[SCSI] seperate max_sectors from max_hw_sectors

- export __blk_put_request and blk_execute_rq_nowait
needed for async REQ_BLOCK_PC requests
- seperate max_hw_sectors and max_sectors for block/scsi_ioctl.c and
SG_IO bio.c helpers per Jens's last comments. Since block/scsi_ioctl.c SG_IO was
already testing against max_sectors and SCSI-ml was setting max_sectors and
max_hw_sectors to the same value this does not change any scsi SG_IO behavior. It only
prepares ll_rw_blk.c, scsi_ioctl.c and bio.c for when SCSI-ml begins to set
a valid max_hw_sectors for all LLDs. Today if a LLD does not set it
SCSI-ml sets it to a safe default and some LLDs set it to a artificial low
value to overcome memory and feedback issues.

Note: Since we now cap max_sectors to BLK_DEF_MAX_SECTORS, which is 1024,
drivers that used to call blk_queue_max_sectors with a large value of
max_sectors will now see the fs requests capped to BLK_DEF_MAX_SECTORS.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

6 files changed, 42 insertions, 21 deletions

diff --git a/block/ll_rw_blk.c b/block/ll_rw_blk.c
index c525b5a..d4beb9a 100644
--- a/block/ll_rw_blk.c
+++ b/block/ll_rw_blk.c
@@ -239,7 +239,7 @@ void blk_queue_make_request(request_queue_t * q, make_request_fn * mfn)
 	q->backing_dev_info.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
 	q->backing_dev_info.state = 0;
 	q->backing_dev_info.capabilities = BDI_CAP_MAP_COPY;
-	blk_queue_max_sectors(q, MAX_SECTORS);
+	blk_queue_max_sectors(q, SAFE_MAX_SECTORS);
 	blk_queue_hardsect_size(q, 512);
 	blk_queue_dma_alignment(q, 511);
 	blk_queue_congestion_threshold(q);
@@ -555,7 +555,12 @@ void blk_queue_max_sectors(request_queue_t *q, unsigned short max_sectors)
 		printk("%s: set to minimum %d\n", __FUNCTION__, max_sectors);
 	}
 
-	q->max_sectors = q->max_hw_sectors = max_sectors;
+	if (BLK_DEF_MAX_SECTORS > max_sectors)
+		q->max_hw_sectors = q->max_sectors = max_sectors;
+ 	else {
+		q->max_sectors = BLK_DEF_MAX_SECTORS;
+		q->max_hw_sectors = max_sectors;
+	}
 }
 
 EXPORT_SYMBOL(blk_queue_max_sectors);
@@ -657,8 +662,8 @@ EXPORT_SYMBOL(blk_queue_hardsect_size);
 void blk_queue_stack_limits(request_queue_t *t, request_queue_t *b)
 {
 	/* zero is "infinity" */
-	t->max_sectors = t->max_hw_sectors =
-		min_not_zero(t->max_sectors,b->max_sectors);
+	t->max_sectors = min_not_zero(t->max_sectors,b->max_sectors);
+	t->max_hw_sectors = min_not_zero(t->max_hw_sectors,b->max_hw_sectors);
 
 	t->max_phys_segments = min(t->max_phys_segments,b->max_phys_segments);
 	t->max_hw_segments = min(t->max_hw_segments,b->max_hw_segments);
@@ -1293,9 +1298,15 @@ static inline int ll_new_hw_segment(request_queue_t *q,
 static int ll_back_merge_fn(request_queue_t *q, struct request *req, 
 			    struct bio *bio)
 {
+	unsigned short max_sectors;
 	int len;
 
-	if (req->nr_sectors + bio_sectors(bio) > q->max_sectors) {
+	if (unlikely(blk_pc_request(req)))
+		max_sectors = q->max_hw_sectors;
+	else
+		max_sectors = q->max_sectors;
+
+	if (req->nr_sectors + bio_sectors(bio) > max_sectors) {
 		req->flags |= REQ_NOMERGE;
 		if (req == q->last_merge)
 			q->last_merge = NULL;
@@ -1325,9 +1336,16 @@ static int ll_back_merge_fn(request_queue_t *q, struct request *req,
 static int ll_front_merge_fn(request_queue_t *q, struct request *req, 
 			     struct bio *bio)
 {
+	unsigned short max_sectors;
 	int len;
 
-	if (req->nr_sectors + bio_sectors(bio) > q->max_sectors) {
+	if (unlikely(blk_pc_request(req)))
+		max_sectors = q->max_hw_sectors;
+	else
+		max_sectors = q->max_sectors;
+
+
+	if (req->nr_sectors + bio_sectors(bio) > max_sectors) {
 		req->flags |= REQ_NOMERGE;
 		if (req == q->last_merge)
 			q->last_merge = NULL;
@@ -2144,7 +2162,7 @@ int blk_rq_map_user(request_queue_t *q, struct request *rq, void __user *ubuf,
 	struct bio *bio;
 	int reading;
 
-	if (len > (q->max_sectors << 9))
+	if (len > (q->max_hw_sectors << 9))
 		return -EINVAL;
 	if (!len || !ubuf)
 		return -EINVAL;
@@ -2259,7 +2277,7 @@ int blk_rq_map_kern(request_queue_t *q, struct request *rq, void *kbuf,
 {
 	struct bio *bio;
 
-	if (len > (q->max_sectors << 9))
+	if (len > (q->max_hw_sectors << 9))
 		return -EINVAL;
 	if (!len || !kbuf)
 		return -EINVAL;
diff --git a/block/scsi_ioctl.c b/block/scsi_ioctl.c
index 382dea7..4e390df 100644
--- a/block/scsi_ioctl.c
+++ b/block/scsi_ioctl.c
@@ -233,7 +233,7 @@ static int sg_io(struct file *file, request_queue_t *q,
 	if (verify_command(file, cmd))
 		return -EPERM;
 
-	if (hdr->dxfer_len > (q->max_sectors << 9))
+	if (hdr->dxfer_len > (q->max_hw_sectors << 9))
 		return -EIO;
 
 	if (hdr->dxfer_len)
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index a6d3baa..a6f2dc6 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -638,7 +638,7 @@ int dm_split_args(int *argc, char ***argvp, char *input)
 static void check_for_valid_limits(struct io_restrictions *rs)
 {
 	if (!rs->max_sectors)
-		rs->max_sectors = MAX_SECTORS;
+		rs->max_sectors = SAFE_MAX_SECTORS;
 	if (!rs->max_phys_segments)
 		rs->max_phys_segments = MAX_PHYS_SEGMENTS;
 	if (!rs->max_hw_segments)
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 54a72f1..14ad2a7 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -462,6 +462,7 @@ int scsi_execute_async(struct scsi_device *sdev, const unsigned char *cmd,
 	req = blk_get_request(sdev->request_queue, write, gfp);
 	if (!req)
 		goto free_sense;
+	req->flags |= REQ_BLOCK_PC | REQ_QUIET;
 
 	if (use_sg)
 		err = scsi_req_map_sg(req, buffer, use_sg, bufflen, gfp);
@@ -477,7 +478,6 @@ int scsi_execute_async(struct scsi_device *sdev, const unsigned char *cmd,
 	req->sense_len = 0;
 	req->timeout = timeout;
 	req->retries = retries;
-	req->flags |= REQ_BLOCK_PC | REQ_QUIET;
 	req->end_io_data = sioc;
 
 	sioc->data = privdata;
diff --git a/fs/bio.c b/fs/bio.c
index 4d21ee3..38d3e80 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -313,7 +313,8 @@ int bio_get_nr_vecs(struct block_device *bdev)
 }
 
 static int __bio_add_page(request_queue_t *q, struct bio *bio, struct page
-			  *page, unsigned int len, unsigned int offset)
+			  *page, unsigned int len, unsigned int offset,
+			  unsigned short max_sectors)
 {
 	int retried_segments = 0;
 	struct bio_vec *bvec;
@@ -327,7 +328,7 @@ static int __bio_add_page(request_queue_t *q, struct bio *bio, struct page
 	if (bio->bi_vcnt >= bio->bi_max_vecs)
 		return 0;
 
-	if (((bio->bi_size + len) >> 9) > q->max_sectors)
+	if (((bio->bi_size + len) >> 9) > max_sectors)
 		return 0;
 
 	/*
@@ -401,7 +402,7 @@ static int __bio_add_page(request_queue_t *q, struct bio *bio, struct page
 int bio_add_pc_page(request_queue_t *q, struct bio *bio, struct page *page,
 		    unsigned int len, unsigned int offset)
 {
-	return __bio_add_page(q, bio, page, len, offset);
+	return __bio_add_page(q, bio, page, len, offset, q->max_hw_sectors);
 }
 
 /**
@@ -420,8 +421,8 @@ int bio_add_pc_page(request_queue_t *q, struct bio *bio, struct page *page,
 int bio_add_page(struct bio *bio, struct page *page, unsigned int len,
 		 unsigned int offset)
 {
-	return __bio_add_page(bdev_get_queue(bio->bi_bdev), bio, page,
-			      len, offset);
+	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
+	return __bio_add_page(q, bio, page, len, offset, q->max_sectors);
 }
 
 struct bio_map_data {
@@ -533,7 +534,7 @@ struct bio *bio_copy_user(request_queue_t *q, unsigned long uaddr,
 			break;
 		}
 
-		if (__bio_add_page(q, bio, page, bytes, 0) < bytes) {
+		if (bio_add_pc_page(q, bio, page, bytes, 0) < bytes) {
 			ret = -EINVAL;
 			break;
 		}
@@ -647,7 +648,8 @@ static struct bio *__bio_map_user_iov(request_queue_t *q,
 			/*
 			 * sorry...
 			 */
-			if (__bio_add_page(q, bio, pages[j], bytes, offset) < bytes)
+			if (bio_add_pc_page(q, bio, pages[j], bytes, offset) <
+					    bytes)
 				break;
 
 			len -= bytes;
@@ -820,8 +822,8 @@ static struct bio *__bio_map_kern(request_queue_t *q, void *data,
 		if (bytes > len)
 			bytes = len;
 
-		if (__bio_add_page(q, bio, virt_to_page(data), bytes,
-				   offset) < bytes)
+		if (bio_add_pc_page(q, bio, virt_to_page(data), bytes,
+				    offset) < bytes)
 			break;
 
 		data += bytes;
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 509e9a0..a18500d 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -702,7 +702,8 @@ extern int blkdev_issue_flush(struct block_device *, sector_t *);
 
 #define MAX_PHYS_SEGMENTS 128
 #define MAX_HW_SEGMENTS 128
-#define MAX_SECTORS 255
+#define SAFE_MAX_SECTORS 255
+#define BLK_DEF_MAX_SECTORS 1024
 
 #define MAX_SEGMENT_SIZE	65536