From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?ISO-8859-1?Q?Andreas_Bie=DFmann?= Date: Mon, 02 Jul 2012 12:23:26 +0200 Subject: [U-Boot] [PATCH v2] arm : Atmel : add at91sam9x5ek board support In-Reply-To: <4FF171D0.2000309@gmail.com> References: <1341200122-18329-1-git-send-email-voice.shen@atmel.com> <4FF14744.7040802@googlemail.com> <4FF15D75.70003@gmail.com> <4FF16E4B.4020601@gmail.com> <4FF16F9C.3050101@atmel.com> <4FF171D0.2000309@gmail.com> Message-ID: <4FF1769E.8020703@gmail.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Dear all, On 02.07.2012 12:02, Andreas Bie?mann wrote: > Dear all, > > On 02.07.2012 11:53, Nicolas Ferre wrote: >> On 07/02/2012 11:47 AM, Andreas Bie?mann : >>> On 02.07.2012 10:36, Bo Shen wrote: >> >> [..] >> >>>>>> +/* bootstrap + u-boot + env + linux in nandflash */ >>>>>> +#define CONFIG_ENV_IS_IN_NAND >>>>>> +#define CONFIG_ENV_OFFSET 0xc0000 >>>>>> +#define CONFIG_ENV_OFFSET_REDUND 0x100000 >>>>>> +#define CONFIG_ENV_SIZE 0x20000 /* 1 sector = 128 kB */ >>>>>> +#define CONFIG_BOOTCOMMAND "nand read.jffs2 " \ >>>>>> + "0x22000000 0x200000 0x300000; " \ >>>>>> + "bootm 0x22000000" >>>>> >>>>> Well, you read from jffs but work with ubifs later on? Is that ok? >>>>> BTW: I do not know 'nand read.jffs2', does this subcommand exist really? >>>> >>>> Using the read with subfix .jffs2 it to skip the bad block. This >>>> partition is for Linux kernel. >>>> I will remove the .jffs2 to avoid confusing. >>> >>> Well, it is ok with me to use read.jffs2 (if it works). I just do not >>> know about that subcommand and I can only find some references about >>> that in two other boards. I guess these boards are just forgotten while >>> removing the 'nand read.jffs2' command. Please check, if that command >>> works, if yes let it as is, if not rewrite. >> >> I am in favor of using read.jffs2 because it is able to skip bad blocks. >> As the programming tool is also skipping bad blocks the same way, we >> have a configuration that is perfectly working even in case of bad block >> in the middle of the kernel image. >> >> So please, keep using read.jffs2 (or equivalent)! > > I have the same opinion here, we should skip bad blocks. But I think > 'nand read.jffs2' is not available any longer. > > from current common/cmd_nand.c: > ---8<--- > "nand read - addr off|partition size\n" > "nand write - addr off|partition size\n" > " read/write 'size' bytes starting at offset 'off'\n" > " to/from memory address 'addr', skipping bad blocks.\n" > "nand read.raw - addr off|partition [count]\n" > "nand write.raw - addr off|partition [count]\n" > " Use read.raw/write.raw to avoid ECC and access the flash as-is.\n" > --->8--- Ok, some investigation showed up that a) the default read/write should skip bad blocks since 984e03cdf1431bb593aeaa1b74c445d616f955d3 (mid of 2008!) b) the .jffs2 subcommand is still available, but not in the help -> in line 645: if (!s || !strcmp(s, ".jffs2") || Since the read.jffs2 seems to be a leftover from older days I tend to use just 'nand read' here. But as said before, I will also take the 'nand read.jffs2' if it works. My first objection was that I doubt it works at all. Best regards Andreas Bie?mann