* No subject
@ 2013-06-19 10:57 Ben Dooks
2013-06-19 10:57 ` [PATCH] atags_to_fdt: take into account root's #size and #cells Ben Dooks
0 siblings, 1 reply; 5+ messages in thread
From: Ben Dooks @ 2013-06-19 10:57 UTC (permalink / raw)
To: linux-arm-kernel
I found this one whilst testing versatile-express with a -rc kernel
and was considering if this is the correct way to fix this.
[PATCH] atags_to_fdt: take into account root's #size and #cells
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH] atags_to_fdt: take into account root's #size and #cells
2013-06-19 10:57 No subject Ben Dooks
@ 2013-06-19 10:57 ` Ben Dooks
2013-06-19 11:11 ` Thomas Petazzoni
2013-06-19 11:12 ` Gregory CLEMENT
0 siblings, 2 replies; 5+ messages in thread
From: Ben Dooks @ 2013-06-19 10:57 UTC (permalink / raw)
To: linux-arm-kernel
When creating entries for /memory, the root #size and #cells must be taken
into account otherwise the node will not be recongnized when the OF code
scans the nodes.
On some systems, the default is now to be 2 for each of #size and #cells
which means that appending a kernrel and having it update from ATAGS does
not work properly.
---
arch/arm/boot/compressed/atags_to_fdt.c | 22 +++++++++++++++++++++-
1 file changed, 21 insertions(+), 1 deletion(-)
diff --git a/arch/arm/boot/compressed/atags_to_fdt.c b/arch/arm/boot/compressed/atags_to_fdt.c
index 74e193e..da36840 100644
--- a/arch/arm/boot/compressed/atags_to_fdt.c
+++ b/arch/arm/boot/compressed/atags_to_fdt.c
@@ -1,5 +1,6 @@
#include <asm/setup.h>
#include <libfdt.h>
+#include <linux/of.h>
#if defined(CONFIG_ARM_ATAG_DTB_COMPAT_CMDLINE_EXTEND)
#define do_extend_cmdline 1
@@ -95,9 +96,12 @@ static void merge_fdt_bootargs(void *fdt, const char *fdt_cmdline)
int atags_to_fdt(void *atag_list, void *fdt, int total_space)
{
struct tag *atag = atag_list;
- uint32_t mem_reg_property[2 * NR_BANKS];
+ uint32_t mem_reg_property[2 * 2 * NR_BANKS];
+ const __be32 *prop;
int memcount = 0;
int ret;
+ u32 dt_root_size_cells;
+ u32 dt_root_addr_cells;
/* make sure we've got an aligned pointer */
if ((u32)atag_list & 0x3)
@@ -118,6 +122,17 @@ int atags_to_fdt(void *atag_list, void *fdt, int total_space)
if (ret < 0)
return ret;
+ dt_root_size_cells = OF_ROOT_NODE_SIZE_CELLS_DEFAULT;
+ dt_root_addr_cells = OF_ROOT_NODE_ADDR_CELLS_DEFAULT;
+
+ prop = getprop(fdt, "/", "#size-cells", NULL);
+ if (prop)
+ dt_root_size_cells = be32_to_cpup(prop);
+
+ prop = getprop(fdt, "/", "#address-cells", NULL);
+ if (prop)
+ dt_root_addr_cells = be32_to_cpup(prop);
+
for_each_tag(atag, atag_list) {
if (atag->hdr.tag == cpu_to_atag32(ATAG_CMDLINE)) {
/* Append the ATAGS command line to the device tree
@@ -137,7 +152,12 @@ int atags_to_fdt(void *atag_list, void *fdt, int total_space)
continue;
if (!atag32_to_cpu(atag->u.mem.size))
continue;
+
+ for (ret = 0; ret < dt_root_addr_cells - 1; ret++)
+ mem_reg_property[memcount++] = cpu_to_fdt32(0);
mem_reg_property[memcount++] = cpu_to_fdt32(atag32_to_cpu(atag->u.mem.start));
+ for (ret = 0; ret < dt_root_size_cells - 1; ret++)
+ mem_reg_property[memcount++] = cpu_to_fdt32(0);
mem_reg_property[memcount++] = cpu_to_fdt32(atag32_to_cpu(atag->u.mem.size));
} else if (atag->hdr.tag == cpu_to_atag32(ATAG_INITRD2)) {
uint32_t initrd_start, initrd_size;
--
1.7.10.4
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH] atags_to_fdt: take into account root's #size and #cells
2013-06-19 10:57 ` [PATCH] atags_to_fdt: take into account root's #size and #cells Ben Dooks
@ 2013-06-19 11:11 ` Thomas Petazzoni
2013-06-19 11:42 ` Ben Dooks
2013-06-19 11:12 ` Gregory CLEMENT
1 sibling, 1 reply; 5+ messages in thread
From: Thomas Petazzoni @ 2013-06-19 11:11 UTC (permalink / raw)
To: linux-arm-kernel
Dear Ben Dooks,
On Wed, 19 Jun 2013 11:57:51 +0100, Ben Dooks wrote:
> When creating entries for /memory, the root #size and #cells must be taken
> into account otherwise the node will not be recongnized when the OF code
> scans the nodes.
>
> On some systems, the default is now to be 2 for each of #size and #cells
> which means that appending a kernrel and having it update from ATAGS does
> not work properly.
> ---
> arch/arm/boot/compressed/atags_to_fdt.c | 22 +++++++++++++++++++++-
> 1 file changed, 21 insertions(+), 1 deletion(-)
Isn't this the same
https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/commit/?id=faefd550c45d8d314e8f260f21565320355c947f
which is already in linux-next?
Best regards,
Thomas
--
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH] atags_to_fdt: take into account root's #size and #cells
2013-06-19 10:57 ` [PATCH] atags_to_fdt: take into account root's #size and #cells Ben Dooks
2013-06-19 11:11 ` Thomas Petazzoni
@ 2013-06-19 11:12 ` Gregory CLEMENT
1 sibling, 0 replies; 5+ messages in thread
From: Gregory CLEMENT @ 2013-06-19 11:12 UTC (permalink / raw)
To: linux-arm-kernel
On 06/19/2013 12:57 PM, Ben Dooks wrote:
> When creating entries for /memory, the root #size and #cells must be taken
> into account otherwise the node will not be recongnized when the OF code
> scans the nodes.
>
> On some systems, the default is now to be 2 for each of #size and #cells
> which means that appending a kernrel and having it update from ATAGS does
> not work properly.
Hi Ben,
There is already a fix for it which is applied in Russell King's git tree, see
http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=7722/1
Regards,
> ---
> arch/arm/boot/compressed/atags_to_fdt.c | 22 +++++++++++++++++++++-
> 1 file changed, 21 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm/boot/compressed/atags_to_fdt.c b/arch/arm/boot/compressed/atags_to_fdt.c
> index 74e193e..da36840 100644
> --- a/arch/arm/boot/compressed/atags_to_fdt.c
> +++ b/arch/arm/boot/compressed/atags_to_fdt.c
> @@ -1,5 +1,6 @@
> #include <asm/setup.h>
> #include <libfdt.h>
> +#include <linux/of.h>
>
> #if defined(CONFIG_ARM_ATAG_DTB_COMPAT_CMDLINE_EXTEND)
> #define do_extend_cmdline 1
> @@ -95,9 +96,12 @@ static void merge_fdt_bootargs(void *fdt, const char *fdt_cmdline)
> int atags_to_fdt(void *atag_list, void *fdt, int total_space)
> {
> struct tag *atag = atag_list;
> - uint32_t mem_reg_property[2 * NR_BANKS];
> + uint32_t mem_reg_property[2 * 2 * NR_BANKS];
> + const __be32 *prop;
> int memcount = 0;
> int ret;
> + u32 dt_root_size_cells;
> + u32 dt_root_addr_cells;
>
> /* make sure we've got an aligned pointer */
> if ((u32)atag_list & 0x3)
> @@ -118,6 +122,17 @@ int atags_to_fdt(void *atag_list, void *fdt, int total_space)
> if (ret < 0)
> return ret;
>
> + dt_root_size_cells = OF_ROOT_NODE_SIZE_CELLS_DEFAULT;
> + dt_root_addr_cells = OF_ROOT_NODE_ADDR_CELLS_DEFAULT;
> +
> + prop = getprop(fdt, "/", "#size-cells", NULL);
> + if (prop)
> + dt_root_size_cells = be32_to_cpup(prop);
> +
> + prop = getprop(fdt, "/", "#address-cells", NULL);
> + if (prop)
> + dt_root_addr_cells = be32_to_cpup(prop);
> +
> for_each_tag(atag, atag_list) {
> if (atag->hdr.tag == cpu_to_atag32(ATAG_CMDLINE)) {
> /* Append the ATAGS command line to the device tree
> @@ -137,7 +152,12 @@ int atags_to_fdt(void *atag_list, void *fdt, int total_space)
> continue;
> if (!atag32_to_cpu(atag->u.mem.size))
> continue;
> +
> + for (ret = 0; ret < dt_root_addr_cells - 1; ret++)
> + mem_reg_property[memcount++] = cpu_to_fdt32(0);
> mem_reg_property[memcount++] = cpu_to_fdt32(atag32_to_cpu(atag->u.mem.start));
> + for (ret = 0; ret < dt_root_size_cells - 1; ret++)
> + mem_reg_property[memcount++] = cpu_to_fdt32(0);
> mem_reg_property[memcount++] = cpu_to_fdt32(atag32_to_cpu(atag->u.mem.size));
> } else if (atag->hdr.tag == cpu_to_atag32(ATAG_INITRD2)) {
> uint32_t initrd_start, initrd_size;
>
--
Gregory Clement, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH] atags_to_fdt: take into account root's #size and #cells
2013-06-19 11:11 ` Thomas Petazzoni
@ 2013-06-19 11:42 ` Ben Dooks
0 siblings, 0 replies; 5+ messages in thread
From: Ben Dooks @ 2013-06-19 11:42 UTC (permalink / raw)
To: linux-arm-kernel
On 19/06/13 12:11, Thomas Petazzoni wrote:
> Dear Ben Dooks,
>
> On Wed, 19 Jun 2013 11:57:51 +0100, Ben Dooks wrote:
>> When creating entries for /memory, the root #size and #cells must be taken
>> into account otherwise the node will not be recongnized when the OF code
>> scans the nodes.
>>
>> On some systems, the default is now to be 2 for each of #size and #cells
>> which means that appending a kernrel and having it update from ATAGS does
>> not work properly.
>> ---
>> arch/arm/boot/compressed/atags_to_fdt.c | 22 +++++++++++++++++++++-
>> 1 file changed, 21 insertions(+), 1 deletion(-)
>
> Isn't this the same
> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/commit/?id=faefd550c45d8d314e8f260f21565320355c947f
> which is already in linux-next?
Ah, it is almost the same although it only takes into
account the #size-cells but not the number of them.
--
Ben Dooks http://www.codethink.co.uk/
Senior Engineer Codethink - Providing Genius
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2013-06-19 11:42 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-06-19 10:57 No subject Ben Dooks
2013-06-19 10:57 ` [PATCH] atags_to_fdt: take into account root's #size and #cells Ben Dooks
2013-06-19 11:11 ` Thomas Petazzoni
2013-06-19 11:42 ` Ben Dooks
2013-06-19 11:12 ` Gregory CLEMENT
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).