[Patch ARM] Fix up some issues with init-cpu code in T32 state.

Richard Earnshaw rearnsha@arm.com
Mon Oct 14 15:16:00 GMT 2013

On 10/10/13 10:19, Ramana Radhakrishnan wrote:
> Hi,
> 	This fixes up some issues uncovered while building the cpu-init code in 
> a default T32 state, apparently the cpu-init code was always only built 
> in ARM state and not T32 state which is strange. Additionally it wasn't 
> always built for the right multilib sometimes and I tracked that down to 
> the CPU_INIT_OBJS rule in the makefile.in fragment not having the right 
> * Turn rdimon-aem.S into an empty file for M class cores.
> * Additionally not call _rdimon_hw_init_hook from crt0.S
> * Force everything into ARM state anyway as it really doesn't matter.
> <DATE>  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
> 	 * libgloss/arm/cpu-init/rdimon-aem.S: Disable for M class cores.
> 	 * libgloss/arm/crt0.S: Don't call _rdimon_hw_init_hook for non-A class 
> cores.
> 	 * libgloss/arm/cpu-init/Makefile.in (CPU_INIT_OBJS): Use CFLAGS.
> Tested on arm-none-eabi with a set of multilibs with no regressions and 
> Joey confirms that it fixes his issues.
> Ok ? if so can someone commit this for me please ?

I've put this in.  For next time, libgloss has its own ChangeLog file,
so libgloss/ as a prefix is not needed.


More information about the Newlib mailing list