On 17.09.2012 13:10, Paul Parsons wrote:
On 17/09/12 12:41, Paul Parsons wrote:
command line is:

mtdparts=ipaq:0x00100000@0x00000000(bootldr)ro,0x00200000@0x00100000(kernel),0x04CC0000@0x00300000(root),0x03000000@0x04FC0000(home),0x00040000@0x07FC0000(asset)

  root=/dev/mtdblock2 rw rootfstype=jffs2 console=ttyS0,115200n8
linuxEntryPoint=A0008000
Booting Linux image
Uncompressing Linux... done, booting the kernel.
--------- snip ---------

...and that's all she wrote. I can only assume there is some problem in the way I am building the kernel (using a recipe with OpenEmbedded).

hmm, my wild guess would be that the kernel is not receiving the
correct partition information. I don't know how HaRET communicates the partition information, but the SDG bootloader uses the mtdparts kernel
parameter, which in turn requires CONFIG_MTD_CMDLINE_PARTS to be
defined in the kernel config file.

Also if you define CONFIG_VT_CONSOLE you should see some kernel output
on the LCD screen.

I know what the (other?) problem is: the mtd-id "ipaq" is not
recognized by the kernel's command line partition table parser. It's
expecting the mtd-id "physmap-flash" instead.

There are 2 solutions:

1. Change the SDG bootloader to pass in the mtd-id "physmap-flash"
instead of "ipaq". That is what I do.

2. Hack the kernel to expect the mtd-id "ipaq" instead of
"physmap-flash":

diff --git a/drivers/mtd/cmdlinepart.c b/drivers/mtd/cmdlinepart.c
index 4558e0f..c4519fc 100644
--- a/drivers/mtd/cmdlinepart.c
+++ b/drivers/mtd/cmdlinepart.c
@@ -320,7 +320,7 @@ static int parse_cmdline_partitions(struct mtd_info *master,
        unsigned long offset;
        int i;
        struct cmdline_mtd_partition *part;
-       const char *mtd_id = master->name;
+       const char *mtd_id = "ipaq";

        /* parse command line */
        if (!cmdline_parsed)

This is obviously a grubby hack, but it's the quickest thing to try.

OK, so I enabled CONFIG_VT_CONSOLE and applied the change above, as well as enabling CONFIG_FS_JFFS2 which was disabled (oops). Unfortunately after reflashing the kernel the result was the same. It's also worth noting that after the bootloader's blue progress bar reaches the end there is no change in the display at all.

Cheers,
Paul

_______________________________________________
Angstrom-distro-devel mailing list
Angstrom-distro-devel@linuxtogo.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel

Reply via email to