efikamx:linux-kernel.git
7 years agodrivers/net/usb/asix.c: Fix unaligned accesses
Neil Jones [Tue, 18 May 2010 00:18:28 +0000 (17:18 -0700)]
drivers/net/usb/asix.c: Fix unaligned accesses

Using this driver can cause unaligned accesses in the IP layer
This has been fixed by aligning the skb data correctly using the
spare room left over by the 4 byte header inserted between packets
by the device.

Signed-off-by: Neil Jones <NeilJay@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agoasix: fix setting mac address for AX88772
Jussi Kivilinna [Tue, 9 Mar 2010 12:24:38 +0000 (12:24 +0000)]
asix: fix setting mac address for AX88772

Setting new MAC address only worked when device was set to promiscuous mode.
Fix MAC address by writing new address to device using undocumented command
AX_CMD_READ_NODE_ID+1. Patch is tested with AX88772 device.

Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
Acked-by: David Hollis <dhollis@davehollis.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
7 years agoARM: 6468/1: backtrace: fix calculation of thread stack base
Will Deacon [Thu, 4 Nov 2010 17:22:51 +0000 (18:22 +0100)]
ARM: 6468/1: backtrace: fix calculation of thread stack base

When unwinding stack frames we must take care not to unwind
areas of memory that lie outside of the known extent of the stack.

This patch fixes an incorrect calculation of the stack base where
THREAD_SIZE is added to the stack pointer after it has already
been aligned to this value. Since the ALIGN macro performs this
addition internally, we end up overshooting the base by 8k.

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
7 years ago[ARM] 5613/1: implement CALLER_ADDRESSx
Uwe Kleine-König [Tue, 21 Jul 2009 08:56:27 +0000 (09:56 +0100)]
[ARM] 5613/1: implement CALLER_ADDRESSx

From: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>

As __builtin_return_address(n) doesn't work for ARM with n > 0, the
kernel needs its own implementation.

This fixes many warnings saying:

warning: unsupported argument to '__builtin_return_address'

The new methods and walk_stackframe must not be instrumented because
CALLER_ADDRESSx is used in the various tracers and tracing the tracer is
a bad idea.

What's currently missing is an implementation using unwind tables.  This
is not fatal though, it's just that the tracers don't get enough
information to be really useful.

Note that if both ARM_UNWIND and FRAME_POINTER are enabled,
walk_stackframe uses unwind information.  So in this case the same
implementation is used as when FRAME_POINTER is disabled.

Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
7 years agoARM: 6341/1: unwind - optimise linked-list searches for modules
Phil Carmody [Thu, 19 Aug 2010 14:20:37 +0000 (15:20 +0100)]
ARM: 6341/1: unwind - optimise linked-list searches for modules

With several sections per module, and dozens of modules, the
searches down the linked list of sections would dominate the
lookup time, dwarfing any savings from the binary search
within the section.

A simple move-to-front optimisation exploits the commonality
of the code paths taken, and in simple real-world tests reduces
the number of steps in the search to barely more than 1.

Signed-off-by: Phil Carmody <ext-phil.2.carmody@nokia.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
7 years agoARM: 6140/1: silence a bogus sparse warning in unwind.c
Alexander Shishkin [Fri, 21 May 2010 11:32:07 +0000 (12:32 +0100)]
ARM: 6140/1: silence a bogus sparse warning in unwind.c

The check for compiler which is supposed to miscompile unwind tables
clearly has nothing to do with sparse (which does not define necessary
macros anyway), so simply silence it.

Signed-off-by: Alexander Shishkin <virtuoso@slind.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
7 years agoARM: 5776/1: Check compiler version and EABI support when adding ARM unwind support.
Claudio Scordino [Fri, 30 Oct 2009 11:06:05 +0000 (12:06 +0100)]
ARM: 5776/1: Check compiler version and EABI support when adding ARM unwind support.

ARM unwind is known to compile only with EABI and not-buggy compilers.
The problem is not the unwinding information but the -fno-frame-pointer
option added as a result of !CONFIG_FRAME_POINTER.  Now we check the
compiler and raise a #warning in case of wrong compiler.

Signed-off-by: Claudio Scordino <claudio@evidence.eu.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
7 years agosched: update load count only once per cpu in 10 tick update window (v2 -> v3)
Matt Sealey [Sat, 22 Jan 2011 04:04:22 +0000 (22:04 -0600)]
sched: update load count only once per cpu in 10 tick update window (v2 -> v3)

7 years agoIPU updates from latest FSL BSP, seemingly all related to burst sizes and so on....
Matt Sealey [Sat, 22 Jan 2011 03:43:30 +0000 (21:43 -0600)]
IPU updates from latest FSL BSP, seemingly all related to burst sizes and so on. New pixel format.

7 years agosiihdmi: remove sii9022 from Kconfig for video for Smartbook too
Matt Sealey [Fri, 21 Jan 2011 21:57:14 +0000 (15:57 -0600)]
siihdmi: remove sii9022 from Kconfig for video for Smartbook too

7 years agoefikamx: remove redundant crap from display platform file
Matt Sealey [Fri, 21 Jan 2011 21:40:57 +0000 (15:40 -0600)]
efikamx: remove redundant crap from display platform file

7 years agobump to .14 for PD release efikasb-10.08.00-20110121
Matt Sealey [Fri, 21 Jan 2011 21:23:06 +0000 (15:23 -0600)]
bump to .14 for PD release

7 years agosiihdmi: native panel resolution and mode culling support
Matt Sealey [Fri, 21 Jan 2011 21:19:15 +0000 (15:19 -0600)]
siihdmi: native panel resolution and mode culling support

* Will now parse the mode list to:
 - Remove pixel clocks above 133MHz which the MX51 IPU DI cannot display (PLL limitation)
 - Remove modes from mode list that are interlaced (cannot display)
 - Find and save the best (native) panel mode and try and match a mode as close as possible instead of
   arbitrarily trying 1280x720 (unless EDID failed for some reason)
* Reduce bus timeout default to 10ms as it always takes less (<4ms or so, but we want to handle slow monitors)
* Comment out instrumentation until it returns reasonable values (jiffies aren't accurate enough)
* Some variable renames for clearer code

7 years agodvfs-per: fix references to pixel clocks
Matt Sealey [Fri, 21 Jan 2011 21:17:55 +0000 (15:17 -0600)]
dvfs-per: fix references to pixel clocks

7 years agosiihdmi: fix edid reference, use usecs for instrumentation
Matt Sealey [Thu, 20 Jan 2011 18:57:01 +0000 (12:57 -0600)]
siihdmi: fix edid reference, use usecs for instrumentation

7 years agosiihdmi: inline siihdmi_detect_sink
Saleem Abdulrasool [Thu, 20 Jan 2011 17:42:41 +0000 (09:42 -0800)]
siihdmi: inline siihdmi_detect_sink

The framebuffer routines also work with the cea861 data.  Move the edid reading
into the function itself, which avoids us having to read the EDID data multiple
times and allows us to pass the full EDID data to the framebuffer routines.

We always have to read the first block to determine the actual length of the
EDID data.  In order to keep the reading logic the same all the time, we simply
re-read the EDID data always (even if there are no extension blocks).  This
hopefully will help keep timing issues similar across all devices.

7 years agoMerge branch 'efikasb-10.08.00' of gitorious.org:efikamx/linux-kernel into efikasb...
Matt Sealey [Thu, 20 Jan 2011 17:04:13 +0000 (11:04 -0600)]
Merge branch 'efikasb-10.08.00' of gitorious.org:efikamx/linux-kernel into efikasb-10.08.00

7 years agoSome backports from FSL tree: lost the ENGR numbers but in summary:
Matt Sealey [Thu, 20 Jan 2011 16:23:28 +0000 (10:23 -0600)]
Some backports from FSL tree: lost the ENGR numbers but in summary:

* fix bus frequency driver check for frequency (> instead of !=)
* fix VPU lockup due to misassigning a clock
* stop dvfs before reboots (otherwise it might not reboot)
* small cleanups in DVFS code for frequency changing and added spinlock

7 years agosiihdmi: fix up cea861 header parsing
Saleem Abdulrasool [Thu, 20 Jan 2011 01:47:49 +0000 (17:47 -0800)]
siihdmi: fix up cea861 header parsing

The data block collection can be arbitrarily ordered.  We need to iterate over
all the blocks and check them.

7 years agoENGR00135971 GPU: move clk_enable/disable out of timer handler
Richard Zhao [Mon, 6 Dec 2010 02:50:52 +0000 (10:50 +0800)]
ENGR00135971 GPU: move clk_enable/disable out of timer handler

For timer based power autogating, we have to move clk_enable/disable out of
timer handler, because they become may sleep.

Signed-off-by: Richard Zhao <richard.zhao@freescale.com>
7 years agoENGR00134150 GPU: destroy device in workqueue for error interrupts
Jie Zhou [Thu, 25 Nov 2010 07:19:42 +0000 (15:19 +0800)]
ENGR00134150 GPU: destroy device in workqueue for error interrupts

If GPU error interrupt detected, related device will be destroyed.
In the destory function, device_idle and mutex_lock may cause sleep,
which may cause kernel crash, so delay the device destroy to the
workqueue to avoid this.

Signed-off-by: Jie Zhou <b30303@freescale.com>
7 years agoconfig: re-enable DVFS-PER
Matt Sealey [Mon, 17 Jan 2011 20:28:54 +0000 (14:28 -0600)]
config: re-enable DVFS-PER

7 years agomtl017: fix mode data for AUO panel in SB, upper margin 24 -> 21
Matt Sealey [Mon, 17 Jan 2011 16:36:13 +0000 (10:36 -0600)]
mtl017: fix mode data for AUO panel in SB, upper margin 24 -> 21

7 years agosiihdmi: various cleanups
Saleem Abdulrasool [Sun, 16 Jan 2011 22:31:20 +0000 (14:31 -0800)]
siihdmi: various cleanups

Cleanup the siihdmi header (which really is pointless to split out...), redefine
the fields for the sii info frame packets, and restructure sizes to be in terms
of registers.

7 years agosiihdmi: fixup tmds_state handling
Saleem Abdulrasool [Sun, 16 Jan 2011 22:04:25 +0000 (14:04 -0800)]
siihdmi: fixup tmds_state handling

tmds_state is really a boolean, so rename it to tmds_enabled.  Furthermore, use
boolean values to interact with the boolean type.

7 years agosiihdmi: simplify resolution setting
Saleem Abdulrasool [Sun, 16 Jan 2011 22:02:00 +0000 (14:02 -0800)]
siihdmi: simplify resolution setting

The allocations are no longer done in the info frame setting, DVI can transmit
the info frames, and the cost is simply the i2c writes.  This is an infrequent
operation, so just absorb the cost regardless of the connection type.

7 years agocea861: improve CEA EDID extension block definition
Saleem Abdulrasool [Sun, 16 Jan 2011 21:26:13 +0000 (13:26 -0800)]
cea861: improve CEA EDID extension block definition

Improve the CEA861 EDID block definition so that we can use it ot properly
identify a HDMI sink.

7 years agoedid: improve edid parsing
Saleem Abdulrasool [Sun, 16 Jan 2011 20:45:20 +0000 (12:45 -0800)]
edid: improve edid parsing

Use a proper edid structure definition rather than indexing into a data blob.
This also adds additional helper functions to work with the returned edid data.

There is much room for further improvements still.  We should simply read the
first block, then use that to actually determine the number of blocks we need to
read from the device.  We cant simply read all the potential blocks as that can
be up to 32K of data.

Once we are reading all the EDID extensions, we should iterate over the blocks
to find the CEA block rather than assuming that the device is sending the CEA
block as the first block (especially given that the VESA spec calls for block 1
to be an extension map if there are more than 1 block present).

7 years agomake bus_timeout a module parameter
Saleem Abdulrasool [Sat, 15 Jan 2011 23:05:14 +0000 (15:05 -0800)]
make bus_timeout a module parameter

bus_timeout depends somewhat on the output device for certain operations.  As
such, it can be useful to actually change the value at runtime.  To that end,
make the value a module parameter.

Bonus whitespace cleanups and centralises the value to a single location.

7 years agorefactor info frame management
Saleem Abdulrasool [Sat, 15 Jan 2011 18:40:45 +0000 (10:40 -0800)]
refactor info frame management

info frame checksumming is defined by the CEA861-E standard, and is not specific
to SII9xxx, so move that into the CEA861 header.

Don't bother allocating the info frame structures off the heap.  That causes
unnecessary fragmentation and complexity.  Allocate them off the stack as they
are small enough.  This also helps keep everything simple.

As a bonus, fixed up the SII HDMI structures and the CEA861 InfoFrame structures
in the process as well.  Also clean up the magic number in the AVI InfoFrame
assertion.

7 years agocleanup unused variable warning
Saleem Abdulrasool [Sat, 15 Jan 2011 17:57:12 +0000 (09:57 -0800)]
cleanup unused variable warning

7 years agofix irq leak
Saleem Abdulrasool [Thu, 16 Dec 2010 06:12:05 +0000 (22:12 -0800)]
fix irq leak

7 years agoavoid unnecessary string allocation
Saleem Abdulrasool [Sun, 12 Dec 2010 02:44:31 +0000 (18:44 -0800)]
avoid unnecessary string allocation

7 years agomove edid definitions into an external header
Saleem Abdulrasool [Sun, 12 Dec 2010 02:40:37 +0000 (18:40 -0800)]
move edid definitions into an external header

7 years agosome whitespace fixes
Saleem Abdulrasool [Sun, 12 Dec 2010 02:40:16 +0000 (18:40 -0800)]
some whitespace fixes

7 years agosiihdmi: instrument the i2c timeouts so that we can reduce the timeout values. We...
Matt Sealey [Sat, 15 Jan 2011 17:09:14 +0000 (11:09 -0600)]
siihdmi: instrument the i2c timeouts so that we can reduce the timeout values. We need to pull down some of the 100ms ones more importantly but this is a good start on working out how fast your monitor responds.

7 years agosiihdmi: make hotplugging and interrupt usage a configuration option, and turn it...
Matt Sealey [Sat, 15 Jan 2011 03:07:12 +0000 (21:07 -0600)]
siihdmi: make hotplugging and interrupt usage a configuration option, and turn it off while it's broken

7 years agosiihdmi: embed delayed_work into the structure instead of trying to allocate it
Matt Sealey [Fri, 14 Jan 2011 20:18:18 +0000 (14:18 -0600)]
siihdmi: embed delayed_work into the structure instead of trying to allocate it
also make kmalloc/memset into kzalloc in probe

7 years agosiihdmi: improve irq handler, edge trigger it. change the way the workqueue works...
Matt Sealey [Fri, 14 Jan 2011 18:16:07 +0000 (12:16 -0600)]
siihdmi: improve irq handler, edge trigger it. change the way the workqueue works. retries->timeouts. general cleanliness updates.

7 years agosiihdmi: disable irq handler as it is blocking boot now that the interrupt is it...
Matt Sealey [Thu, 13 Jan 2011 23:39:19 +0000 (17:39 -0600)]
siihdmi: disable irq handler as it is blocking boot now that the interrupt is it's own and it does not clear the SII9022

7 years agoAdded missing include mach/irqs.h
Matt Sealey [Thu, 13 Jan 2011 23:15:07 +0000 (17:15 -0600)]
Added missing include mach/irqs.h

7 years agoCorrect hdmi irq value (TO_IRQ not TO_GPIO). I owe Michal Schultz a beer.
Matt Sealey [Thu, 13 Jan 2011 22:59:35 +0000 (16:59 -0600)]
Correct hdmi irq value (TO_IRQ not TO_GPIO). I owe Michal Schultz a beer.

7 years agoARM: VFP: Fix vfp_put_double() for d16-d31
Matt Sealey [Thu, 13 Jan 2011 03:14:14 +0000 (21:14 -0600)]
ARM: VFP: Fix vfp_put_double() for d16-d31

vfp_put_double() takes the double value in r0,r1 not r1,r2.

Reported-by: Tarun Kanti DebBarma <tarun.kanti@ti.com>
Cc: <stable@kernel.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
7 years agoata: cleanup init and exit for Smartbook to unify the code
Matt Sealey [Wed, 12 Jan 2011 02:35:29 +0000 (20:35 -0600)]
ata: cleanup init and exit for Smartbook to unify the code

7 years agoconfig: move to new hdmi driver
Matt Sealey [Tue, 11 Jan 2011 03:17:37 +0000 (21:17 -0600)]
config: move to new hdmi driver

7 years agonaive attempt at tuning for cortex-a8
Matt Sealey [Tue, 11 Jan 2011 03:14:50 +0000 (21:14 -0600)]
naive attempt at tuning for cortex-a8

7 years agonew year, new configs
Matt Sealey [Tue, 11 Jan 2011 02:45:57 +0000 (20:45 -0600)]
new year, new configs
put pata_fsl and mx_sdhci back in the kernel because module probing is a bitch
add dual fifo for audio

7 years agoplatforms -> smarttop for this kernel version, it'll merge back later
Matt Sealey [Tue, 11 Jan 2011 02:41:34 +0000 (20:41 -0600)]
platforms -> smarttop for this kernel version, it'll merge back later

7 years agopata_platform: set up ATA before registering the device
Matt Sealey [Thu, 30 Dec 2010 17:51:30 +0000 (11:51 -0600)]
pata_platform: set up ATA before registering the device

7 years agobattery: SBS gives times in minutes, power_supply class specifies seconds. Multiple...
Matt Sealey [Thu, 23 Dec 2010 17:42:34 +0000 (11:42 -0600)]
battery: SBS gives times in minutes, power_supply class specifies seconds. Multiple values by 60.

7 years agopata_fsl: Don't disable "regulators" that are in fact NULL pointers.
Matt Sealey [Thu, 23 Dec 2010 02:06:15 +0000 (20:06 -0600)]
pata_fsl: Don't disable "regulators" that are in fact NULL pointers.

7 years agopata_platform: fix platform resource values for ATA SFF standard registers (maybe?)
Matt Sealey [Wed, 22 Dec 2010 08:45:20 +0000 (02:45 -0600)]
pata_platform: fix platform resource values for ATA SFF standard registers (maybe?)

7 years agospi_imx: free gpio for chipselects before registering platform driver or it will...
Matt Sealey [Wed, 22 Dec 2010 07:54:24 +0000 (01:54 -0600)]
spi_imx: free gpio for chipselects before registering platform driver or it will fail
(cherry picked from commit 7909dd00e5ef8f6dd16d736bf43b80ad3a1f5fde)

7 years agomx ata: add support for evil pata_platform while pata_fsl is not working. Note: this...
Matt Sealey [Wed, 22 Dec 2010 07:53:41 +0000 (01:53 -0600)]
mx ata: add support for evil pata_platform while pata_fsl is not working. Note: this also does not work yet.
(cherry picked from commit 38b490f7a71d83f63fa7fc43cb16bf0f55a7b48e)

7 years agosiihdmi: Only initialize the DISP3 BG framebuffer to stop it from setting resolution...
Matt Sealey [Wed, 22 Dec 2010 07:52:52 +0000 (01:52 -0600)]
siihdmi: Only initialize the DISP3 BG framebuffer to stop it from setting resolution twice on boot.
(cherry picked from commit fd46562598a3355ecbd8f568ad7b2745bfad8bfa)

7 years agoata: remove redundant active/inactive functions as they do nothing
Matt Sealey [Wed, 22 Dec 2010 04:39:38 +0000 (22:39 -0600)]
ata: remove redundant active/inactive functions as they do nothing
(cherry picked from commit 365802eb1356f9aa42fd4ae781315fe81247fa7f)

7 years agopata_fsl: sync changes to 10.10 BSP. Note that the new DMA scripts seem to have broke...
Matt Sealey [Wed, 22 Dec 2010 02:04:47 +0000 (20:04 -0600)]
pata_fsl: sync changes to 10.10 BSP. Note that the new DMA scripts seem to have broken the driver before and after this change...
(cherry picked from commit f95119c0759e8c682b8277eb03bb7eb0ef67879a)

7 years agomx audio: restore headphone detect IOMUX setting (accidentally removed by clock pin...
Matt Sealey [Wed, 22 Dec 2010 01:19:55 +0000 (19:19 -0600)]
mx audio: restore headphone detect IOMUX setting (accidentally removed by clock pin change)
(cherry picked from commit 5734b93a12058164f8517eba0bc232af265de46b)

7 years agosiihdmi: add a 10ms pause before setting resolution
Matt Sealey [Wed, 22 Dec 2010 01:19:20 +0000 (19:19 -0600)]
siihdmi: add a 10ms pause before setting resolution

7 years agoaudio: request gpio 1_9 before using it since iomux doesn't cover this pin as gpio
Matt Sealey [Tue, 21 Dec 2010 23:51:37 +0000 (17:51 -0600)]
audio: request gpio 1_9 before using it since iomux doesn't cover this pin as gpio

7 years agoipu: don't look for lvds bridge clocks on systems that aren't using the lvds bridge
Matt Sealey [Tue, 21 Dec 2010 23:47:30 +0000 (17:47 -0600)]
ipu: don't look for lvds bridge clocks on systems that aren't using the lvds bridge

7 years agosb: remove absolutely useless idle timer code
Matt Sealey [Tue, 21 Dec 2010 23:12:45 +0000 (17:12 -0600)]
sb: remove absolutely useless idle timer code

7 years agooverlay driver: remove unused platform device (it's in mach-mx5/devices.c now)
Matt Sealey [Tue, 21 Dec 2010 23:08:58 +0000 (17:08 -0600)]
overlay driver: remove unused platform device (it's in mach-mx5/devices.c now)

7 years agohid: remove egregious idle timer stuff and clean up some whitespace
Matt Sealey [Tue, 21 Dec 2010 23:03:59 +0000 (17:03 -0600)]
hid: remove egregious idle timer stuff and clean up some whitespace

7 years agosiihdmi: add CEA Data Block Collection definitions so we can parse them properly...
Matt Sealey [Tue, 21 Dec 2010 18:43:08 +0000 (12:43 -0600)]
siihdmi: add CEA Data Block Collection definitions so we can parse them properly and accurately detect an HDMI sink

7 years agoStaging: usbip: Fix typo "Contoroller".
Robert P. J. Day [Fri, 13 Nov 2009 20:13:43 +0000 (15:13 -0500)]
Staging: usbip: Fix typo "Contoroller".

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
7 years agoStaging: usb-ip: vhci_hdc:Fix the returned error value
Shan Wei [Fri, 24 Jul 2009 08:57:35 +0000 (16:57 +0800)]
Staging: usb-ip: vhci_hdc:Fix the returned error value

Compared with other drivers, the "ret" should be nagative and
returned. But in vhci_hdc, it always return 0;

I dont't use the driver, and I'm not familiar with the code.
Hope the patch is helpful.

Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
7 years agoStaging: USB-IP code cleanup
Brian G. Merrell [Tue, 21 Jul 2009 06:46:13 +0000 (00:46 -0600)]
Staging: USB-IP code cleanup

This includes fixes for all of the legit checkpatch.pl errors and
warnings.  I have also included several of the suggestions from the
linux-kernel mailing list when the USB-IP code was first added.

Signed-off-by: Brian G. Merrell <bgmerrell@novell.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
7 years agoadd and register mxc_v4l2out_device
Matt Sealey [Sat, 18 Dec 2010 08:27:11 +0000 (02:27 -0600)]
add and register mxc_v4l2out_device

7 years agodrag mxc_v4l2_output video overlay code into the modern world (BSP 10.11.01)
Matt Sealey [Sat, 18 Dec 2010 07:09:45 +0000 (01:09 -0600)]
drag mxc_v4l2_output video overlay code into the modern world (BSP 10.11.01)
drag mxc_ipuv3_fb with it (remove final efikamx specific stuff)
update some IPU stuff and ioctls for it all to interoperate

notes:
that line_length bug is still there, I really can't find why it's 0...
doesn't register because the platform_device isn't in a/a/p-mxc/devices.?

7 years agoBring IPU up to latest BSP, including these fixes:
Matt Sealey [Tue, 14 Dec 2010 06:45:28 +0000 (00:45 -0600)]
Bring IPU up to latest BSP, including these fixes:

ENGR00133954 ipuv3: remove FB_SYNC_EXT flag
ENGR00133034 ipuv3: fix black flash issue during high resolution video playback
ENGR00134076 ipuv3: should not set burst size again
ENGR00133345 IPUv3 VDI:Support bottom field first setting
ENGR00131660 ipuv3: left stripe output column should be set to zero

This actually fixes the clock sync problem with the AOC HDMI monitor we have (perfectly
working display but "input not support" floating around). Happily, the boot splash
actually works, too (previously it would not work until the second set of the display
when GDM booted). Hopefully IPU support is nice and stable now.

7 years agosgtl5000: correct audio clock pin to schematics (the Pegatron, the)
Matt Sealey [Tue, 14 Dec 2010 06:27:14 +0000 (00:27 -0600)]
sgtl5000: correct audio clock pin to schematics (the Pegatron, the)

7 years agosgtl5000: audio clock gating platform support to fit ENGR00132525
Matt Sealey [Tue, 14 Dec 2010 06:18:53 +0000 (00:18 -0600)]
sgtl5000: audio clock gating platform support to fit ENGR00132525

7 years agosgtl5000: drag to latest BSP
Matt Sealey [Tue, 14 Dec 2010 06:09:06 +0000 (00:09 -0600)]
sgtl5000: drag to latest BSP

includes patches:
ENGR00132153 sgtl5000: do not enforce symmetry if rate is unknown
ENGR00132525 sgtl5000: audio clock gating
ENGR00132580 allow prediv by 1 for ext1 clk

(audio clock gating on Efika will be patched in later)

7 years agosdma script code: remove unused file
Matt Sealey [Mon, 13 Dec 2010 21:06:49 +0000 (15:06 -0600)]
sdma script code: remove unused file

7 years agoENGR00131932 upgrade SDMA scripts to 1.1, dual SSI FIFO support
Zeng Zhaoming [Mon, 13 Dec 2010 21:05:29 +0000 (15:05 -0600)]
ENGR00131932 upgrade SDMA scripts to 1.1, dual SSI FIFO support

7 years agoENGR00132971 Workaround for imx5x M4IF burst len errata for armv7 neon
Sammy He [Mon, 13 Dec 2010 20:01:20 +0000 (14:01 -0600)]
ENGR00132971 Workaround for imx5x M4IF burst len errata for armv7 neon

Workaround for imx5x M4IF burst len errata when armv7 neon enabled, use
pgprot_writecombine instead of pgprot_noncached, to avoid burst len > 8.

Signed-off-by: Sammy He <r62914@freescale.com>
7 years agoENGR00133590 UART: let user control HW flow control
Zhang Jiejing [Mon, 13 Dec 2010 19:58:29 +0000 (13:58 -0600)]
ENGR00133590 UART: let user control HW flow control

HW flow control should be enable/disable by userspace.
It can be set by tty option: options.c_cflag of  CRTSCTS
bit.

This bug is found by testing UART BT AR3001 chip,
this chip will wake up the chip by set RTS to high,
it needs frist disable HW flow control, set RTS,
then disable HW flow control.

Signed-off-by: Zhang Jiejing <jiejing.zhang@freescale.com>
7 years agoMatch this patch from Freescale:
Matt Sealey [Mon, 13 Dec 2010 18:29:08 +0000 (12:29 -0600)]
Match this patch from Freescale:

ENGR00133429 MX51 DVFS_PER: Set LP voltage to 1.25 V

MX51 DVFS_PER: Set high setpoint and low setpoint of LP voltage
from 1.2 V to 1.25 V.

Signed-off-by: Nancy Chen <Nancy.Chen@freescale.com>
7 years agoENGR00133681 VPU:Fix kernel dump when VPU doesn't use IRAM
Matt Sealey [Mon, 13 Dec 2010 18:14:16 +0000 (12:14 -0600)]
ENGR00133681 VPU:Fix kernel dump when VPU doesn't use IRAM

If VPU doesn't use IRAM, there will be a kernel dump when
we enter VPU driver probe function.
This issue is caused by allocating IRAM with zero buffer
size.

Signed-off-by: Liu Ying <b17645@freescale.com>
7 years agoENGR00133737 MXC SDMA: fix system hangs when play audio with irq threaded
Matt Sealey [Mon, 13 Dec 2010 18:11:51 +0000 (12:11 -0600)]
ENGR00133737 MXC SDMA: fix system hangs when play audio with irq threaded

When apply rt patch or turn on hardirq threaded in kernel,
play audio by application with realtime schedule policy hangs the system.

It is caused by:
Requesting SDMA channel, a specific channel is used to load script
and context from memory.

This load phase also uses sdma to transfer, then poll
a complete flag within a while(1) loop.

When application with realtime schedule policy request sdma channel,
it will preempt threaded irqs and prevent complete flag to be set.

So a deadlock appears.

Signed-off-by: Zeng Zhaoming <b32542@freescale.com>
7 years agoENGR00132543 v4l2 output: fill black in streamoff
Matt Sealey [Mon, 13 Dec 2010 15:57:03 +0000 (09:57 -0600)]
ENGR00132543 v4l2 output: fill black in streamoff

Fill black data in streamoff, otherwise the last frame will keep in fb.

Signed-off-by: Jason Chen <b02280@freescale.com>
7 years agosiihdmi: add hdmi hotplug event handler (only registers the event, doesn't do anythin...
Matt Sealey [Mon, 13 Dec 2010 15:11:52 +0000 (09:11 -0600)]
siihdmi: add hdmi hotplug event handler (only registers the event, doesn't do anything yet)

7 years agoexperimental: more generic sii hdmi driver
Matt Sealey [Sat, 11 Dec 2010 21:46:40 +0000 (15:46 -0600)]
experimental: more generic sii hdmi driver

7 years agomx: remove mxcfb_dump_modelines as the sii driver includes it.
Matt Sealey [Sat, 11 Dec 2010 21:31:44 +0000 (15:31 -0600)]
mx: remove mxcfb_dump_modelines as the sii driver includes it.

7 years agoEnable section mismatch debug as it's very useful
Matt Sealey [Sat, 11 Dec 2010 21:31:20 +0000 (15:31 -0600)]
Enable section mismatch debug as it's very useful

7 years agoenable unionfs: add Makefile, Kconfig and updated defconfigs to include it
Matt Sealey [Sat, 11 Dec 2010 21:30:44 +0000 (15:30 -0600)]
enable unionfs: add Makefile, Kconfig and updated defconfigs to include it

7 years agoMerge branch 'efikasb-10.08.00' of gitorious.org:efikamx/linux-kernel into efikasb...
Matt Sealey [Sat, 11 Dec 2010 21:29:30 +0000 (15:29 -0600)]
Merge branch 'efikasb-10.08.00' of gitorious.org:efikamx/linux-kernel into efikasb-10.08.00

7 years agoremove unnecessary magic number usage
Saleem Abdulrasool [Fri, 10 Dec 2010 06:02:39 +0000 (22:02 -0800)]
remove unnecessary magic number usage

7 years agofs: unionfs 2.5.7 for linux 2.6.31.14
Matt Sealey [Tue, 7 Dec 2010 13:10:30 +0000 (07:10 -0600)]
fs: unionfs 2.5.7 for linux 2.6.31.14

7 years agobump to .13 finally
Matt Sealey [Tue, 7 Dec 2010 12:34:09 +0000 (06:34 -0600)]
bump to .13 finally

7 years agosii9022: rename notifier block structure item from "fb" to "nb"
Matt Sealey [Tue, 7 Dec 2010 12:33:41 +0000 (06:33 -0600)]
sii9022: rename notifier block structure item from "fb" to "nb"

7 years agoconfigs: add ecryptfs, configfs, security key retention.
Matt Sealey [Tue, 7 Dec 2010 12:32:36 +0000 (06:32 -0600)]
configs: add ecryptfs, configfs, security key retention.
also pulls in requirements for udlfb framebuffer operations

7 years agodev_get_platdata is static
Matt Sealey [Mon, 6 Dec 2010 06:32:03 +0000 (00:32 -0600)]
dev_get_platdata is static

7 years agospi platform data: fix spi platform device to name the mx51 unit and not send the...
Matt Sealey [Mon, 6 Dec 2010 06:21:00 +0000 (00:21 -0600)]
spi platform data: fix spi platform device to name the mx51 unit and not send the driver into autodetect (which cannot autodetect an mx51!?)

7 years agospi_imx: clock hack
Matt Sealey [Mon, 6 Dec 2010 06:20:38 +0000 (00:20 -0600)]
spi_imx: clock hack

7 years agospi_imx: add dev_get_platdata from more recent kernels
Matt Sealey [Mon, 6 Dec 2010 06:09:27 +0000 (00:09 -0600)]
spi_imx: add dev_get_platdata from more recent kernels

7 years agomc13892: remove "connectivity" option as it only applies to USB chargers etc.
Matt Sealey [Mon, 6 Dec 2010 06:04:36 +0000 (00:04 -0600)]
mc13892: remove "connectivity" option as it only applies to USB chargers etc.

7 years agoframebuffer: Remove pixel clock setting hack
Matt Sealey [Mon, 6 Dec 2010 02:09:24 +0000 (20:09 -0600)]
framebuffer: Remove pixel clock setting hack

7 years agosii9022: Add a tiny delay to the setup and mode change in the hope we avoid IPU_INT_S...
Matt Sealey [Mon, 6 Dec 2010 02:06:52 +0000 (20:06 -0600)]
sii9022: Add a tiny delay to the setup and mode change in the hope we avoid IPU_INT_STAT errors