In message <[EMAIL PROTECTED]> you wrote:
> Configuration changes added to support ADS5121 Rev 3 version of the board.
> This includes changes to the ref clock, memory size (from 256 to 512), the
> addition of PCI functionality and Makefile options to build either the Rev 3
> or Rev 2 U-Boot. The new README explains these options.
Do we really need a separate Makefile target and thus separate images
for this? Isn't there a way to detect the hardware revision
automatically in the software?
>
> Signed-off-by: Martha J Marx <[EMAIL PROTECTED]>
> ---
Makefile | 9 ++++-----
> board/ads5121/README | 9 +++++++++
> board/ads5121/ads5121.c | 34 ++++++++++++++++++++++++++++++----
> cpu/mpc512x/cpu.c | 8 ++++++--
> include/configs/ads5121.h | 27 +++++++++++++++++++++++----
> 5 files changed, 72 insertions(+), 15 deletions(-)
> create mode 100644 board/ads5121/README
Your patch cannot be applied because it was corrupted (line wrapped)
by your mailer (if that is a valid name for something as Outlook). I
strongly recommend to use git-send-email to post patches.
> diff --git a/Makefile b/Makefile
> index ac0a17f..f91c75b 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -742,12 +742,11 @@ motionpro_config: unconfig
> ## MPC512x Systems
> #########################################################################
> ads5121_config \
> -ads5121_PCI_config \
> - : unconfig
> +ads5121_256_config \
...
> +To configure for the Rev 2 ADS5121 type
> +make ads5121_256_config
That's not exactly intuitive - when you call it "Rev 2" hadrware, then
also use that name as the make target.
> @@ -50,7 +52,6 @@ int board_early_init_f (void)
> {
> volatile immap_t *im = (immap_t *) CFG_IMMR;
> u32 lpcaw;
> -
> /*
> * Initialize Local Window for the CPLD registers access (CS2
> selects
^^^^^^^^^^^^^^^^^^^^^^
Line wrapped.
> * the CPLD chip)
> @@ -58,7 +59,6 @@ int board_early_init_f (void)
> im->sysconf.lpcs2aw = CSAW_START(CFG_CPLD_BASE) |
> CSAW_STOP(CFG_CPLD_BASE, CFG_CPLD_SIZE);
> im->lpc.cs_cfg[2] = CFG_CS2_CFG;
> -
> /*
> * According to MPC5121e RM, configuring local access windows should
> * be followed by a dummy read of the config register that was
I don't like these white space changes. Please leave the blank lines
as it was before.
> * Without this the flash identification routine fails, as it needs
> to issue
^^^^^^^^^^^^^^^^^^^^^^
Line wrapped.
>
> +#ifdef CONFIG_ADS5121_256
> + *((volatile u8 *)(CFG_CPLD_BASE + 0x08)) = 0xC1;
> +#else
> + if (*((u8 *)(CFG_CPLD_BASE + 0x08)) & 0x04) {
> +
> + *((volatile u8 *)(CFG_CPLD_BASE + 0x08)) = 0xC1;
> + } else {
> + /* running from Backup flash */
^^^^^^^^^^^^^^^^^^^^^
Indentation wrong.
> + *((volatile u8 *)(CFG_CPLD_BASE + 0x08)) = 0x32;
> + }
> +#endif
> + /*
> + * Configure Flash Speed
> + */
> + *((volatile u32 *)(CFG_IMMR + LPC_OFFSET + CS0_CONFIG)) =
> CFG_CS0_CFG;
^^^^^^^^^^^^^^^^^^^^^^
Line wrapped.
> @@ -143,6 +144,9 @@ void ft_cpu_setup(void *blob, bd_t *bd)
> do_fixup_by_path(blob, eth_path, "local-mac-address",
> bd->bi_enetaddr, 6, 0);
>
> /* this is so old kernels with old device trees will boot */
> +
Do NOT add an empty line here.
> do_fixup_by_path_u32(blob, "/" OF_SOC_OLD, "bus-frequency",
> bd->bi_ipsfreq, 0);
^^^^^^^^^^^^^^^^^^^^^^
Line wrapped.
> + do_fixup_by_path(blob, eth_path_old, "local-mac-address",
> bd->bi_enetaddr, 6, 0);
^^^^^^^^^^^^^^^^^^^^^^
Line wrapped.
> + do_fixup_by_path(blob, eth_path_old, "address", bd->bi_enetaddr, 6,
> 0);
^^^^^^^^^^^^^^^^^^^^^^
Line wrapped.
> }
> #endif
> diff --git a/include/configs/ads5121.h b/include/configs/ads5121.h
> index 81e7c1e..5f82855 100644
> --- a/include/configs/ads5121.h
> +++ b/include/configs/ads5121.h
> @@ -48,9 +48,15 @@
>
> /* CONFIG_PCI is defined at config time */
>
> +#ifdef CONFIG_ADS5121_256
> #define CFG_MPC512X_CLKIN 66000000 /* in Hz */
> +#else
> +#define CFG_MPC512X_CLKIN 33333333 /* in Hz */
Is this 33.333333 MHz or 33.0 MHz ?
> @@ -60,7 +66,11 @@
> /*
> * DDR Setup - manually set all parameters as there's no SPD etc.
> */
> +#ifdef CONFIG_ADS5121_256
> #define CFG_DDR_SIZE 256 /* MB */
> +#else
> +#define CFG_DDR_SIZE 512 /* MB */
> +#endif
We should use auto-sizing for the RAM like all "correct" ports do.
> @@ -175,8 +191,11 @@
>
> #define CFG_SRAM_BASE 0x30000000
> #define CFG_SRAM_SIZE 0x00020000 /* 128 KB */
> -
Do not delete this empty line.
> +#ifdef ADS5121_256_CONFIG
> #define CFG_CS0_CFG 0x05059310 /* ALE active low, data size
> 4bytes */
^^^^^^^^^^^^^^^^^^^^
Line wrapped.
> +#else
> +#define CFG_CS0_CFG 0x05059310 /* ALE active low, data size
> 4bytes */
^^^^^^^^^^^^^^^^^^^^
Line wrapped.
> +#endif
> #define CFG_CS2_CFG 0x05059010 /* ALE active low, data size
> 1byte */
^^^^^^^^^^^^^^^^^^^^
Line wrapped.
Best regards,
Wolfgang Denk
--
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: [EMAIL PROTECTED]
Eureka! -- Archimedes
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
U-Boot-Users mailing list
U-Boot-Users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/u-boot-users
