Blackfin arch: Fix BUG - Kernel does not link when CONFIG_CPU_FREQ_STAT=m

to reproduce:
$ make BF537-STAMP_defconfig
$ make menuconfig
  CPU Frequency scaling  --->
    [*] CPU Frequency scaling
    <M>   CPU frequency translation statistics
$ make

arch/blackfin/mach-common/built-in.o: In function `bfin_target':
arch/blackfin/mach-common/cpufreq.c:(.text+0xf1e): undefined reference to
`cpufreq_frequency_table_target'
arch/blackfin/mach-common/built-in.o: In function `bfin_verify_speed':
arch/blackfin/mach-common/cpufreq.c:(.text+0x1056): undefined reference to
`cpufreq_frequency_table_verify'
arch/blackfin/mach-common/built-in.o: In function `__bfin_cpu_init':
arch/blackfin/mach-common/cpufreq.c:(.init.text+0x554): undefined reference to
`cpufreq_frequency_table_get_attr'
arch/blackfin/mach-common/cpufreq.c:(.init.text+0x56a): undefined reference to
`cpufreq_frequency_table_cpuinfo'
arch/blackfin/mach-common/built-in.o: In function `sys_call_table':
(.data+0x83c): undefined reference to
`cpufreq_freq_attr_scaling_available_freqs'
make: *** [.tmp_vmlinux1] Error 1

Blackfin wants CPU_FREQ_TABLE

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: rename cache_lock() to bfin_cache_lock()

rename cache_lock() to bfin_cache_lock() to avoid namespace collision
with common code

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: use generic xor.h

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: cleanup the time.c code

 - make the code a bit more readable
 - kill of warnings/ifdef mess a bit

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: fix building with CONFIG_DEBUG_DOUBLEFAULT

arch/blackfin/mach-common/entry.S:465: Error: pcrel too far
BFD_RELOC_BFIN_10

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: do not define decode_instruction if hwtrace is turned off

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: rename irq_flags to bfin_irq_flags

rename irq_flags to bfin_irq_flags to avoid namespace
collision with common code

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: move out irq related functions

move irq related functions into asm/irq.h and out of the mondo asm/system.h

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: move the PORT registers down

Move the PORT registers down as well since they may utilize
asm/system.h code

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: introduce an IM_MEM macro to kgdb

create an IN_MEM() macro to simplify comparing an address in an on-chip
region of memory and make things readable

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: fix off-by-one errors on end of memory range

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: delete unused vars and add parenthesis to fixup warnings

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: move fixed code into init section

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: remove superficial check on blackfin watchdog

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: set PM_WAKEUP_GPIO_NUMBER default to 2 for everyone

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: delay PLL_CTL/VR_CTL wrappers

Delay PLL_CTL/VR_CTL wrappers as much as possible to avoid
the inter-dependency problems with cdef and common headers

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: dont export ack_bad_irq as no one else does

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: move SIC_SYSIRQ() to the one file that actually uses it

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: move ack_bad_irq() prototype to hardirq.h to match other arches

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: delete old comments and NO_IRQ define

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: overlay thread.usp over PT_USP

overlay thread.usp over PT_USP when getting the whole regfile to
match PT_USP behavior

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: rewrite blackfin_invalidate_entire_dcache function

rewrite blackfin_invalidate_entire_dcache() in C for easier management,
better optimization, and so we take all SSYNC anomalies into account

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: make ADI board names consistent

Cc: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: Fix bug - hardware breakpoint doesn't always work in kgdb

Hardware breakpoint doesn't always work in kgdb. It works at the first
two times, but if you repeatedly trigger that hardware breakpoint, it
would slip over that point once in two times.

Fix it by always setting hw bp skip to 0. gdb does skip after hw bp trap.

Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: fix unused warning for some blackfin derivatives

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: Use GPIO_BANKSIZE macro to replace const number 16 for GPIO_BANK_NUM macro caculating

Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: SMP supporting patchset: some other misc code

Blackfin dual core BF561 processor can support SMP like features.
https://docs.blackfin.uclinux.org/doku.php?id=linux-kernel:smp-like

In this patch, we provide SMP extend to some other misc code

Singed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: SMP supporting patchset: Blackfin kernel and memory management code

Blackfin dual core BF561 processor can support SMP like features.
https://docs.blackfin.uclinux.org/doku.php?id=linux-kernel:smp-like

In this patch, we provide SMP extend to Blackfin kernel and memory management code

Singed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: SMP supporting patchset: Blackfin CPLB related code

Blackfin dual core BF561 processor can support SMP like features.
https://docs.blackfin.uclinux.org/doku.php?id=linux-kernel:smp-like

In this patch, we provide SMP extend to Blackfin CPLB related code

Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: SMP supporting patchset: Blackfin header files and machine common code

Blackfin dual core BF561 processor can support SMP like features.
https://docs.blackfin.uclinux.org/doku.php?id=linux-kernel:smp-like

In this patch, we provide SMP extend to Blackfin header files
and machine common code

Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: SMP supporting patchset: BF561 related code

Blackfin dual core BF561 processor can support SMP like features.
https://docs.blackfin.uclinux.org/doku.php?id=linux-kernel:smp-like

In this patch, we provide SMP extend to BF561 kernel code

Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: use physmap in board resources rather than legacy config method

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: Convert Blackfin GPIO driver to use common gpiolib/gpiochip infrastructure

 - This patch adds support for ARCH_WANT_OPTIONAL_GPIOLIB.
 - It may be changed in future to ARCH_REQUIRE_GPIOLIB.
 - Change GPIO_BANK_NUM use DIV_ROUND_UP( , ) macro

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: Cleanup and unify Blackfin IRQ and GPIO IRQ handling

 - Remove SSYNC()
 - Use irq_to_gpio where applicable
 - Remove gpio_edge_triggered bitfield, check irq_desc fields instead.
 - Remove gpio_both_edge_triggeredb bitfield, check irq_desc fields
   instead.
 - Use BITMAP and bitops on gpio_enabled
 - Preferably use 32-bit
 - Looking at the disassembly this indeed saves quite a few instructions.

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: workaround bug: spi_mmc driver working in dma mode may hang the system

Disable spi_mmc DMA by default.

Signed-off-by: Yi Li <yi.li@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: remove useless SSYNC() in irq priority code

 - remove SSYNC() left over from irq init split
 - do not force SSYNC() when masking/unmasking IRQs in the SIC
   as any order enforced by the hardware should already be enforced
   by software

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: fix bug - gpio_bank() macros messed up bank number caculating with positioning a gpio

The whole story:

Before BF51x merged, all the MAX_BLACKFIN_GPIOS are integral multiple of GPIO_BANKSIZE (= 16).
But BF51x provides MAX_BLACKFIN_GPIOS = 40 which includes 3 banks and the 3rd bank has only 8
GPIO pins.

Therefore, gpio_bank() macros is correct when you try to find a GPIO in which bank (GPIO_35 is
in bank 2). But on BF51x gpio_bank(MAX_BLACKFIN_GPIOS) only gives out 2 banks instead of 3
banks for some static array initialization.

This patch add a new macros gpio_bank_n() and GPIO_BANK_NUM to do bank number caculating and
remain the gpio_bank() macros for positioning a gpio in which bank.

Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: enable i2c_bfin_twi0_device in cm_bf548 board

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: Enable ISP1760 USB Host Driver in platform device initialization code.

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: fixup get_user() macros

 - to avoid uninitialized errors
 - make invalid sizes a build error

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: do not bother initializing the first 4k

our kernel should be sane now, and we want to catch NULL bugs,
not ignore them

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: use sti to set the mask rather than banging on imask

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: don't assume a specific SPI flash part - take whatever you probe

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: update defconfig file for all boards

 - do not bother generating deprecated /sys files by default now since
   mdev does not need it
 - Don't built-in char sport driver and build it as a module in defconfig
 - disable CONFIG_DEVKMEM by default
 - enable spi flash driver on boards that have one
 - switch config to the NAND platfrom driver rather than the bfin async one
 - do not make BFIN_DMA_5XX optional since a large portion of our code relies
   on dma functions existing

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: remove most BUG_ON channel checks

keep BUG_ON in DMA request, free and set_dma_callback.

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: only add IFLUSH nop padding when anomaly 443 is enabled

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: Remove useless SSYNCs in DMA code

Tons of SSYNC operation will impact the DMA performance

Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: lookup channel2irq() only once

Add irq to struct dma_channel lookup channel2irq() only once,
since channel2irq() is fairly large on some Blackfin derivatives.

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: update anomaly lists to match latest sheets

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: refine the gpio check

refine the gpio check in peripheral_request() so that it only
checks pins that can be used as both GPIO and a peripheral

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: unify port_setup() to reduce arch differences

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: unify peripheral_request() to reduce arch differences

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: fix cmp_label() so it doesnt incorrectly accept partial leading matches

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: unify check_gpio() to reduce arch differences

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: remove unused local define

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: add support for Blackfin latest processor family BF51x

Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: put quotes around error to silence a ton of warnings from gcc-4.3

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: fix bug - kernel with SMP patch can not bootup

The original code defined _exception_stack but not alloc space for the exception
stack. In exception, this area is over written by exception stack. Common kernel
luckly boot up, but SMP kernel stuck.

Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: drop redundant BFIN_DMA_5XX depends

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: drop unused exports and comment remaining exports

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: add an option to allow people to stick exception stack into L1 scratch

allow people to stick exception stack into L1 scratch
and make sure it gets placed into .bss sections rather than .data

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: move EXPORT_SYMBOL to the place where it is actually defined

 - kernel_thread
 - irq_flags
 - checksum

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: BF538/9 Linux kernel Support

Add supporing for Blackfin BF538 and BF539 processors.

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Blackfin arch: unify duplicated bss init code

Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>

Fix up 64-bit byte swaps for most 32-bit architectures

The __SWAB_64_THRU_32__ case of a 64-bit byte swap was depending on the
no-longer-existant ___swab32() method (three underscores).  We got rid
of some of the worst indirection and complexity, and now it should just
use the 32-bit swab function that was defined right above it.

Reported-and-tested-by: Nicolas Pitre <nico@cam.org>
Reported-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

byteorder: remove the now unused byteorder.h

This implementation caused problems in userspace which can, and does
define _both_ __LITTLE_ENDIAN and __BIG_ENDIAN.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mn10300: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

frv: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

m32r: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

m68knommu: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

h8300: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ia64: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

blackfin: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

arm: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

alpha: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

m68k: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

x86: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sparc: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sh: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

s390: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

powerpc: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mips: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

avr32: introduce asm/swab.h

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

byteorder: only use linux/swab.h

The first step to make swab.h a regular header that will
include an asm/swab.h with arch overrides.

Avoid the gratuitous differences introduced in the new
linux/swab.h by naming the ___constant_swabXX bits and
__fswabXX bits exactly as found in the old implementation
in byteorder/swab[b].h

Use this new swab.h in byteorder/[big|little]_endian.h and
remove the two old swab headers.

Although the inclusion of asm/byteorder.h looks strange in
linux/swab.h, this will allow each arch to move the actual
arch overrides for the swab bits in an asm file and then
the includes can be cleaned up without requiring a flag day
for all arches at once.

Keep providing __fswabXX in case some userspace was using them
directly, but the revised __swabXX should be used instead in
any new code and will always do constant folding not dependent
on the optimization level, which means the __constant versions
can be phased out in-kernel.

Arches that use the old-style arch macros will lose their
optimized versions until they move to the new style, but at
least they will still compile.  Many arches have already moved
and the patches to move the remaining arches are trivial.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge branch 'drm-next' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm: fix ordering of driver unload vs agp unload.
  drm/i915: Respect the other stolen memory sizes we know of.
  drm/i915: Non-mobile parts don't have integrated TV-out.
  drm/i915: Add support for integrated HDMI on G4X hardware.
  drm/i915: Pin cursor bo and unpin old bo when setting cursor.
  drm/i915: Don't allow objects to get bound while VT switched.

drm: fix ordering of driver unload vs agp unload.

For KMS drivers, we really need to cleanup the driver before disabling
the AGP subsystem.

Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/i915: Respect the other stolen memory sizes we know of.

fd.o bug #19336.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@linux.ie>

drm/i915: Non-mobile parts don't have integrated TV-out.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@linux.ie>

drm/i915: Add support for integrated HDMI on G4X hardware.

This is ported directly from the userland 2D driver code.  The HDMI audio bits
aren't hooked up yet.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@linux.ie>

drm/i915: Pin cursor bo and unpin old bo when setting cursor.

We also didn't track the cursor bo before and would leak a reference
when the cursor image was change.

Signed-off-by: Kristian Høgsberg <krh@redhat.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@linux.ie>

drm/i915: Don't allow objects to get bound while VT switched.

This avoids a BUG_ON in the enter_vt path due to objects being in the GTT
when we shouldn't have ever let them be (as we're not supposed to touch the
device during that time).

This was triggered by a change in the 2D driver to use the GTT mapping of
objects after pinning them to improve software fallback performance.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@linux.ie>

Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input: (29 commits)
  Input: i8042 - add Dell Vostro 1510 to nomux list
  Input: gtco - use USB endpoint API
  Input: add support for Maple controller as a joystick
  Input: atkbd - broaden the Dell DMI signatures
  Input: HIL drivers - add MODULE_ALIAS()
  Input: map_to_7segment.h - convert to __inline__ for userspace
  Input: add support for enhanced rotary controller on pxa930 and pxa935
  Input: add support for trackball on pxa930 and pxa935
  Input: add da9034 touchscreen support
  Input: ads7846 - strict_strtoul takes unsigned long
  Input: make some variables and functions static
  Input: add tsc2007 based touchscreen driver
  Input: psmouse - add module parameters to control OLPC touchpad delays
  Input: i8042 - add Gigabyte M912 netbook to noloop exception table
  Input: atkbd - Samsung NC10 key repeat fix
  Input: atkbd - add keyboard quirk for HP Pavilion ZV6100 laptop
  Input: libps2 - handle 0xfc responses from devices
  Input: add support for Wacom W8001 penabled serial touchscreen
  Input: synaptics - report multi-taps only if supported by the device
  Input: add joystick driver for Walkera WK-0701 RC transmitter
  ...

Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/security-testing-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
  CRED: Fix regression in cap_capable() as shown up by sys_faccessat() [ver #3]
  Revert "CRED: Fix regression in cap_capable() as shown up by sys_faccessat() [ver #2]"
  SELinux: shrink sizeof av_inhert selinux_class_perm and context
  CRED: Fix regression in cap_capable() as shown up by sys_faccessat() [ver #2]
  keys: fix sparse warning by adding __user annotation to cast
  smack: Add support for unlabeled network hosts and networks
  selinux: Deprecate and schedule the removal of the the compat_net functionality
  netlabel: Update kernel configuration API

Merge branch 'timers-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  hrtimer: splitout peek ahead functionality, fix
  hrtimer: fixup comments
  hrtimer: fix recursion deadlock by re-introducing the softirq
  hrtimer: simplify hotplug migration
  hrtimer: fix HOTPLUG_CPU=n compile warning
  hrtimer: splitout peek ahead functionality

Merge branch 'sched-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched: fix section mismatch
  sched: fix double kfree in failure path
  sched: clean up arch_reinit_sched_domains()
  sched: mark sched_create_sysfs_power_savings_entries() as __init
  getrusage: RUSAGE_THREAD should return ru_utime and ru_stime
  sched: fix sched_slice()
  sched_clock: prevent scd->clock from moving backwards, take #2
  sched: sched.c declare variables before they get used

Merge branch 'irq-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'irq-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  genirq: provide irq_to_desc() to non-genirq architectures too

Merge branch 'core-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  rcu: fix rcutorture bug
  rcu: eliminate synchronize_rcu_xxx macro
  rcu: make treercu safe for suspend and resume
  rcu: fix rcutree grace-period-latency bug on small systems
  futex: catch certain assymetric (get|put)_futex_key calls
  futex: make futex_(get|put)_key() calls symmetric
  locking, percpu counters: introduce separate lock classes
  swiotlb: clean up EXPORT_SYMBOL usage
  swiotlb: remove unnecessary declaration
  swiotlb: replace architecture-specific swiotlb.h with linux/swiotlb.h
  swiotlb: add support for systems with highmem
  swiotlb: store phys address in io_tlb_orig_addr array
  swiotlb: add hwdev to swiotlb_phys_to_bus() / swiotlb_sg_to_bus()

Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ASoC: Fix SND_SOC_ALL_CODECS handling of dual SPI and I2C control buses
  ASoC: Use snd_soc_dapm_nc_pin() in at91sam9g20ek
  ASoC: TWL4030: Convert the bitfield enums to VALUE_ENUM type
  ASoC: New enum type: value_enum
  pxa2xx-ac97: switch AC unit to correct state before probing
  ASoC: Clocking fixes for davinci-evm.c
  ASoC: Mark non-connected TWL4030 pins for pandora
  ASoC: OMAP: Select OMAP pin multiplexing when using Nokia N810 ASoC drivers

Merge git://git./linux/kernel/git/gregkh/staging-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (237 commits)
  Staging: android: binder: fix build errors
  Staging: android: add lowmemorykiller driver
  Staging: android: remove dummy android.c driver
  Staging: android: timed_gpio: Rename android_timed_gpio to timed_gpio
  Staging: android: add timed_gpio driver
  Staging: android: add ram_console driver
  Staging: android: add logging driver
  staging: android: binder: Fix use of euid
  Staging: android: binder: Fix gcc warnings about improper format specifiers for size_t in printk
  Staging: android: add binder driver
  Staging: add android framework
  Staging: epl: fix netdev->priv b0rkage
  Staging: epl: hr timers all run in hard irq context now
  Staging: epl: run Lindent on *.c files
  Staging: epl: run Lindent on *.h files
  Staging: epl: run Lindent on all user/*.h files
  Staging: epl: run Lindent on all kernel/*.h files
  Staging: add epl stack
  Staging: frontier: fix compiler warnings
  Staging: frontier: remove unused alphatrack_sysfs.c file
  ...

Merge git://git./linux/kernel/git/gregkh/driver-core-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6: (60 commits)
  uio: make uio_info's name and version const
  UIO: Documentation for UIO ioport info handling
  UIO: Pass information about ioports to userspace (V2)
  UIO: uio_pdrv_genirq: allow custom irq_flags
  UIO: use pci_ioremap_bar() in drivers/uio
  arm: struct device - replace bus_id with dev_name(), dev_set_name()
  libata: struct device - replace bus_id with dev_name(), dev_set_name()
  avr: struct device - replace bus_id with dev_name(), dev_set_name()
  block: struct device - replace bus_id with dev_name(), dev_set_name()
  chris: struct device - replace bus_id with dev_name(), dev_set_name()
  dmi: struct device - replace bus_id with dev_name(), dev_set_name()
  gadget: struct device - replace bus_id with dev_name(), dev_set_name()
  gpio: struct device - replace bus_id with dev_name(), dev_set_name()
  gpu: struct device - replace bus_id with dev_name(), dev_set_name()
  hwmon: struct device - replace bus_id with dev_name(), dev_set_name()
  i2o: struct device - replace bus_id with dev_name(), dev_set_name()
  IA64: struct device - replace bus_id with dev_name(), dev_set_name()
  i7300_idle: struct device - replace bus_id with dev_name(), dev_set_name()
  infiniband: struct device - replace bus_id with dev_name(), dev_set_name()
  ISDN: struct device - replace bus_id with dev_name(), dev_set_name()
  ...