On Tue, Dec 13, 2011 at 12:44 PM, Ilya Yanok yanok@emcraft.com wrote:
Hi Tom,
On 13.12.2011 23:30, Tom Rini wrote:
+#define CONFIG_SYS_TEXT_BASE 0x80008000
[snip]
+#define CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR 0x300 /* address 0x60000 */ +#define CONFIG_SYS_MMC_SD_FAT_BOOT_PARTITION 1 +#define CONFIG_SPL_FAT_LOAD_PAYLOAD_NAME "u-boot.img"
You're able to have CONFIG_SYS_TEXT_BASE that low and not have the corruption problem devkit8000 had?
No, I can't see any problems. I have to say I can't do _real_ MMC boot, as my board lacks the required jumper, so I'm testing it but starting SPL from NAND, rewriting omap3_boot_device to be MMC with BDI and running it further. It reads u-boot.img from MMC and starts it without any problem.
Well, is it just your board that's missing the jumper, or the MCX HTWK lacks the jumper? If the latter, just drop the MMC bits out :)
I can't ;) this was an explicit requirement.
As for the devkit8000 boot problem, I have BSS set up close to the top of the RAM. Probably that's the reason I don't see any problems.
Unfortunately, I can't recall why I had moved BSS... It clashed with something but I don't remember the details.
OK. I _want_ to put this into my /next branch (and thanks for working over all of the issues that've popped up since you first posted the series) but I also would like to try and keep differences for difference sake out of the SPL implementations since I know other folks are starting to look at the code and port their custom boards or in-tree boards and I'm sure they'll go "why does X place things here and Y there and my board isn't working?". So if you can try and recall what the clash was (and we can comment it, or maybe it was the FAT thing :)) I'd really appreciate it.