Commit Graph

549 Commits

Author SHA1 Message Date
Magnus Damm
b58d134c7f sh: missing symbol fix for sh4-202
This patch adds a plat_irq_setup() symbol for sh4-202. Without
this fix it is impossible to build a working kernel using the
microdev_defconfig.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-22 13:32:07 +09:00
Paul Mundt
da2f5f7bf4 sh: Fix DSP opcode regression for SH3-DSP parts.
Older versions of binutils do not support -Wa,-isa= tuning, which is
something we rely on for enabling DSP opcode support on the newer
parts. SH3-DSP parts can still be handled with -Wa,-dsp even if the
newer parts require the newer versions of binutils for supporting the
new opcodes.

This was broken in -rc1 when the SH4AL-DSP support was being reworked,
and is needed to get SH3-DSP working with older toolchains again.

Reported-by: Markus Brunner <super.firetwister@gmail.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-21 16:56:56 +09:00
Andy Whitcroft
96989d9d75 sh: remove extraneous ; on scif_sercon_putc wait loop
It seems we have gained an extraneous trailing ';' on one of the
wait loops in scif_sercon_putc().  Although this is completely
benign as the apparent payload is also the empty statement, it
invites error in the future.  Clean it up now.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-17 01:25:34 +09:00
Paul Mundt
5556410e22 sh: panic on machvec section misalignment.
Now that __initmv references the machvec section unconditionally
there have been cases where folks have been mistakenly flagging
non-machvec structures with the machvec section attribute (presumably
to shut up modpost). This leads to obscure breakage in earlyprintk
amongst other places and is rather non-obvious.

Add a simple sanity check to try and catch __initmv misuse and
panic early.

Reported-by: Markus Brunner <super.firetwister@gmail.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-07 19:13:23 +09:00
Paul Mundt
662ae2169a sh: Fix PTRACE_PEEKTEXT/PEEKDATA fallout from generic_ptrace_peekdata().
When generic_ptrace_peekdata() was merged, the break for these cases
ended up getting dropped, which lead to each PEEKTEXT/PEEKDATA op leaking
in to PEEKUSR and get_user_pages() always -EFAULTing. Add the break back
in.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-04 13:39:21 +09:00
Magnus Damm
cc6f33db30 sh: fix defconfigs for sh7751r boards
This patch fixes up the defconfig for various sh7751r based boards
by updating them to the single cpu subtype CONFIG_CPU_SUBTYPE_SH7751R.
The following sh4 boards are updated: hs7751rvoip, landisk, lboxre2,
systemh, titan.

The current defconfigs with two subtypes defined trigger a configuration
bug which result in kernel configurations with missing board support
code. We end up with kernels without board code and with generic machvec
only.

So we need this patch to make sure the board code gets compiled in.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-03 10:17:47 +09:00
Magnus Damm
5ca95c48f1 sh: fix cf support on r2d boards
This patch makes sure cf support is enabled on R2D-PLUS but disabled
on R2D-1. Without this fix R2D-1 boards hang on bootup.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-03 10:16:29 +09:00
Magnus Damm
a3b3e23015 sh: update r2d defconfig
- Disable CONFIG_SH_DMA to avoid boot up freeze on R2D-1
- Disable CONFIG_SH_STANDARD_BIOS to support R2D-PLUS boot loaders

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-03 10:15:53 +09:00
David McCullough
49dd3ff3a5 sh: update snapgear defconfig.
Updated the snapgear defconfig to get a booting kernel.

Signed-off-by: David McCullough <david_mccullough@au.securecomputing.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-03 10:14:09 +09:00
Paul Mundt
e0a3647fd7 sh: Fix SH-X3 FPU exception handling.
SH-X3 has the FPU exceptions on different vectors completely,
patch in do_fpu_state_restore() to the proper vectors.

Results in a much happier userspace.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-01 16:55:07 +09:00
Paul Mundt
06f862c8ce sh: Fix pgd mismatch from cached TTB in unhandled fault.
When reading the cached TTB value and extracting the pgd, we
accidentally applied a __va() to it and bumped it off in to bogus
space which ended up causing multiple faults in the error path.

Fix it up so unhandled faults don't do strange and highly unorthodox
things when oopsing.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-01 16:39:51 +09:00
Paul Mundt
56c74c733c sh: Don't include fault-nommu on SH-2/SH-2A.
fault-nommu defines the page fault handler stubs for SH-3/4 parts,
but is not needed on SH-2/SH-2A now that the entry code has been
logically separated.

Add it in for SH-3 and SH-4 explicitly.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-01 16:26:21 +09:00
Paul Mundt
c71799433a sh: Fix irqflags tracing for SH-3/4 nommu.
We were missing the trace_hardirqs_on() instrumentation in the nommu
case, resync with the MMU version of the page fault handler to have
this behaving consistently. Also explicitly re-enable IRQs now that
the assembly code isn't doing it for us any more.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-01 16:19:49 +09:00
Paul Mundt
c347d12cd1 sh: Fix lockdep debugging oops on SH-3/4.
In the SH-3/4 TLB access violation path we were enabling IRQs before
the call in to trace_hardirqs_on(), which ended up triggering:

        if (DEBUG_LOCKS_WARN_ON(!irqs_disabled()))
                return;

in kernel/lockdep.c:2031. Fix this up by removing the early re-enable,
we were already re-enabling IRQs post-trace_hardirqs_on() already, so
the semantics are now as was initially intended.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-08-01 16:17:07 +09:00
Paul Mundt
e06c4e5775 sh: Fix fs.h removal from mm.h regressions.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-31 13:01:43 +09:00
David McCullough
c64ac9f058 sh: fix get_wchan() for SH kernels without framepointers
Do not follow the frame pointers (/proc/X/task/1/stat) unless we were
compiled with them.

Signed-off-by: David McCullough <david_mccullough@au.securecomputing.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-26 17:46:07 +09:00
David McCullough
1f25756a11 sh: arch/sh/boot - fix shell usage
Fix the shell call to explicitly use bash, since they are bash
specific and not all systems have bash as the default.

Signed-off-by: David McCullough <david_mccullough@au.securecomputing.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-26 17:43:41 +09:00
Magnus Damm
d89ddd1c84 sh: remove support for sh7300 and solution engine 7300
This patch removes old dead code:
- kill off sh7300 cpu support
- get rid of broken solution engine 7300 board support

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-26 15:37:57 +09:00
Paul Mundt
e257ad062a sh: Kill off virt_to_bus()/bus_to_virt().
Wire up ARCH_NO_VIRT_TO_BUS, and kill off the remaining users. The
dma-mapping code really wanted virt_to_phys()/phys_to_virt() anyways,
there are no inherently special bus addresses.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-26 15:37:40 +09:00
Paul Mundt
ac79fd58a2 sh: Restrict DSP support to specific CPUs.
Not all CPUs support the DSP, and this leads to problems when mixing
and matching CPU types and DSP opcodes. Fix this up by only allowing
CONFIG_SH_DSP to be enabled for the CPUs that explicitly have such a
block.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-26 15:35:24 +09:00
Paul Mundt
b067c50a7f sh: Silence sq compile warning on sh4 nommu.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-26 15:35:17 +09:00
Paul Mundt
3c6b6c7fb7 sh: Kill the rest of the SE73180 cruft.
There was a stray header, and the mach-type removal was also
missed.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-26 15:34:56 +09:00
Magnus Damm
870e8a2438 sh: remove support for sh73180 and solution engine 73180
This patch removes old dead code:
- kill off sh73180 cpu support
- get rid of broken solution engine 73180 board support

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-25 10:49:21 +09:00
Magnus Damm
347b9bdddb sh: remove old broken pint code
The code in arch/sh/kernel/cpu/irq/pint.c doesn't compile, so let's
get rid of it to make space for a future pint implementation on top
of intc.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-25 10:46:32 +09:00
Paul Mundt
f0b859e3d6 sh: Reclaim beginning of P3 space for vmalloc area.
The first 1MB of P3 space was reserved and used for page colouring,
as we've reworked that to use fixmaps, we can reclaim the space and
hand it back to VMALLOC_START.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-25 10:43:47 +09:00
Adrian McMenamin
eb695dbf00 sh: Fix Dreamcast DMA issues.
The current SH DMA API is somewhat broken, not correctly matching
virtual channel to the correct SH DMAC. This wasn't noticeable when
using g2 DMA for the sound driver - one channel 0 is as good as any
other! - but caused the pvr2 driver to fail.

This patch fixes the pvr2 problem and consequently fixes the sound
driver to ensure it continues to function.

Signed-off by: Adrian McMenamin <adrian@mcmen.demon.co.uk>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-24 13:30:55 +09:00
Paul Mundt
8cf1a74305 sh: Add kmap_coherent()/kunmap_coherent() interface for SH-4.
This wires up kmap_coherent() and kunmap_coherent() on SH-4, and
moves away from the p3map_mutex and reserved P3 space, opting to
use fixmaps for colouring instead.

The copy_user_page()/clear_user_page() implementations are moved
to this, which fixes the nasty blowups with spinlock debugging
as a result of having some of these calls nested under the page
table lock.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-24 13:28:26 +09:00
Thomas Gleixner
18de5bc4c1 clockevents: fix resume logic
We need to make sure, that the clockevent devices are resumed, before
the tick is resumed. The current resume logic does not guarantee this.

Add CLOCK_EVT_MODE_RESUME and call the set mode functions of the clock
event devices before resuming the tick / oneshot functionality.

Fixup the existing users.

Thanks to Nigel Cunningham for tracking down a long standing thinko,
which affected the jinxed VAIO.

[akpm@linux-foundation.org: xen build fix]
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: john stultz <johnstul@us.ibm.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-21 17:49:15 -07:00
Magnus Damm
56386f6424 sh: intc - add support for SH7750 and its variants
This patch converts the cpu specific 7750 setup code to use the
new intc controller. Many new vectors are added and multiple
processor variants including 7091, 7750, 7750s, 7750r, 7751 and
7751r should all have the correct vectors hooked up.

IRLM interrupts can be enabled using ipr_irq_enable_irlm() which
now is marked as __init.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 18:44:49 +09:00
Paul Mundt
339547bf5d sh: Move entry point code to .text.head.
Follow Al Viro's m68k change from l-k:

	i.e. tell modpost that entry point code (that has to be outside
	of .init.text for external reasons) is OK to refer to .init.*

Shuts up some section mismatch warnings from modpost.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 17:40:03 +09:00
Paul Mundt
bf73af2314 sh: heartbeat: Shut up resource size warning.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 17:35:42 +09:00
Magnus Damm
ee54505199 sh: update r2d defconfig and fix SH7751R pci compliation
This patch updates the r2d board support in a few ways:

- CPU_SUBTYPE_SH7751R is selected in the defconfig to play well
  with the r2d board Kconfig entry. Without this the defconfig
  results in no board enabled.

- Enable EARLY_PRINTK.
- Enable SH_STANDARD_BIOS
	- this works well for early printk on the r2d board.

- Add "earlyprink=bios" to the cmdline for early serial port
  output by default.

- CONFIG_SUBTYPE_SH7751R support is added to the sh-specific
  pci makefile.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 17:25:48 +09:00
Paul Mundt
98d877c487 sh: Many symbol exports for nommu allmodconfig.
allmodconfig generates a lot of interesting code, a lot of the
generated symbols we've never exported before, so this fixes
those up. Verified with both GCC3 and GCC4 toolchains.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 16:59:49 +09:00
Magnus Damm
3a353824df sh: zero terminate 8250 platform data for r2d board
struct plat_serial8250_port should contain a terminating zero entry

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 16:34:22 +09:00
Paul Mundt
6894e0a943 sh: cpufreq: Fix up the build for SH-2.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 16:03:02 +09:00
Paul Mundt
fc10020292 sh: Make on-chip DMA channel selection explicit.
Currently this has a prompt to allow users to change it. There's
no reason to do this, and it has caused breakage and confusion
in the past, so remove it entirely.

We'll get rid of this when the whole driver is tidied for
the driver model.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 14:24:57 +09:00
Paul Mundt
3e7678333a sh: Fix up CPU dependencies for on-chip DMAC.
We only handle SH-3 and SH-4 at present, don't expose it to the
other CPUs.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 14:23:04 +09:00
Paul Mundt
da9f0ac2f1 Merge branch 'clkfwk' 2007-07-20 13:38:49 +09:00
Paul Mundt
cb5ec75b8b sh: cpufreq: clock framework support.
This gets the SH cpufreq working again. We follow the changes
in the AVR32 implementation for wrapping in to the clock framework.
CPUs that wish to use this are required to define rate rounding
primitives in order to satisfy clk_round_rate().

This works well enough for the common case, though we should
look at unifying this driver across all of the platforms that
implement clock framework support in one capacity or another.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 13:38:19 +09:00
Paul Mundt
fe04d7798f sh: Support rate rounding for SH7722 FRQCR clocks.
Now that the round_rate() op is supported, hook it up on SH7722
for the FRQCR (CPU, PCLK, etc.) clocks.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 13:30:07 +09:00
Paul Mundt
f6991b0456 sh: Implement clk_round_rate() in the clock framework.
This is an optional component of the clock framework. However,
as we're going to be using this in the cpufreq drivers, add
support for it to the framework.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 13:29:09 +09:00
Paul Mundt
bd5f0d1c25 sh: Fix up PCI section mismatch warnings.
Flag pcibios_setup() and pcibios_fixup_bus() as __devinit.
Follows the sh64 change.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 13:22:47 +09:00
Paul Mundt
0c99adb0a6 sh: Wire up fallocate() syscall.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:27:09 +09:00
Magnus Damm
39c7aa9ea9 sh: intc - add support for 7780
This patch converts the cpu specific 7780 setup code to use the
new intc controller. Many new vectors are added and also support for
external interrupt sense configuration. So with this patch it is now
possible to configure external interrupt pins as edge or level
triggered using set_irq_type().

No external interrupts are registered by default.
Use plat_irq_setup_pins() to select between IRQ or IRL mode.

This patch also fixes the Alarm IRQ for the RTC.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:18:21 +09:00
Magnus Damm
680c45981a sh: intc - improve group support
This patch improves intc group support, ie it makes it possible to
group interrupts together and mask / unmask the entire group. This
also works with priorities, so setting a priority for an entire group
is also possible. This patch is needed to properly support certain
processors such as the 7780.

Fixes for NULL pointers in DECLARE_INTC_DESC() are also included.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:18:21 +09:00
Paul Mundt
e5e3dfc80a sh: push-switch: Correct license string.
This was accidentally set as "GPLv2", whereas the kernel expects v2
to be written "GPL v2", this caused complaints regarding the use
of the platform device APIs when built as a module.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:18:20 +09:00
Paul Mundt
792e947a29 sh: cpufreq: Fix driver dependencies and flag as broken.
This is only supported on SH-4, so don't expose it for the other
CPUs. Additionally, it's suffered some bitrot, so add a BROKEN
dependency as well until we fix it up.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:18:20 +09:00
Magnus Damm
90015c8938 sh: IPR/INTC2 IRQ setup consolidation.
This patch unifies the cpu specific interrupt setup functions for
interrupt controller blocks such as ipr, intc2 and intc. There is no
point in having separate functions for each interrupt controller, so
let's clean this up.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:18:20 +09:00
Magnus Damm
493a358e0a sh: clean up interrupt code for solution engine 7722 board
This patch cleans up solution engine 7722 specific interrupt code.
The main purpose is to replace the mux function with use of
set_irq_chained_handler() and replace hard coded register poking
code with set_irq_type(). The board specific interrupts are also
moved to start from SE7722_FPGA_IRQ_BASE.

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:18:20 +09:00
Magnus Damm
1b06428ee5 sh: intc - add support for 7722 processor
This patch converts the cpu specific 7722 setup code to use the
new intc controller. Many new vectors are added and also support
for external interrupt sense configuration. So with this patch
it is now possible to configure external interrupt pins as edge
or level triggered using set_irq_type().

Signed-off-by: Magnus Damm <damm@igel.co.jp>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2007-07-20 12:18:20 +09:00