Daniel Drake wrote:
> Alan Cox wrote:
>> Lots of *other* problems occur instead. Daniel is reporting that if he
>> makes a stupid request to a buggy drive he gets a reset and the system
>> continues happily. Even that reset being a reset not a new command issue
>> is actually us being excessively paranoid.
>
> Sorry if I'm sprouting nonsense here -- I don't have the knowledge that
> you all do. However, just wanted to point out that I looked up the info
> about that mode page in the MMC specs.
>
> http://www.t10.org/ftp/t10/drafts/mmc4/mmc4r05a.pdf page 513.
> "E.3.3 MM Capabilities and Mechanical Status Page (Page Code 2Ah)"
>
> The length of this mode page varies from drive to drive, so there is no
> "one size" that you can supply to the SG_IO command (unless you want to
> use a stupidly large buffer) to retrieve all the data at once. Instead,
> as Tejun describes, you put a short read request in first, look at byte
> 1 of the page which tells you the length, and then read the whole lot.
>
> Again, ignore me if I'm not contributing anything useful, but I'm
> increasingly thinking that the SG_IO command block in question is
> perfectly valid, and doing a short read of the mode page in question
> (and probably others too) is in fact required before you can know its
> true size to do a full read anyway.
Yeap, the SG command is fine. The drive is being weird tho. The
allocation length field says 10 bytes, so it should just have
transferred 10 bytes without causing HSM violation.
Can you please apply the attached patch and report what the kernel says
after triggering the error condition?
--
tejun
diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c
index 8ee56e5..5424a59 100644
--- a/drivers/ata/libata-core.c
+++ b/drivers/ata/libata-core.c
@@ -5251,9 +5251,13 @@ fsm_start:
if (likely(status & (ATA_ERR | ATA_DF)))
/* device stops HSM for abort/error */
qc->err_mask |= AC_ERR_DEV;
- else
+ else {
/* HSM violation. Let EH handle this */
+ ata_dev_printk(qc->dev, KERN_WARNING,
+ "HSM violation: HSM_ST_FIRST: "
+ "!DRQ && !ERR && !DF\n");
qc->err_mask |= AC_ERR_HSM;
+ }
ap->hsm_task_state = HSM_ST_ERR;
goto fsm_start;
@@ -5344,13 +5348,17 @@ fsm_start:
if (likely(status & (ATA_ERR | ATA_DF)))
/* device stops HSM for abort/error */
qc->err_mask |= AC_ERR_DEV;
- else
+ else {
/* HSM violation. Let EH handle this.
* Phantom devices also trigger this
* condition. Mark hint.
*/
+ ata_dev_printk(qc->dev, KERN_WARNING,
+ "HSM violation: HSM_ST: "
+ "!DRQ && !ERR && !DF\n");
qc->err_mask |= AC_ERR_HSM |
AC_ERR_NODEV_HINT;
+ }
ap->hsm_task_state = HSM_ST_ERR;
goto fsm_start;
@@ -5375,8 +5383,12 @@ fsm_start:
status = ata_wait_idle(ap);
}
- if (status & (ATA_BUSY | ATA_DRQ))
+ if (status & (ATA_BUSY | ATA_DRQ)) {
+ ata_dev_printk(qc->dev, KERN_WARNING,
+ "HSM violation: HSM_ST: "
+ "ERR/DF w/ BUSY/DRQ\n");
qc->err_mask |= AC_ERR_HSM;
+ }
/* ata_pio_sectors() might change the
* state to HSM_ST_LAST. so, the state
diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c
index 496edaf..8caaa77 100644
--- a/drivers/ata/libata-eh.c
+++ b/drivers/ata/libata-eh.c
@@ -1394,6 +1394,8 @@ static void ata_eh_analyze_serror(struct ata_link *link)
action |= ATA_EH_SOFTRESET;
}
if (serror & SERR_PROTOCOL) {
+ ata_link_printk(link, KERN_WARNING,
+ "HSM violation: analyze_serror: SERR_PROTOCOL\n");
err_mask |= AC_ERR_HSM;
action |= ATA_EH_SOFTRESET;
}
@@ -1504,6 +1506,8 @@ static unsigned int ata_eh_analyze_tf(struct ata_queued_cmd *qc,
u8 stat = tf->command, err = tf->feature;
if ((stat & (ATA_BUSY | ATA_DRQ | ATA_DRDY)) != ATA_DRDY) {
+ ata_dev_printk(qc->dev, KERN_WARNING,
+ "HSM violation: eh_analyze_tf: BUSY|DRQ\n");
qc->err_mask |= AC_ERR_HSM;
return ATA_EH_SOFTRESET;
}
diff --git a/drivers/ata/libata.h b/drivers/ata/libata.h
