On 1/15/25 1:29 PM, Shin'ichiro Kawasaki wrote:

The current null_blk implementation checks if any bad blocks exist in
the target blocks of each IO. If so, the IO fails and data is not
transferred for all of the IO target blocks. However, when real storage
devices have bad blocks, the devices may transfer data partially up to
the first bad blocks (e.g., SAS drives). Especially, when the IO is a
write operation, such partial IO leaves partially written data on the
device.

To simulate such partial IO using null_blk, introduce the new parameter
'badblocks_partial_io'. When this parameter is set,
null_handle_badblocks() returns the number of the sectors for the
partial IO as its third pointer argument. Pass the returned number of
sectors to the following calls to null_handle_memory_backend() in
null_process_cmd() and null_zone_write().

Signed-off-by: Shin'ichiro Kawasaki <redacted>

A couple of nits below. Otherwise looks OK to me.

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

+/*
+ * Check if the command should fail for the badblocks. If so, return
+ * BLK_STS_IOERR and return number of partial I/O sectors.

...and return the number of sectors that were read or written, which may be less
than the requested number of sectors.

+ *
+ * @cmd:        The command to handle.
+ * @sector:     The start sector for I/O.
+ * @nr_sectors: The caller specifies number of sectors to write or read.
+ *              Returns number of sectors to be written or read for partial I/O.

@nr_sectors: Specifies number of sectors to read or write and returns the number
	     of sectors that were read or written.

+ */
 blk_status_t null_handle_badblocks(struct nullb_cmd *cmd, sector_t sector,
-				   sector_t nr_sectors)
+				   unsigned int *nr_sectors)
 {
 	struct badblocks *bb = &cmd->nq->dev->badblocks;
+	struct nullb_device *dev = cmd->nq->dev;
+	unsigned int block_sectors = dev->blocksize >> SECTOR_SHIFT;
 	sector_t first_bad;
 	int bad_sectors;
+	unsigned int partial_io_sectors = 0;
 
-	if (!badblocks_check(bb, sector, nr_sectors, &first_bad, &bad_sectors))
+	if (!badblocks_check(bb, sector, *nr_sectors, &first_bad, &bad_sectors))
 		return BLK_STS_OK;
 
 	if (cmd->nq->dev->badblocks_once)
 		badblocks_clear(bb, first_bad, bad_sectors);
 
+	if (cmd->nq->dev->badblocks_partial_io) {
+		if (!IS_ALIGNED(first_bad, block_sectors))
+			first_bad = ALIGN_DOWN(first_bad, block_sectors);
+		if (sector < first_bad)
+			partial_io_sectors = first_bad - sector;
+	}
+	*nr_sectors = partial_io_sectors;
+
 	return BLK_STS_IOERR;
 }

-- 
Damien Le Moal
Western Digital Research