[U-Boot] Fwd: [PATCH] nand booting support (SPL) for phyCORE-i.MX31
Hello, Magnus
29.11.08, 12:59, "Magnus Lilja" lilja.magnus@gmail.com:
Thank for time that you spend to review this patch.
Hi 2008/11/29 user scn1874@yandex.ru:
nand booting support (SPL) for phyCORE-i.MX31 ... Patch is applicable on v2008.10 release tag master branch.
stgit couldn't apply the patch, but the standard 'patch' command could (it said 'applied with fuzz' on start.S).
May be it's my native utf-8 charset doesn't recode to latin-8859-1? Anything, I try this again.
Your patch seems to have some trailing whitespaces here and there.
It is require to correct in according codestyle?
#define CCM_MPCTL (CCM_BASE + 0x10) -#define CCM_UPCTL (CCM_BASE + 0x10) +#define CCM_UPCTL (CCM_BASE + 0x14)
The above fix should really go into the arm-tree immediately since it's a separate bug fix (IMHO).
What can I make with it?
+/*
- Set INT to 0, FCMD to 1, rest to 0 in NFC_CONFIG2 Register for Command
Change "rest" to "reset" (for all #define's below as well).
More correctly to change "rest" to "the rest".
/* Block of NAND flash to be unlocked *//*NFC_UNLOCKSTART_BLKADDR = 0x0;*//*NFC_UNLOCKEND_BLKADDR = 0x4000;*//* Unlock NAND Flash Block Command for given address range *//*NFC_WRPROT = 0x4;*/If the above is not needed, remove it.
You are right.
mx31_read_page(from, buf);from ++;Remove space before ++;
buf = buf + CFG_NAND_PAGE_SIZE;}return 0;+}
Insert empty line. Shouldn't the code check if the current block is marked as bad before copying all the pages from that block into SDRAM? The first block is typically guaranteed to be good (if I understand correctly) but any subsequent block may be marked bad already from the factory and should be skipped. The entire U-boot won't fit in the first block, atleast for small page NAND devices.
Yes, it's so. I wrote it for guaranteed first block chip. But for phycore board it's not so. I'll add bb verify feature.
Best wishes, Maxim
Hi
2008/12/1 Maxim Artamonov scn1874@yandex.ru:
Hello, Magnus
29.11.08, 12:59, "Magnus Lilja" lilja.magnus@gmail.com:
Your patch seems to have some trailing whitespaces here and there.
It is require to correct in according codestyle?
Yes, I think so.
#define CCM_MPCTL (CCM_BASE + 0x10) -#define CCM_UPCTL (CCM_BASE + 0x10) +#define CCM_UPCTL (CCM_BASE + 0x14)
The above fix should really go into the arm-tree immediately since it's a separate bug fix (IMHO).
What can I make with it?
You can move this part of the patch to a separate stand-alone patch and submit it to the mailing list.
+/*
- Set INT to 0, FCMD to 1, rest to 0 in NFC_CONFIG2 Register for Command
Change "rest" to "reset" (for all #define's below as well).
More correctly to change "rest" to "the rest".
Ah, true.
mx31_read_page(from, buf);from ++;Remove space before ++;
buf = buf + CFG_NAND_PAGE_SIZE;}return 0;+}
Insert empty line. Shouldn't the code check if the current block is marked as bad before copying all the pages from that block into SDRAM? The first block is typically guaranteed to be good (if I understand correctly) but any subsequent block may be marked bad already from the factory and should be skipped. The entire U-boot won't fit in the first block, atleast for small page NAND devices.
Yes, it's so. I wrote it for guaranteed first block chip. But for phycore board it's not so. I'll add bb verify feature.
Excellent!
Thanks, Magnus
participants (2)
-
Magnus Lilja -
Maxim Artamonov