summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorTejun Heo <tj@kernel.org>2009-01-29 20:31:34 +0900
committerJeff Garzik <jgarzik@redhat.com>2009-02-02 23:03:34 -0500
commitc2c7a89c5eabaea8c0c2aa0c1069e510144513ab (patch)
tree4cec432ebd36a103e366d56d2d62221674df5a99
parenta07d499b4759881db1359dd8812eecd00b0e0a28 (diff)
downloadlwn-c2c7a89c5eabaea8c0c2aa0c1069e510144513ab.tar.gz
lwn-c2c7a89c5eabaea8c0c2aa0c1069e510144513ab.zip
libata: improve probe failure handling
When link is flaky at high speed, it isn't uncommon for a device to repeatedly fail probing sequence early after successfully negotiating high link speed. This often leads to consecutive hotplug events without successful probing. This patch improves libata EH such that it remembers probing trials and if there have been more than two unsuccessful trials in the past 60 seconds, slows down link speed to 1.5Gbps. As link speed negotiation is the duty of the PHY layer proper, the goal of this fallback mechanism is to provide the last resort when everything else fails, which unfortunately happens not too infrequently, so no fancy 6->3->1.5 speeding down or highest successful transmission speed seen kind of logics (yet). Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
-rw-r--r--drivers/ata/libata-eh.c38
1 files changed, 38 insertions, 0 deletions
diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c
index 685509bc7ff0..caa1e9fde3c9 100644
--- a/drivers/ata/libata-eh.c
+++ b/drivers/ata/libata-eh.c
@@ -82,6 +82,10 @@ enum {
ATA_EH_FASTDRAIN_INTERVAL = 3000,
ATA_EH_UA_TRIES = 5,
+
+ /* probe speed down parameters, see ata_eh_schedule_probe() */
+ ATA_EH_PROBE_TRIAL_INTERVAL = 60000, /* 1 min */
+ ATA_EH_PROBE_TRIALS = 2,
};
/* The following table determines how we sequence resets. Each entry
@@ -2975,9 +2979,24 @@ static int ata_eh_skip_recovery(struct ata_link *link)
return 1;
}
+static int ata_count_probe_trials_cb(struct ata_ering_entry *ent, void *void_arg)
+{
+ u64 interval = msecs_to_jiffies(ATA_EH_PROBE_TRIAL_INTERVAL);
+ u64 now = get_jiffies_64();
+ int *trials = void_arg;
+
+ if (ent->timestamp < now - min(now, interval))
+ return -1;
+
+ (*trials)++;
+ return 0;
+}
+
static int ata_eh_schedule_probe(struct ata_device *dev)
{
struct ata_eh_context *ehc = &dev->link->eh_context;
+ struct ata_link *link = ata_dev_phys_link(dev);
+ int trials = 0;
if (!(ehc->i.probe_mask & (1 << dev->devno)) ||
(ehc->did_probe_mask & (1 << dev->devno)))
@@ -2990,6 +3009,25 @@ static int ata_eh_schedule_probe(struct ata_device *dev)
ehc->saved_xfer_mode[dev->devno] = 0;
ehc->saved_ncq_enabled &= ~(1 << dev->devno);
+ /* Record and count probe trials on the ering. The specific
+ * error mask used is irrelevant. Because a successful device
+ * detection clears the ering, this count accumulates only if
+ * there are consecutive failed probes.
+ *
+ * If the count is equal to or higher than ATA_EH_PROBE_TRIALS
+ * in the last ATA_EH_PROBE_TRIAL_INTERVAL, link speed is
+ * forced to 1.5Gbps.
+ *
+ * This is to work around cases where failed link speed
+ * negotiation results in device misdetection leading to
+ * infinite DEVXCHG or PHRDY CHG events.
+ */
+ ata_ering_record(&dev->ering, 0, AC_ERR_OTHER);
+ ata_ering_map(&dev->ering, ata_count_probe_trials_cb, &trials);
+
+ if (trials > ATA_EH_PROBE_TRIALS)
+ sata_down_spd_limit(link, 1);
+
return 1;
}