Skip to contentSkip to conversation

Avoid ide-scsi from starting DMA too soon

PATCH [email protected] 1 participant 1 message May 15, 2003 SCSI

lore
Share & tools
lore
#0 lore raw

This went into 2.4 with the following comments..

ide-scsi driver starts DMA as soon as it writes the ATAPI PACKET command in command register and before sending the ATAPI command. This will cause problems on many drives. Right way to do it is to start DMA after sending the ATAPI command. I am attaching a patch that fixes this. This patch will allow many more CD-RW drives to work reliably in DMA mode than do today

Alan's comment to this diff previously..
"Thats the least of the 2.5 ide-scsi problems, but yes its probably one to add"

drivers/scsi/ide-scsi.c +8 -4
@@ -78,6 +78,7 @@ typedef struct idescsi_pc_s {
 #define PC_DMA_IN_PROGRESS		0	/* 1 while DMA in progress */
 #define PC_WRITING			1	/* Data direction */
 #define PC_TRANSFORM			2	/* transform SCSI commands */
+#define PC_DMA_OK			4	/* Use DMA */
 
 /*
  *	SCSI command transformation layer
@@ -494,6 +495,10 @@ static ide_startstop_t idescsi_transfer_
 	ide_set_handler(drive, &idescsi_pc_intr, get_timeout(pc), NULL);
 	/* Send the actual packet */
 	atapi_output_bytes(drive, scsi->pc->c, 12);
+	if (test_bit (PC_DMA_OK, &pc->flags)) {
+		set_bit (PC_DMA_IN_PROGRESS, &pc->flags);
+		(void) (HWIF(drive)->ide_dma_begin(drive));
+	}
 	return ide_started;
 }
 
@@ -527,10 +532,9 @@ static ide_startstop_t idescsi_issue_pc 
 	HWIF(drive)->OUTB(bcount.b.high, IDE_BCOUNTH_REG);
 	HWIF(drive)->OUTB(bcount.b.low, IDE_BCOUNTL_REG);
 
-	if (feature.b.dma) {
-		set_bit(PC_DMA_IN_PROGRESS, &pc->flags);
-		(void) (HWIF(drive)->ide_dma_begin(drive));
-	}
+	if (feature.b.dma)
+		set_bit(PC_DMA_OK, &pc->flags);
+
 	if (test_bit(IDESCSI_DRQ_INTERRUPT, &scsi->flags)) {
 		if (HWGROUP(drive)->handler != NULL)
 			BUG();