Contributing to PyraOS



I think that's the one.

I think things broke with the commit after it, which happened a few days later: https://git.ti.com/cgit/graphics/om...39&id=4cdbb6b192fc5cac53695faca3c3f8be16ca871
(and had me wondering why stuff didn't work for me, but did for the others)

We also patch the lib in the package post install script, so not everything on the repos matches what's in the image/tgz ( at offset 0x98eb in libsrv_um.so we change d0 to e0 )
I think this skipped a test since we're not using the exact cpu/gpu that driver expects.
 
  • Like
Reactions: rSl
Exactly. And nice to know that TI finally has a GitWeb for this. Didn't know...
I think things broke with the commit after it, which happened a few days later: https://git.ti.com/cgit/graphics/om...39&id=4cdbb6b192fc5cac53695faca3c3f8be16ca871
(and had me wondering why stuff didn't work for me, but did for the others)
and that was the reason why I could not find the binary as any of the branch heads at https://git.ti.com/cgit/graphics/omap5-sgx-ddk-um-linux/
We also patch the lib in the package post install script, so not everything on the repos matches what's in the image/tgz ( at offset 0x98eb in libsrv_um.so we change d0 to e0 )
I think this skipped a test since we're not using the exact cpu/gpu that driver expects.
Yes. It is needed in addition so that the omap5432 machine name is not rejected.
The offset works only if the underlaying binary is exactly the right one.
Anyways, we're looking at the mess of the runtime libs and binaries a little better. :)
 
  • Like
Reactions: rSl
Back
Top