libata: fix a bunch of PMP related problems

The @online out parameter is supposed to set to true iff link is online and reset succeeded as advertised in the function description and callers are coded expecting that. However, sata_link_reset() didn't behave this way on device readiness test failure. Fix it.

---
 drivers/ata/libata-core.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)

index 927b692..c6c316f 100644
@@ -3653,9 +3653,13 @@ int sata_link_hardreset(struct ata_link *link, const unsigned long *timing,
 	if (check_ready)
 		rc = ata_wait_ready(link, deadline, check_ready);
  out:
-	if (rc && rc != -EAGAIN)
+	if (rc && rc != -EAGAIN) {
+		/* online is set iff link is online && reset succeeded */
+		if (online)
+			*online = false;
 		ata_link_printk(link, KERN_ERR,
 				"COMRESET failed (errno=%d)\n", rc);
+	}
 	DPRINTK("EXIT, rc=%d\n", rc);
 	return rc;
 }

1.5.2.4

Reorganize ata_eh_reset() such that @prereset() is called even when no reset method is available and if block is used instead of goto to skip actual reset. This makes no reset case behave better (readiness wait) and future changes easier.

---
 drivers/ata/libata-eh.c |  102 ++++++++++++++++++++++++-----------------------
 1 files changed, 52 insertions(+), 50 deletions(-)

index 62e0331..a34adc2 100644
@@ -2098,7 +2098,9 @@ int ata_eh_reset(struct ata_link *link, int classify,
 	u32 sstatus;
 	int rc;
 
-	/* about to reset */
+	/*
+	 * Prepare to reset
+	 */
 	spin_lock_irqsave(ap->lock, flags);
 	ap->pflags |= ATA_PFLAG_RESETTING;
 	spin_unlock_irqrestore(ap->lock, flags);

@@ -2124,16 +2126,8 @@ int ata_eh_reset(struct ata_link *link, int classify,
 			ap->ops->set_piomode(ap, dev);
 	}
 
-	if (!softreset && !hardreset) {
-		if (verbose)
-			ata_link_printk(link, KERN_INFO, "no reset method "
-					"available, skipping reset\n");
-		if (!(lflags & ATA_LFLAG_ASSUME_CLASS))
-			lflags |= ATA_LFLAG_ASSUME_ATA;
-		goto done;
-	}
-
 	/* prefer hardreset */
+	reset = NULL;
 	ehc->i.action &= ~ATA_EH_RESET;
 	if (hardreset) {
 		reset = hardreset;

@@ -2141,11 +2135,6 @@ int ata_eh_reset(struct ata_link *link, int classify,
 	} else if (softreset) {
 		reset = softreset;
 		ehc->i.action = ATA_EH_SOFTRESET;
-	} else {
-		ata_link_printk(link, KERN_ERR, "BUG: no reset method, "
-				"please report to [email protected]\n");
-		dump_stack();
-		return -EINVAL;
 	}
 
 	if (prereset) {

@@ -2165,55 +2154,68 @@ int ata_eh_reset(struct ata_link *link, int classify,
 					"prereset failed (errno=%d)\n", rc);
 			goto out;
 		}
-	}
 
-	/* prereset() might have cleared ATA_EH_RESET */
-	if (!(ehc->i.action & ATA_EH_RESET)) {
-		/* prereset told us not to reset, bang classes and return */
-		ata_link_for_each_dev(dev, link)
-			classes[dev->devno] = ATA_DEV_NONE;
-		rc = 0;
-		goto out;
+		/* prereset() might have cleared ATA_EH_RESET.  If so,
+		 * bang classes and return.
+		 */
+		if (reset && !(ehc->i.action & ATA_EH_RESET)) {
+			ata_link_for_each_dev(dev, link)
+				classes[dev->devno] = ATA_DEV_NONE;
+			rc = 0;
+			goto out;
+		}
 	}
 
  retry:
+	/*
+	 * Perform reset
+	 */
 	deadline = jiffies + ata_eh_reset_timeouts[try++];
 
-	/* shut up during boot probing */
-	if (verbose)
-		ata_link_printk(link, KERN_INFO, "%s resetting link\n",
-				reset == softreset ? "soft" : "hard");
+	if (reset) {
+		if (verbose)
+			ata_link_printk(link, KERN_INFO, "%s resetting link\n",
+					reset == softreset ? "soft" : "hard");
 
-	/* mark that this EH session started with reset */
-	if (reset == hardreset)
-		ehc->i.flags |= ATA_EHI_DID_HARDRESET;
-	else
-		ehc->i.flags |= ATA_EHI_DID_SOFTRESET;
+		/* mark that this EH session started with reset */
+		if (reset == hardreset)
+			ehc->i.flags |= ATA_EHI_DID_HARDRESET;
+		else
+			ehc->i.flags |= ATA_EHI_DID_SOFTRESET;
 
-	rc = ata_do_reset(link, reset, classes, deadline);
+		rc = ata_do_reset(link, reset, classes, deadline);
 
-	if (reset == hardreset &&
-	    ata_eh_followup_srst_needed(link, rc, classify, classes)) {
-		/* okay, let's do follow-up softreset */
-		reset = softreset;
+		if (reset == hardreset &&
+		    ata_eh_followup_srst_needed(link, rc, classify, classes)) {
+			/* okay, let's do follow-up softreset */
+			reset = softreset;
 
-		if (!reset) {
-			ata_link_printk(link, KERN_ERR,
-					"follow-up softreset required "
-					"but no softreset avaliable\n");
-			rc = -EINVAL;
-			goto fail;
+			if (!reset) {
+				ata_link_printk(link, KERN_ERR,
+						"follow-up softreset required "
+						"but no softreset avaliable\n");
+				rc = -EINVAL;
+				goto fail;
+			}
+
+			ata_eh_about_to_do(link, NULL, ATA_EH_RESET);
+			rc = ata_do_reset(link, reset, classes, deadline);
 		}
 
-		ata_eh_about_to_do(link, NULL, ATA_EH_RESET);
-		rc = ata_do_reset(link, reset, classes, deadline);
+		/* -EAGAIN can happen if we skipped followup SRST */
+		if (rc && rc != -EAGAIN)
+			goto fail;
+	} else {
+		if (verbose)
+			ata_link_printk(link, KERN_INFO, "no reset method "
+					"available, skipping reset\n");
+		if (!(lflags & ATA_LFLAG_ASSUME_CLASS))
+			lflags |= ATA_LFLAG_ASSUME_ATA;
 	}
 
-	/* -EAGAIN can happen if we skipped followup SRST */
-	if (rc && rc != -EAGAIN)
-		goto fail;
-
- done:
+	/*
+	 * Post-reset processing
+	 */
 	ata_link_for_each_dev(dev, link) {
 		/* After the reset, the device state is PIO 0 and the
 		 * controller state is undefined.  Reset also wakes up

1.5.2.4

Hello,

This patchset is started from investigating sil24 - mv4140 PMP detection / NCQ problems. It soon turned out there are quite a number of issues lurking in the current PMP implementation.

* We still have a hole during reset where hotplug can go unnoticed.

* PMP r/w timeout is too short. It's currently 250ms. This is from SIMG PMPs and was mainly to avoid longish sequential multiple timeouts when access PMP regs via polled commands which is not the case anymore and mv 4140 and SIMG4726 needs more time occasionally.

* SDB Notify is not masked during fan-out port resets if PMP hasn't been reset in that pass. This causes PMP register access failure because SDB Notify is received while PMP register write is in progress for fan-out reset. I missed this because most of my testing was before PMP register access was converted to IRQ driven.

* ata_lpm_schedule() is called during probe right after SCSI scan is complete which schedules EH immediately, reliably triggering above two on certain hardware combinations. This by itself should be okay but makes other problems more painful.

* sil3124/32 and mv4140 combination for some reason can't do NCQ reliably. I don't know why. I'll blacklist it for the time being and contact both companies about this.

* recovered errors shouldn't trigger resets.

* SIMG4726 config device still is a real pain in the ass.

This patchset contains ten patches addressing all the above issues. Although this patchset is rather large, it's basically bug fixes.

The patchset fixes all hotplug problems I can reproduce including the hotplug problems on inic162x and JMB ahcis. The only remaining issue is that sometimes device detection gets delayed by 30sec IDENTIFY timeout. I'll prep a patchset to make EH command timeouts more intelligent (for #upstream).

The sad part is that with PMP in the mix and the current host controller designs, we have inherent race condition during reset (no reliable way to wait for the initial D2H FIS after hardresetting fan-out port, so SRST races with D2H FIS) and this seems to lead to problems during or after reset sometimes. We have to resort to intelligently timed retries to make it work bearably.

Thanks.

 drivers/ata/libata-core.c |   40 ++++----
 drivers/ata/libata-eh.c   |  207 +++++++++++++++++++++++++---------------------
 drivers/ata/libata-pmp.c  |   44 ++++-----
 drivers/ata/libata-scsi.c |    6 -
 drivers/ata/sata_sil24.c  |   11 ++
 include/linux/libata.h    |    4
 6 files changed, 170 insertions(+), 142 deletions(-)

--
tejun

Previously reset freeze/thaw handling lived outside of ata_eh_reset() mainly because the original PMP reset code needed the port frozen while resetting all the fan-out ports, which is no longer the case.

This patch moves freeze/thaw handling into ata_eh_reset(). @prereset() and @postreset() are now called w/o freezing the port although @prereset() an be called frozen if the port is frozen prior to entering ata_eh_reset().

This makes code simpler and will help removing hotplug event related races.

---
 drivers/ata/libata-eh.c  |   46 ++++++++++++++++++----------------------------
 drivers/ata/libata-pmp.c |    4 ----
 2 files changed, 18 insertions(+), 32 deletions(-)

index a34adc2..06a92c5 100644
@@ -2170,6 +2170,9 @@ int ata_eh_reset(struct ata_link *link, int classify,
 	/*
 	 * Perform reset
 	 */
+	if (ata_is_host_link(link))
+		ata_eh_freeze_port(ap);
+
 	deadline = jiffies + ata_eh_reset_timeouts[try++];
 
 	if (reset) {

@@ -2238,6 +2241,10 @@ int ata_eh_reset(struct ata_link *link, int classify,
 	if (sata_scr_read(link, SCR_STATUS, &sstatus) == 0)
 		link->sata_spd = (sstatus >> 4) & 0xf;
 
+	/* thaw the port */
+	if (ata_is_host_link(link))
+		ata_eh_thaw_port(ap);
+
 	if (postreset)
 		postreset(link, classes);
 

@@ -2589,7 +2596,7 @@ int ata_eh_recover(struct ata_port *ap, ata_prereset_fn_t prereset,
 	struct ata_link *link;
 	struct ata_device *dev;
 	int nr_failed_devs, nr_disabled_devs;
-	int reset, rc;
+	int rc;
 	unsigned long flags;
 
 	DPRINTK("ENTER\n");

@@ -2632,7 +2639,6 @@ int ata_eh_recover(struct ata_port *ap, ata_prereset_fn_t prereset,
 	rc = 0;
 	nr_failed_devs = 0;
 	nr_disabled_devs = 0;
-	reset = 0;
 
 	/* if UNLOADING, finish immediately */
 	if (ap->pflags & ATA_PFLAG_UNLOADING)

@@ -2646,40 +2652,24 @@ int ata_eh_recover(struct ata_port *ap, ata_prereset_fn_t prereset,
 		if (ata_eh_skip_recovery(link))
 			ehc->i.action = 0;
 
-		/* do we need to reset? */
-		if (ehc->i.action & ATA_EH_RESET)
-			reset = 1;
-
 		ata_link_for_each_dev(dev, link)
 			ehc->classes[dev->devno] = ATA_DEV_UNKNOWN;
 	}
 
 	/* reset */
-	if (reset) {
-		/* if PMP is attached, this function only deals with
-		 * downstream links, port should stay thawed.
-		 */
-		if (!sata_pmp_attached(ap))
-			ata_eh_freeze_port(ap);
-
-		ata_port_for_each_link(link, ap) {
-			struct ata_eh_context *ehc = &link->eh_context;
+	ata_port_for_each_link(link, ap) {
+		struct ata_eh_context *ehc = &link->eh_context;
 
-			if (!(ehc->i.action & ATA_EH_RESET))
-				continue;
+		if (!(ehc->i.action & ATA_EH_RESET))
+			continue;
 
-			rc = ata_eh_reset(link, ata_link_nr_vacant(link),
-					  prereset, softreset, hardreset,
-					  postreset);
-			if (rc) {
-				ata_link_printk(link, KERN_ERR,
-						"reset failed, giving up\n");
-				goto out;
-			}
+		rc = ata_eh_reset(link, ata_link_nr_vacant(link),
+				  prereset, softreset, hardreset, postreset);
+		if (rc) {
+			ata_link_printk(link, KERN_ERR,
+					"reset failed, giving up\n");
+			goto out;
 		}
-
-		if (!sata_pmp_attached(ap))
-			ata_eh_thaw_port(ap);
 	}
 
 	/* the rest */

index ff1822a..f3ad024 100644
@@ -700,8 +700,6 @@ static int sata_pmp_eh_recover_pmp(struct ata_port *ap,
 	if (ehc->i.action & ATA_EH_RESET) {
 		struct ata_link *tlink;
 
-		ata_eh_freeze_port(ap);
-
 		/* reset */
 		rc = ata_eh_reset(link, 0, prereset, softreset, hardreset,
 				  postreset);

@@ -711,8 +709,6 @@ static int sata_pmp_eh_recover_pmp(struct ata_port *ap,
 			goto fail;
 		}
 
-		ata_eh_thaw_port(ap);
-
 		/* PMP is reset, SErrors cannot be trusted, scan all */
 		ata_port_for_each_link(tlink, ap) {
 			struct ata_eh_context *ehc = &tlink->eh_context;

1.5.2.4

Originally, whole reset processing was done while the port is frozen and SError was cleared during @postreset(). This had two race conditions. 1: hotplug could occur after reset but before SError is cleared and libata won't know about it. 2: hotplug could occur after all the reset is complete but before the port is thawed. As all events are cleared on thaw, the hotplug event would be lost.

Commit ac371987a81c61c2efbd6931245cdcaf43baad89 kills the first race by clearing SError during link resume but before link onlineness test. However, this doesn't fix race #2 and in some cases clearing SError after SRST is a good idea.

This patch solves this problem by cross checking link onlineness with classification result after SError is cleared and port is thawed. Reset is retried if link is online but all devices attached to the link are unknown. As all devices will be revalidated, this one-way check is enough to ensure that all devices are detected and revalidated reliably.

This, luckily, also fixes the cases where host controller returns bogus status while harddrive is spinning up after hotplug making classification run before the device sends the first FIS and thus causes misdetection.

Low level drivers can bypass the logic by setting class explicitly to ATA_DEV_NONE if ever necessary (currently none requires this).

---
 drivers/ata/libata-core.c |   21 +++++++-----------
 drivers/ata/libata-eh.c   |   52 ++++++++++++++++++++++++++++++++++++--------
 2 files changed, 50 insertions(+), 23 deletions(-)

index c6c316f..ffc689d 100644
@@ -3490,22 +3490,11 @@ int sata_link_resume(struct ata_link *link, const unsigned long *params,
 	if ((rc = sata_link_debounce(link, params, deadline)))
 		return rc;
 
-	/* Clear SError.  PMP and some host PHYs require this to
-	 * operate and clearing should be done before checking PHY
-	 * online status to avoid race condition (hotplugging between
-	 * link resume and status check).
-	 */
+	/* clear SError, some PHYs require this even for SRST to work */
 	if (!(rc = sata_scr_read(link, SCR_ERROR, &serror)))
 		rc = sata_scr_write(link, SCR_ERROR, serror);
-	if (rc == 0 || rc == -EINVAL) {
-		unsigned long flags;
 
-		spin_lock_irqsave(link->ap->lock, flags);
-		link->eh_info.serror = 0;
-		spin_unlock_irqrestore(link->ap->lock, flags);
-		rc = 0;
-	}
-	return rc;
+	return rc != -EINVAL ? rc : 0;
 }
 
 /**

@@ -3704,8 +3693,14 @@ int sata_std_hardreset(struct ata_link *link, unsigned int *class,
  */
 void ata_std_postreset(struct ata_link *link, unsigned int *classes)
 {
+	u32 serror;
+
 	DPRINTK("ENTER\n");
 
+	/* reset complete, clear SError */
+	if (!sata_scr_read(link, SCR_ERROR, &serror))
+		sata_scr_write(link, SCR_ERROR, serror);
+
 	/* print link status */
 	sata_print_link_status(link);
 

1.5.2.4

There's no reason to schedule LPM action after probing is complete causing another EH iteration. Just schedule it together with probing itself.

---
 drivers/ata/libata-core.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

index ffc689d..a12a27e 100644
@@ -5615,7 +5615,7 @@ int ata_host_register(struct ata_host *host, struct scsi_host_template *sht)
 			spin_lock_irqsave(ap->lock, flags);
 
 			ehi->probe_mask |= ATA_ALL_DEVICES;
-			ehi->action |= ATA_EH_RESET;
+			ehi->action |= ATA_EH_RESET | ATA_EH_LPM;
 			ehi->flags |= ATA_EHI_NO_AUTOPSY | ATA_EHI_QUIET;
 
 			ap->pflags &= ~ATA_PFLAG_INITIALIZING;

@@ -5648,7 +5648,6 @@ int ata_host_register(struct ata_host *host, struct scsi_host_template *sht)
 		struct ata_port *ap = host->ports[i];
 
 		ata_scsi_scan_host(ap, 1);
-		ata_lpm_schedule(ap, ap->pm_policy);
 	}
 
 	return 0;

1.5.2.4

This timeout was set low because previously PMP register access was done via polling and register access timeouts could stack up. This is no longer the case. One timeout will make all following accesses fail immediately.

In rare cases both marvell and SIMG PMPs need almost a second. Bump it to 3s.

While at it, rename it to SATA_PMP_RW_TIMEOUT. It's not specific to SCR access.

---
 drivers/ata/libata-pmp.c |    4 ++--
 include/linux/libata.h   |    2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

index f3ad024..04a486a 100644
@@ -48,7 +48,7 @@ static unsigned int sata_pmp_read(struct ata_link *link, int reg, u32 *r_val)
 	tf.device = link->pmp;
 
 	err_mask = ata_exec_internal(pmp_dev, &tf, NULL, DMA_NONE, NULL, 0,
-				     SATA_PMP_SCR_TIMEOUT);
+				     SATA_PMP_RW_TIMEOUT);
 	if (err_mask)
 		return err_mask;
 

@@ -88,7 +88,7 @@ static unsigned int sata_pmp_write(struct ata_link *link, int reg, u32 val)
 	tf.lbah = (val >> 24) & 0xff;
 
 	return ata_exec_internal(pmp_dev, &tf, NULL, DMA_NONE, NULL, 0,
-				 SATA_PMP_SCR_TIMEOUT);
+				 SATA_PMP_RW_TIMEOUT);
 }
 
 /**

index 0f17643..95e4169 100644
@@ -341,7 +341,7 @@ enum {
 	ATA_EH_PMP_TRIES	= 5,
 	ATA_EH_PMP_LINK_TRIES	= 3,
 
-	SATA_PMP_SCR_TIMEOUT	= 250,
+	SATA_PMP_RW_TIMEOUT	= 3000,		/* PMP read/write timeout */
 
 	/* Horkage types. May be set by libata or controller on drives
 	   (some horkage may be drive/controller pair dependant */

1.5.2.4

No reason to get overzealous about recovered comm and data errors. Some PHYs habitually sets them w/o no good reason and being draconian about these soft error conditions doesn't seem to help anybody.

If need ever rises, we might need to add soft PHY error condition, say AC_ERR_MAYBE_ATA_BUS and use it only to determine whether speed down is necessary but I don't think that's very likely to happen. It's far more likely we'll get timeouts or fatal transmission errors if recovered errors are so prominent that they hamper operation.

---
 drivers/ata/libata-eh.c |    7 +------
 1 files changed, 1 insertions(+), 6 deletions(-)

index 751dad0..7894d83 100644
@@ -1308,12 +1308,7 @@ static void ata_eh_analyze_serror(struct ata_link *link)
 	unsigned int err_mask = 0, action = 0;
 	u32 hotplug_mask;
 
-	if (serror & SERR_PERSISTENT) {
-		err_mask |= AC_ERR_ATA_BUS;
-		action |= ATA_EH_RESET;
-	}
-	if (serror &
-	    (SERR_DATA_RECOVERED | SERR_COMM_RECOVERED | SERR_DATA)) {
+	if (serror & (SERR_PERSISTENT | SERR_DATA)) {
 		err_mask |= AC_ERR_ATA_BUS;
 		action |= ATA_EH_RESET;
 	}

1.5.2.4

PMP notification during reset can make some controllers fail reset processing and needs to be turned off during resets. PMP attach and full-revalidation path did this via sata_pmp_configure() but the quick revalidation wasn't. Move the notification disable code right above fan-out port recovery so that it's always turned off.

This fixes obscure reset failures.

---
 drivers/ata/libata-pmp.c |   36 ++++++++++++++++++++----------------
 1 files changed, 20 insertions(+), 16 deletions(-)

index 04a486a..0f9386d 100644
@@ -257,19 +257,6 @@ static int sata_pmp_configure(struct ata_device *dev, int print_info)
 		goto fail;
 	}
 
-	/* turn off notification till fan-out ports are reset and configured */
-	if (gscr[SATA_PMP_GSCR_FEAT_EN] & SATA_PMP_FEAT_NOTIFY) {
-		gscr[SATA_PMP_GSCR_FEAT_EN] &= ~SATA_PMP_FEAT_NOTIFY;
-
-		err_mask = sata_pmp_write(dev->link, SATA_PMP_GSCR_FEAT_EN,
-					  gscr[SATA_PMP_GSCR_FEAT_EN]);
-		if (err_mask) {
-			rc = -EIO;
-			reason = "failed to write GSCR_FEAT_EN";
-			goto fail;
-		}
-	}
-
 	if (print_info) {
 		ata_dev_printk(dev, KERN_INFO, "Port Multiplier %s, "
 			       "0x%04x:0x%04x r%d, %d ports, feat 0x%x/0x%x\n",

@@ -860,6 +847,7 @@ static int sata_pmp_eh_recover(struct ata_port *ap)
 	struct ata_link *pmp_link = &ap->link;
 	struct ata_device *pmp_dev = pmp_link->device;
 	struct ata_eh_context *pmp_ehc = &pmp_link->eh_context;
+	u32 *gscr = pmp_dev->gscr;
 	struct ata_link *link;
 	struct ata_device *dev;
 	unsigned int err_mask;

@@ -897,6 +885,22 @@ static int sata_pmp_eh_recover(struct ata_port *ap)
 	if (rc)
 		goto pmp_fail;
 
+	/* PHY event notification can disturb reset and other recovery
+	 * operations.  Turn it off.
+	 */
+	if (gscr[SATA_PMP_GSCR_FEAT_EN] & SATA_PMP_FEAT_NOTIFY) {
+		gscr[SATA_PMP_GSCR_FEAT_EN] &= ~SATA_PMP_FEAT_NOTIFY;
+
+		err_mask = sata_pmp_write(pmp_link, SATA_PMP_GSCR_FEAT_EN,
+					  gscr[SATA_PMP_GSCR_FEAT_EN]);
+		if (err_mask) {
+			ata_link_printk(pmp_link, KERN_WARNING,
+				"failed to disable NOTIFY (err_mask=0x%x)\n",
+				err_mask);
+			goto pmp_fail;
+		}
+	}
+
 	/* handle disabled links */
 	rc = sata_pmp_eh_handle_disabled_links(ap);
 	if (rc)

@@ -919,10 +923,10 @@ static int sata_pmp_eh_recover(struct ata_port *ap)
 
 	/* enable notification */
 	if (pmp_dev->flags & ATA_DFLAG_AN) {
-		pmp_dev->gscr[SATA_PMP_GSCR_FEAT_EN] |= SATA_PMP_FEAT_NOTIFY;
+		gscr[SATA_PMP_GSCR_FEAT_EN] |= SATA_PMP_FEAT_NOTIFY;
 
-		err_mask = sata_pmp_write(pmp_dev->link, SATA_PMP_GSCR_FEAT_EN,
-					  pmp_dev->gscr[SATA_PMP_GSCR_FEAT_EN]);
+		err_mask = sata_pmp_write(pmp_link, SATA_PMP_GSCR_FEAT_EN,
+					  gscr[SATA_PMP_GSCR_FEAT_EN]);
 		if (err_mask) {
 			ata_dev_printk(pmp_dev, KERN_ERR, "failed to write "
 				       "PMP_FEAT_EN (Emask=0x%x)\n", err_mask);

1.5.2.4

When 4140 PMP is attached to sil24, NCQ commands to fan out port 1 and 2 (0 based) often stall if commands are in progress to other ports. I've tried a number of things but can't tell what's going on. It never happens w/ ahci and reportedly sata_mv which can issue NCQ commands to multiple devices simultaneously like sil24 does.

Disable NCQ for devices behind 4140 PMP for the time being.

---
 drivers/ata/sata_sil24.c |   11 +++++++++++
 1 files changed, 11 insertions(+), 0 deletions(-)

index 27a1101..8ee6b5b 100644
@@ -899,14 +899,25 @@ static bool sil24_qc_fill_rtf(struct ata_queued_cmd *qc)
 
 static void sil24_pmp_attach(struct ata_port *ap)
 {
+	u32 *gscr = ap->link.device->gscr;
+
 	sil24_config_pmp(ap, 1);
 	sil24_init_port(ap);
+
+	if (sata_pmp_gscr_vendor(gscr) == 0x11ab &&
+	    sata_pmp_gscr_devid(gscr) == 0x4140) {
+		ata_port_printk(ap, KERN_INFO,
+			"disabling NCQ support due to sil24-mv4140 quirk\n");
+		ap->flags &= ~ATA_FLAG_NCQ;
+	}
 }
 
 static void sil24_pmp_detach(struct ata_port *ap)
 {
 	sil24_init_port(ap);
 	sil24_config_pmp(ap, 0);
+
+	ap->flags |= ATA_FLAG_NCQ;
 }
 
 static int sil24_pmp_hardreset(struct ata_link *link, unsigned int *class,

1.5.2.4

I was hoping ATA_HORKAGE_NODMA | ATA_HORKAGE_SKIP_PM could keep it happy but no even this doesn't work under certain configurations and it's not like we can do anything useful with the cofig device anyway. Replace ATA_HORKAGE_SKIP_PM with ATA_HORKAGE_DISABLE and use it for the config device. This makes the device completely ignored by libata.

---
 drivers/ata/libata-core.c |   10 ++++++++--
 drivers/ata/libata-scsi.c |    6 ------
 include/linux/libata.h    |    2 +-
 3 files changed, 9 insertions(+), 9 deletions(-)

index a12a27e..3c89f20 100644
@@ -2126,6 +2126,13 @@ int ata_dev_configure(struct ata_device *dev)
 	dev->horkage |= ata_dev_blacklisted(dev);
 	ata_force_horkage(dev);
 
+	if (dev->horkage & ATA_HORKAGE_DISABLE) {
+		ata_dev_printk(dev, KERN_INFO,
+			       "unsupported device, disabling\n");
+		ata_dev_disable(dev);
+		return 0;
+	}
+
 	/* let ACPI work its magic */
 	rc = ata_acpi_on_devcfg(dev);
 	if (rc)

@@ -3893,8 +3900,7 @@ static const struct ata_blacklist_entry ata_device_blacklist [] = {
 	{ "SAMSUNG CD-ROM SN-124", "N001",	ATA_HORKAGE_NODMA },
 	{ "Seagate STT20000A", NULL,		ATA_HORKAGE_NODMA },
 	/* Odd clown on sil3726/4726 PMPs */
-	{ "Config  Disk",	NULL,		ATA_HORKAGE_NODMA |
-						ATA_HORKAGE_SKIP_PM },
+	{ "Config  Disk",	NULL,		ATA_HORKAGE_DISABLE },
 
 	/* Weird ATAPI devices */
 	{ "TORiSAN DVD-ROM DRD-N216", NULL,	ATA_HORKAGE_MAX_SEC_128 },

index 3ce4392..aeb6e01 100644
@@ -1082,12 +1082,6 @@ static unsigned int ata_scsi_start_stop_xlat(struct ata_queued_cmd *qc)
 	if (((cdb[4] >> 4) & 0xf) != 0)
 		goto invalid_fld;       /* power conditions not supported */
 
-	if (qc->dev->horkage & ATA_HORKAGE_SKIP_PM) {
-		/* the device lacks PM support, finish without doing anything */
-		scmd->result = SAM_STAT_GOOD;
-		return 1;
-	}
-
 	if (cdb[4] & 0x1) {
 		tf->nsect = 1;	/* 1 sector, lba=0 */
 

index 95e4169..95a2000 100644
@@ -351,7 +351,7 @@ enum {
 	ATA_HORKAGE_NONCQ	= (1 << 2),	/* Don't use NCQ */
 	ATA_HORKAGE_MAX_SEC_128	= (1 << 3),	/* Limit max sects to 128 */
 	ATA_HORKAGE_BROKEN_HPA	= (1 << 4),	/* Broken HPA */
-	ATA_HORKAGE_SKIP_PM	= (1 << 5),	/* Skip PM operations */
+	ATA_HORKAGE_DISABLE	= (1 << 5),	/* Disable it */
 	ATA_HORKAGE_HPA_SIZE	= (1 << 6),	/* native size off by one */
 	ATA_HORKAGE_IPM		= (1 << 7),	/* Link PM problems */
 	ATA_HORKAGE_IVB		= (1 << 8),	/* cbl det validity bit bugs */

1.5.2.4

Git tree available at...

http://git.kernel.org/?p=linux/kernel/git/tj/libata-dev.git;a=shortlog;h=hotplug-fixes
git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata-dev.git hotplug-fixes

Brian, the original thread is

http://thread.gmane.org/gmane.linux.ide/31572

and it might fix your device detection problem. Please test. If you don't git, patch against v2.6.26-rc2 is available at...

http://master.kernel.org/~tj/v2.6.26-rc2-to-hotplug-fixes.patch

Thanks.

tejun

..

Mmm.. I wonder if sil24 is somehow triggering the "proprietary switching" mode of the 4140 -- this mode allows command-based-switching controllers to actually do something similar to FIS-based switching with the 4140.

But I don't know how the 4140 decides to enter that mode..

I'll ping the Marvell folks again on that.

Cheers

applied 1-10

Tejun-

The hotplug fixes didn't solve the problem. After adding some "logic monitoring" statements, I found that the successful discovery of the drives depends on the response from sil24_exec_polled_cmd.

If the call from ata_wait_register is 327680, the drives are recognized; if the response is 262144 the drives have timed out.

I've also noticed that the drives respond to the EH entering with DevExch vs. PHY RDY changed.

Bottom line, the drives seem to be recognized almost every time I attach them to the PMP after the EH has completed processing the empty PMP. If the drives are attached to the PMP before the PMP is attached to the controller, the drives are almost never recognized (EH times out).

I've attached some more dmesg output (with my logic monitoring statements included).

-Brian

-----Original Message-----
From: Tejun Heo [mailto:[email protected]]
Sent: Sunday, May 18, 2008 12:29 PM
To: [email protected]; [email protected]
Cc: [email protected]; Brian & Chamaigne Scamman
Subject: Re: [PATCHSET #upstream-fixes] git tree available

Git tree available at...

http://git.kernel.org/?p=linux/kernel/git/tj/libata-dev.git;a=shortlog;h=hot
plug-fixes
git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata-dev.git
hotplug-fixes

Brian, the original thread is

http://thread.gmane.org/gmane.linux.ide/31572

and it might fix your device detection problem. Please test. If you don't git, patch against v2.6.26-rc2 is available at...

http://master.kernel.org/~tj/v2.6.26-rc2-to-hotplug-fixes.patch

Thanks.

tejun

ata3: exception Emask 0x10 SAct 0x0 SErr 0x0 action 0xe frozen
ata3: irq stat 0x00b40090, PHY RDY changed
ata3: BJS: Prep for recovery
ata3: BJS: dev 0
ata3: BJS: Prep for EH
ata3: BJS: Reset
ata3: hard resetting link
ata3: BJS: follow-up softreset required
ata3: BJS: orig timeout (7812)
ata3: BJS: IRQ port (11862016)
ata3: BJS: rc=0
ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3: BJS: The Rest...
ata3.15: Port Multiplier 1.1, 0x1095:0x3726 r23, 6 ports, feat 0x1/0x9 ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.00: hard resetting link
ata3.00: BJS: follow-up softreset required
ata3.00: BJS: orig timeout (9680)
ata3: BJS: IRQ port (262144)
ata3: BJS: IRQ !complete
ata3.00: softreset failed (timeout)
ata3.00: BJS: rc=-5
ata3.15: qc timeout (cmd 0xe4)
ata3.00: failed to read SCR 0 (Emask=0x4)
ata3.00: reset failed, giving up
ata3.15: hard resetting link
ata3.15: BJS: follow-up softreset required
ata3.15: BJS: orig timeout (7816)
ata3: BJS: IRQ port (11862016)
ata3.15: BJS: rc=0
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.00: hard resetting link
ata3.00: BJS: follow-up softreset required
ata3.00: BJS: orig timeout (9680)
ata3: BJS: IRQ port (262144)
ata3: BJS: IRQ complete
ata3.00: softreset failed (timeout)
ata3.00: BJS: rc=-5
ata3.15: qc timeout (cmd 0xe4)
ata3.00: failed to read SCR 0 (Emask=0x4)
ata3.00: reset failed, giving up
ata3.15: hard resetting link
ata3.15: BJS: follow-up softreset required
ata3.15: BJS: orig timeout (7816)
ata3: BJS: IRQ port (11862016)
ata3.15: BJS: rc=0
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.00: hard resetting link
ata3.00: BJS: follow-up softreset required
ata3.00: BJS: orig timeout (9680)
ata3: BJS: IRQ port (262144)
ata3: BJS: IRQ !complete
ata3.00: softreset failed (timeout)
ata3.00: BJS: rc=-5
ata3.15: qc timeout (cmd 0xe4)
ata3.00: failed to read SCR 0 (Emask=0x4)
ata3.00: reset failed, giving up
ata3.00: failed to recover link after 3 tries, disabling
ata3: failed to recover PMP, retrying in 5 secs
ata3.15: hard resetting link
ata3.15: BJS: follow-up softreset required
ata3.15: BJS: orig timeout (7816)
ata3: BJS: IRQ port (11862016)
ata3.15: BJS: rc=0
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.01: BJS: Reset
ata3.01: hard resetting link
ata3.01: SATA link down (SStatus 0 SControl 320)
ata3.02: BJS: Reset
ata3.02: hard resetting link
ata3.02: SATA link down (SStatus 0 SControl 320)
ata3.03: BJS: Reset
ata3.03: hard resetting link
ata3.03: SATA link down (SStatus 0 SControl 320)
ata3.04: BJS: Reset
ata3.04: hard resetting link
ata3.04: SATA link down (SStatus 0 SControl 320)
ata3.05: BJS: Reset
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: BJS: The Rest...
ata3.01: BJS: The Rest...
ata3.02: BJS: The Rest...
ata3.03: BJS: The Rest...
ata3.04: BJS: The Rest...
ata3.05: BJS: The Rest...
ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.01: BJS: Reset
ata3.02: BJS: Reset
ata3.03: BJS: Reset
ata3.04: BJS: Reset
ata3.05: BJS: Reset
ata3.00: BJS: The Rest...
ata3.01: BJS: The Rest...
ata3.02: BJS: The Rest...
ata3.03: BJS: The Rest...
ata3.04: BJS: The Rest...
ata3.05: BJS: The Rest...
ata3: EH complete
ata3.00: exception Emask 0xl0 SAct 0x0 SErr 0x4010000 action 0xf
ata3: SError: { PHYRdyChg DevExch }
ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.00: hard resetting link
ata3.00: SATA link down (SStatus 0 SControl 320)
ata3.01: BJS: Reset
ata3.02: BJS: Reset
ata3.03: BJS: Reset
ata3.04: BJS: Reset
ata3.05: BJS: Reset
ata3.00: BJS: The Rest...
ata3.01: BJS: The Rest...
ata3.02: BJS: The Rest...
ata3.03: BJS: The Rest...
ata3.04: BJS: The Rest...
ata3.05: BJS: The Rest...
ata3: EH complete
ata3.00: exception Emask 0xl0 SAct 0x0 SErr 0x4040000 action 0xf
ata3: SError: { CommWake DevExch }
ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.00: hard resetting link
ata3.00: BJS: follow-up softreset required
ata3.00: BJS: orig timeout (9276)
ata3: BJS: IRQ port (327680)
ata3.00: BJS: rc=0
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: BJS: Reset
ata3.02: BJS: Reset
ata3.03: BJS: Reset
ata3.04: BJS: Reset
ata3.05: BJS: Reset
ata3.00: BJS: The Rest...
ata3.00: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.00: 127923200 sectors, multi 1: LBA
ata3.00: applying bridge limits
ata3.00: configured for UDMA/100
ata3.01: BJS: The Rest...
ata3.02: BJS: The Rest...
ata3.03: BJS: The Rest...
ata3.04: BJS: The Rest...
ata3.05: BJS: The Rest...
ata3: EH complete
scsi 2:0:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:0:0:0: [sda] 127923200 512-byte hardware sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA sd 2:0:0:0: [sda] 127923200 512-byte hardware sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA
sda: sda1
sd 2:0:0:0: [sda] Attached SCSI disk
sd 2:0:0:0: Attached scsi generic sg0 type 0

ata3: exception Emask 0x10 SAct 0x0 SErr 0x10000 action 0xe frozen
ata3: irq_stat 0x00a40080, device exchanged
ata3: SError: { PHYRdyChg }
ata3: BJS: Prep for recovery
ata3: BJS: dev 0
ata3: BJS: Prep for EH
ata3: BJS: Reset
ata3: hard resetting link
ata3: BJS: follow-up softreset required
ata3: BJS: orig timeout (7816)
ata3: BJS: IRQ port (11862016)
ata3: BJS: rc=0
ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3: BJS: The Rest...
ata3.15: Port Multiplier 1.1, 0x1095:0x3726 r23, 6 ports, feat 0x1/0x9 ata3.00: BJS: Prep for recovery
ata3.00: BJS: dev 0
ata3.01: BJS: Prep for recovery
ata3.01: BJS: dev 0
ata3.02: BJS: Prep for recovery
ata3.02: BJS: dev 0
ata3.03: BJS: Prep for recovery
ata3.03: BJS: dev 0
ata3.04: BJS: Prep for recovery
ata3.04: BJS: dev 0
ata3.05: BJS: Prep for recovery
ata3.05: BJS: dev 0
ata3.00: BJS: Prep for EH
ata3.01: BJS: Prep for EH
ata3.02: BJS: Prep for EH
ata3.03: BJS: Prep for EH
ata3.04: BJS: Prep for EH
ata3.05: BJS: Prep for EH
ata3.00: BJS: Reset
ata3.00: hard resetting link
ata3.00: BJS: follow-up softreset required
ata3.00: BJS: orig timeout (9680)
ata3: BJS: IRQ port (327680)
ata3.00: BJS: rc=0
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: BJS: Reset
ata3.01: hard resetting link
ata3.01: SATA link down (SStatus 0 SControl 320)
ata3.02: BJS: Reset
ata3.02: hard resetting link
ata3.02: SATA link down (SStatus 0 SControl 320)
ata3.03: BJS: Reset
ata3.03: hard resetting link
ata3.03: SATA link down (SStatus 0 SControl 320)
ata3.04: BJS: Reset
ata3.04: hard resetting link
ata3.04: SATA link down (SStatus 0 SControl 320)
ata3.05: BJS: Reset
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: BJS: The Rest...
ata3.00: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.00: 127923200 sectors, multi 1: LBA
ata3.00: applying bridge limits
ata3.00: configured for UDMA/100
ata3.01: BJS: The Rest...
ata3.02: BJS: The Rest. .
ata3.03: BJS: The Rest...
ata3.04: BJS: The Rest. .
ata3.05: BJS: The Rest...
ata3: EH complete
scsi 2:0:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:0:0:0: [sda] 127923200 512-byte hardware Sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA sd 2:0:0:0: [sda] 127923200 512-byte hardware sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA
sda: sdal
sd 2:0:0:0: [sda] Attached SCSI disk
sd 2:0:0:0: Attached scsi generic sg0 type 0

..

hp-polling is beginning to look better and better.. again. :)

Can you please describe what you did exactly? sil3726/4726 has some problems when its first fan out port goes online and offline while it's powered up. You want to keep it occupied at all times.

Also, does putting sslep(5) right before followup-SRST help?

hp-polling wouldn't really help here. What's failing is not hotplug event detection, the reset protocol is failing.

tejun

Can you please test whether the attached patch fixes the detection problem?

tejun

index 3374ec5..b65db30 100644
@@ -322,9 +322,12 @@ static void sata_pmp_quirks(struct ata_port *ap)
 	if (vendor == 0x1095 && devid == 0x3726) {
 		/* sil3726 quirks */
 		ata_port_for_each_link(link, ap) {
-			/* class code report is unreliable */
+			/* Class code report is unreliable and SRST
+			 * times out under certain configurations.
+			 */
 			if (link->pmp < 5)
-				link->flags |= ATA_LFLAG_ASSUME_ATA;
+				link->flags |= ATA_LFLAG_NO_SRST |
+					       ATA_LFLAG_ASSUME_ATA;
 
 			/* port 5 is for SEMB device and it doesn't like SRST */
 			if (link->pmp == 5)

Here's the current setup:
- PMP powered via computer's power supply; on with computer
- SSD and bridges powered via computer's power supply; on with computer
- FC9 with new 2.6.26-rc2-hotplug-patches installed and booted
- Once computer is booted, login, open a terminal and dmesg -c For the case which fails:
- plug SSD drive/bridge into PMP port 0
- plug eSata cable into PMP
- plug eSata cable into eSata port on controller
- watch the softreset timeout
For the case which works:
- plug eSata cable into PMP
- plug eSata cable into eSata port on controller
- watch the PMP become recognized
- plug SSD drive/bridge into PMP port 0
- watch the SSD drive become recognized

Steps above are repeatable. Previously I was applying power to the PMP while the drives were attached (PMP, bridges and drives all receiving power at the same time) and then plugging in the eSata cable, but this case always failed as well.

I'll look at adding the sslep line and the new patch (independently).

-Brian

Please try the patch. I think that should work.

tejun

Tejun-
The patch worked great; see dmesg log below...
I didn't test the sslep(5) command since the patch worked; do you still need me to test it?

One remaining issue is the "failed to IDENTIFY (I/O error, err mask=0xll)" message. If the PMP, bridges and drives are powered up before I insert the eSATA cable in to the controller, I don't receive the message. The message appears when only when I have the eSATA cable attached to the controller and then simultaneously power on the PMP and drives. In the attached dmesg log, the drive is eventually recognized; however there have been a couple of times when the drives aren't recognized. An eSATA cable remove/re-insert fixes the problem.

-Brian

ata3: exception Emask 0x10 SAct 0x0 SErr 0x0 action 0xe frozen
ata3: irq_stat 0x00b40090, PHY RDY changed
ata3: hard resetting link
ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.15: Port Multiplier 1.1, 0x1095:0x3726 r23, 6 ports, feat 0xl/0x9 ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: hard resetting link
ata3.01: SATA link down (SStatus 0 SControl 320)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.03: hard resetting link
ata3.03: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.04: hard resetting link
ata3.04: SATA link down (SStatus 0 SControl 320)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: failed to IDENTIFY (I/O error, err mask=0xll)
ata3.15: hard resetting link
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.03: hard resetting link
ata3.03: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.00: 127923200 sectors, multi 1: LBA
ata3.00: applying bridge limits
ata3.00: configured for UDMA/100
ata3.02: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.02: 127923200 sectors, multi 1: LBA
ata3.02: applying bridge limits
ata3.02: configured for UDMA/100
ata3.03: failed to IDENTIFY (I/O error, err mask=0xll)
ata3.15: hard resetting link
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: hard resetting link
ata3.01: SATA link down (SStatus 0 SControl 320)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.03: hard resetting link
ata3.03: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.04: hard resetting link
ata3.04: SATA link down (SStatus 0 SControl 320)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: configured for UDMA/100
ata3.02: configured for UDMA/100
ata3.03: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.03: 127923200 sectors, multi 1: LBA
ata3.03: applying bridge limits
ata3.03: configured for UDMA/100
ata3: EH complete
scsi 2:0:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:0:0:0: [sda] 127923200 512-byte hardware sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA sd 2:0:0:0: [sda] 127923200 512-byte hardware sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA
sda: sda1
sd 2:0:0:0: [sda] Attached SCSI disk
sd 2:0:0:0: Attached scsi generic sg0 type 0
scsi 2:2:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:2:0:0: [sdb] 127923200 512-byte hardware sectors (65497 MB)
sd 2:2:0:0: [sdb] Write Protect is off
sd 2:2:0:0: [sdb] Mode Sense: 00 3a 00 00
sd 2:2:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA sd 2:2:0:0: [sdb] 127923200 512-byte hardware sectors (65497 MB)
sd 2:2:0:0: [sdb] Write Protect is off
sd 2:2:0:0: [sdb] Mode Sense: 00 3a 00 00
sd 2:2:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA
sdb: sdb1
sd 2:2:0:0: [sdb] Attached SCSI disk
sd 2:2:0:0: Attached scsi generic sg1 type 0
scsi 2:3:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:3:0:0: [sdc] 127923200 512-byte hardware sectors (65497 MB)
sd 2:3:0:0: [sdc] Write Protect is off
sd 2:3:0:0: [sdc] Mode Sense: 00 3a 00 00
sd 2:3:0:0: [sdc] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA sd 2:3:0:0: [sdc] 127923200 512-byte hardware sectors (65497 MB)
sd 2:3:0:0: [sdc] Write Protect is off
sd 2:3:0:0: [sdc] Mode Sense: 00 3a 00 00
sd 2:3:0:0: [sdc] Write cache: disabled, read cache: enabled, doesn�t support DPO or FUA
sdc: sdc1
sd 2:3:0:0: [sdc] Attached SCSI disk
sd 2:3:0:0: Attached scsi generic sg2 type 0

Nope.

Hmm... can you please post the failing log?

tejun

mask=0xll)"

the

message

and

log,

In my previous post, there are two cases where the drive wasn't initially recognized, but did respond after a PMP link reset (ata3.00 and ata3.03). I'll post a log today were one of the drives is never recognized (doesn't happen that often)

-Brian

mask=0xll)"

Here's a dmesg log where one of the drives is never recognized, but the others eventually are...

-Brian

ata3: exception Emask 0x10 SAct 0x0 SErr 0x0 action 0xe frozen
ata3: irq_stat 0x00b40090, PHY RDY changed
ata3: hard resetting link
ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.15: Port Multiplier 1.1, 0x1095:0x3726 r23, 6 ports, feat 0x1/0x9 ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: hard resetting link
ata3.01: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.03: hard resetting link
ata3.03: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.04: hard resetting link
ata3.04: SATA link down (SStatus 0 SControl 320)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: failed to IDENTIFY (I/O error, err_mask=0x11)
ata3.15: hard resetting link
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: hard resetting link
ata3.01: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.03: hard resetting link
ata3.03: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.00: 127923200 sectors, multi 1: LBA
ata3.00: applying bridge limits
ata3.00: configured for UDMA/100
ata3.01: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.01: 127923200 sectors, multi 1: LBA
ata3.01: applying bridge limits
ata3.01: configured for UDMA/100
ata3.02: ATA-6: Super Talent Tech, Rev 2.11, max UDMA/133
ata3.02: 127923200 sectors, multi 1: LBA
ata3.02: applying bridge limits
ata3.02: configured for UDMA/100
ata3.03: failed to IDENTIFY (I/O error, err_mask=0x11)
ata3.15: hard resetting link
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: hard resetting link
ata3.01: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.03: hard resetting link
ata3.03: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.04: hard resetting link
ata3.04: SATA link down (SStatus 0 SControl 320)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: configured for UDMA/100
ata3.01: configured for UDMA/100
ata3.02: configured for UDMA/100
ata3.03: failed to IDENTIFY (I/O error, err_mask=0x11)
ata3.15: hard resetting link
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: hard resetting link
ata3.01: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.03: hard resetting link
ata3.03: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: configured for UDMA/100
ata3.01: configured for UDMA/100
ata3.02: configured for UDMA/100
ata3.03: failed to IDENTIFY (I/O error, err_mask=0x11)
ata3.03: failed to recover link after 3 tries, disabling
ata3: failed to recover PMP, retrying in 5 secs
ata3.15: hard resetting link
ata3.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
ata3.00: hard resetting link
ata3.00: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.01: hard resetting link
ata3.01: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.02: hard resetting link
ata3.02: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata3.04: hard resetting link
ata3.04: SATA link down (SStatus 0 SControl 320)
ata3.05: hard resetting link
ata3.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata3.00: configured for UDMA/100
ata3.01: configured for UDMA/100
ata3.02: configured for UDMA/100
ata3: EH complete
scsi 2:0:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:0:0:0: [sda] 127923200 512-byte hardware sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA sd 2:0:0:0: [sda] 127923200 512-byte hardware sectors (65497 MB)
sd 2:0:0:0: [sda] Write Protect is off
sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
sda: sda1
sd 2:0:0:0: [sda] Attached SCSI disk
sd 2:0:0:0: Attached scsi generic sg0 type 0
scsi 2:1:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:1:0:0: [sdb] 127923200 512-byte hardware sectors (65497 MB)
sd 2:1:0:0: [sdb] Write Protect is off
sd 2:1:0:0: [sdb] Mode Sense: 00 3a 00 00
sd 2:1:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA sd 2:1:0:0: [sdb] 127923200 512-byte hardware sectors (65497 MB)
sd 2:1:0:0: [sdb] Write Protect is off
sd 2:1:0:0: [sdb] Mode Sense: 00 3a 00 00
sd 2:1:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
sdb: sdb1
sd 2:1:0:0: [sdb] Attached SCSI disk
sd 2:1:0:0: Attached scsi generic sg1 type 0
scsi 2:2:0:0: Direct-Access ATA Super Talent Tec Rev PQ: 0 ANSI: 5
sd 2:2:0:0: [sdc] 127923200 512-byte hardware sectors (65497 MB)
sd 2:2:0:0: [sdc] Write Protect is off
sd 2:2:0:0: [sdc] Mode Sense: 00 3a 00 00
sd 2:2:0:0: [sdc] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA sd 2:2:0:0: [sdc] 127923200 512-byte hardware sectors (65497 MB)
sd 2:2:0:0: [sdc] Write Protect is off
sd 2:2:0:0: [sdc] Mode Sense: 00 3a 00 00
sd 2:2:0:0: [sdc] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
sdc: sdc1
sd 2:2:0:0: [sdc] Attached SCSI disk
sd 2:2:0:0: Attached scsi generic sg2 type 0

Can you please put ssleep(5) right after the first ata_do_reset() call in drivers/ata/libata-eh.c::ata_eh_reset() and see whether the problem goes away?

Thanks.

tejun

This actually caused more problems. The "failed to IDENTIFY (I/O error, err_mask=0x11)" was still printed, plus there was a 5 second delay for each port on the PMP during each of the 3 attempts to read the drive. In the end, neither of the two SSD drives I tried ended up working. Changing ssleep(5) to msleep(5) didn't remove the IDENTIFY errors, but still allowed the drives to work.

Is there anyway to control the length of the reset pulse? I've heard that some of the SSD's require the reset pulse to be held longer than normal.

-Brian

For sata_sil24, it's determined by the controller. Hmmm... Please wait a bit, I have another thing to try.

tejun

Which didn't work out too well. :-(

I'm sorry but I'm out of ideas. I'm gonna ask SIMG about it. Do you mind being cc'd there?

Thanks.

tejun

Not at all.

-----Original Message-----
From: Tejun Heo [mailto:[email protected]]
Sent: Wednesday, May 28, 2008 11:07 PM
To: Brian & Chamaigne Scamman
Cc: 'Mark Lord'; [email protected]; [email protected]
Subject: Re: [PATCHSET #upstream-fixes] git tree available

Which didn't work out too well. :-(

I'm sorry but I'm out of ideas. I'm gonna ask SIMG about it. Do you mind being cc'd there?

Thanks.

tejun