SMP:  Fix timer related issues:  Round robin and sporadic scheduling were only being performed for tasks running on the CPU that processes the system timer interrupt.  Similary, CPU load measurements were only be processed for running on the CPU that receives the sampling interrupt.

Shared Libraries:  In the FLAT build mode, kernel modules may be used to provide minimal shared library functionality.

Fix a compile error: in sched_cpuload.c:Line136, the variables ts and secs are not defined if CONFIG_CPULOAD_ONESHOT_ENTROPY = 0. However, these variables are used regardless of CONFIG_CPULOAD_ONESHOT_ENTROPY at lines~180:onwards.

ARM memcpy():  Use DWord vs. HWord offset.  ARM 32-bit instructions must be aligned to DWord boundaries and this gives us more range in the jump tables.

Adding missing CONFIG_ prefix

Move optimized ARM memcpy functions from arch/arm/src/ to libc/machine/.  This is necessary for the PROTECTED and KERNEL build modes.  Otherwise, memcpy() will be built in to kernel space and not accessible to applications.

Eliminate CONFIG_ARCH_OPTIMIZED_FUNCTIONS.  Move options to select architectur-specific C library options from libc/Kconfig to libc/machine/Kconfig and rename.

SAM3/4: Add support for ATSAM4S4C.

STM32L4:  Port fix for multiple oneshot timers from STM32.  Also fixes a few issues with original STM32 implementation.