James:
I found some small errors in the sr driver while working on something
else (patch for that to follow). In this patch (as571):
Update the s_size variable when the block length is changed.
This seems like such an obvious mistake, I have to wonder if
I have misunderstood the code.
Set SCpnt->underflow to the sector size instead of the transfer
size. If even one sector is transferred successfully, we
should accept it and try to move forward -- don't decide that
the whole thing failed.
Alan Stern
Signed-off-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
Index: usb-2.6/drivers/scsi/sr.c
===================================================================
--- usb-2.6.orig/drivers/scsi/sr.c
+++ usb-2.6/drivers/scsi/sr.c
@@ -352,9 +352,10 @@ static int sr_init_command(struct scsi_c
*/
s_size = cd->device->sector_size;
if (s_size > 2048) {
- if (!in_interrupt())
+ if (!in_interrupt()) {
sr_set_blocklength(cd, 2048);
- else
+ s_size = 2048;
+ } else
printk("sr: can't switch blocksize: in interrupt\n");
}
@@ -431,8 +432,7 @@ static int sr_init_command(struct scsi_c
* host adapter, it's safe to assume that we can at least transfer
* this many bytes between each connect / disconnect.
*/
- SCpnt->transfersize = cd->device->sector_size;
- SCpnt->underflow = this_count << 9;
+ SCpnt->transfersize = SCpnt->underflow = s_size;
queue:
SCpnt->allowed = MAX_RETRIES;
-
: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
