Re: dtb for pandaboard

[Dennis Gilmore <dennis@ausil.us>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 11 Jul 2012 00:42:33 -0700
Tony Lindgren <tony@atomide.com> wrote:

> * Dennis Gilmore <dennis@ausil.us> [120710 16:16]:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > 
> > On Mon, 9 Jul 2012 02:08:49 -0700
> > Tony Lindgren <tony@atomide.com> wrote:
> > 
> > > * Dennis Gilmore <dennis@ausil.us> [120708 16:11]:
> > > > -----BEGIN PGP SIGNED MESSAGE-----
> > > > Hash: SHA1
> > > > 
> > > > Hi does anyone happen to have a working dtb file for the
> > > > pandaboard? trying to test booting the fedora rawhide kernel
> > > > which is 3.5.0-rc5 with a dtb file I made from the matching
> > > > sources I dont get past decompressing the kernel. if i dont use
> > > > a dtb file the system boots further but it i'm unable to find
> > > > the root filesystem since it seems that the platform data has
> > > > been messed up. modprobing omap results in no registration of
> > > > the mmc block devices. 
> > > > 
> > > > i've not tested on a pandaboard ES or beagleboard yet  or any
> > > > other type of omap system. 
> > > 
> > > FYI, I booted my panda es recently just fine with dtb append
> > > support and the following .config addition using the plain
> > > arch/arm/boot/dts/omap4-panda.dts file:
> > > 
> > > CONFIG_ARM_APPENDED_DTB=y
> > > CONFIG_ARM_ATAG_DTB_COMPAT=y
> > > CONFIG_EXT4_FS=y
> > > 
> > > Can you please enable DEBUG_LL, EARLY_PRINTK and add earlyprintk
> > > to your cmdline and post what you get?
> > > 
> > > Regards,
> > > 
> > > Tony
> > 
> > Loaded environment from uEnv.txt
> > Importing environment from mmc ...
> > reading uImage-omap
> > 
> > 3988040 bytes read
> > reading uInitrd-omap
> > 
> > 1190 bytes read
> > reading omap4-panda.dtb
> > 
> > 5483 bytes read
> > Booting from mmc ...
> > ## Booting kernel from Legacy Image at 80300000 ...
> >    Image Name:   3.5.0-0.rc6.git0.3.fc18.armv7hl.
> >    Image Type:   ARM Linux Kernel Image (uncompressed)
> >    Data Size:    3987976 Bytes = 3.8 MiB
> >    Load Address: 80008000
> >    Entry Point:  80008000
> >    Verifying Checksum ... OK
> > ## Loading init Ramdisk from Legacy Image at 81600000 ...
> >    Image Name:   initramfs
> >    Image Type:   ARM Linux RAMDisk Image (uncompressed)
> >    Data Size:    1126 Bytes = 1.1 KiB
> >    Load Address: 00000000
> >    Entry Point:  00000000
> >    Verifying Checksum ... OK
> > ## Flattened Device Tree blob at 82600000
> >    Booting using the fdt blob at 0x82600000
> >    Loading Kernel Image ... OK
> > OK
> >    reserving fdt memory region: addr=9d000000 size=3000000
> >    Loading Ramdisk to bfefa000, end bfefa466 ... OK
> >    Loading Device Tree to bfef5000, end bfef956a ... OK
> > 
> > Starting kernel ...
> > 
> > Uncompressing Linux... done, booting the kernel.
> > Warning: Neither atags nor dtb found
> > 
> > is as far as i am getting.
> 
> Sounds like it's some kind of issue with dtb getting overwritten
> by something. We had an issue where kernel BSS was overlapping dtb
> in some cases, but those should be fixed.

seems that they are not all fixed appending the dtb allows me to boot i
could try loading the dtb at a different address. 


> Maybe try to leave out ARCH_OMAP2 and ARCH_OMAP3 and maybe CONFIG_NET
> from your .config to make the kernel smaller and see if that makes
> a difference?
> 
> If that works, then moving the dtb address in uEnv.txt should help.
> 
> Also, please check if the same issue happens with appended dtb:
with the appended dtb image im back to where i was not using a dtb file
at all.  that is that omap is not being autoloaded. and the sdcard so
rootfs never shows up.  i get dropped to a dracut rescue shell where if
i manually modprobe omap  nothing is happening. 
additionally i get over and over 

[  349.573883] BUG: sleeping function called from invalid context at include/linux/freezer.h:46
[  349.582763] in_atomic(): 0, irqs_disabled(): 128, pid: 180, name: dracut-initqueu
[  349.590606] no locks held by dracut-initqueu/180.
[  349.595550] irq event stamp: 559882
[  349.599212] hardirqs last  enabled at (559881): [<c051dba4>] _raw_spin_unlock_irq+0x2c/0x4c
[  349.607971] hardirqs last disabled at (559882): [<c000e4ec>] ret_fast_syscall+0xc/0x3c
[  349.616302] softirqs last  enabled at (558920): [<c00455fc>] irq_exit+0x60/0xb8
[  349.623992] softirqs last disabled at (558907): [<c00455fc>] irq_exit+0x60/0xb8
[  349.631713] [<c0016790>] (unwind_backtrace+0x0/0x124) from [<c0052fc8>] (get_signal_to_deliver+0x50/0x81c)
[  349.641845] [<c0052fc8>] (get_signal_to_deliver+0x50/0x81c) from [<c00115f0>] (do_signal+0x8c/0x4c4)
[  349.651428] [<c00115f0>] (do_signal+0x8c/0x4c4) from [<c0011e78>] (do_notify_resume+0x20/0x5c)
[  349.660491] [<c0011e78>] (do_notify_resume+0x20/0x5c) from [<c000e548>] (work_pending+0x28/0x2c)


so im not sure if the dtb is not defining the platform correctly or if
something in the omap driver is just not right and its not able to
register correctly.

Dennis
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.18 (GNU/Linux)

iEYEARECAAYFAk/9hEcACgkQkSxm47BaWfennQCeMMEvYkvM25xblEcdgkG+3uK+
Ag0AmQGad8XeAHzJmqUBuivtZWz/HF9c
=qdRv
-----END PGP SIGNATURE-----