Hi!
I installed an Angstrom image downloaded from their site and have been
working with it for a while. Now I need to recompile the kernel for
drivers and MUX setting, and I am trying to get a hold of the exact
same version from the git repository so that I don't have to reinstall
everything on a different kernel build.
I used the image called:
[ ] uImage-2.6.32-r61+gitrac07af55554567b6226ba8dd623756c630e93103-
beagleboard.bin
And after cloning the git archine, i did:
git checkout -b my_dev_branch ac07af55554567b6226ba8dd623756c630e93103
Unfortunately that didnt work.
Any ideas?
Tallak
I'm seeking to do the same thing but so far have been unsuccessful. I
did manage to find out what bitbake recipe is being used to build
those images however.
If you look in the archive of modules-2.6.32-
r61+gitrac07af55554567b6226ba8dd623756c630e93103-beagleboard.tgz
you'll see a hardlink pointing to /OE/angstrom-dev/work/beagleboard-
angstrom-linux-gnueabi/linux-omap-psp-2.6.32-
r61+gitrac07af55554567b6226ba8dd623756c630e93103/git. So it looks
like these demo images are using the linux-omap-psp-2.6.32.bb recipe.
I tried simply running "bitbake linux-omap-psp-2.6.32" and this
compiles a semi-working kernel. The problem is that when I boot into
my beagleboard with the new kernel and then run depmod, it freezes on
the next reboot. Also, the new kernel ends up being linux-omap-
psp-2.6.32-r51 and not linux-omap-psp-2.6.32-r61. There are also
quite a few differences in the defconfigs for the new kernel and the
demo kernel.
Wish I could help more, but I'm looking for the same thing you are!
-Kameron
KLarsen <kroylar@gmail.com> [2010-03-18 02:36:57]:
[..snip..]
So it looks like these demo images are using the linux-omap-psp-2.6.32.bb
recipe.
You can find all the basic information in recipes/linux/linux-omap-psp-2.6.32.bb:
[..snip..]
SRC_URI = "git://arago-project.org/git/people/sriram/ti-psp-omap.git;protocol=git;branch=master
[..snip..]
ynezz@wizard:~/git/ti-psp-omap.git$ git log ac07af55554567b6226ba8dd623756c630e93103 -1
commit ac07af55554567b6226ba8dd623756c630e93103
Author: Sriramakrishnan <srk@ti.com>
Update PSP version entry in PROCFS
Update the release version string for
PSP 03.00.00.05 release.
Signed-off-by: Sriramakrishnan <srk@ti.com>
Thank you Petr. I've tried changing the SRCREV variable to
"ac07af55554567b6226ba8dd623756c630e93103" but when I do this, there
is a compile error about undefined references to a few variables. I'm
thinking that Koen (the guy who made the Angstrom demo image) has
added more patches to recipes/linux/linux-omap-psp-2.6.32.bb to get it
to compile and has not yet committed them to the repository. He has
also changed recipes/linux/linux-omap-psp_2.6.32/beagleboard/defconfig
as the configuration in the demo uImage is not the same as that in OE
git.
Oh well, hopefully he pushes those changes out sometime.
-Kameron
KLarsen <kroylar@gmail.com> [2010-03-18 14:07:58]:
Thank you Petr. I've tried changing the SRCREV variable to
"ac07af55554567b6226ba8dd623756c630e93103" but when I do this, there
is a compile error about undefined references to a few variables.
I wonder if you can do that easily by just changing SRCREV, but I don't think
so. You would need to revert state of whole(maybe just few recipes, dirs) OE
back to the revision where there was
SRCREV="ac07af55554567b6226ba8dd623756c630e93103" in linux-omap-psp_2.6.32.bb.
Then it might build.
thinking that Koen (the guy who made the Angstrom demo image) has
added more patches to recipes/linux/linux-omap-psp-2.6.32.bb to get it
to compile and has not yet committed them to the repository. He has
also changed recipes/linux/linux-omap-psp_2.6.32/beagleboard/defconfig
as the configuration in the demo uImage is not the same as that in OE
git.
Yes, he's evil guy, just blame him! 
Oh well, hopefully he pushes those changes out sometime.
No, you need to go back in time with OE to build that kernel revision and have same
defconfig. BTW, why do you need just that particular revision?
-- ynezz
KLarsen <kroy...@gmail.com> [2010-03-18 14:07:58]:
> Thank you Petr. I've tried changing the SRCREV variable to
> "ac07af55554567b6226ba8dd623756c630e93103" but when I do this, there
> is a compile error about undefined references to a few variables.
I wonder if you can do that easily by just changing SRCREV, but I don't think
so. You would need to revert state of whole(maybe just few recipes, dirs) OE
back to the revision where there was
SRCREV="ac07af55554567b6226ba8dd623756c630e93103" in linux-omap-psp_2.6.32.bb.
Then it might build.
Well, the current SRCREV is older (Nov '09) than this new (Feb '10)
"ac07af5555..." SRCREV that I've changed it to. So I don't think a
linux-omap-psp_2.6.32.bb has ever been committed where
SRCREV="ac07af55554567b6226ba8dd623756c630e93103". But then again,
they could have gone back to the older sources for some other reason.
> thinking that Koen (the guy who made the Angstrom demo image) has
> added more patches to recipes/linux/linux-omap-psp-2.6.32.bb to get it
> to compile and has not yet committed them to the repository. He has
> also changed recipes/linux/linux-omap-psp_2.6.32/beagleboard/defconfig
> as the configuration in the demo uImage is not the same as that in OE
> git.
Yes, he's evil guy, just blame him!
Haha! Aren't all developers evil? 
> Oh well, hopefully he pushes those changes out sometime.
No, you need to go back in time with OE to build that kernel revision and have same
defconfig. BTW, why do you need just that particular revision?
I'd just like to start from a known good configuration. I know that
the demo image works very well for me so I would be nice to start by
recompiling it exactly before making any changes.
But speaking of going back in time, I'm new to git. What would be the
easiest way to find out if SRCREV was ever equal to
"ac07af55554567b6226ba8dd623756c630e93103"?
-Kameron
It seems I misunderstood the whole thing.
The git revision in the image file is the revision in the linux kernel
git repository, not the revision in the openemebedded repository.
Oh, and yes I also discovered that the demo image is linux-omap-psp...
not linux-omap... so many variables to consider...
- Tallak