Backport ARM Cortex A9 scheduling changes

[Andrew Stubbs]

ARM are planning to implement Cortex A9 specific scheduling support. Only they can do this because it requires access to confidential micro-architecture documents.The new patches will be applied to GCC trunk.

This task is to backport these patches to the Linaro toolchain if/when they are available. The resultant tools must then be tested and benchmarked to ensure that the patches have worked as desired.

[Andrew Stubbs (ams-codesourcery)]

The patches should be backported to the GCC 4.5 branch only. GCC 4.4 is thought to be too old.

[Richard Earnshaw (richard-earnshaw)]

Patch to upstream GCC.

http://gcc.gnu.org/ml/gcc-patches/2010-08/msg00481.html

[Andrew Stubbs (ams-codesourcery)]

The patch has now been backported.

I also had to backport an additional prerequisite patch:
 http://gcc.gnu.org/ml/gcc-patches/2010-04/msg01093.html

[Andrew Stubbs (ams-codesourcery)]

Reopening as there are more patches to come from ARM (as yet still unposted upstream).

[Loïc Minier (lool)]

"Incomplete" is when the submitter didn't provide enough information to reproduce his bug.

I think we should close this bug and open another one for whatever other scheduling changes remain to be posted; a more descriptive name for them would be ideal.