FWIW, I appended the attached changeset to the 'upstream-linus' branch
of libata-dev.git (and also forwarded it to stable@xxxxxxxxxx).
FWIW 2, except for fixes like the attached, 'upstream-linus' is a frozen
branch, and normally isn't touched until you pull. Development
continues in the separate 'upstream' branch, which ensures there are no
surprise csets when 'upstream-linus' is pulled. 'upstream-linus' is a
branch that behaves like a tag. Most of the time.
Jeff
commit cd85f6e2f58282186ad720fc18482be228f0b972
Author: Jeff Garzik <jeff@xxxxxxxxxx>
Date: Mon Mar 20 19:49:54 2006 -0500
[libata] sata_mv: fix irq port status usage
Interrupt handler did not properly initialize a variable on a per-port
basis, leading to incorrect behavior on ports other than port 0.
Bug caught and fixed by Mark Lord.
Signed-off-by: Jeff Garzik <jeff@xxxxxxxxxx>
cd85f6e2f58282186ad720fc18482be228f0b972
diff --git a/drivers/scsi/sata_mv.c b/drivers/scsi/sata_mv.c
index aceaf56..e561281 100644
--- a/drivers/scsi/sata_mv.c
+++ b/drivers/scsi/sata_mv.c
@@ -1353,7 +1353,6 @@ static void mv_host_intr(struct ata_host
u32 hc_irq_cause;
int shift, port, port0, hard_port, handled;
unsigned int err_mask;
- u8 ata_status = 0;
if (hc == 0) {
port0 = 0;
@@ -1371,6 +1370,7 @@ static void mv_host_intr(struct ata_host
hc,relevant,hc_irq_cause);
for (port = port0; port < port0 + MV_PORTS_PER_HC; port++) {
+ u8 ata_status = 0;
ap = host_set->ports[port];
hard_port = port & MV_PORT_MASK; /* range 0-3 */
handled = 0; /* ensure ata_status is set if handled++ */
