Make the HDIO_DRIVE_CMD ioctl in libata (ATA command pass through)
return a few ATA registers to userspace, following the same convention
as the drivers/ide implementation of the same ioctl. This is needed to
support ATA commands like CHECK POWER MODE, which return information
in nsectors.
This fixes "hdparm -C" on SATA drives.
Forcing the sense data read via the cc flag causes spurious check
conditions, so we filter these out (following the
ATA command pass-through specification T10/04-262r7).
Signed-off-by: Eran Tromer <[email protected]>
---
Applies to git/jgarzik/libata-dev.git master and git/torvalds/linux-2.6.git
master. To apply to 2.6.18, change drivers/ata/ to drivers/scsi/.
On 2006-08-29 05:39, Tejun Heo wrote to [email protected]:
>> 2) hdparm -C for all 4 drives always shows "drive state is: standby"
>> even when I'm certain that the drives are active.
>
> hdparm -C says the same thing for my drive. I think it's safe to
> ignore. Hmmm... it needs to be tracked down. Maybe some problem in
> HDIO ioctl implementation in libata.
Yes, and fixed by this patch.
drivers/ata/libata-scsi.c | 46 +++++++++++++++++++++++++++++++++++++++-------
1 file changed, 39 insertions(+), 7 deletions(-)
diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
--- a/drivers/ata/libata-scsi.c
+++ b/drivers/ata/libata-scsi.c
@@ -164,10 +164,10 @@ int ata_cmd_ioctl(struct scsi_device *sc
{
int rc = 0;
u8 scsi_cmd[MAX_COMMAND_SIZE];
- u8 args[4], *argbuf = NULL;
+ u8 args[4], *argbuf = NULL, *sensebuf = NULL;
int argsize = 0;
- struct scsi_sense_hdr sshdr;
enum dma_data_direction data_dir;
+ int cmd_result;
if (arg == NULL)
return -EINVAL;
@@ -175,6 +175,10 @@ int ata_cmd_ioctl(struct scsi_device *sc
if (copy_from_user(args, arg, sizeof(args)))
return -EFAULT;
+ sensebuf = kzalloc(SCSI_SENSE_BUFFERSIZE, GFP_NOIO);
+ if (!sensebuf)
+ return -ENOMEM;
+
memset(scsi_cmd, 0, sizeof(scsi_cmd));
if (args[3]) {
@@ -191,7 +195,7 @@ int ata_cmd_ioctl(struct scsi_device *sc
data_dir = DMA_FROM_DEVICE;
} else {
scsi_cmd[1] = (3 << 1); /* Non-data */
- /* scsi_cmd[2] is already 0 -- no off.line, cc, or data xfer */
+ scsi_cmd[2] = 0x20; /* cc but no off.line or data xfer */
data_dir = DMA_NONE;
}
@@ -210,18 +214,46 @@ int ata_cmd_ioctl(struct scsi_device *sc
/* Good values for timeout and retries? Values below
from scsi_ioctl_send_command() for default case... */
- if (scsi_execute_req(scsidev, scsi_cmd, data_dir, argbuf, argsize,
- &sshdr, (10*HZ), 5)) {
+ cmd_result = scsi_execute(scsidev, scsi_cmd, data_dir, argbuf, argsize,
+ sensebuf, (10*HZ), 5, 0);
+
+ if ((cmd_result>>24) == DRIVER_SENSE) { /* sense data available */
+ u8 *desc = sensebuf + 8;
+ cmd_result &= ~(0xFF<<24); /* DRIVER_SENSE is not an error */
+
+ /* If we set cc then ATA pass-through will cause a
+ * check condition even if no error. Filter that. */
+ if (cmd_result & SAM_STAT_CHECK_CONDITION) {
+ struct scsi_sense_hdr sshdr;
+ scsi_normalize_sense(sensebuf, SCSI_SENSE_BUFFERSIZE,
+ &sshdr);
+ if (sshdr.sense_key==0 &&
+ sshdr.asc==0 && sshdr.ascq==0)
+ cmd_result &= ~SAM_STAT_CHECK_CONDITION;
+ }
+
+ /* Send userspace a few ATA registers (same as drivers/ide) */
+ if (sensebuf[0] == 0x72 && /* format is "descriptor" */
+ desc[0] == 0x09 ) { /* code is "ATA Descriptor" */
+ args[0] = desc[13]; /* status */
+ args[1] = desc[3]; /* error */
+ args[2] = desc[5]; /* sector count (0:7) */
+ if (copy_to_user(arg, args, sizeof(args)))
+ rc = -EFAULT;
+ }
+ }
+
+
+ if (cmd_result) {
rc = -EIO;
goto error;
}
- /* Need code to retrieve data from check condition? */
-
if ((argbuf)
&& copy_to_user(arg + sizeof(args), argbuf, argsize))
rc = -EFAULT;
error:
+ kfree(sensebuf);
kfree(argbuf);
return rc;
}
Hello, Eran Tromer.
On Wed, Sep 27, 2006 at 07:06:07PM +0300, Eran Tromer wrote:
> > hdparm -C says the same thing for my drive. I think it's safe to
> > ignore. Hmmm... it needs to be tracked down. Maybe some problem in
> > HDIO ioctl implementation in libata.
>
> Yes, and fixed by this patch.
Great. Things look good at the first glance. Just a few comments.
> @@ -210,18 +214,46 @@ int ata_cmd_ioctl(struct scsi_device *sc
>
> /* Good values for timeout and retries? Values below
> from scsi_ioctl_send_command() for default case... */
> - if (scsi_execute_req(scsidev, scsi_cmd, data_dir, argbuf, argsize,
> - &sshdr, (10*HZ), 5)) {
> + cmd_result = scsi_execute(scsidev, scsi_cmd, data_dir, argbuf, argsize,
> + sensebuf, (10*HZ), 5, 0);
> +
> + if ((cmd_result>>24) == DRIVER_SENSE) { /* sense data available */
driver_byte() seems more appropriate.
> + u8 *desc = sensebuf + 8;
> + cmd_result &= ~(0xFF<<24); /* DRIVER_SENSE is not an error */
> +
> + /* If we set cc then ATA pass-through will cause a
> + * check condition even if no error. Filter that. */
> + if (cmd_result & SAM_STAT_CHECK_CONDITION) {
> + struct scsi_sense_hdr sshdr;
> + scsi_normalize_sense(sensebuf, SCSI_SENSE_BUFFERSIZE,
> + &sshdr);
> + if (sshdr.sense_key==0 &&
> + sshdr.asc==0 && sshdr.ascq==0)
> + cmd_result &= ~SAM_STAT_CHECK_CONDITION;
> + }
> +
> + /* Send userspace a few ATA registers (same as drivers/ide) */
> + if (sensebuf[0] == 0x72 && /* format is "descriptor" */
> + desc[0] == 0x09 ) { /* code is "ATA Descriptor" */
> + args[0] = desc[13]; /* status */
> + args[1] = desc[3]; /* error */
> + args[2] = desc[5]; /* sector count (0:7) */
> + if (copy_to_user(arg, args, sizeof(args)))
> + rc = -EFAULT;
> + }
> + }
> +
> +
> + if (cmd_result) {
I wonder whether we need to fake default error ATA regsiters here.
Anyways, it's unrelated to this patch.
> rc = -EIO;
> goto error;
> }
>
> - /* Need code to retrieve data from check condition? */
> -
> if ((argbuf)
> && copy_to_user(arg + sizeof(args), argbuf, argsize))
> rc = -EFAULT;
> error:
> + kfree(sensebuf);
> kfree(argbuf);
> return rc;
Acked-by: Tejun Heo <[email protected]>
--
tejun