Patchwork [v1,1/1] Fix 64bit Maple Host Bridge Address and Size Nodes

login
register
mail settings
Submitter Tiejun Chen
Date April 3, 2009, 10:46 a.m.
Message ID <1238755594-18869-2-git-send-email-tiejun.china@gmail.com>
Download mbox | patch
Permalink /patch/25571/
State Superseded
Headers show

Comments

Tiejun Chen - April 3, 2009, 10:46 a.m.
In the DTB tree created by firmware on some Maple 64bit targets, such as ATCA6101, 
these two properties, address&size, should be 2. But the actual corresponding values 
of host bridge node are set 1 incorrectly by the firmware, we have to provide one 
fixup function to fix that.

Signed-off-by: Tiejun Chen <tiejun.china@gmail.com>
---
 arch/powerpc/kernel/prom_init.c |   51 ++++++++++++++++++++++++++++++++++++++-
 1 files changed, 50 insertions(+), 1 deletions(-)
Segher Boessenkool - April 14, 2009, 8:11 a.m.
> In the DTB tree created by firmware on some Maple 64bit targets,  
> such as ATCA6101,
> these two properties, address&size, should be 2.

And they are.

> But the actual corresponding values
> of host bridge node are set 1 incorrectly by the firmware, we have  
> to provide one
> fixup function to fix that.

It is the "reg" property that is incorrectly encoded as-if
#address-cells and #size-cells were 1.

Please fix the commit message (and subject line).

> +	name = "/hostbridge@f8000000";
> +	hb = call_prom("finddevice", 1, 1, ADDR(name));

It's interesting that the firmware can find your node this way,
it won't match the "reg"!  Looks like there is more than one
bug in there.

Code looks good,

Reviewed-by: Segher Boessenkool <segher@kernel.crashing.org>


Segher
Tiejun Chen - April 14, 2009, 10:11 a.m.
Segher,

Thank you very much for your review.

I already re-send this patch as you expect and please check this:
[v2 PATCH 1/1] Fix 64bit Maple Host Bridge 'reg' property

Best Regards
Tiejun

2009/4/14 Segher Boessenkool <segher@kernel.crashing.org>

> In the DTB tree created by firmware on some Maple 64bit targets, such as
>> ATCA6101,
>> these two properties, address&size, should be 2.
>>
>
> And they are.
>
>  But the actual corresponding values
>> of host bridge node are set 1 incorrectly by the firmware, we have to
>> provide one
>> fixup function to fix that.
>>
>
> It is the "reg" property that is incorrectly encoded as-if
> #address-cells and #size-cells were 1.
>
> Please fix the commit message (and subject line).
>
>  +       name = "/hostbridge@f8000000";
>> +       hb = call_prom("finddevice", 1, 1, ADDR(name));
>>
>
> It's interesting that the firmware can find your node this way,
> it won't match the "reg"!  Looks like there is more than one
> bug in there.
>
> Code looks good,
>
> Reviewed-by: Segher Boessenkool <segher@kernel.crashing.org>
>
>
> Segher
>
>

Patch

diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
index 2445945..7848b45 100644
--- a/arch/powerpc/kernel/prom_init.c
+++ b/arch/powerpc/kernel/prom_init.c
@@ -1907,14 +1907,63 @@  static void __init flatten_device_tree(void)
 }
 
 #ifdef CONFIG_PPC_MAPLE
+/* On ATCA6101 64bit target host bridge parent node has specified address and
+ * size properties to be 2. But the actual "cell" value for host bridge node 
+ * is 1 since early MOTLoad internal bug. */
+static void __init fixup_device_tree_atca6101(void)
+{
+	phandle hb;
+	u32 hb_ranges[4];
+	u32 size_cell, addr_cell;
+	struct prom_t *_prom = &RELOC(prom);
+	char *name;
+
+	name = "/hostbridge@f8000000";
+	hb = call_prom("finddevice", 1, 1, ADDR(name));
+	if (!PHANDLE_VALID(hb))
+		return;
+
+	if (prom_getproplen(hb, "reg") != 8)
+		return;
+
+	if (prom_getprop(hb, "reg", hb_ranges, (sizeof(hb_ranges))/2) 
+		== PROM_ERROR)
+		return;
+
+	prom_getprop(_prom->root, "#address-cells", &addr_cell, sizeof(addr_cell));
+	prom_getprop(_prom->root, "#size-cells", &size_cell, sizeof(size_cell));
+	
+	if ((addr_cell != 2) || (size_cell != 2) || 
+		(hb_ranges[0] != 0xf8000000))
+		return;
+
+	prom_printf("Fixing up bogus HOSTBRIDGE reg on ATCA6101...\n");
+
+	hb_ranges[3] = hb_ranges[1];
+	hb_ranges[1] = hb_ranges[0];
+	hb_ranges[0] = hb_ranges[2] = 0;
+	prom_setprop(hb, name, "reg",
+	hb_ranges, sizeof(hb_ranges));
+}
 /* PIBS Version 1.05.0000 04/26/2005 has an incorrect /ht/isa/ranges property.
- * The values are bad, and it doesn't even have the right number of cells. */
+ * The values are bad, and it doesn't even have the right number of cells. 
+ * Additionally, the early MOTLoad generate incorrect address&size cells on 
+ * some Maple platform such as ATCA6101. */ 
 static void __init fixup_device_tree_maple(void)
 {
 	phandle isa;
 	u32 rloc = 0x01002000; /* IO space; PCI device = 4 */
 	u32 isa_ranges[6];
 	char *name;
+	u32 node;
+	char prop[64];
+	int model;
+	
+	/*  FIXME: This may be used for more Maple targets not only ATCA6101. */
+	node = call_prom("finddevice", 1, 1, ADDR("/"));
+	model = prom_getprop(node, "model", prop, sizeof(prop));
+	if (model != PROM_ERROR && (strcmp(prop, "Motorola,ATCA-6101") == 0))
+                fixup_device_tree_atca6101();
 
 	name = "/ht@0/isa@4";
 	isa = call_prom("finddevice", 1, 1, ADDR(name));