sh: update r7780rp interrupt code

This patch updates the board specific irq code for r7780rp. The new code is
very similar to the other highlander implementations, with the exception that
the r7780rp handles pci interrupts using IRL. To simplify the pci code and
use the same interrupt numbers as r7780mp and r7785rp we hook in to the
cpu specific pci vectors.

The pci interrupts and the push switch all work well with and without this
patch. CF and AX88796 are not ok though and the source of the problem is
unknown at this point. The AX88796 does for not detect it's proper mac
address (IPL gets it right) and the kernel hangs on CF access. As a workaround
this patch removes the CF and the AX88796 from the platform datain case of
r7780rp.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: remove consistent alloc stuff from the machine vector

Now with the voyagergx cruft gone and the dreamcast using declared
coherent memory for pci there are no users of the consistent alloc and
free functions pointers in the machine vector.

So this little patch simply removes these function pointers from the macvec.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: use declared coherent memory for dreamcast pci ethernet adapter

This patch makes the dreamcast use the recently added declared coherent
memory functions to point out the memory window suitable for dma.

Apart from cleaning up, this gives the dreamcast a proper memory allocator
for pci dma memory.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: declared coherent memory support V2

This patch adds declared coherent memory support to the sh architecture. All
functions are based on the x86 implementation. Header files are adjusted to
use the new functions instead of the former consistent_alloc() code.

This version includes the few changes what were included in the fix patch
together with modifications based on feedback from Paul.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add support for SDK7780 board.

Add support for Renesas Technology Europe SDK7780 board.

Signed-off-by: Nicholas Beck <nbeck@mpc-data.co.uk>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: constify function pointer tables

Signed-off-by: Jan Engelhardt <jengelh@computergmbh.de>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off -traditional for linker script.

Some of Sam's new work in the kbuild queue depend on ## concatenation
within the linker script, which doesn't work when -traditional is
enabled. -traditional is a legacy remnant anyways, and we no longer
require it for anything, so kill it off completely.

Noted-by: Andreas Schwab <schwab@suse.de>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

cdrom: Add support for Sega Dreamcast GD-ROM.

This patch adds support for the GD-Rom drive, SEGA's proprietary
implementation of an IDE CD Rom for the SEGA Dreamcast. This driver
implements Sega's Packet Interface (SPI) - at least partially. It will
also read disks in SEGA's propreitary GD format.

Unlike previous drivers (which were never in mainline) this uses DMA and
not PIO to read disks. It is a new driver, not a refactoring of old
drivers.

Signed-off by: Adrian McMenamin <adrian@mcmen.demon.co.uk>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off hs7751rvoip reference from arch/sh/Kconfig.

Remove reference to board deleted in commit 758e06ded4c48024835ef0a14627afcde2e25929

Signed-off-by: Adrian McMenamin <adrian@mcmen.demon.co.uk>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Drop r7780rp_defconfig, use r7780mp_defconfig as kbuild default.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off dead HS771RVoIP board support.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: r7785rp: Fix up DECLARE_INTC_DESC() arg mismatch.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: r7785rp: Hook up the rest of the HL7785 FPGA IRQ vectors.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: r2d - enable sm501 usb host function

This patch tells the sm501 mfd driver to build platform data for the
sm501 usb host driver.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: remove voyagergx

This patch removes redundant irq handling code together with unused
consistent alloc code. R2D uart setup code is changed to use
sm501-regs.h and unused header files are removed.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: r2d - add lcd planel timings to sm501 platform data

This patch adds timings to drive a Sharp LQ104V1DG21 lcd panel that can
be hooked up to R2D-1 or R2D-PLUS. The sm501fb driver should leave the
pins FPEN and VBIASEN alone, and this patch instructs the driver to do
so by not setting flags flags for these pins.

This patch works best together with the patch posted to the
linux-fbdev-devel list "sm501fb: control panel pin usage with platform
data flags", but this patch can be merged independently.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add OHCI and UDC platform devices for SH7720.

Signed-off-by: Yoshihiro Shimoda <shimoda.yoshihiro@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: intc - remove default interrupt priority tables

This patch removes interrupt priority tables from the intc code.
Optimal priority assignment varies with embedded application anyway,
so keeping the interrupt priority tables together with cpu-specific
code doesn't make sense.

The function intc_set_priority() should be used instead to set the
desired interrupt priority level.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Correct pte size mismatch for X2 TLB.

Fixes up a build warning/error in arch/sh/mm/fault_32.c.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix posix_types.h userspace breakage from sh64 merge.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Update SE7712 PCLK definition.

This patch provides a correct value for CONFIG_SH_PCLK_FREQ for the
SH7712 solution engine when used with the board's default factory
settings. This results in the board running at its maximum CPU clock
rate (200 MHz).

Signed-off-by: Andrew Murray <amurray@mpc-data.co.uk>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: sh7712 clock support

This patch provides specific clock support for the SH7712.

Signed-off-by: Andrew Murray <amurray@mpc-data.co.uk>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add support for SH7763 CPU subtype.

Signed-off-by: Yoshihiro Shimoda <shimoda.yoshihiro@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix get_user()/put_user() build error.

Fixes the build error caused by -Werror on gcc 3.x compilers:

arch/sh/kernel/signal_32.c: In function `sys_sigaction':
arch/sh/kernel/signal_32.c:66: warning: initialization discards qualifiers from pointer target type
arch/sh/kernel/signal_32.c:67: warning: initialization discards qualifiers from pointer target type
arch/sh/kernel/signal_32.c:69: warning: initialization discards qualifiers from pointer target type
arch/sh/kernel/signal_32.c:70: warning: initialization discards qualifiers from pointer target type

The mismatch in question was introduced by commit-id
9c5a4eec79b3eb8876d2e7fddfa1e040a7650e55.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add support for SH7721 CPU subtype.

Signed-off-by: Yoshihiro Shimoda <shimoda.yoshihiro@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Provide a stubbed __set_fixmap() for nommu.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Always use CONFIG_HZ for HZ.

Currently the wdt forces HZ=1000 and sidesteps CONFIG_HZ completely. This
is a remnant from when HZ was hardcoded and before CONFIG_HZ was
introduced. Additionally, not all of the timers have this requirement
these days, so it's also an artificial limitation. Just kill it off and
use CONFIG_HZ directly.

Reported-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: include/asm-sh/: Spelling fixes.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: arch/sh/: Spelling fixes.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up KERNEL_ENTRY calculation for uImage.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up binfmt_flat compile warnings.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off pgtable.h from scatterlist.h.

Fixes up the mmc build.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up switch_to() type casts.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Stub in page_table_range_init() on nommu.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Use def_bool where possible.

Change occurances of:
bool
default X

to:
def_bool X

Change ocurances of:
bool "Foo"
default X

to:
def_bool X
prompt "Foo"

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off superfluous __KERNEL__ check in asm/elf.h.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Provide the FPSCR init through AT_FPUCW.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Use utsname()->machine for ELF_PLATFORM.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Encode L1/L2 cache shape in auxvt.

This adds in the L1I/L1D/L2 cache shape support to their respective
entries in the ELF auxvt, based on the Alpha implementation. We use
this on the userspace libc side for calculating a tightly packed
SHMLBA amongst other things.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Support denormalization on SH-4 FPU.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up uImage target entry point.

This patch changes the uImage target so that it generates a wrapped
compressed vmlinux, rather than a wrapped zImage. The previous version
matched the ARM, this version matches the PPC. However I would question
how useful a self decompressing image is with a boot loader which does
decompression, so I think this is more useful. I also feel it matches
the descrition in the help text ("Compressed kernel image") better.

Signed-off-by: Thomas Betker <thomas.betker@5etech.eu>
Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: get_user fixes and nommu consolidation.

When a get_user(to, from++) is called the pointer increment is performed
after its first usage, in the specific after the __add_ok invokation.
This causes a wrong get_user return value, putting a wrong character
in the destination variable. This patch solves the problem using a new
temporary pointer.

Additionally this reworks the use of the register banks, allowing for
consolidation between the MMU and nommu implementations.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
Signed-off-by: Giuseppe Condorelli <giuseppe.condorelli@st.com>
Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Document PTEL 31:29 use on PTEA-wielding parts.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Populate swapper_pg_dir with fixmap range.

This saves us from having to use kmalloc() for the fixmap entries,
which is needed early for the uncached fixmap.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Clean up places that make 29-bit physical assumptions.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Preparation for uncached jumps through PMB.

Presently most of the 29-bit physical parts do P1/P2 segmentation
with a 1:1 cached/uncached mapping, jumping between the two to
control the caching behaviour. This provides the basic infrastructure
to maintain this behaviour on 32-bit physical parts that don't map
P1/P2 at all, using a shiny new linker section and corresponding
fixmap entry.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Explicit alignment for PAGE_SIZE in copy/clear_page().

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: GUSA atomic rollback support.

This implements kernel-level atomic rollback built on top of gUSA,
as an alternative non-IRQ based atomicity method. This is generally
a faster method for platforms that are lacking the LL/SC pairs that
SH-4A and later use, and is only supportable on legacy cores.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix compile error of arch/sh/mm/pmb.c

Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: comment tidying for sh64->sh migration.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: syscall auditing for sh5, too.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

rtc: rtc-sh: Split out the CPU defs to asm/cpu/.

With all of the different CPU types this was getting a but unwieldly.
Since sh64 is now integrated, we don't have to worry about multiple
architectures caring about the header definitions.

Split out the defs for each asm/cpu/ to make rtc-sh slightly less
visually offensive.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: CCR1->CCR renaming for SH-2 parts.

Avoid namespace collision with a CCR1 definition. The general
SH code always expects CCR anyways, so there's no point in keeping
the CCR1 naming around.

Fixes up synclink collisions:

drivers/char/pcmcia/synclink_cs.c:283:1: warning: "CCR1" redefined
In file included from include/asm/cache.h:13,
                 from include/asm/processor_32.h:15,
                 from include/asm/processor.h:60,
                 from include/linux/prefetch.h:14,
                 from include/linux/list.h:8,
                 from include/linux/module.h:9,
                 from drivers/char/pcmcia/synclink_cs.c:38:
include/asm/cpu/cache.h:21:1: warning: this is the location of the previous definition

Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Bump up ARCH_KMALLOC_MINALIGN for DMA cases.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Invalidate the TLB after applying PMB mappings.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Provide a 29/32-bit physical hint for bootloaders.

Shoves a magic word in to the empty_zero_page section for the
bootloader to work out whether to start the kernel in 29-bit
or 32-bit mode.

[ Renesas CPUs already take care of the initial PMB mappings entirely
  in hardware and decide on 29-bit/32-bit physical depending on which
  pin powered up the CPU, so this is mostly for ST parts. -- PFM ].

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add a printk() to warn legacy mem= growers.

mem= can't be used to grow the size of kernel memory, so provide a
warning to that effect.

Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: SH-2A FPU support.

Signed-off-by: Kieran Bingham <kbingham@mpc-data.co.uk>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add SH7263 CPU support.

This adds support for the SH7263 (SH-2A) CPU.

This particular CPU is a superset of SH7203, adding some additional
peripheral blocks and hooking up additional (reserved on SH7203)
vectors in the INTC block.

No visibly nasty surprises, yet..

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix the arch/sh/kernel/traps.c build for sh32.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add SH7203 CPU support.

This adds support for the SH7203 (SH-2A) CPU.

Signed-off-by: Kieran Bingham <kbingham@mpc-data.co.uk>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

rtc: rtc-sh: SH-2A support.

Trivial support for the SH-2A on-chip RTC.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: syscall audit support.

Support syscall auditing..

Signed-off-by: Yuichi Nakamura <ynakam@hitachisoft.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up uname -m matching for native sh64.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Move mach-cayman in with the rest of the boards.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off the mach-harp and mach-sim machtypes.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up bug trap handler build for sh32.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Build fixes for lib32 clear_page.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up default zImage target for sh32.

This was using the absolute path, which was confusing the make target.
Switch it to just 'zImage', as per powerpc.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Bring the SH-5 FPU in line with the SH-4 FPU API.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Get the SH-5 PCI support building.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Bump number of quicklists for SH-5.

Sync up with the SH definitions.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add SH-5 subtypes to check_bugs() for utsname.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off the rest of arch/sh64/kernel/.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Get the mach-cayman IRQ support building.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off the last of the sh64 headers.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: clear/copy_page renames in lib and lib64.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Rename the _32 and _64 TLB flush variants.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Consolidate sh/sh64 Makefiles.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Tidy up KBUILD_IMAGE for sh64.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Remove duplicate SUPERH64 symbol.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Generic BUG only works for sh32.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Add onchip remap prototypes, kill old sh64 io.h.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Copy over special CFLAGS for SH-5 TLB miss.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up user_fpu_struct typo for SH-5.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Mark some IRQ debug options sh32 only.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Special layout for SH-5 stat.h and user.h.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Purge dead sh64 headers.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off arch/sh64/mm.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up proc ASIDs for CPU-local ASID cache accessors.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Move over SH-5 TLB and cache support code.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Merge sh64 Kconfig.debug options.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off a stray proc tlb reference in fault_64.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Merge the sh64 zImage bits.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: SH-5 uses a 64-bit PTE_MAGNITUDE, as X2 TLB.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Use existing CONFIG_CACHE_xx options in head_64.S.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Fix up generic BUG build for SH-5.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Consolidate Kconfig.sh/Kconfig.sh64.

Fold in the sh64-specific bits in to the main Kconfig.sh, and move
this back as arch/sh/Kconfig.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Bump the MEMORY_SIZE default to something reasonable.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

sh: Kill off some dead sh64 kconfig symbols.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>