I've made yet another update to my dead function optimisation patch to the
2.2.x series kernel, and I'd really appreciate anyone who could take a little
while to try it on their platform. It's generally useful to everyone for
reducing kernel size, and is particularly handy for emdedded systems people.
This update fixes the following problems I missed in the last one:
- Ralf Baechle pointed out that get_wchan relied on function ordering, and
hence was broken by my patch due to the linker section-reordering bug. I
changed the mechanism used to identify scheduler functions so it no longer
relies on function ordering.
- I had neglected to patch the arm/vmlinux*.lds files
- I had neglected to include .rodata.* sections for static const data
- I had neglected to change a .text.lock to .lock.text
See the patch header for all the gory details.
The updated patch is available at:
I think/hope we're getting close to something that could be accepted into the
official kernel now. Once again, any feedback is most welcome!
Principal Hardware/Software Engineer
Canon Information Systems Research Australia
Ph: +61 2 9805 2909 Fax: +61 2 9805 2929