summaryrefslogtreecommitdiffstats
path: root/drivers/ata/libata-eh.c
diff options
context:
space:
mode:
authorTejun Heo <tj@kernel.org>2009-11-19 15:36:45 +0900
committerJeff Garzik <jgarzik@redhat.com>2009-12-03 02:46:35 -0500
commit6013efd8860bf15c1f86f365332642cfe557152f (patch)
tree9d7fe6d92ed9e067cda3b2e62d69f50175450e97 /drivers/ata/libata-eh.c
parentfd6c29e3dec9e44ecbcba3c57efa08af70a10f1e (diff)
downloadlinux-6013efd8860bf15c1f86f365332642cfe557152f.tar.bz2
libata: retry failed FLUSH if device didn't fail it
If ATA device failed FLUSH, it means that the device failed to write out some amount of data and the error needs to be reported to upper layers. As retries can't recover the lost data, FLUSH failures need to be reported immediately in general. However, if FLUSH fails due to transmission errors, the FLUSH needs to be retried; otherwise, filesystems may switch to RO mode and/or raid array may drop a drive for a random transmission glitch. This condition can be rather easily reproduced on certain ahci controllers which go through a PHY event after powersave mode switch + ext4 combination. Powersave mode switch is often closely followed by flush from the filesystem failing the FLUSH with ATA bus error which makes the filesystem code believe that data is lost and drop to RO mode. This was reported in the following bugzilla bug. http://bugzilla.kernel.org/show_bug.cgi?id=14543 This patch makes libata EH retry FLUSH if it wasn't failed by the device. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Andrey Vihrov <andrey.vihrov@gmail.com> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Diffstat (limited to 'drivers/ata/libata-eh.c')
-rw-r--r--drivers/ata/libata-eh.c94
1 files changed, 94 insertions, 0 deletions
diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c
index bba2ae5df1c2..0ea97c942ced 100644
--- a/drivers/ata/libata-eh.c
+++ b/drivers/ata/libata-eh.c
@@ -110,6 +110,13 @@ static const unsigned long ata_eh_identify_timeouts[] = {
ULONG_MAX,
};
+static const unsigned long ata_eh_flush_timeouts[] = {
+ 15000, /* be generous with flush */
+ 15000, /* ditto */
+ 30000, /* and even more generous */
+ ULONG_MAX,
+};
+
static const unsigned long ata_eh_other_timeouts[] = {
5000, /* same rationale as identify timeout */
10000, /* ditto */
@@ -147,6 +154,8 @@ ata_eh_cmd_timeout_table[ATA_EH_CMD_TIMEOUT_TABLE_SIZE] = {
.timeouts = ata_eh_other_timeouts, },
{ .commands = CMDS(ATA_CMD_INIT_DEV_PARAMS),
.timeouts = ata_eh_other_timeouts, },
+ { .commands = CMDS(ATA_CMD_FLUSH, ATA_CMD_FLUSH_EXT),
+ .timeouts = ata_eh_flush_timeouts },
};
#undef CMDS
@@ -3112,6 +3121,82 @@ static int atapi_eh_clear_ua(struct ata_device *dev)
return 0;
}
+/**
+ * ata_eh_maybe_retry_flush - Retry FLUSH if necessary
+ * @dev: ATA device which may need FLUSH retry
+ *
+ * If @dev failed FLUSH, it needs to be reported upper layer
+ * immediately as it means that @dev failed to remap and already
+ * lost at least a sector and further FLUSH retrials won't make
+ * any difference to the lost sector. However, if FLUSH failed
+ * for other reasons, for example transmission error, FLUSH needs
+ * to be retried.
+ *
+ * This function determines whether FLUSH failure retry is
+ * necessary and performs it if so.
+ *
+ * RETURNS:
+ * 0 if EH can continue, -errno if EH needs to be repeated.
+ */
+static int ata_eh_maybe_retry_flush(struct ata_device *dev)
+{
+ struct ata_link *link = dev->link;
+ struct ata_port *ap = link->ap;
+ struct ata_queued_cmd *qc;
+ struct ata_taskfile tf;
+ unsigned int err_mask;
+ int rc = 0;
+
+ /* did flush fail for this device? */
+ if (!ata_tag_valid(link->active_tag))
+ return 0;
+
+ qc = __ata_qc_from_tag(ap, link->active_tag);
+ if (qc->dev != dev || (qc->tf.command != ATA_CMD_FLUSH_EXT &&
+ qc->tf.command != ATA_CMD_FLUSH))
+ return 0;
+
+ /* if the device failed it, it should be reported to upper layers */
+ if (qc->err_mask & AC_ERR_DEV)
+ return 0;
+
+ /* flush failed for some other reason, give it another shot */
+ ata_tf_init(dev, &tf);
+
+ tf.command = qc->tf.command;
+ tf.flags |= ATA_TFLAG_DEVICE;
+ tf.protocol = ATA_PROT_NODATA;
+
+ ata_dev_printk(dev, KERN_WARNING, "retrying FLUSH 0x%x Emask 0x%x\n",
+ tf.command, qc->err_mask);
+
+ err_mask = ata_exec_internal(dev, &tf, NULL, DMA_NONE, NULL, 0, 0);
+ if (!err_mask) {
+ /*
+ * FLUSH is complete but there's no way to
+ * successfully complete a failed command from EH.
+ * Making sure retry is allowed at least once and
+ * retrying it should do the trick - whatever was in
+ * the cache is already on the platter and this won't
+ * cause infinite loop.
+ */
+ qc->scsicmd->allowed = max(qc->scsicmd->allowed, 1);
+ } else {
+ ata_dev_printk(dev, KERN_WARNING, "FLUSH failed Emask 0x%x\n",
+ err_mask);
+ rc = -EIO;
+
+ /* if device failed it, report it to upper layers */
+ if (err_mask & AC_ERR_DEV) {
+ qc->err_mask |= AC_ERR_DEV;
+ qc->result_tf = tf;
+ if (!(ap->pflags & ATA_PFLAG_FROZEN))
+ rc = 0;
+ }
+ }
+ return rc;
+}
+
static int ata_link_nr_enabled(struct ata_link *link)
{
struct ata_device *dev;
@@ -3455,6 +3540,15 @@ int ata_eh_recover(struct ata_port *ap, ata_prereset_fn_t prereset,
}
}
+ /* retry flush if necessary */
+ ata_for_each_dev(dev, link, ALL) {
+ if (dev->class != ATA_DEV_ATA)
+ continue;
+ rc = ata_eh_maybe_retry_flush(dev);
+ if (rc)
+ goto dev_fail;
+ }
+
/* configure link power saving */
if (ehc->i.action & ATA_EH_LPM)
ata_for_each_dev(dev, link, ALL)