Re: building meta-ti + poky HOWTO?

[Jack Mitchell <ml@communistcode.co.uk>]

On 05/04/2012 21:43, Russell Senior wrote:
>>>>>> "Denys" == Denys Dmytriyenko<denis@denix.org>  writes:
> Russell>  I have been unable to get a booting image out of Angstrom in
> Russell>  about three weeks of trying and several unanswered queries to
> Russell>  the mailing list and elsewhere.  In my continuing effort to
> Russell>  get a build system that produces an image that boots on my
> Russell>  beaglebone, I'd like to try building meta-ti with poky, just
> Russell>  to see if that fails for me too.
>
> Denys>  As far as I remember, you are having problems with rootfs
> Denys>  images that generate an entire SD card disk image with multiple
> Denys>  partitions, e.g. Cloud9 image.  There are known issues with the
> Denys>  tools used to do that, like sometimes fdisk not making the
> Denys>  correct MBR and partition table that BootROM can boot from. So
> Denys>  far it seems to depend on the host distro and their tools. This
> Denys>  area still needs further investigation and probably the use of
> Denys>  different tools...
>
> I am building on a Debian/unstable box with a core i7 cpu and 24 gig of RAM.
>
> Denys>  But you should not be stuck on this - why don't you use the
> Denys>  individual pieces, like kernel image, rootfs and load them on
> Denys>  SD card yourself, manually? Try working around the issue...
>
> I have tried taking the pre-baked Angstrom image from here:
>
>    http://www.angstrom-distribution.org/demo/beaglebone/Angstrom-Cloud9-IDE-eglibc-ipk-v2012.02-core-beaglebone-2012.02.14.img.xz
>
> (which works for me), and replacing the uImage on the vfat partition
> by copying over the existing one with one I built.  That boots, but the
> rootfs isn't aligned (different kernel/module versions?).
>
> If I then try dropping in my rootfs, by reformatting the ext4fs and
> untarring the rootfs tarball I built, I get 0.6 seconds into the boot
> and then the boot stalls with some udevd messages, like so:
>
>    U-Boot SPL 2011.09-00010-g81c8c79 (Feb 13 2012 - 14:48:03)
>    Texas Instruments Revision detection unimplemented
>    No AC power, disabling frequency switch
>    OMAP SD/MMC: 0
>    reading u-boot.img
>    reading u-boot.img
>
>
>    U-Boot 2011.09-00010-g81c8c79 (Feb 13 2012 - 14:48:03)
>
>    I2C:   ready
>    DRAM:  256 MiB
>    No daughter card present
>    NAND:  HW ECC Hamming Code selected
>    nand_get_flash_type: unknown NAND device: Manufacturer ID: 0x10, Chip ID: 0x10
>    No NAND device found!!!
>    0 MiB
>    MMC:   OMAP SD/MMC: 0
>    *** Warning - readenv() failed, using default environment
>
>    Net:   cpsw
>    Hit any key to stop autoboot:  0
>    SD/MMC found on device 0
>    reading uEnv.txt
>
>    33 bytes read
>    Loaded environment from uEnv.txt
>    Importing environment from mmc ...
>    reading uImage
>
>    3140320 bytes read
>    ## Booting kernel from Legacy Image at 80007fc0 ...
>       Image Name:   Angstrom/3.2/beaglebone
>       Image Type:   ARM Linux Kernel Image (uncompressed)
>       Data Size:    3140256 Bytes = 3 MiB
>       Load Address: 80008000
>       Entry Point:  80008000
>       Verifying Checksum ... OK
>       XIP Kernel Image ... OK
>    OK
>
>    Starting kernel ...
>
>    Uncompressing Linux... done, booting the kernel.
>    [    0.120087] cpuidle-am33xx cpuidle-am33xx.0: failed to register driver
>    [    0.251953] _omap_mux_get_by_name: Could not find signal leds-gpio
>    [    0.635798] omap2_set_init_voltage: unable to get clk dpll1_ck
>    [    0.642040] omap2_set_init_voltage: unable to set vdd_mpu_iva
>    [    0.648112] omap2_set_init_voltage: unable to get clk l3_ick
>    [    0.654099] omap2_set_init_voltage: unable to set vdd_core
>    udevd[56]: worker [79] did not accept message -1 (Connection refused), kill it
>    ^M
>    udevd[56]: worker [80] did not accept message -1 (Connection refused), kill it
>    ^M
>    udevd[56]: worker [81] did not accept message -1 (Connection refused), kill it
>    ^M
>    udevd[56]: worker [85] did not accept message -1 (Connection refused), kill it
>    ^M
>    udevd[56]: worker [86] did not accept message -1 (Connection refused), kill it
>    ^M
>    udevd[56]: worker [87] did not accept message -1 (Connection refused), kill it
>    ^M
>
> I don't know what that means, nor where to go from there.
>
>
Russel,

I took the time to post a brief guide[1] on how to build Angstrom for 
the BeagleBone, it takes you through the steps and also uses the 
separate images built rather than the big 'sd card partition' image it 
has been noted you are having problems with.

I don't know if it will help in your situation but it's something else 
you can inspect.

Regards,
Jack.

[1] http://communistcode.co.uk/blog/blogPost.php?blogPostID=2