Add internal SRAM sections with cache enabled to early MMU settings.
The internal SRAM initialization will use the early MMU settings. This
allows applications to use cache inhibited areas in the internal SRAM.
The increase of the stack size by CPU_STACK_ALIGNMENT in
_Thread_Stack_Allocate() is disadvantageous. This may lead to a huge
over allocation for specialized stack allocators. The
CPU_STACK_ALIGNMENT is at most 16 on all current RTEMS CPU ports. The
mimimum stack size ensured by _Stack_Ensure_minimum() must be
considerable larger than this value, otherwise stack overflows will
likely occur. Thus the _Stack_Adjust_size() is also superfluous.
The comment in configure.ac is probably inherited from the LEON2
BSP. The LEON3 console driver implements a "flush" mechanism on
console_close() in order to solve the problem described with
sis/LEON2.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
The drivers are updated to use the new AMBA layer, however the
backwards-compatibility interface (ambapp_old) is used.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
The old layer had some limitations/problems for multiple AHB
buses since the data structure containing all AMBA devices
were allocated before scanning.
The new layer create devices as they are found and memory is
allocated using malloc() or bsp_early_malloc() during booting.
The old 8 functions for finding a specific AHB-Slave or
APB-Slave device has been replaced with one function,
ambapp_for_each(), which iterates over all devices matching
the specified search options and calls a user provided
function. The new way lowers the footprint and makes searching
more flexible.
The frequency information is now supported, if the frequency
of one device is reported by the user.
More AHB-to-AHB bridges are supported.
The API has been split into several parts in order to lower the
footprint.
The API also introduces the AMBAPP CORE concept, where one
ambapp_core can be created from one AHB Master, AHB Slave
and one APB Slave, at least one device is required for creating
a core.
Signed-off-by: Daniel Hellstrom <daniel@gaisler.com>
