Please pull from 'upstream-linus' branch of
master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev.git upstream-linus
to receive the following updates:
drivers/ata/libata-scsi.c | 2 +-
drivers/ata/pata_artop.c | 2 +-
drivers/ata/pata_hpt37x.c | 19 ++++++++++++++++---
3 files changed, 18 insertions(+), 5 deletions(-)
Alan Cox:
hpt37x: Check the enablebits
Alexey Dobriyan:
pata_artop: fix "& (1 >>" typo
Darrick J. Wong:
libata: fix double-completion on error
diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
index 7af2a4b..5c1fc46 100644
--- a/drivers/ata/libata-scsi.c
+++ b/drivers/ata/libata-scsi.c
@@ -1612,9 +1612,9 @@ early_finish:
err_did:
ata_qc_free(qc);
-err_mem:
cmd->result = (DID_ERROR << 16);
done(cmd);
+err_mem:
DPRINTK("EXIT - internal\n");
return 0;
diff --git a/drivers/ata/pata_artop.c b/drivers/ata/pata_artop.c
index 690828e..96a0980 100644
--- a/drivers/ata/pata_artop.c
+++ b/drivers/ata/pata_artop.c
@@ -92,7 +92,7 @@ static int artop6260_pre_reset(struct at
return -ENOENT;
pci_read_config_byte(pdev, 0x49, &tmp);
- if (tmp & (1 >> ap->port_no))
+ if (tmp & (1 << ap->port_no))
ap->cbl = ATA_CBL_PATA40;
else
ap->cbl = ATA_CBL_PATA80;
diff --git a/drivers/ata/pata_hpt37x.c b/drivers/ata/pata_hpt37x.c
index 7350443..fce3fcd 100644
--- a/drivers/ata/pata_hpt37x.c
+++ b/drivers/ata/pata_hpt37x.c
@@ -25,7 +25,7 @@ #include <scsi/scsi_host.h>
#include <linux/libata.h>
#define DRV_NAME "pata_hpt37x"
-#define DRV_VERSION "0.5"
+#define DRV_VERSION "0.5.1"
struct hpt_clock {
u8 xfer_speed;
@@ -453,7 +453,13 @@ static int hpt37x_pre_reset(struct ata_p
{
u8 scr2, ata66;
struct pci_dev *pdev = to_pci_dev(ap->host->dev);
-
+ static const struct pci_bits hpt37x_enable_bits[] = {
+ { 0x50, 1, 0x04, 0x04 },
+ { 0x54, 1, 0x04, 0x04 }
+ };
+ if (!pci_test_config_bits(pdev, &hpt37x_enable_bits[ap->port_no]))
+ return -ENOENT;
+
pci_read_config_byte(pdev, 0x5B, &scr2);
pci_write_config_byte(pdev, 0x5B, scr2 & ~0x01);
/* Cable register now active */
@@ -488,10 +494,17 @@ static void hpt37x_error_handler(struct
static int hpt374_pre_reset(struct ata_port *ap)
{
+ static const struct pci_bits hpt37x_enable_bits[] = {
+ { 0x50, 1, 0x04, 0x04 },
+ { 0x54, 1, 0x04, 0x04 }
+ };
u16 mcr3, mcr6;
u8 ata66;
-
struct pci_dev *pdev = to_pci_dev(ap->host->dev);
+
+ if (!pci_test_config_bits(pdev, &hpt37x_enable_bits[ap->port_no]))
+ return -ENOENT;
+
/* Do the extra channel work */
pci_read_config_word(pdev, 0x52, &mcr3);
pci_read_config_word(pdev, 0x56, &mcr6);
diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
index 7af2a4b..5c1fc46 100644
--- a/drivers/ata/libata-scsi.c
+++ b/drivers/ata/libata-scsi.c
@@ -1612,9 +1612,9 @@ early_finish:
err_did:
ata_qc_free(qc);
-err_mem:
cmd->result = (DID_ERROR << 16);
done(cmd);
+err_mem:
DPRINTK("EXIT - internal\n");
return 0;
This doesn't look correct to me, but I did miss out on the original discussion(?).
Any time we return 0 from queuecommand, the SCSI mid-layer expects us
to also take care of invoking the done() function. Where does this now
happen for this case (err_mem) ???
Cheers
Mark Lord wrote:
> Any time we return 0 from queuecommand, the SCSI mid-layer expects us
> to also take care of invoking the done() function. Where does this now
> happen for this case (err_mem) ???
It _already_ happened in the error path of ata_scsi_qc_new(), which is
why this is a double-completion bug fix.
Jeff
Jeff Garzik wrote:
> Mark Lord wrote:
>> Any time we return 0 from queuecommand, the SCSI mid-layer expects us
>> to also take care of invoking the done() function. Where does this now
>> happen for this case (err_mem) ???
>
> It _already_ happened in the error path of ata_scsi_qc_new(), which is
> why this is a double-completion bug fix.
Ack.