diff mbox

[U-Boot] m28evk: Use GENERATED_GBL_DATA_SIZE

Message ID 1325165738-9155-1-git-send-email-festevam@gmail.com
State Accepted
Commit 1084606c972ea5f1d89f69bdbd978b867d0ee521
Delegated to: Stefano Babic
Headers show

Commit Message

Fabio Estevam Dec. 29, 2011, 1:35 p.m. UTC
Use GENERATED_GBL_DATA_SIZE for calculating CONFIG_SYS_INIT_SP_OFFSET.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
---
 include/configs/m28evk.h |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

Comments

Marek Vasut Dec. 29, 2011, 3:22 p.m. UTC | #1
> Use GENERATED_GBL_DATA_SIZE for calculating CONFIG_SYS_INIT_SP_OFFSET.
> 
> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>

Maybe I'd be better to consolidate the i.MX28 config to a common config file and 
have only board specific stuff in the board-specific configs, which would in 
turn include the file?

M

> ---
>  include/configs/m28evk.h |    9 ++++++++-
>  1 files changed, 8 insertions(+), 1 deletions(-)
> 
> diff --git a/include/configs/m28evk.h b/include/configs/m28evk.h
> index 39c841f..2b4871e 100644
> --- a/include/configs/m28evk.h
> +++ b/include/configs/m28evk.h
> @@ -92,7 +92,14 @@
>  #define	CONFIG_SYS_MEMTEST_END		0x40400000	/* 4 MB RAM test 
*/
>  #define	CONFIG_SYS_SDRAM_BASE		PHYS_SDRAM_1
>  /* Point initial SP in SRAM so SPL can use it too. */
> -#define	CONFIG_SYS_INIT_SP_ADDR		0x00002000
> +
> +#define CONFIG_SYS_INIT_RAM_ADDR	0x00002000
> +#define CONFIG_SYS_INIT_RAM_SIZE	(128 * 1024)
> +
> +#define CONFIG_SYS_INIT_SP_OFFSET \
> +	(CONFIG_SYS_INIT_RAM_SIZE - GENERATED_GBL_DATA_SIZE)
> +#define CONFIG_SYS_INIT_SP_ADDR \
> +	(CONFIG_SYS_INIT_RAM_ADDR + CONFIG_SYS_INIT_SP_OFFSET)
>  /*
>   * We need to sacrifice first 4 bytes of RAM here to avoid triggering some
>   * strange BUG in ROM corrupting first 4 bytes of RAM when loading U-Boot
Fabio Estevam Dec. 29, 2011, 3:38 p.m. UTC | #2
On Thu, Dec 29, 2011 at 1:22 PM, Marek Vasut <marek.vasut@gmail.com> wrote:

> Maybe I'd be better to consolidate the i.MX28 config to a common config file and
> have only board specific stuff in the board-specific configs, which would in
> turn include the file?

Sounds like a good idea.

Regards,

Fabio Estevam
Stefano Babic Dec. 29, 2011, 3:47 p.m. UTC | #3
On 29/12/2011 16:22, Marek Vasut wrote:
>> Use GENERATED_GBL_DATA_SIZE for calculating CONFIG_SYS_INIT_SP_OFFSET.
>>
>> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
> 
> Maybe I'd be better to consolidate the i.MX28 config to a common config file and 
> have only board specific stuff in the board-specific configs, which would in 
> turn include the file?

But this issue is not strictly related only to MX28. The same is for any
SOC in u-boot, and most of CONFIG_ are duplicated in the configuration
files for boards having the same SOC.

At the moment, we do this only for slightly different boards (imx27lite
and magnesium, for example) or for boards having the same SOM.

Agree we can rationalize better, but it should be discussed if we plan
to introduce some hierarchy in the configuration files - the board
configuration file can include a SOC configuration file, that can
include a ARCH configuration file....

Apart of that, I think it is not bad to merge this Fabio's patch - it
makes the board consistent with the other ones.

Best regards,
Stefano Babic
Marek Vasut Dec. 29, 2011, 6:22 p.m. UTC | #4
> On 29/12/2011 16:22, Marek Vasut wrote:
> >> Use GENERATED_GBL_DATA_SIZE for calculating CONFIG_SYS_INIT_SP_OFFSET.
> >> 
> >> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
> > 
> > Maybe I'd be better to consolidate the i.MX28 config to a common config
> > file and have only board specific stuff in the board-specific configs,
> > which would in turn include the file?
> 
> But this issue is not strictly related only to MX28. The same is for any
> SOC in u-boot, and most of CONFIG_ are duplicated in the configuration
> files for boards having the same SOC.
> 
> At the moment, we do this only for slightly different boards (imx27lite
> and magnesium, for example) or for boards having the same SOM.
> 
> Agree we can rationalize better, but it should be discussed if we plan
> to introduce some hierarchy in the configuration files - the board
> configuration file can include a SOC configuration file, that can
> include a ARCH configuration file....
> 
> Apart of that, I think it is not bad to merge this Fabio's patch - it
> makes the board consistent with the other ones.

Fine by me, but I can't test it.

M

> 
> Best regards,
> Stefano Babic
Stefano Babic Jan. 4, 2012, 1:53 p.m. UTC | #5
On 29/12/2011 14:35, Fabio Estevam wrote:
> Use GENERATED_GBL_DATA_SIZE for calculating CONFIG_SYS_INIT_SP_OFFSET.
> 
> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
> ---

Applied to u-boot-imx, thanks.

Best regards,
Stefano Babic
Marek Vasut April 2, 2012, 4:17 a.m. UTC | #6
Dear Fabio Estevam,

I think this patch is wrong ... did I ack this one?

> Use GENERATED_GBL_DATA_SIZE for calculating CONFIG_SYS_INIT_SP_OFFSET.
> 
> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
> ---
>  include/configs/m28evk.h |    9 ++++++++-
>  1 files changed, 8 insertions(+), 1 deletions(-)
> 
> diff --git a/include/configs/m28evk.h b/include/configs/m28evk.h
> index 39c841f..2b4871e 100644
> --- a/include/configs/m28evk.h
> +++ b/include/configs/m28evk.h
> @@ -92,7 +92,14 @@
>  #define	CONFIG_SYS_MEMTEST_END		0x40400000	/* 4 MB RAM test 
*/
>  #define	CONFIG_SYS_SDRAM_BASE		PHYS_SDRAM_1
>  /* Point initial SP in SRAM so SPL can use it too. */
> -#define	CONFIG_SYS_INIT_SP_ADDR		0x00002000
> +
> +#define CONFIG_SYS_INIT_RAM_ADDR	0x00002000

This 0x2000 is wrong, it should be 0x0. Otherwise your maths below overflow and 
SP is located at 0x21f80.

> +#define CONFIG_SYS_INIT_RAM_SIZE	(128 * 1024)
> +
> +#define CONFIG_SYS_INIT_SP_OFFSET \
> +	(CONFIG_SYS_INIT_RAM_SIZE - GENERATED_GBL_DATA_SIZE)
> +#define CONFIG_SYS_INIT_SP_ADDR \
> +	(CONFIG_SYS_INIT_RAM_ADDR + CONFIG_SYS_INIT_SP_OFFSET)
>  /*
>   * We need to sacrifice first 4 bytes of RAM here to avoid triggering some
>   * strange BUG in ROM corrupting first 4 bytes of RAM when loading U-Boot

Best regards,
Marek Vasut
diff mbox

Patch

diff --git a/include/configs/m28evk.h b/include/configs/m28evk.h
index 39c841f..2b4871e 100644
--- a/include/configs/m28evk.h
+++ b/include/configs/m28evk.h
@@ -92,7 +92,14 @@ 
 #define	CONFIG_SYS_MEMTEST_END		0x40400000	/* 4 MB RAM test */
 #define	CONFIG_SYS_SDRAM_BASE		PHYS_SDRAM_1
 /* Point initial SP in SRAM so SPL can use it too. */
-#define	CONFIG_SYS_INIT_SP_ADDR		0x00002000
+
+#define CONFIG_SYS_INIT_RAM_ADDR	0x00002000
+#define CONFIG_SYS_INIT_RAM_SIZE	(128 * 1024)
+
+#define CONFIG_SYS_INIT_SP_OFFSET \
+	(CONFIG_SYS_INIT_RAM_SIZE - GENERATED_GBL_DATA_SIZE)
+#define CONFIG_SYS_INIT_SP_ADDR \
+	(CONFIG_SYS_INIT_RAM_ADDR + CONFIG_SYS_INIT_SP_OFFSET)
 /*
  * We need to sacrifice first 4 bytes of RAM here to avoid triggering some
  * strange BUG in ROM corrupting first 4 bytes of RAM when loading U-Boot