From: Swarna Prabhu <sw.prabhu6@gmail.com>

The WRITE SAME(16) and WRITE SAME(10) scsi commands uses
a page from a dedicated mempool('sd_page_pool') for its
payload. This pool was initialized to allocate single
pages, which was sufficient as long as the device sector
size did not exceed the PAGE_SIZE.

Given that block layer now supports block size upto
64K ie beyond PAGE_SIZE, initialize large page pool in
'sd_probe()' if a higher sector device is attached ensuring
atomicity. Adapt 'sd_set_special_bvec()' to use large page
pool when a higher sector size device is attached. Hence
enable sector sizes > PAGE_SIZE in scsi sd driver.

Signed-off-by: Swarna Prabhu <s.prabhu@samsung.com>
Co-developed-by: Pankaj Raghav <p.raghav@samsung.com>
Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>
---
 drivers/scsi/sd.c | 80 ++++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 68 insertions(+), 12 deletions(-)

diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index d76996d6cbc9..ae6ccfed79b9 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -107,8 +107,11 @@ static void sd_config_write_same(struct scsi_disk *sdkp,
 static void  sd_revalidate_disk(struct gendisk *);
 
 static DEFINE_IDA(sd_index_ida);
+static DEFINE_MUTEX(sd_mutex_lock);
 
 static mempool_t *sd_page_pool;
+static mempool_t *sd_large_page_pool;
+static atomic_t sd_large_page_pool_users = ATOMIC_INIT(0);
 static struct lock_class_key sd_bio_compl_lkclass;
 
 static const char *sd_cache_types[] = {
@@ -116,6 +119,33 @@ static const char *sd_cache_types[] = {
 	"write back, no read (daft)"
 };
 
+static int sd_large_pool_create_lock(void)
+{
+	mutex_lock(&sd_mutex_lock);
+	if (!sd_large_page_pool) {
+		sd_large_page_pool = mempool_create_page_pool(
+			SD_MEMPOOL_SIZE, get_order(BLK_MAX_BLOCK_SIZE));
+		if (!sd_large_page_pool) {
+			printk(KERN_ERR "sd: can't create large page mempool\n");
+			mutex_unlock(&sd_mutex_lock);
+			return -ENOMEM;
+		}
+	}
+	atomic_inc(&sd_large_page_pool_users);
+	mutex_unlock(&sd_mutex_lock);
+	return 0;
+}
+
+static void sd_large_pool_destroy_lock(void)
+{
+	mutex_lock(&sd_mutex_lock);
+	if (atomic_dec_and_test(&sd_large_page_pool_users)) {
+		mempool_destroy(sd_large_page_pool);
+		sd_large_page_pool = NULL;
+	}
+	mutex_unlock(&sd_mutex_lock);
+}
+
 static void sd_disable_discard(struct scsi_disk *sdkp)
 {
 	sdkp->provisioning_mode = SD_LBP_DISABLE;
@@ -928,14 +958,24 @@ static unsigned char sd_setup_protect_cmnd(struct scsi_cmnd *scmd,
 	return protect;
 }
 
-static void *sd_set_special_bvec(struct request *rq, unsigned int data_len)
+static void *sd_set_special_bvec(struct scsi_cmnd *cmd, unsigned int data_len)
 {
 	struct page *page;
+	struct request *rq = scsi_cmd_to_rq(cmd);
+	struct scsi_device *sdp = cmd->device;
+	unsigned sector_size = sdp->sector_size;
+	unsigned int nr_pages = DIV_ROUND_UP(sector_size, PAGE_SIZE);
+	int n;
 
-	page = mempool_alloc(sd_page_pool, GFP_ATOMIC);
+	if (sector_size > PAGE_SIZE)
+		page = mempool_alloc(sd_large_page_pool, GFP_ATOMIC);
+	else
+		page = mempool_alloc(sd_page_pool, GFP_ATOMIC);
 	if (!page)
 		return NULL;
-	clear_highpage(page);
+
+	for (n = 0; n < nr_pages; n++)
+		clear_highpage(page + n);
 	bvec_set_page(&rq->special_vec, page, data_len, 0);
 	rq->rq_flags |= RQF_SPECIAL_PAYLOAD;
 	return bvec_virt(&rq->special_vec);
@@ -951,7 +991,7 @@ static blk_status_t sd_setup_unmap_cmnd(struct scsi_cmnd *cmd)
 	unsigned int data_len = 24;
 	char *buf;
 
-	buf = sd_set_special_bvec(rq, data_len);
+	buf = sd_set_special_bvec(cmd, data_len);
 	if (!buf)
 		return BLK_STS_RESOURCE;
 
@@ -1040,7 +1080,7 @@ static blk_status_t sd_setup_write_same16_cmnd(struct scsi_cmnd *cmd,
 	u32 nr_blocks = sectors_to_logical(sdp, blk_rq_sectors(rq));
 	u32 data_len = sdp->sector_size;
 
-	if (!sd_set_special_bvec(rq, data_len))
+	if (!sd_set_special_bvec(cmd, data_len))
 		return BLK_STS_RESOURCE;
 
 	cmd->cmd_len = 16;
@@ -1067,7 +1107,7 @@ static blk_status_t sd_setup_write_same10_cmnd(struct scsi_cmnd *cmd,
 	u32 nr_blocks = sectors_to_logical(sdp, blk_rq_sectors(rq));
 	u32 data_len = sdp->sector_size;
 
-	if (!sd_set_special_bvec(rq, data_len))
+	if (!sd_set_special_bvec(cmd, data_len))
 		return BLK_STS_RESOURCE;
 
 	cmd->cmd_len = 10;
@@ -1513,9 +1553,15 @@ static blk_status_t sd_init_command(struct scsi_cmnd *cmd)
 static void sd_uninit_command(struct scsi_cmnd *SCpnt)
 {
 	struct request *rq = scsi_cmd_to_rq(SCpnt);
+	struct scsi_device *sdp = SCpnt->device;
+	unsigned sector_size = sdp->sector_size;
 
-	if (rq->rq_flags & RQF_SPECIAL_PAYLOAD)
-		mempool_free(rq->special_vec.bv_page, sd_page_pool);
+	if (rq->rq_flags & RQF_SPECIAL_PAYLOAD) {
+		if (sector_size > PAGE_SIZE)
+			mempool_free(rq->special_vec.bv_page, sd_large_page_pool);
+		else
+			mempool_free(rq->special_vec.bv_page, sd_page_pool);
+	}
 }
 
 static bool sd_need_revalidate(struct gendisk *disk, struct scsi_disk *sdkp)
@@ -2912,10 +2958,7 @@ sd_read_capacity(struct scsi_disk *sdkp, struct queue_limits *lim,
 			  "Sector size 0 reported, assuming 512.\n");
 	}
 
-	if (sector_size != 512 &&
-	    sector_size != 1024 &&
-	    sector_size != 2048 &&
-	    sector_size != 4096) {
+	if (blk_validate_block_size(sector_size)) {
 		sd_printk(KERN_NOTICE, sdkp, "Unsupported sector size %d.\n",
 			  sector_size);
 		/*
@@ -4043,6 +4086,12 @@ static int sd_probe(struct scsi_device *sdp)
 	sdkp->max_medium_access_timeouts = SD_MAX_MEDIUM_TIMEOUTS;
 
 	sd_revalidate_disk(gd);
+	if (sdp->sector_size > PAGE_SIZE) {
+		if (sd_large_pool_create_lock()) {
+			error = -ENOMEM;
+			goto out_free_index;
+		}
+	}
 
 	if (sdp->removable) {
 		gd->flags |= GENHD_FL_REMOVABLE;
@@ -4060,6 +4109,8 @@ static int sd_probe(struct scsi_device *sdp)
 	if (error) {
 		device_unregister(&sdkp->disk_dev);
 		put_disk(gd);
+		if (sdp->sector_size > PAGE_SIZE)
+			sd_large_pool_destroy_lock();
 		goto out;
 	}
 
@@ -4212,6 +4263,9 @@ static void sd_remove(struct scsi_device *sdp)
 		sd_shutdown(sdp);
 
 	put_disk(sdkp->disk);
+
+	if (sdp->sector_size > PAGE_SIZE)
+		sd_large_pool_destroy_lock();
 }
 
 static inline bool sd_do_start_stop(struct scsi_device *sdev, bool runtime)
@@ -4435,6 +4489,8 @@ static void __exit exit_sd(void)
 
 	scsi_unregister_driver(&sd_template);
 	mempool_destroy(sd_page_pool);
+	if (sd_large_page_pool)
+		mempool_destroy(sd_large_page_pool);
 
 	class_unregister(&sd_disk_class);
 
-- 
2.39.5

On 2/14/26 10:18, sw.prabhu6@gmail.com wrote:
> From: Swarna Prabhu <sw.prabhu6@gmail.com>
> 
> The WRITE SAME(16) and WRITE SAME(10) scsi commands uses
> a page from a dedicated mempool('sd_page_pool') for its
> payload. This pool was initialized to allocate single
> pages, which was sufficient as long as the device sector
> size did not exceed the PAGE_SIZE.
> 
> Given that block layer now supports block size upto
> 64K ie beyond PAGE_SIZE, initialize large page pool in
> 'sd_probe()' if a higher sector device is attached ensuring
> atomicity. Adapt 'sd_set_special_bvec()' to use large page
> pool when a higher sector size device is attached. Hence
> enable sector sizes > PAGE_SIZE in scsi sd driver.
> 
> Signed-off-by: Swarna Prabhu <s.prabhu@samsung.com>
> Co-developed-by: Pankaj Raghav <p.raghav@samsung.com>
> Signed-off-by: Pankaj Raghav <p.raghav@samsung.com>

One nit below. With that fixed, feel free to add:

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

> ---
>  drivers/scsi/sd.c | 80 ++++++++++++++++++++++++++++++++++++++++-------
>  1 file changed, 68 insertions(+), 12 deletions(-)
> 
> diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
> index d76996d6cbc9..ae6ccfed79b9 100644
> --- a/drivers/scsi/sd.c
> +++ b/drivers/scsi/sd.c
> @@ -107,8 +107,11 @@ static void sd_config_write_same(struct scsi_disk *sdkp,
>  static void  sd_revalidate_disk(struct gendisk *);
>  
>  static DEFINE_IDA(sd_index_ida);
> +static DEFINE_MUTEX(sd_mutex_lock);
>  
>  static mempool_t *sd_page_pool;
> +static mempool_t *sd_large_page_pool;
> +static atomic_t sd_large_page_pool_users = ATOMIC_INIT(0);
>  static struct lock_class_key sd_bio_compl_lkclass;
>  
>  static const char *sd_cache_types[] = {
> @@ -116,6 +119,33 @@ static const char *sd_cache_types[] = {
>  	"write back, no read (daft)"
>  };
>  
> +static int sd_large_pool_create_lock(void)

I do not see the point of the "_lock" suffix since you do not have a "no lock"
variant of this function. So let's drop that suffix.

> +static void sd_large_pool_destroy_lock(void)

Same here.

-- 
Damien Le Moal
Western Digital Research