Patchwork Ultra1 ESP detection problem

login
register
mail settings
Submitter David Miller
Date April 17, 2009, 11:10 a.m.
Message ID <20090417.041021.125958522.davem@davemloft.net>
Download mbox | patch
Permalink /patch/26119/
State Superseded
Delegated to: David Miller
Headers show

Comments

David Miller - April 17, 2009, 11:10 a.m.
From: David Miller <davem@davemloft.net>
Date: Sun, 15 Mar 2009 20:32:36 -0700 (PDT)

> Your change will make the 'qe' nodes match SBUS as a bus.  That's
> wrong, and we need to use na = 2 and ns = 2 for this case.
> 
> So this doesn't work as a mechanism to bypass intermediate parent
> nodes up to the SBUS.  You need to fix this while preserving the
> invariants and expectations of how this translation system works.
> 
> For now, perhaps we can add those use_1to1_mapping() checks to
> of_bus_sbus_match() as a quick fix that won't break other cases like
> the 'qec' mentioned above.

Ok, following through on this... Here is how I would like to
fix this bug.

sparc: Fix bus type probing for ESP and LE devices.

If there is a dummy "espdma" or "ledma" parent device above ESP scsi
or LE ethernet device nodes, we have to match the bus as SBUS.

Otherwise the address and size cell counts are wrong and we don't
calculate the final physical device resource values correctly at all.

Commit 5280267c1dddb8d413595b87dc406624bb497946 ("sparc: Fix handling
of LANCE and ESP parent nodes in of_device.c") was meant to fix this
problem, but that only influences the inner loop of
build_device_resources().  We need this logic to also kick in at the
beginning of build_device_resources() as well, when we make the first
attempt to determine the device's immediate parent bus type for 'reg'
property element extraction.

Based almost entirely upon a patch by Friedrich Oslage.

Signed-off-by: David S. Miller <davem@davemloft.net>

--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Meelis Roos - April 19, 2009, 9:16 a.m.
> Ok, following through on this... Here is how I would like to
> fix this bug.
> 
> sparc: Fix bus type probing for ESP and LE devices.
> 
> If there is a dummy "espdma" or "ledma" parent device above ESP scsi
> or LE ethernet device nodes, we have to match the bus as SBUS.

Unfortunately it does not work, onboard ESP is still not detected on my 
Ultra 1.
David Miller - April 19, 2009, 10:18 a.m.
From: Meelis Roos <mroos@linux.ee>
Date: Sun, 19 Apr 2009 12:16:21 +0300 (EEST)

>> Ok, following through on this... Here is how I would like to
>> fix this bug.
>> 
>> sparc: Fix bus type probing for ESP and LE devices.
>> 
>> If there is a dummy "espdma" or "ledma" parent device above ESP scsi
>> or LE ethernet device nodes, we have to match the bus as SBUS.
> 
> Unfortunately it does not work, onboard ESP is still not detected on my 
> Ultra 1.

Thanks for testing.  I'll try to figure out why it doesn't work.
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/arch/sparc/kernel/of_device_32.c b/arch/sparc/kernel/of_device_32.c
index 0a83bd7..a7164c5 100644
--- a/arch/sparc/kernel/of_device_32.c
+++ b/arch/sparc/kernel/of_device_32.c
@@ -246,8 +246,25 @@  static unsigned long of_bus_pci_get_flags(const u32 *addr, unsigned long flags)
 
 static int of_bus_sbus_match(struct device_node *np)
 {
-	return !strcmp(np->name, "sbus") ||
-		!strcmp(np->name, "sbi");
+	struct device_node *dp = np;
+
+	while (dp) {
+		if (!strcmp(np->name, "sbus") ||
+		    !strcmp(np->name, "sbi"))
+			return 1;
+
+		/* Have a look at use_1to1_mapping().  We're trying
+		 * to match SBUS if that's the top-level bus and we
+		 * don't have some intervening real bus that provides
+		 * ranges based translations.
+		 */
+		if (of_find_property(dp, "ranges", NULL) != NULL)
+			break;
+
+		dp = dp->parent;
+	}
+
+	return 0;
 }
 
 static void of_bus_sbus_count_cells(struct device_node *child,
diff --git a/arch/sparc/kernel/of_device_64.c b/arch/sparc/kernel/of_device_64.c
index 27381f1..734f7ba 100644
--- a/arch/sparc/kernel/of_device_64.c
+++ b/arch/sparc/kernel/of_device_64.c
@@ -300,8 +300,25 @@  static unsigned long of_bus_pci_get_flags(const u32 *addr, unsigned long flags)
 
 static int of_bus_sbus_match(struct device_node *np)
 {
-	return !strcmp(np->name, "sbus") ||
-		!strcmp(np->name, "sbi");
+	struct device_node *dp = np;
+
+	while (dp) {
+		if (!strcmp(np->name, "sbus") ||
+		    !strcmp(np->name, "sbi"))
+			return 1;
+
+		/* Have a look at use_1to1_mapping().  We're trying
+		 * to match SBUS if that's the top-level bus and we
+		 * don't have some intervening real bus that provides
+		 * ranges based translations.
+		 */
+		if (of_find_property(dp, "ranges", NULL) != NULL)
+			break;
+
+		dp = dp->parent;
+	}
+
+	return 0;
 }
 
 static void of_bus_sbus_count_cells(struct device_node *child,