sg: add SG_FLAG_Q_AT_TAIL flag

PATCH Superseded Douglas Gilbert 2 participants 2 messages Jun 4, 2014

Subsystems (2)

SCSI SCSI SG driver

lore

Share & tools

lore

Douglas Gilbert

author maintainer

2014-06-04 01:37 UTC #0 lore raw

PATCH 1/1 sg: add SG_FLAG_Q_AT_TAIL flag

When the SG_IO ioctl was copied into the block layer and later into the bsg driver, subtle differences emerged.

One difference is the way injected commands are queued through the block layer (i.e. this is not SCSI device queueing nor SATA NCQ). Summarizing:
- SG_IO in the block layer: blk_exec*(at_head=false)
- sg SG_IO: at_head=true
- bsg SG_IO: at_head=true

Some time ago Boaz Harrosh introduced a sg v4 flag called BSG_FLAG_Q_AT_TAIL to override the bsg driver default. This patch does the equivalent for the sg driver.

ChangeLog:
Introduce SG_FLAG_Q_AT_TAIL flag to cause commands to be injected into the block layer with at_head=false.

Signed-off-by Douglas Gilbert <[email protected]>

drivers/scsi/sg.c +3 -2

index 177f755..1b7d4f6 100644
@@ -740,7 +740,7 @@ static int
 sg_common_write(Sg_fd * sfp, Sg_request * srp,
 		unsigned char *cmnd, int timeout, int blocking)
 {
-	int k, data_dir;
+	int k, data_dir, at_head;
 	Sg_device *sdp = sfp->parentdp;
 	sg_io_hdr_t *hp = &srp->header;

@@ -784,11 +784,12 @@ sg_common_write(Sg_fd * sfp, Sg_request * srp,
 		break;
 	}
 	hp->duration = jiffies_to_msecs(jiffies);
+	at_head = !(hp->interface_id && (SG_FLAG_Q_AT_TAIL & hp->flags));
 
 	srp->rq->timeout = timeout;
 	kref_get(&sfp->f_ref); /* sg_rq_end_io() does kref_put(). */
 	blk_execute_rq_nowait(sdp->device->request_queue, sdp->disk,
-			      srp->rq, 1, sg_rq_end_io);
+			      srp->rq, at_head, sg_rq_end_io);
 	return 0;
 }

include/scsi/sg.h +1 -0

index d8c0c43..9859355 100644
@@ -86,6 +86,7 @@ typedef struct sg_io_hdr
 #define SG_FLAG_MMAP_IO 4       /* request memory mapped IO */
 #define SG_FLAG_NO_DXFER 0x10000 /* no transfer of kernel buffers to/from */
 				/* user space (debug indirect IO) */
+#define SG_FLAG_Q_AT_TAIL 0x10  /* default is Q_AT_HEAD */
 
 /* following 'info' values are "or"-ed together */
 #define SG_INFO_OK_MASK 0x1

Christoph Hellwig

2014-06-04 08:58 UTC #1 lore raw

Quoted 35 lines · Douglas Gilbert

When the SG_IO ioctl was copied into the block layer and later into the bsg driver, subtle differences emerged.
One difference is the way injected commands are queued through the block layer (i.e. this is not SCSI device queueing nor SATA NCQ). Summarizing:
- SG_IO in the block layer: blk_exec*(at_head=false)
- sg SG_IO: at_head=true
- bsg SG_IO: at_head=true
Some time ago Boaz Harrosh introduced a sg v4 flag called BSG_FLAG_Q_AT_TAIL to override the bsg driver default. This patch does the equivalent for the sg driver.
ChangeLog:
Introduce SG_FLAG_Q_AT_TAIL flag to cause commands to be injected into the block layer with at_head=false.
Signed-off-by Douglas Gilbert <[email protected]>
drivers/scsi/sg.c +1 -1
index 177f755..1b7d4f6 100644
@@ -740,7 +740,7 @@ static int
 sg_common_write(Sg_fd * sfp, Sg_request * srp,
 		unsigned char *cmnd, int timeout, int blocking)
 {
-	int k, data_dir;
+	int k, data_dir, at_head;
 	Sg_device *sdp = sfp->parentdp;
 	sg_io_hdr_t *hp = &srp->header;
 

drivers/scsi/sg.c sg_common_write(Sg_fd * sfp, Sg_request * srp, · line 787

Quoted 4 lines

drivers/scsi/sg.c +1 -0

@@ -784,11 +784,12 @@ sg_common_write(Sg_fd * sfp, Sg_request * srp,
 		break;
 	}
 	hp->duration = jiffies_to_msecs(jiffies);
+	at_head = !(hp->interface_id && (SG_FLAG_Q_AT_TAIL & hp->flags));

It seems like all other occurances of ->interface_id compare to either '\0' or 'S', so this looks odd. Also I think this would really benefit from beeing untangled into and if/else as it took me literally two minutes to parse..