openwrt/staging/blogic.git
18 years agoV4L/DVB (3723): Avoid unnecessary firmware re-loads in or51132 frontend
Trent Piepho [Tue, 4 Apr 2006 12:30:29 +0000 (09:30 -0300)]
V4L/DVB (3723): Avoid unnecessary firmware re-loads in or51132 frontend

As QAM_64, QAM_256, and QAM_AUTO all use the same firmware, switching
between these modulations doesn't require a firmware re-load.  This also
fixes a mishandled error condition, in which the firmware file is loaded
into the kernel, the clock mode is changed, but then the firmware upload
to the device fails.  The modulation change is aborted, but the clock
mode would still be changed.

Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Andrew de Quincey <adq_dvb@lidskialf.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3721): Fix spelling
Trent Piepho [Tue, 4 Apr 2006 04:56:30 +0000 (01:56 -0300)]
V4L/DVB (3721): Fix spelling

It's "Terrestrial"

Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3720): Fix spelling error / typo in comments
Michael Krufky [Mon, 3 Apr 2006 04:29:09 +0000 (01:29 -0300)]
V4L/DVB (3720): Fix spelling error / typo in comments

Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3718): Removed trailing newlines
Michael Krufky [Sun, 2 Apr 2006 22:11:10 +0000 (19:11 -0300)]
V4L/DVB (3718): Removed trailing newlines

Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3715): Change all emails to the currently used one.
Mauro Carvalho Chehab [Mon, 3 Apr 2006 10:53:40 +0000 (07:53 -0300)]
V4L/DVB (3715): Change all emails to the currently used one.

Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3712): Fix video input setting of em28xx, use _INT_S_VIDEO_ROUTING in tvp5150
Hans Verkuil [Sun, 2 Apr 2006 16:35:00 +0000 (13:35 -0300)]
V4L/DVB (3712): Fix video input setting of em28xx, use _INT_S_VIDEO_ROUTING in tvp5150

- Use new routing input defines in em28xx-cards.c
- Fix S-Video settings for tvp5150-based cards (input was copied from saa7115
  based cards and worked only because S-Video was selected in the default: case)
- Replace VIDIOC_S_INPUT by VIDIOC_INT_S_VIDEO_ROUTING in em28xx-video.c
- Remove the now obsolete VIDIOC_S_INPUT handler in saa7115.c
- Add VIDIOC_INT_G/S_VIDEO_ROUTING in tvp5150.c
- Add new media/tvp5150.h with the routing defines.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3711): Add support for VIDIOC_INT_S_CRYSTAL_FREQ internal command.
Hans Verkuil [Sun, 2 Apr 2006 15:50:42 +0000 (12:50 -0300)]
V4L/DVB (3711): Add support for VIDIOC_INT_S_CRYSTAL_FREQ internal command.

Some saa7115-based cards use a different crystal frequency and a different
audio clock generation. Add a new VIDIOC_INT_S_CRYSTAL_FREQ command to be
able to set these values.
Also change the default APLL setting to 0. It makes no sense to have the
audio clock independent from the video clock, this can lead to audio/video
synchronization problems. Setting this to 0 is also consistent with the old
saa7114.c source and the way the Hauppauge Windows driver sets it.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3709): Improve line-in handling
Hans Verkuil [Sun, 2 Apr 2006 11:21:02 +0000 (08:21 -0300)]
V4L/DVB (3709): Improve line-in handling

- improve handling of the EXTERN input: don't start an unnecessary carrier scan
- improve the LOG_STATUS output
- ensure that a carrier scan is started again when switching back to the tuner.
- set correct prescale for L-NICAM

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoV4L/DVB (3708a): Clean up some cruft in or51132 frontend
Trent Piepho [Sun, 2 Apr 2006 23:40:33 +0000 (20:40 -0300)]
V4L/DVB (3708a): Clean up some cruft in or51132 frontend

There is some old code in the or51133 firmware loading function that
has no effect.  Left a comment to in case it helps someone trying
to reverse engineer the chip.

Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Andrew de Quincey <adq_dvb@lidskialf.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
18 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Sun, 25 Jun 2006 00:48:14 +0000 (17:48 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm: (25 commits)
  [ARM] 3648/1: Update struct ucontext layout for coprocessor registers
  [ARM] Add identifying number for non-rt sigframe
  [ARM] Gather common sigframe saving code into setup_sigframe()
  [ARM] Gather common sigframe restoration code into restore_sigframe()
  [ARM] Re-use sigframe within rt_sigframe
  [ARM] Merge sigcontext and sigmask members of sigframe
  [ARM] Replace extramask with a full copy of the sigmask
  [ARM] Remove rt_sigframe puc and pinfo pointers
  [ARM] 3647/1: S3C24XX: add Osiris to the list of simtec pm machines
  [ARM] 3645/1: S3C2412: irq support for external interrupts
  [ARM] 3643/1: S3C2410: Add new usb clocks
  [ARM] 3642/1: S3C24XX: Add machine SMDK2413
  [ARM] 3641/1: S3C2412: Fixup gpio register naming
  [ARM] 3640/1: S3C2412: Use S3C24XX_DCLKCON instead of S3C2410_DCLKCON
  [ARM] 3639/1: S3C2412: serial port support
  [ARM] 3638/1: S3C2412: core clocks
  [ARM] 3637/1: S3C24XX: Add mpll clock, and set as fclk parent
  [ARM] 3636/1: S3C2412: Add selection of CPU_ARM926
  [ARM] 3635/1: S3C24XX: Add S3C2412 core cpu support
  [ARM] 3633/1: S3C24XX: s3c2410 gpio bugfix - wrong pin nos
  ...

18 years agoRevert "[PATCH] usb: drivers/usb/core/devio.c dereferences a userspace pointer"
Linus Torvalds [Sun, 25 Jun 2006 00:47:09 +0000 (17:47 -0700)]
Revert "[PATCH] usb: drivers/usb/core/devio.c dereferences a userspace pointer"

This reverts commit 786dc1d3d7333f269e17d742886eac2188a2d9cc.

As Al so eloquently points out, the patch is crap. The old code was fine,
the new code was bogus.

It never dereferenced a user pointer, the "->" operator was to an array
member, which gives the _address_ of the member (in user space), not an
actual dereference at all.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge signal handler branch
Russell King [Sat, 24 Jun 2006 22:53:01 +0000 (23:53 +0100)]
Merge signal handler branch

18 years ago[ARM] 3648/1: Update struct ucontext layout for coprocessor registers
Daniel Jacobowitz [Sat, 24 Jun 2006 22:46:21 +0000 (23:46 +0100)]
[ARM] 3648/1: Update struct ucontext layout for coprocessor registers

Patch from Daniel Jacobowitz

In order for userspace to find saved coprocessor registers, move them from
struct rt_sigframe into struct ucontext.  Also allow space for glibc's
sigset_t, so that userspace and kernelspace can use the same ucontext
layout.  Define the magic numbers for iWMMXt in the header file for easier
reference.  Include the size of the coprocessor data in the magic numbers.

Also define magic numbers and layout for VFP, not yet saved.

Signed-off-by: Daniel Jacobowitz <dan@codesourcery.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years agoMerge branch 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Sat, 24 Jun 2006 21:48:43 +0000 (14:48 -0700)]
Merge branch 'drm-patches' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm: radeon constify radeon microcode
  Add i915 ioctls to configure pipes for vblank interrupt.
  drm: update radeon to 1.25 add r200 vertex program support
  drm: radeon add a tcl state flush before accessing tcl vector space

18 years agoMerge branch 'intelfb-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airli...
Linus Torvalds [Sat, 24 Jun 2006 21:48:35 +0000 (14:48 -0700)]
Merge branch 'intelfb-patches' of git://git./linux/kernel/git/airlied/intelfb-2.6

* 'intelfb-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/intelfb-2.6:
  intelfb delousing

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Sat, 24 Jun 2006 21:48:24 +0000 (14:48 -0700)]
Merge /pub/scm/linux/kernel/git/davem/sparc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6: (40 commits)
  [SPARC64]: Update defconfig.
  [SPARC64]: Make auxio a real driver.
  [PARPORT] sunbpp: Convert to new SBUS device framework.
  [Documentation]: Update probing info in sbus_drivers.txt
  [SCSI] qlogicpti: Convert to new SBUS device framework.
  [SCSI] esp: Fix bug in esp_remove_common.
  [NET] sunhme: Kill useless loop over sdevs in quattro_sbus_find().
  [NET] myri_sbus: Kill unused next_module struct member.
  [NET] myri_sbus: Convert to new SBUS device layer.
  [NET] sunqe: Convert to new SBUS driver layer.
  [NET] sunbmac: Convert over to new SBUS device framework.
  [NET] sunlance: Convert to new SBUS driver framework.
  [NET] sunhme: Convert to new SBUS driver framework.
  [NET] sunhme: Kill __sparc__ and __sparc_v9__ ifdefs.
  [SCSI] sparc: Port esp to new SBUS driver layer.
  [SOUND] sparc: Port amd7930 to new SBUS device layer.
  [SBUS]: Rewrite and plug into of_device framework.
  [SPARC]: Port of_device layer and make ebus use it.
  [SPARC]: Port sparc64 in-kernel device tree code to sparc32.
  [SPARC64]: Add of_device layer and make ebus/isa use it.
  ...

18 years agoEnable minimal per-device resume tracing
Linus Torvalds [Sat, 24 Jun 2006 21:32:18 +0000 (14:32 -0700)]
Enable minimal per-device resume tracing

This is the minimal resume trace code to find which device resume (if
any) results in problems.  Usually, you'd use the information this
generates as a starting point to adding more fine-grained trace event
points.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoAdd some basic resume trace facilities
Linus Torvalds [Sat, 24 Jun 2006 21:27:42 +0000 (14:27 -0700)]
Add some basic resume trace facilities

Considering that there isn't a lot of hw we can depend on during resume,
this is about as good as it gets.

This is x86-only for now, although the basic concept (and most of the
code) will certainly work on almost any platform.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[ARM] Add identifying number for non-rt sigframe
Russell King [Sat, 24 Jun 2006 21:41:09 +0000 (22:41 +0100)]
[ARM] Add identifying number for non-rt sigframe

GDB couldn't reliably tell the difference between the old and new
non-rt sigframes, so provide it with a number at the beginning which
will never appear in the old sigframe, and hence provide gdb with a
reliable way to tell the two apart.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years agoMerge S3C24xx branch
Russell King [Sat, 24 Jun 2006 21:21:49 +0000 (22:21 +0100)]
Merge S3C24xx branch

18 years agoMerge ep93xx branch
Russell King [Sat, 24 Jun 2006 21:21:35 +0000 (22:21 +0100)]
Merge ep93xx branch

18 years ago[ARM] 3647/1: S3C24XX: add Osiris to the list of simtec pm machines
Ben Dooks [Sat, 24 Jun 2006 20:22:35 +0000 (21:22 +0100)]
[ARM] 3647/1: S3C24XX: add Osiris to the list of simtec pm machines

Patch from Ben Dooks

Enable power management for Simtec Osiris

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3645/1: S3C2412: irq support for external interrupts
Ben Dooks [Sat, 24 Jun 2006 20:21:37 +0000 (21:21 +0100)]
[ARM] 3645/1: S3C2412: irq support for external interrupts

Patch from Ben Dooks

Move the decoding of the IRQ_EXT4 and above out of
the entry macro, and into an chained irq handler
as the EXTINT registers move depending on the CPU
being used.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3643/1: S3C2410: Add new usb clocks
Ben Dooks [Sat, 24 Jun 2006 20:21:36 +0000 (21:21 +0100)]
[ARM] 3643/1: S3C2410: Add new usb clocks

Patch from Ben Dooks

Make the S3C2410 use the same usb clock naming
as the S3C2412

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3642/1: S3C24XX: Add machine SMDK2413
Ben Dooks [Sat, 24 Jun 2006 20:21:35 +0000 (21:21 +0100)]
[ARM] 3642/1: S3C24XX: Add machine SMDK2413

Patch from Ben Dooks

Add basic support for the Samsung/Aiji SMDK2413

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3641/1: S3C2412: Fixup gpio register naming
Ben Dooks [Sat, 24 Jun 2006 20:21:33 +0000 (21:21 +0100)]
[ARM] 3641/1: S3C2412: Fixup gpio register naming

Patch from Ben Dooks

The current S3C2412 has used to moving S3C24XX_
for the generic form of an register has been
moved from the S3C2410.

Fixup S3C2410_EXTINTx and S3C2410_EINFLTx to
S3C24XX_EXTINTx and S3C24XX_EXTINTx

Depends on Patch #3635/1, Patch #3640/1

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3640/1: S3C2412: Use S3C24XX_DCLKCON instead of S3C2410_DCLKCON
Ben Dooks [Sat, 24 Jun 2006 20:21:32 +0000 (21:21 +0100)]
[ARM] 3640/1: S3C2412: Use S3C24XX_DCLKCON instead of S3C2410_DCLKCON

Patch from Ben Dooks

The current S3C2412 support has moved to using
S3C24XX_DCLKCON unless the specific DCLKCON is
required (S3C2412_DCLKCON or S3C2410_DKCLKCON)

Move the few places using S3C2410_DCLKCON to
S3C24XX_DCLKCON

Depends on Patch #3635/1

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3639/1: S3C2412: serial port support
Ben Dooks [Sat, 24 Jun 2006 20:21:32 +0000 (21:21 +0100)]
[ARM] 3639/1: S3C2412: serial port support

Patch from Ben Dooks

Serial port support for the on-board UART blocks
on the Samsung S3C2412 and S3C2413 UARTs.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3638/1: S3C2412: core clocks
Ben Dooks [Sat, 24 Jun 2006 20:21:31 +0000 (21:21 +0100)]
[ARM] 3638/1: S3C2412: core clocks

Patch from Ben Dooks

Clock support for the clocks on the Samsung S3C2412
and S3C2413 SoCs. This provides clock enables and
parent selection for all the standard clocks.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3637/1: S3C24XX: Add mpll clock, and set as fclk parent
Ben Dooks [Sat, 24 Jun 2006 20:21:29 +0000 (21:21 +0100)]
[ARM] 3637/1: S3C24XX: Add mpll clock, and set as fclk parent

Patch from Ben Dooks

Update the clocks with the MPLL clock, and
use it as the parent. Also export these to
the rest of arch/arm/mach-s3c2410

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3636/1: S3C2412: Add selection of CPU_ARM926
Ben Dooks [Sat, 24 Jun 2006 20:21:28 +0000 (21:21 +0100)]
[ARM] 3636/1: S3C2412: Add selection of CPU_ARM926

Patch from Ben Dooks

Select CONFIG_CPU_ARM926 when CONFIG_CPU_S3C2412 is
selected.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3635/1: S3C24XX: Add S3C2412 core cpu support
Ben Dooks [Sat, 24 Jun 2006 20:21:27 +0000 (21:21 +0100)]
[ARM] 3635/1: S3C24XX: Add S3C2412 core cpu support

Patch from Ben Dooks

Add support for the Samsung S3C2412 and S3C2413 range
of SoCs. This patch contains the core identification,
debug macros, and basic register updates to get these
to build.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3633/1: S3C24XX: s3c2410 gpio bugfix - wrong pin nos
Ben Dooks [Sat, 24 Jun 2006 20:21:26 +0000 (21:21 +0100)]
[ARM] 3633/1: S3C24XX: s3c2410 gpio bugfix - wrong pin nos

Patch from Ben Dooks

The s3c2410 gpio functions have a pair of bugs where
the code is using the pin function definitions instead
of the pin gpio numbers.

Also remove the changelog

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3631/1: Remove legacy __mem_isa() definitions
Andrew Victor [Sat, 24 Jun 2006 08:58:14 +0000 (09:58 +0100)]
[ARM] 3631/1: Remove legacy __mem_isa() definitions

Patch from Andrew Victor

Remove the remaining legacy __mem_isa() definitions for the ARM
platforms.

Signed-off-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3644/1: ixp2000: export gpio_line_config()
Lennert Buytenhek [Sat, 24 Jun 2006 08:57:14 +0000 (09:57 +0100)]
[ARM] 3644/1: ixp2000: export gpio_line_config()

Patch from Lennert Buytenhek

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3646/1: ep93xx: instantiate ep93xx-ohci platform device
Lennert Buytenhek [Sat, 24 Jun 2006 09:33:03 +0000 (10:33 +0100)]
[ARM] 3646/1: ep93xx: instantiate ep93xx-ohci platform device

Patch from Lennert Buytenhek

The ep93xx ohci bits have been merged into the gregkh-2.6 tree, which
means that they will probably go upstream soon, so make the core ep93xx
code instantiate an appropriate ep93xx-ohci platform device.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3634/1: ep93xx: initial implementation of the clk_* API
Lennert Buytenhek [Sat, 24 Jun 2006 09:33:02 +0000 (10:33 +0100)]
[ARM] 3634/1: ep93xx: initial implementation of the clk_* API

Patch from Lennert Buytenhek

Add an initial implementation of the clk_* API for the cirrus ep93xx
to the tree.  The initial implementation is somewhat minimal, with the
intention of extending it as we go along.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3632/1: iwmmxt: fix up argument comment for iwmmxt_task_restore()
Lennert Buytenhek [Sat, 24 Jun 2006 08:57:13 +0000 (09:57 +0100)]
[ARM] 3632/1: iwmmxt: fix up argument comment for iwmmxt_task_restore()

Patch from Lennert Buytenhek

Commit d6551e884cf66de072b81f8b6d23259462c40baf forgot to update the
description of what goes into r2 when calling iwmmxt_task_restore().

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years agodrm: radeon constify radeon microcode
Dave Airlie [Sat, 24 Jun 2006 07:37:48 +0000 (17:37 +1000)]
drm: radeon constify radeon microcode

From: Tilman (DRM CVS)
Signed-off-by: Dave Airlie <airlied@linux.ie>
18 years agoAdd i915 ioctls to configure pipes for vblank interrupt.
Dave Airlie [Sat, 24 Jun 2006 07:07:34 +0000 (17:07 +1000)]
Add i915 ioctls to configure pipes for vblank interrupt.

i915 vblanks can be generated from either pipe a or b, however a disabled
pipe generates no interrupts. This change allows the X server to select
which pipe generates vblank interrupts.

From: Keith Packard <keith.packard@intel.com> via DRM CVS
Signed-off-by: Dave Airlie <airlied@linux.ie>
18 years agodrm: update radeon to 1.25 add r200 vertex program support
Dave Airlie [Sat, 24 Jun 2006 07:04:07 +0000 (17:04 +1000)]
drm: update radeon to 1.25 add r200 vertex program support

Add support for r200 vertex programs (R200_EMIT_VAP_PVS_CNTL, and new
packet type for making it possible to address whole tcl vector space
and have a larger count)

From: Roland Scheidegger (DRM CVS)
Signed-off-by: Dave Airlie <airlied@linux.ie>
18 years agodrm: radeon add a tcl state flush before accessing tcl vector space
Dave Airlie [Sat, 24 Jun 2006 06:55:34 +0000 (16:55 +1000)]
drm: radeon add a tcl state flush before accessing tcl vector space

Do a tcl state flush before accessing tcl vector space. This fixes some
more problems with flickering (bug #6637). drm may not be appropriate
place for this, since doing that flush there might both be overkill and
insufficient in some cases. However, it's hard to figure out when that
flush is needed, so this has to suffice. There does not seem to be a
performance penalty associated with it.

From: Roland Scheidegger (DRM CVS)
Signed-off-by: Dave Airlie <airlied@linux.ie>
18 years agointelfb delousing
Al Viro [Fri, 23 Jun 2006 05:05:39 +0000 (06:05 +0100)]
intelfb delousing

ring_head is offset in card memory, not iomem pointer.  Fixed, removed
fuckloads of amazingly bogus casts somebody had sprinkled all over the
place.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Dave Airlie <airlied@linux.ie>
18 years ago[SPARC64]: Update defconfig.
David S. Miller [Sat, 24 Jun 2006 06:17:48 +0000 (23:17 -0700)]
[SPARC64]: Update defconfig.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Make auxio a real driver.
David S. Miller [Sat, 24 Jun 2006 05:46:49 +0000 (22:46 -0700)]
[SPARC64]: Make auxio a real driver.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PARPORT] sunbpp: Convert to new SBUS device framework.
David S. Miller [Sat, 24 Jun 2006 05:22:13 +0000 (22:22 -0700)]
[PARPORT] sunbpp: Convert to new SBUS device framework.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[Documentation]: Update probing info in sbus_drivers.txt
David S. Miller [Sat, 24 Jun 2006 05:12:39 +0000 (22:12 -0700)]
[Documentation]: Update probing info in sbus_drivers.txt

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SCSI] qlogicpti: Convert to new SBUS device framework.
David S. Miller [Sat, 24 Jun 2006 05:00:34 +0000 (22:00 -0700)]
[SCSI] qlogicpti: Convert to new SBUS device framework.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SCSI] esp: Fix bug in esp_remove_common.
David S. Miller [Sat, 24 Jun 2006 04:58:34 +0000 (21:58 -0700)]
[SCSI] esp: Fix bug in esp_remove_common.

Do not try to kfree(esp), scsi_host_put() takes care of that.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] sunhme: Kill useless loop over sdevs in quattro_sbus_find().
David S. Miller [Sat, 24 Jun 2006 04:36:46 +0000 (21:36 -0700)]
[NET] sunhme: Kill useless loop over sdevs in quattro_sbus_find().

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] myri_sbus: Kill unused next_module struct member.
David S. Miller [Sat, 24 Jun 2006 04:33:44 +0000 (21:33 -0700)]
[NET] myri_sbus: Kill unused next_module struct member.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] myri_sbus: Convert to new SBUS device layer.
David S. Miller [Sat, 24 Jun 2006 04:32:48 +0000 (21:32 -0700)]
[NET] myri_sbus: Convert to new SBUS device layer.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] sunqe: Convert to new SBUS driver layer.
David S. Miller [Sat, 24 Jun 2006 04:23:01 +0000 (21:23 -0700)]
[NET] sunqe: Convert to new SBUS driver layer.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] sunbmac: Convert over to new SBUS device framework.
David S. Miller [Sat, 24 Jun 2006 01:48:04 +0000 (18:48 -0700)]
[NET] sunbmac: Convert over to new SBUS device framework.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] sunlance: Convert to new SBUS driver framework.
David S. Miller [Sat, 24 Jun 2006 01:39:40 +0000 (18:39 -0700)]
[NET] sunlance: Convert to new SBUS driver framework.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] sunhme: Convert to new SBUS driver framework.
David S. Miller [Sat, 24 Jun 2006 01:21:02 +0000 (18:21 -0700)]
[NET] sunhme: Convert to new SBUS driver framework.

And make it a real PCI driver too.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET] sunhme: Kill __sparc__ and __sparc_v9__ ifdefs.
David S. Miller [Sat, 24 Jun 2006 00:31:12 +0000 (17:31 -0700)]
[NET] sunhme: Kill __sparc__ and __sparc_v9__ ifdefs.

Use CONFIG_SPARC and CONFIG_SPARC64 instead.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SCSI] sparc: Port esp to new SBUS driver layer.
David S. Miller [Sat, 24 Jun 2006 00:29:28 +0000 (17:29 -0700)]
[SCSI] sparc: Port esp to new SBUS driver layer.

This also turns the driver into a new-style scsi driver.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SOUND] sparc: Port amd7930 to new SBUS device layer.
David S. Miller [Fri, 23 Jun 2006 23:23:12 +0000 (16:23 -0700)]
[SOUND] sparc: Port amd7930 to new SBUS device layer.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SBUS]: Rewrite and plug into of_device framework.
David S. Miller [Fri, 23 Jun 2006 22:55:45 +0000 (15:55 -0700)]
[SBUS]: Rewrite and plug into of_device framework.

I severely apologize, I was still learning how to program
in C when I wrote this stuff 10 years ago...

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC]: Port of_device layer and make ebus use it.
David S. Miller [Fri, 23 Jun 2006 22:55:17 +0000 (15:55 -0700)]
[SPARC]: Port of_device layer and make ebus use it.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC]: Port sparc64 in-kernel device tree code to sparc32.
David S. Miller [Fri, 23 Jun 2006 22:53:31 +0000 (15:53 -0700)]
[SPARC]: Port sparc64 in-kernel device tree code to sparc32.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Add of_device layer and make ebus/isa use it.
David S. Miller [Fri, 23 Jun 2006 08:44:10 +0000 (01:44 -0700)]
[SPARC64]: Add of_device layer and make ebus/isa use it.

Sparcspkr and power drivers are converted, to make sure it works.
Eventually the SBUS device layer will use this as a sub-class.

I really cannot cut loose on that bit until sparc32 is given the
same infrastructure.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Expand of_*() interfaces some more.
David S. Miller [Fri, 23 Jun 2006 05:08:58 +0000 (22:08 -0700)]
[SPARC64]: Expand of_*() interfaces some more.

Import some more stuff from powerpc.

Add of_device_is_compatible(), and of_find_compatible_node().
Export some more of the other routines to modules.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Kill unused local vars in map_prom_timers().
David S. Miller [Fri, 23 Jun 2006 03:23:56 +0000 (20:23 -0700)]
[SPARC64]: Kill unused local vars in map_prom_timers().

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Kill off some more prom_getproperty() remnants.
David S. Miller [Fri, 23 Jun 2006 03:21:22 +0000 (20:21 -0700)]
[SPARC64]: Kill off some more prom_getproperty() remnants.

The remaining ones occur before we have imported the
device tree.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Convert Cheetah memory controller driver to in-kernel PROM tree.
David S. Miller [Fri, 23 Jun 2006 03:04:30 +0000 (20:04 -0700)]
[SPARC64]: Convert Cheetah memory controller driver to in-kernel PROM tree.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Convert central bus layer to in-kernel PROM device tree.
David S. Miller [Fri, 23 Jun 2006 02:53:24 +0000 (19:53 -0700)]
[SPARC64]: Convert central bus layer to in-kernel PROM device tree.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Kill ebus/isa range and interrupt mapping struct members.
David S. Miller [Fri, 23 Jun 2006 02:31:11 +0000 (19:31 -0700)]
[SPARC64]: Kill ebus/isa range and interrupt mapping struct members.

Unused outside of initial bus probe scan.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Use in-kernel PROM tree for EBUS and ISA.
David S. Miller [Fri, 23 Jun 2006 02:12:03 +0000 (19:12 -0700)]
[SPARC64]: Use in-kernel PROM tree for EBUS and ISA.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Convert sparc64 PCI layer to in-kernel device tree.
David S. Miller [Thu, 22 Jun 2006 23:18:54 +0000 (16:18 -0700)]
[SPARC64]: Convert sparc64 PCI layer to in-kernel device tree.

One thing this change pointed out was that we really should
pull the "get 'local-mac-address' property" logic into a helper
function all the network drivers can call.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Must run smp_setup_cpu_possible_map() after paging_init()
David S. Miller [Thu, 22 Jun 2006 07:49:15 +0000 (00:49 -0700)]
[SPARC64]: Must run smp_setup_cpu_possible_map() after paging_init()

Otherwise the in-kernel PROM device tree isn't built yet,
and therefore the present cpu bits don't get set properly.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Fix for Niagara memory corruption.
David S. Miller [Thu, 22 Jun 2006 07:01:56 +0000 (00:01 -0700)]
[SPARC64]: Fix for Niagara memory corruption.

On some sun4v systems, after netboot the ethernet controller and it's
DMA mappings can be left active.  The net result is that the kernel
can end up using memory the ethernet controller will continue to DMA
into, resulting in corruption.

To deal with this, we are more careful about importing IOMMU
translations which OBP has left in the IO-TLB.  If the mapping maps
into an area the firmware claimed was free and available memory for
the kernel to use, we demap instead of import that IOMMU entry.

This is going to cause the network chip to take a PCI master abort on
the next DMA it attempts, if it has been left going like this.  All
tests show that this is handled properly by the PCI layer and the e1000
drivers.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Minor bug fix to obp_read_memory().
David S. Miller [Thu, 22 Jun 2006 07:00:00 +0000 (00:00 -0700)]
[SPARC64]: Minor bug fix to obp_read_memory().

If we end up zero'ing out the size of one of the entries,
pop it out of the array completely because some code that
examines these things cannot handle a zero length element
properly.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Convert cpu_find_by_*() interface to in-kernel PROM device tree.
David S. Miller [Thu, 22 Jun 2006 06:34:02 +0000 (23:34 -0700)]
[SPARC64]: Convert cpu_find_by_*() interface to in-kernel PROM device tree.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Add of_getintprop_default().
David S. Miller [Thu, 22 Jun 2006 06:07:29 +0000 (23:07 -0700)]
[SPARC64]: Add of_getintprop_default().

This encodes a common idiomatic coding pattern used when
dealing with integer properties.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Convert sun4v virtual-device layer to in-kernel PROM device tree.
David S. Miller [Thu, 22 Jun 2006 05:56:20 +0000 (22:56 -0700)]
[SPARC64]: Convert sun4v virtual-device layer to in-kernel PROM device tree.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Rate limited kernel unaligned trap logging, ala IA64.
David S. Miller [Thu, 22 Jun 2006 05:31:08 +0000 (22:31 -0700)]
[SPARC64]: Rate limited kernel unaligned trap logging, ala IA64.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Disable verbose PCI IRQ probing messages by default.
David S. Miller [Thu, 22 Jun 2006 05:07:21 +0000 (22:07 -0700)]
[SPARC64]: Disable verbose PCI IRQ probing messages by default.

Allow them to be enabled with "pci=irq_verbose" on the
boot command line.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Use in-kernel OBP device tree for PCI controller probing.
David S. Miller [Thu, 22 Jun 2006 01:18:47 +0000 (18:18 -0700)]
[SPARC64]: Use in-kernel OBP device tree for PCI controller probing.

It can be pushed even further down, but this is a first step.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Add of_find_node_by_{name,type}().
David S. Miller [Wed, 21 Jun 2006 23:33:54 +0000 (16:33 -0700)]
[SPARC64]: Add of_find_node_by_{name,type}().

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SPARC64]: Import OBP device tree into kernel data structures.
David S. Miller [Wed, 21 Jun 2006 22:35:28 +0000 (15:35 -0700)]
[SPARC64]: Import OBP device tree into kernel data structures.

The basic framework is based on the PowerPC OF code.

This code even tries to get the device addressing components
correct in the full path names.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[SBUS]: Start cleaning up generic sbus support layer.
David S. Miller [Tue, 20 Jun 2006 22:23:28 +0000 (15:23 -0700)]
[SBUS]: Start cleaning up generic sbus support layer.

In particular, move the IRQ probing out to sparc32/sparc64
arch specific code where it belongs.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PATCH] fix typo in acpi video brightness changes.
Dave Jones [Sat, 24 Jun 2006 04:33:08 +0000 (00:33 -0400)]
[PATCH] fix typo in acpi video brightness changes.

Prevent possible null dereference due to misplaced ;

Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Fri, 23 Jun 2006 22:58:44 +0000 (15:58 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev: (258 commits)
  [libata] conversion to new debug scheme, part 1 of $N
  [PATCH] libata: Add ata_scsi_dev_disabled
  [libata] Add host lock to struct ata_port
  [PATCH] libata: implement per-dev EH action mask eh_info->dev_action[]
  [PATCH] libata-dev: move the CDB-intr DMA blacklisting
  [PATCH] ahci: disable NCQ support on vt8251
  [libata] ahci: add JMicron PCI IDs
  [libata] sata_nv: add PCI IDs
  [libata] ahci: Add NVIDIA PCI IDs.
  [PATCH] libata: convert several bmdma-style controllers to new EH, take #3
  [PATCH] sata_via: convert to new EH, take #3
  [libata] sata_nv: s/spin_lock_irqsave/spin_lock/ in irq handler
  [PATCH] sata_nv: add hotplug support
  [PATCH] sata_nv: convert to new EH
  [PATCH] sata_nv: better irq handlers
  [PATCH] sata_nv: simplify constants
  [PATCH] sata_nv: kill struct nv_host_desc and nv_host
  [PATCH] sata_nv: kill not-working hotplug code
  [libata] Update docs to reflect current driver API
  [PATCH] libata: add host_set->next for legacy two host_sets case, take #3
  ...

18 years ago[IA64] fix ia64 build (fadt_descriptor)
Tony Luck [Fri, 23 Jun 2006 20:49:25 +0000 (13:49 -0700)]
[IA64] fix ia64 build (fadt_descriptor)

arch/ia64/kernel/acpi.c got forgotten when include/acpi/actbl.h
got a cleanup.

Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoAuto-update from upstream
Tony Luck [Fri, 23 Jun 2006 20:46:23 +0000 (13:46 -0700)]
Auto-update from upstream

18 years ago[PATCH] fix silly ARM non-EABI build error
Nicolas Pitre [Fri, 23 Jun 2006 15:40:53 +0000 (11:40 -0400)]
[PATCH] fix silly ARM non-EABI build error

My bad.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] s390_hypfs filesystem: get_sb_single() fix
Andrew Morton [Fri, 23 Jun 2006 11:01:35 +0000 (04:01 -0700)]
[PATCH] s390_hypfs filesystem: get_sb_single() fix

Update hypfs for dhowells API changes.

Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Ingo Oeser <ioe-lkml@rameria.de>
Cc: Joern Engel <joern@wohnheim.fh-wedel.de>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Holzheu <holzheu@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] cpufreq build fix
Andrew Morton [Fri, 23 Jun 2006 10:31:19 +0000 (03:31 -0700)]
[PATCH] cpufreq build fix

drivers/cpufreq/cpufreq_ondemand.c: In function 'do_dbs_timer':
drivers/cpufreq/cpufreq_ondemand.c:374: warning: implicit declaration of function 'lock_cpu_hotplug'
drivers/cpufreq/cpufreq_ondemand.c:381: warning: implicit declaration of function 'unlock_cpu_hotplug'
drivers/cpufreq/cpufreq_conservative.c: In function 'do_dbs_timer':
drivers/cpufreq/cpufreq_conservative.c:425: warning: implicit declaration of function 'lock_cpu_hotplug'
drivers/cpufreq/cpufreq_conservative.c:432: warning: implicit declaration of function 'unlock_cpu_hotplug'

Cc: Dave Jones <davej@codemonkey.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[BLOCK] Fix bounce limit address check
Andi Kleen [Wed, 21 Jun 2006 12:48:09 +0000 (14:48 +0200)]
[BLOCK] Fix bounce limit address check

Do a safer check for when to enable DMA. Currently we enable ISA DMA
for cases that do not need it, resulting in OOM conditions when ZONE_DMA
runs out of space.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] rbtree: support functions used by the io schedulers
Jens Axboe [Wed, 21 Jun 2006 07:36:18 +0000 (09:36 +0200)]
[PATCH] rbtree: support functions used by the io schedulers

They all duplicate macros to check for empty root and/or node, and
clearing a node. So put those in rbtree.h.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] splice: retrieve mapping after locking the page
Jens Axboe [Tue, 20 Jun 2006 13:01:12 +0000 (15:01 +0200)]
[PATCH] splice: retrieve mapping after locking the page

Otherwise we could be racing with truncate/mapping removal.

Problem found/fixed by Nick Piggin <npiggin@suse.de>, logic rewritten
by me.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] cfq-iosched: rq update fixes
Jens Axboe [Fri, 16 Jun 2006 13:35:39 +0000 (15:35 +0200)]
[PATCH] cfq-iosched: rq update fixes

- Remember to set ->last_sector so that the cfq_choose_req() logic
  works correctly.

- Remove redundant call to cfq_choose_req()

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] cfq-iosched: many performance fixes
Jens Axboe [Fri, 16 Jun 2006 09:23:00 +0000 (11:23 +0200)]
[PATCH] cfq-iosched: many performance fixes

This is a collection of patches that greatly improve CFQ performance
in some circumstances.

- Change the idling logic to only kick in after a request is done and we
  are deciding what to do. Before the idling included the request service
  time, so it was hard to adjust. Now it's true think/idle time.

- Take advantage of TCQ/NCQ/queueing for seeky sync workloads, but keep
  it in control for sync and sequential (or close to) workloads.

- Expire queues immediately and move on to other busy queues, if we are
  not going to idle after the current one finishes.

- Don't rearm idle timer if there are no busy queues. Just leave the
  system idle.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] cfq-iosched: correctly set ioprio on both targets
Jens Axboe [Wed, 14 Jun 2006 07:10:45 +0000 (09:10 +0200)]
[PATCH] cfq-iosched: correctly set ioprio on both targets

Patch originally from Vasily Tarasov <vtaras@sw.ru>

If you set io-priority of process 1 using sys_ioprio_set system call by
another process 2 (like ionice do), then cfq_init_prio_data() function
sets priority of process 2 (current) on queue of process 1 and clears
the flag, that designates change of ioprio.  So the process  1 will work
like with priority of process 2.

I propose not to call cfq_init_prio_data() on io-priority change, but
only mark queue as queue with changed prority.  Every time when new
request comes cfq-scheduler checks for this flag and atomaticaly changes
priority of queue to new value.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] Make CFQ the default IO scheduler
Jens Axboe [Mon, 19 Jun 2006 08:06:48 +0000 (10:06 +0200)]
[PATCH] Make CFQ the default IO scheduler

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] Rearrange a few struct request members
Jens Axboe [Tue, 13 Jun 2006 07:02:34 +0000 (09:02 +0200)]
[PATCH] Rearrange a few struct request members

This saves 8 bytes of data in 64-bit archs.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] Get rid of struct request request_pm_state member
Jens Axboe [Tue, 13 Jun 2006 06:46:57 +0000 (08:46 +0200)]
[PATCH] Get rid of struct request request_pm_state member

The IDE power management can just use the ->end_io_data member to store
it's data.

Signed-off-by: Jens Axboe <axboe@suse.de>
18 years ago[PATCH] Kill PF_SYNCWRITE flag
Jens Axboe [Tue, 13 Jun 2006 06:26:10 +0000 (08:26 +0200)]
[PATCH] Kill PF_SYNCWRITE flag

A process flag to indicate whether we are doing sync io is incredibly
ugly. It also causes performance problems when one does a lot of async
io and then proceeds to sync it. Part of the io will go out as async,
and the other part as sync. This causes a disconnect between the
previously submitted io and the synced io. For io schedulers such as CFQ,
this will cause us lost merges and suboptimal behaviour in scheduling.

Remove PF_SYNCWRITE completely from the fsync/msync paths, and let
the O_DIRECT path just directly indicate that the writes are sync
by using WRITE_SYNC instead.

Signed-off-by: Jens Axboe <axboe@suse.de>