As of now we have 2 flavors of ARC SDP boards:
 1) AXS101 - with ARC770 in ASIC
 2) AXS103 - with ARC HS38 in FPGA

Both options share exactly the same base-board and only differ with
CPU-tiles in use. That means all peripherals are the same (they are
implemented in FPGA on the base-board) and so generic board could be
used for both.

While at it:
 * Recreated defconfigs with savedefconfig
 * In include/configs/axs10x.h numerical sizes replaced with
defines from linux/sizes.h for better readability.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Reviewed-by: Tom Rini <trini@konsulko.com>

ARCangel was one of the main development boards back in the day but
now it's gone and replaced by other boards like ARC SDP.

But we also used to have simulation platform very similar to ARCangel4
in terms of CPU settings as well as basic IO like UART. Even though
ARCangel4 is long gone now we have a replacement for simulation which is
a plain or stand-alone nSIM and Free nSIM.

Note Free nSIM is available for download here:
https://www.synopsys.com/cgi-bin/dwarcnsim/req1.cgi

And while at it:
 * Finally switch hex numerical values in nsim.h to defines from
   include/linux/sizes.h
 * Add defconfigs with ARC HS38 cores
 * Recreated all defconfigs with savedefconfig

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

Starting from arc-2016.03 GNU tools linker properly works with
symbols defined in linker script and so external declarations
are no longer required, dump them.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

Initially IVT for ARCv2 was simply copypasted from ARCompact
with some selected fixes so basic stuff works.

Now we update it with more ARCv2 specific vectors like
 * Software Interrupt
 * Division by zero
 * Data cache consistency error
 * Misaligned access

Also normal interrupts are now implemented properly and extened to
all possible 240 items.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

This might be useful to make sure relocation fixups really
happen. And since this info gets printed only in DEBUG
build it doesn't really hurt normal execution.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

Currently on attempt to use global_data.h in an assembly file following
will happen:
-------------------->8-----------------
./arch/arc/include/asm/global_data.h: Assembler messages:
./arch/arc/include/asm/global_data.h:11: Error: bad instruction 'struct arch_global_data{'
./arch/arc/include/asm/global_data.h:12: Error: junk at end of line, first unrecognized character is `}'
scripts/Makefile.build:316: recipe for target 'arch/arc/lib/start.o' failed
-------------------->8-----------------

In this change we disable struct arch_global_data in ASM which fixes
the issue above.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

Newer ARC toolchains don't support "-marchs" option any longer.
Instead "-mcpu=archs" should be used. What's also important older
toiolchains that support ARC HS cores will also happily accept
"-mcpu=archs" so that's a very safe move.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

Memory barriers are proven to be a requirement for both compiler and
real hardware to properly serialize access to critical data.

For example if CPU or data bus it uses may do reordering of data
accesses absence of memory barriers might easily lead to very subtle and
hard to debug data corruptions.

This implementation was heavily borrowed from up to date Linux kernel.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

According to ARC HS databook it is required to flush and disable
caches prior programming IOC registers. Otherwise ongoing coherent
memory operations may not observe the coherency protocols as
expected.

But since in ARC HS v2.1 there's no way to disable SLC (AKA L2 cache)
we're doing our best flushing and invalidating it.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>

invalidate_dcache_all() could be used in different use-cases
and what is especially important most of those cases won't be
related to DMAed data to or from peripherals, i.e. we'll be doing
invalidation of data used purely by CPU cores.

Given that IOC engine only snoops data that goes through DMA
we need to care ourselves about data used only by CPU cores
and so remove dependency on IOC from invalidate_dcache_all()
and always do real invalidation.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>