+
+	while (!list_empty(&packed->list)) {
+		prq = list_entry_rq(packed->list.prev);
+		if (prq->queuelist.prev != &packed->list) {
+			list_del_init(&prq->queuelist);
+			spin_lock_irq(q->queue_lock);
+			blk_requeue_request(mq->queue, prq);
+			spin_unlock_irq(q->queue_lock);
+		} else {
+			list_del_init(&prq->queuelist);
+		}
+	}
+
+	mmc_blk_clear_packed(mq_rq);
+}
+
 static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *rqc)
 {
 	struct mmc_blk_data *md = mq->data;
@@ -1332,10 +1708,15 @@ static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *rqc)
 	struct mmc_queue_req *mq_rq;
 	struct request *req = rqc;
 	struct mmc_async_req *areq;
+	const u8 packed_nr = 2;
+	u8 reqs = 0;
 
 	if (!rqc && !mq->mqrq_prev->req)
 		return 0;
 
+	if (rqc)
+		reqs = mmc_blk_prep_packed_list(mq, rqc);
+
 	do {
 		if (rqc) {
 			/*
@@ -1346,9 +1727,15 @@ static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *rqc)
 			    (card->ext_csd.data_sector_size == 4096)) {
 				pr_err("%s: Transfer size is not 4KB sector size aligned\n",
 					req->rq_disk->disk_name);
+				mq_rq = mq->mqrq_cur;
 				goto cmd_abort;
 			}
-			mmc_blk_rw_rq_prep(mq->mqrq_cur, card, 0, mq);
+
+			if (reqs >= packed_nr)
+				mmc_blk_packed_hdr_wrq_prep(mq->mqrq_cur,
+							    card, mq);
+			else
+				mmc_blk_rw_rq_prep(mq->mqrq_cur, card, 0, mq);
 			areq = &mq->mqrq_cur->mmc_active;
 		} else
 			areq = NULL;
@@ -1372,8 +1759,15 @@ static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *rqc)
 			 * A block was successfully transferred.
 			 */
 			mmc_blk_reset_success(md, type);
-			ret = blk_end_request(req, 0,
+
+			if (mmc_packed_cmd(mq_rq->cmd_type)) {
+				ret = mmc_blk_end_packed_req(mq_rq);
+				break;
+			} else {
+				ret = blk_end_request(req, 0,
 						brq->data.bytes_xfered);
+			}
+
 			/*
 			 * If the blk_end_request function returns non-zero even
 			 * though all data has been transferred and no errors
@@ -1406,7 +1800,8 @@ static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *rqc)
 			err = mmc_blk_reset(md, card->host, type);
 			if (!err)
 				break;
-			if (err == -ENODEV)
+			if (err == -ENODEV ||
+				mmc_packed_cmd(mq_rq->cmd_type))
 				goto cmd_abort;
 			/* Fall through */
 		}
@@ -1437,22 +1832,38 @@ static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *rqc)
 		}
 
 		if (ret) {
-			/*
-			 * In case of a incomplete request
-			 * prepare it again and resend.
-			 */
-			mmc_blk_rw_rq_prep(mq_rq, card, disable_multi, mq);
-			mmc_start_req(card->host, &mq_rq->mmc_active, NULL);
+			if (mmc_packed_cmd(mq_rq->cmd_type)) {
+				if (!mq_rq->packed->retries)
+					goto cmd_abort;
+				mmc_blk_packed_hdr_wrq_prep(mq_rq, card, mq);
+				mmc_start_req(card->host,
+					      &mq_rq->mmc_active, NULL);
+			} else {
+
+				/*
+				 * In case of a incomplete request
+				 * prepare it again and resend.
+				 */
+				mmc_blk_rw_rq_prep(mq_rq, card,
+						disable_multi, mq);
+				mmc_start_req(card->host,
+						&mq_rq->mmc_active, NULL);
+			}
 		}
 	} while (ret);
 
 	return 1;
 
  cmd_abort:
-	if (mmc_card_removed(card))
-		req->cmd_flags |= REQ_QUIET;
-	while (ret)
-		ret = blk_end_request(req, -EIO, blk_rq_cur_bytes(req));
+	if (mmc_packed_cmd(mq_rq->cmd_type)) {
+		mmc_blk_abort_packed_req(mq_rq);
+	} else {
+		if (mmc_card_removed(card))
+			req->cmd_flags |= REQ_QUIET;
+		while (ret)
+			ret = blk_end_request(req, -EIO,
+					blk_rq_cur_bytes(req));
+	}
 
  start_new_req:
 	if (rqc) {
@@ -1460,6 +1871,12 @@ static int mmc_blk_issue_rw_rq(struct mmc_queue *mq, struct request *rqc)
 			rqc->cmd_flags |= REQ_QUIET;
 			blk_end_request_all(rqc, -EIO);
 		} else {
+			/*
+			 * If current request is packed, it needs to put back.
+			 */
+			if (mmc_packed_cmd(mq->mqrq_cur->cmd_type))
+				mmc_blk_revert_packed_req(mq, mq->mqrq_cur);
+
 			mmc_blk_rw_rq_prep(mq->mqrq_cur, card, 0, mq);
 			mmc_start_req(card->host,
 				      &mq->mqrq_cur->mmc_active, NULL);
@@ -1634,6 +2051,14 @@ static struct mmc_blk_data *mmc_blk_alloc_req(struct mmc_card *card,
 		blk_queue_flush(md->queue.queue, REQ_FLUSH | REQ_FUA);
 	}
 
+	if (mmc_card_mmc(card) &&
+	    (area_type == MMC_BLK_DATA_AREA_MAIN) &&
+	    (md->flags & MMC_BLK_CMD23) &&
+	    card->ext_csd.packed_event_en) {
+		if (!mmc_packed_init(&md->queue, card))
+			md->flags |= MMC_BLK_PACKED_CMD;
+	}
+
 	return md;
 
  err_putdisk:
@@ -1742,6 +2167,8 @@ static void mmc_blk_remove_req(struct mmc_blk_data *md)
 
 		/* Then flush out any already in there */
 		mmc_cleanup_queue(&md->queue);
+		if (md->flags & MMC_BLK_PACKED_CMD)
+			mmc_packed_clean(&md->queue);
 		mmc_blk_put(md);
 	}
 }
diff --git a/drivers/mmc/card/queue.c b/drivers/mmc/card/queue.c
index 5e0971016ac..fa4e44ee796 100644
--- a/drivers/mmc/card/queue.c
+++ b/drivers/mmc/card/queue.c
@@ -362,6 +362,49 @@ void mmc_cleanup_queue(struct mmc_queue *mq)
 }
 EXPORT_SYMBOL(mmc_cleanup_queue);
 
+int mmc_packed_init(struct mmc_queue *mq, struct mmc_card *card)
+{
+	struct mmc_queue_req *mqrq_cur = &mq->mqrq[0];
+	struct mmc_queue_req *mqrq_prev = &mq->mqrq[1];
+	int ret = 0;
+
+
+	mqrq_cur->packed = kzalloc(sizeof(struct mmc_packed), GFP_KERNEL);
+	if (!mqrq_cur->packed) {
+		pr_warn("%s: unable to allocate packed cmd for mqrq_cur\n",
+			mmc_card_name(card));
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	mqrq_prev->packed = kzalloc(sizeof(struct mmc_packed), GFP_KERNEL);
+	if (!mqrq_prev->packed) {
+		pr_warn("%s: unable to allocate packed cmd for mqrq_prev\n",
+			mmc_card_name(card));
+		kfree(mqrq_cur->packed);
+		mqrq_cur->packed = NULL;
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	INIT_LIST_HEAD(&mqrq_cur->packed->list);
+	INIT_LIST_HEAD(&mqrq_prev->packed->list);
+
+out:
+	return ret;
+}
+
+void mmc_packed_clean(struct mmc_queue *mq)
+{
+	struct mmc_queue_req *mqrq_cur = &mq->mqrq[0];
+	struct mmc_queue_req *mqrq_prev = &mq->mqrq[1];
+
+	kfree(mqrq_cur->packed);
+	mqrq_cur->packed = NULL;
+	kfree(mqrq_prev->packed);
+	mqrq_prev->packed = NULL;
+}
+
 /**
  * mmc_queue_suspend - suspend a MMC request queue
  * @mq: MMC queue to suspend
@@ -406,6 +449,41 @@ void mmc_queue_resume(struct mmc_queue *mq)
 	}
 }
 
+static unsigned int mmc_queue_packed_map_sg(struct mmc_queue *mq,
+					    struct mmc_packed *packed,
+					    struct scatterlist *sg,
+					    enum mmc_packed_type cmd_type)
+{
+	struct scatterlist *__sg = sg;
+	unsigned int sg_len = 0;
+	struct request *req;
+
+	if (mmc_packed_wr(cmd_type)) {
+		unsigned int hdr_sz = mmc_large_sector(mq->card) ? 4096 : 512;
+		unsigned int max_seg_sz = queue_max_segment_size(mq->queue);
+		unsigned int len, remain, offset = 0;
+		u8 *buf = (u8 *)packed->cmd_hdr;
+
+		remain = hdr_sz;
+		do {
+			len = min(remain, max_seg_sz);
+			sg_set_buf(__sg, buf + offset, len);
+			offset += len;
+			remain -= len;
+			(__sg++)->page_link &= ~0x02;
+			sg_len++;
+		} while (remain);
+	}
+
+	list_for_each_entry(req, &packed->list, queuelist) {
+		sg_len += blk_rq_map_sg(mq->queue, req, __sg);
+		__sg = sg + (sg_len - 1);
+		(__sg++)->page_link &= ~0x02;
+	}
+	sg_mark_end(sg + (sg_len - 1));
+	return sg_len;
+}
+
 /*
  * Prepare the sg list(s) to be handed of to the host driver
  */
@@ -414,14 +492,26 @@ unsigned int mmc_queue_map_sg(struct mmc_queue *mq, struct mmc_queue_req *mqrq)
 	unsigned int sg_len;
 	size_t buflen;
 	struct scatterlist *sg;
+	enum mmc_packed_type cmd_type;
 	int i;
 
-	if (!mqrq->bounce_buf)
-		return blk_rq_map_sg(mq->queue, mqrq->req, mqrq->sg);
+	cmd_type = mqrq->cmd_type;
+
+	if (!mqrq->bounce_buf) {
+		if (mmc_packed_cmd(cmd_type))
+			return mmc_queue_packed_map_sg(mq, mqrq->packed,
+						       mqrq->sg, cmd_type);
+		else
+			return blk_rq_map_sg(mq->queue, mqrq->req, mqrq->sg);
+	}
 
 	BUG_ON(!mqrq->bounce_sg);
 
-	sg_len = blk_rq_map_sg(mq->queue, mqrq->req, mqrq->bounce_sg);
+	if (mmc_packed_cmd(cmd_type))
+		sg_len = mmc_queue_packed_map_sg(mq, mqrq->packed,
+						 mqrq->bounce_sg, cmd_type);
+	else
+		sg_len = blk_rq_map_sg(mq->queue, mqrq->req, mqrq->bounce_sg);
 
 	mqrq->bounce_sg_len = sg_len;
 
diff --git a/drivers/mmc/card/queue.h b/drivers/mmc/card/queue.h
index e20c27b2b8b..031bf6376c9 100644
--- a/drivers/mmc/card/queue.h
+++ b/drivers/mmc/card/queue.h
@@ -12,6 +12,23 @@ struct mmc_blk_request {
 	struct mmc_data		data;
 };
 
+enum mmc_packed_type {
+	MMC_PACKED_NONE = 0,
+	MMC_PACKED_WRITE,
+};
+
+#define mmc_packed_cmd(type)	((type) != MMC_PACKED_NONE)
+#define mmc_packed_wr(type)	((type) == MMC_PACKED_WRITE)
+
+struct mmc_packed {
+	struct list_head	list;
+	u32			cmd_hdr[1024];
+	unsigned int		blocks;
+	u8			nr_entries;
+	u8			retries;
+	s16			idx_failure;
+};
+
 struct mmc_queue_req {
 	struct request		*req;
 	struct mmc_blk_request	brq;
@@ -20,6 +37,8 @@ struct mmc_queue_req {
 	struct scatterlist	*bounce_sg;
 	unsigned int		bounce_sg_len;
 	struct mmc_async_req	mmc_active;
+	enum mmc_packed_type	cmd_type;
+	struct mmc_packed	*packed;
 };
 
 struct mmc_queue {
@@ -49,4 +68,7 @@ extern unsigned int mmc_queue_map_sg(struct mmc_queue *,
 extern void mmc_queue_bounce_pre(struct mmc_queue_req *);
 extern void mmc_queue_bounce_post(struct mmc_queue_req *);
 
+extern int mmc_packed_init(struct mmc_queue *, struct mmc_card *);
+extern void mmc_packed_clean(struct mmc_queue *);
+
 #endif
diff --git a/drivers/mmc/core/mmc_ops.c b/drivers/mmc/core/mmc_ops.c
index 6d8f7012d73..49f04bc9d0e 100644
--- a/drivers/mmc/core/mmc_ops.c
+++ b/drivers/mmc/core/mmc_ops.c
@@ -363,6 +363,7 @@ int mmc_send_ext_csd(struct mmc_card *card, u8 *ext_csd)
 	return mmc_send_cxd_data(card, card->host, MMC_SEND_EXT_CSD,
 			ext_csd, 512);
 }
+EXPORT_SYMBOL_GPL(mmc_send_ext_csd);
 
 int mmc_spi_read_ocr(struct mmc_host *host, int highcap, u32 *ocrp)
 {