firefly-linux-kernel-4.4.55.git
13 years agork29: rfkill: sdk: fix section mismatch warning
黄涛 [Fri, 29 Jul 2011 01:47:51 +0000 (09:47 +0800)]
rk29: rfkill: sdk: fix section mismatch warning

13 years agork29: rfkill: phonepadsdk: fix section mismatch warning
黄涛 [Fri, 29 Jul 2011 01:47:33 +0000 (09:47 +0800)]
rk29: rfkill: phonepadsdk: fix section mismatch warning

13 years agork29: rfkill: phonesdk: fix section mismatch warning
黄涛 [Fri, 29 Jul 2011 01:47:16 +0000 (09:47 +0800)]
rk29: rfkill: phonesdk: fix section mismatch warning

13 years agork29: rfkill: a22: fix section mismatch warning
黄涛 [Fri, 29 Jul 2011 01:47:02 +0000 (09:47 +0800)]
rk29: rfkill: a22: fix section mismatch warning

13 years agonewton: updata irda serial driver
lyx [Fri, 29 Jul 2011 01:26:21 +0000 (18:26 -0700)]
newton: updata irda serial driver
add ioctl get frame length and start send function

13 years agork29: ddr_recofig: add ddr_reconfig.c
黄涛 [Thu, 28 Jul 2011 11:43:10 +0000 (19:43 +0800)]
rk29: ddr_recofig: add ddr_reconfig.c

13 years agork29: add dynamic DDR reconfiguration support, default disabled
黄涛 [Thu, 28 Jul 2011 11:40:06 +0000 (19:40 +0800)]
rk29: add dynamic DDR reconfiguration support, default disabled

13 years agonet:wireless:bcm4329: add country code setting for various radio regulator domains.
lbt [Thu, 28 Jul 2011 10:40:52 +0000 (18:40 +0800)]
net:wireless:bcm4329: add country code setting for various radio regulator domains.

13 years agoRK29Smartphone:core voltage changed from 1.25v to 1.30v, when cpu frequecy is 1G Hz
phc [Thu, 28 Jul 2011 09:45:06 +0000 (17:45 +0800)]
RK29Smartphone:core voltage changed from 1.25v to 1.30v, when cpu frequecy is 1G Hz

13 years agork29: adc: fix error: incompatible type for argument 1 of spin_lock
黄涛 [Thu, 28 Jul 2011 07:20:41 +0000 (15:20 +0800)]
rk29: adc: fix error: incompatible type for argument 1 of spin_lock

13 years agocamera: update ov2655 sensor driver
ddl [Thu, 28 Jul 2011 07:27:26 +0000 (15:27 +0800)]
camera: update ov2655 sensor driver

13 years agoWM831X:enable battery FET
root [Wed, 27 Jul 2011 14:37:54 +0000 (22:37 +0800)]
WM831X:enable battery FET

13 years agork2818: remove all rk2818 stuff, prepare for 2.6.36
黄涛 [Thu, 28 Jul 2011 02:08:21 +0000 (10:08 +0800)]
rk2818: remove all rk2818 stuff, prepare for 2.6.36

13 years agonewton: add support for the pwm led and update board_rk29_newton.c
lmc [Wed, 27 Jul 2011 09:51:04 +0000 (17:51 +0800)]
newton: add support for the pwm led and update board_rk29_newton.c

13 years agocamera: update gc0308 sensor driver from 和天创
ddl [Wed, 27 Jul 2011 08:17:02 +0000 (16:17 +0800)]
camera: update gc0308 sensor driver from 和天创

13 years agofb:support 180 degree rotate
ddl [Wed, 27 Jul 2011 03:24:16 +0000 (11:24 +0800)]
fb:support 180 degree rotate

13 years agocamera: fix ov5640 driver auto focus may be failed
ddl [Thu, 21 Jul 2011 09:43:15 +0000 (17:43 +0800)]
camera: fix ov5640 driver auto focus may be failed

13 years agouvc: fix dqueue ioctl may hold on forever when sensor is dead
ddl [Thu, 21 Jul 2011 02:45:27 +0000 (10:45 +0800)]
uvc: fix dqueue ioctl may hold on forever when sensor is dead

13 years agocamera: fix ov2659 delay 600ms after capture and pmem allocate for 0.3 Mega
ddl [Tue, 19 Jul 2011 03:16:15 +0000 (11:16 +0800)]
camera: fix ov2659 delay 600ms after capture and pmem allocate for 0.3 Mega

13 years agocamera: fix ov5642 display abnormity moment after capture in focus infinity
ddl [Sat, 16 Jul 2011 02:54:41 +0000 (10:54 +0800)]
camera: fix ov5642 display abnormity moment after capture in focus infinity

13 years agocamera: update ov5642 driver
ddl [Tue, 12 Jul 2011 06:43:40 +0000 (14:43 +0800)]
camera: update ov5642 driver

13 years agocamera: fix camera open operation make a difference between O_RDONLY and O_RDWR for...
ddl [Tue, 12 Jul 2011 03:18:42 +0000 (11:18 +0800)]
camera: fix camera open operation make a difference between O_RDONLY and O_RDWR for accerate open device

13 years agocamera: fix 0.0.2 camera driver is not compatible 0.0.1 driver v4l2 format
ddl [Thu, 23 Jun 2011 07:56:52 +0000 (15:56 +0800)]
camera: fix 0.0.2 camera driver is not compatible 0.0.1 driver v4l2 format

13 years agocamera and uvc: support uvc sensor
ddl [Mon, 20 Jun 2011 07:17:23 +0000 (15:17 +0800)]
camera and uvc: support uvc sensor

13 years agonewton:disable usb host 1.1
蔡枫 [Wed, 27 Jul 2011 02:45:43 +0000 (10:45 +0800)]
newton:disable usb host 1.1

13 years agophonesdk: slove the warnimg that the gpio used but not request
xsf@rock-chips.com [Tue, 26 Jul 2011 10:05:42 +0000 (18:05 +0800)]
phonesdk: slove the warnimg that the gpio used but not request

13 years agolzg: close mirror in ov2655 for newton
lzg [Tue, 26 Jul 2011 09:55:42 +0000 (17:55 +0800)]
lzg: close mirror in ov2655 for newton

13 years agofb:src_vir_w and dst_vir_w of ipp request must be setted.
zyc [Tue, 26 Jul 2011 08:01:08 +0000 (01:01 -0700)]
fb:src_vir_w and dst_vir_w of ipp request  must be setted.

13 years agoMerge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop
蔡枫 [Tue, 26 Jul 2011 03:08:08 +0000 (11:08 +0800)]
Merge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop

13 years agonewton:update cs42l52 driver
蔡枫 [Tue, 26 Jul 2011 03:07:48 +0000 (11:07 +0800)]
newton:update cs42l52 driver

13 years agork29: phonesdk: cpufreq add 1G support
黄涛 [Tue, 26 Jul 2011 02:16:20 +0000 (10:16 +0800)]
rk29: phonesdk: cpufreq add 1G support

13 years agork29: clock: more increase aclk_cpu
黄涛 [Mon, 25 Jul 2011 08:57:50 +0000 (16:57 +0800)]
rk29: clock: more increase aclk_cpu

13 years agork29: cpufreq: support limit by voltage
黄涛 [Mon, 25 Jul 2011 08:56:19 +0000 (16:56 +0800)]
rk29: cpufreq: support limit by voltage

13 years agork29: include/linux/regulator/consumer.h: fix warning: no return statement in functio...
黄涛 [Mon, 25 Jul 2011 08:55:47 +0000 (16:55 +0800)]
rk29: include/linux/regulator/consumer.h: fix warning: no return statement in function returning non-void

13 years agork29: pm-vol.h: fix warning: no return statement in function returning non-void
黄涛 [Mon, 25 Jul 2011 08:54:53 +0000 (16:54 +0800)]
rk29: pm-vol.h: fix warning: no return statement in function returning non-void

13 years agocpufreq: ondemand: rk29 ignore nice default
黄涛 [Mon, 25 Jul 2011 08:32:34 +0000 (16:32 +0800)]
cpufreq: ondemand: rk29 ignore nice default

13 years agonewton:modify bu92747 irda_irq type for fir EOF
lyx [Mon, 25 Jul 2011 07:33:35 +0000 (00:33 -0700)]
newton:modify bu92747 irda_irq type for fir EOF

13 years agolzg:update mt9p111 and s5k6aa
lzg [Mon, 25 Jul 2011 07:13:51 +0000 (15:13 +0800)]
lzg:update mt9p111 and s5k6aa

13 years agonewton: update board-rk29-newton.c
lmc [Mon, 25 Jul 2011 06:22:30 +0000 (14:22 +0800)]
newton: update board-rk29-newton.c

13 years agoclockevents: Use u32 for mult and shift factors
Thomas Gleixner [Wed, 11 Nov 2009 14:05:25 +0000 (14:05 +0000)]
clockevents: Use u32 for mult and shift factors

The mult and shift factors of clock events differ in their data type
from those of clock sources for no reason. u32 is sufficient for
both. shift is always <= 32 and mult is limited to 2^32-1 to avoid
64bit multiplication overflows in the conversion.

Preparatory patch for a generic mult/shift factor calculation
function.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Mikael Pettersson <mikpe@it.uu.se>
Acked-by: Ralf Baechle <ralf@linux-mips.org>
Acked-by: Linus Walleij <linus.walleij@stericsson.com>
Cc: John Stultz <johnstul@us.ibm.com>
LKML-Reference: <20091111134229.725664788@linutronix.de>

13 years agophonesdk: step voltage regulator for Vcore
xsf@rock-chips.com [Fri, 22 Jul 2011 10:26:18 +0000 (18:26 +0800)]
phonesdk: step voltage regulator for Vcore

13 years agoa22 and rk29 phonesdk:32K Frequency switch time cut short
张晴 [Fri, 22 Jul 2011 08:55:58 +0000 (01:55 -0700)]
a22 and rk29 phonesdk:32K Frequency switch time cut short

13 years agoRevert "rk29: add hr_msleep/hr_msleep_interruptible/usleep"
黄涛 [Fri, 22 Jul 2011 08:40:22 +0000 (16:40 +0800)]
Revert "rk29: add hr_msleep/hr_msleep_interruptible/usleep"

This reverts commit c4c6281fd76657b320bb68a1698d469a620ccb6b.

13 years agork29: add hr_msleep/hr_msleep_interruptible/usleep
黄涛 [Fri, 22 Jul 2011 08:19:53 +0000 (16:19 +0800)]
rk29: add hr_msleep/hr_msleep_interruptible/usleep

13 years agotimer: Added usleep_range timer
Patrick Pannuto [Mon, 2 Aug 2010 22:01:04 +0000 (15:01 -0700)]
timer: Added usleep_range timer

usleep_range is a finer precision implementations of msleep
and is designed to be a drop-in replacement for udelay where
a precise sleep / busy-wait is unnecessary.

Since an easy interface to hrtimers could lead to an undesired
proliferation of interrupts, we provide only a "range" API,
forcing the caller to think about an acceptable tolerance on
both ends and hopefully avoiding introducing another interrupt.

INTRO

As discussed here ( http://lkml.org/lkml/2007/8/3/250 ), msleep(1) is not
precise enough for many drivers (yes, sleep precision is an unfair notion,
but consistently sleeping for ~an order of magnitude greater than requested
is worth fixing). This patch adds a usleep API so that udelay does not have
to be used. Obviously not every udelay can be replaced (those in atomic
contexts or being used for simple bitbanging come to mind), but there are
many, many examples of

mydriver_write(...)
/* Wait for hardware to latch */
udelay(100)

in various drivers where a busy-wait loop is neither beneficial nor
necessary, but msleep simply does not provide enough precision and people
are using a busy-wait loop instead.

CONCERNS FROM THE RFC

Why is udelay a problem / necessary? Most callers of udelay are in device/
driver initialization code, which is serial...

As I see it, there is only benefit to sleeping over a delay; the
notion of "refactoring" areas that use udelay was presented, but
I see usleep as the refactoring. Consider i2c, if the bus is busy,
you need to wait a bit (say 100us) before trying again, your
current options are:

* udelay(100)
* msleep(1) <-- As noted above, actually as high as ~20ms
on some platforms, so not really an option
* Manually set up an hrtimer to try again in 100us (which
  is what usleep does anyway...)

People choose the udelay route because it is EASY; we need to
provide a better easy route.

Device / driver / boot code is *currently* serial, but every few
months someone makes noise about parallelizing boot, and IMHO, a
little forward-thinking now is one less thing to worry about
if/when that ever happens

udelay's could be preempted

Sure, but if udelay plans on looping 1000 times, and it gets
preempted on loop 200, whenever it's scheduled again, it is
going to do the next 800 loops.

Is the interruptible case needed?

Probably not, but I see usleep as a very logical parallel to msleep,
so it made sense to include the "full" API. Processors are getting
faster (albeit not as quickly as they are becoming more parallel),
so if someone wanted to be interruptible for a few usecs, why not
let them? If this is a contentious point, I'm happy to remove it.

OTHER THOUGHTS

I believe there is also value in exposing the usleep_range option; it gives
the scheduler a lot more flexibility and allows the programmer to express
his intent much more clearly; it's something I would hope future driver
writers will take advantage of.

To get the results in the NUMBERS section below, I literally s/udelay/usleep
the kernel tree; I had to go in and undo the changes to the USB drivers, but
everything else booted successfully; I find that extremely telling in and
of itself -- many people are using a delay API where a sleep will suit them
just fine.

SOME ATTEMPTS AT NUMBERS

It turns out that calculating quantifiable benefit on this is challenging,
so instead I will simply present the current state of things, and I hope
this to be sufficient:

How many udelay calls are there in 2.6.35-rc5?

udealy(ARG) >= | COUNT
1000 | 319
500 | 414
100 | 1146
20 | 1832

I am working on Android, so that is my focus for this. The following table
is a modified usleep that simply printk's the amount of time requested to
sleep; these tests were run on a kernel with udelay >= 20 --> usleep

"boot" is power-on to lock screen
"power collapse" is when the power button is pushed and the device suspends
"resume" is when the power button is pushed and the lock screen is displayed
         (no touchscreen events or anything, just turning on the display)
"use device" is from the unlock swipe to clicking around a bit; there is no
sd card in this phone, so fail loading music, video, camera

ACTION | TOTAL NUMBER OF USLEEP CALLS | NET TIME (us)
boot | 22 | 1250
power-collapse | 9 | 1200
resume | 5 | 500
use device | 59 | 7700

The most interesting category to me is the "use device" field; 7700us of
busy-wait time that could be put towards better responsiveness, or at the
least less power usage.

Signed-off-by: Patrick Pannuto <ppannuto@codeaurora.org>
Cc: apw@canonical.com
Cc: corbet@lwn.net
Cc: arjan@linux.intel.com
Cc: Randy Dunlap <rdunlap@xenotime.net>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
13 years agoMerge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop
蔡枫 [Fri, 22 Jul 2011 07:01:20 +0000 (15:01 +0800)]
Merge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop

13 years agomodify cir send
clb [Fri, 22 Jul 2011 07:02:31 +0000 (15:02 +0800)]
modify cir send

13 years agoMerge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop
蔡枫 [Fri, 22 Jul 2011 06:55:55 +0000 (14:55 +0800)]
Merge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop

13 years agonewton:remove flash usb mass storage
蔡枫 [Fri, 22 Jul 2011 06:55:44 +0000 (14:55 +0800)]
newton:remove flash usb mass storage

13 years agork29: sdk: defconfig: kernel log buffer size 512KB
黄涛 [Fri, 22 Jul 2011 06:23:40 +0000 (14:23 +0800)]
rk29: sdk: defconfig: kernel log buffer size 512KB

13 years agork29: phonesdk: defconfig: kernel log buffer size 512KB
黄涛 [Fri, 22 Jul 2011 06:23:30 +0000 (14:23 +0800)]
rk29: phonesdk: defconfig: kernel log buffer size 512KB

13 years agork29: phonepadsdk: defconfig: kernel log buffer size 512KB
黄涛 [Fri, 22 Jul 2011 06:23:19 +0000 (14:23 +0800)]
rk29: phonepadsdk: defconfig: kernel log buffer size 512KB

13 years agork29: ddr3sdk: defconfig: kernel log buffer size 512KB
黄涛 [Fri, 22 Jul 2011 06:22:59 +0000 (14:22 +0800)]
rk29: ddr3sdk: defconfig: kernel log buffer size 512KB

13 years agofix bug of GIC crash
yangkai [Fri, 22 Jul 2011 01:45:07 +0000 (09:45 +0800)]
fix bug of GIC crash

13 years agomove clk_set_rate out of irq disabled
yangkai [Thu, 21 Jul 2011 11:56:02 +0000 (19:56 +0800)]
move clk_set_rate out of irq disabled

13 years agoa22 and 29phonesdk:sey VCC25 0V when in sleep,29phonesdk:set key gpio pullup/down...
张晴 [Thu, 21 Jul 2011 10:04:52 +0000 (03:04 -0700)]
a22 and 29phonesdk:sey VCC25 0V when in sleep,29phonesdk:set key gpio pullup/down disable

13 years agoupdate hdmi drivers
kfx [Thu, 21 Jul 2011 08:36:50 +0000 (16:36 +0800)]
update hdmi drivers

13 years agork29_fb.c: hdmi_get_scale-->hdmi_get_fbscale
kfx [Thu, 21 Jul 2011 07:34:40 +0000 (15:34 +0800)]
rk29_fb.c: hdmi_get_scale-->hdmi_get_fbscale

13 years agoRK29SmartPhone:sync. SDCard code with XBW, use the old config
phc [Thu, 21 Jul 2011 07:23:31 +0000 (15:23 +0800)]
RK29SmartPhone:sync. SDCard code with XBW, use the old config

13 years agoclose hdmi debug info
kfx [Thu, 21 Jul 2011 01:44:14 +0000 (09:44 +0800)]
close hdmi debug info

13 years agoupdate hdmi driver
kfx [Thu, 21 Jul 2011 01:29:37 +0000 (09:29 +0800)]
update hdmi driver

13 years agoarm: Makefile: better support make boot.img
黄涛 [Thu, 21 Jul 2011 01:18:45 +0000 (09:18 +0800)]
arm: Makefile: better support make boot.img

13 years agowm8994: fix a bug for Bluetooth headsets
邱建斌 [Wed, 20 Jul 2011 11:27:59 +0000 (19:27 +0800)]
wm8994: fix a bug for Bluetooth headsets

13 years agork29_phone: fix heatset main mic Call Features
邱建斌 [Tue, 19 Jul 2011 02:05:32 +0000 (10:05 +0800)]
rk29_phone: fix heatset main mic Call Features

13 years agonewton:update config and board file
蔡枫 [Wed, 20 Jul 2011 09:57:11 +0000 (17:57 +0800)]
newton:update config and board file

13 years agork29: vpu_service: fix bug when delay work and cancel work come at the same time
陈恒明 [Wed, 20 Jul 2011 08:07:00 +0000 (16:07 +0800)]
rk29: vpu_service: fix bug when delay work and cancel work come at the same time

13 years agosupport 3 usb ethernet: dm9620, ax8872b, sr9700 (2)
胡卫国 [Wed, 20 Jul 2011 07:06:20 +0000 (15:06 +0800)]
support 3 usb ethernet: dm9620, ax8872b, sr9700 (2)

13 years agosupport 3 usb ethernet: dm9620, ax8872b, sr9700
胡卫国 [Wed, 20 Jul 2011 07:03:14 +0000 (15:03 +0800)]
support 3 usb ethernet: dm9620, ax8872b, sr9700

13 years agochange AXI frequency division from 3 to 2
陈辉 [Wed, 20 Jul 2011 06:33:20 +0000 (14:33 +0800)]
change AXI frequency division from 3 to 2

13 years agoMerge branch 'develop' of /home/rockchip/kernel into develop
杨云 [Wed, 20 Jul 2011 01:56:48 +0000 (09:56 +0800)]
Merge branch 'develop' of /home/rockchip/kernel into develop

13 years agoMerge remote branch 'origin/develop_xbw_update-SDCard' into develop
杨云 [Wed, 20 Jul 2011 00:50:06 +0000 (08:50 +0800)]
Merge remote branch 'origin/develop_xbw_update-SDCard' into develop

13 years ago Integrate the XBW-version of SDMMC-drver into rk29_ddr3sdk
xbw [Tue, 19 Jul 2011 14:12:51 +0000 (22:12 +0800)]
 Integrate the XBW-version of SDMMC-drver into rk29_ddr3sdk

13 years agoRevert "rk29: vpu_service: fix bug when __cancel_delayed_work and delay work run...
黄涛 [Tue, 19 Jul 2011 10:10:54 +0000 (18:10 +0800)]
Revert "rk29: vpu_service: fix bug when __cancel_delayed_work and delay work run at the same time"

This reverts commit c41dd263406cad12ec15d39986a530be1d23ad11.

13 years agonewton:add gt819 firmware
蔡枫 [Tue, 19 Jul 2011 10:04:28 +0000 (18:04 +0800)]
newton:add gt819 firmware

13 years agoMerge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop
蔡枫 [Tue, 19 Jul 2011 10:00:33 +0000 (18:00 +0800)]
Merge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop

13 years agonewton:add gt819 firmware
蔡枫 [Tue, 19 Jul 2011 10:00:14 +0000 (18:00 +0800)]
newton:add gt819 firmware

13 years agork29: vpu_service: fix bug when __cancel_delayed_work and delay work run at the same...
陈恒明 [Tue, 19 Jul 2011 09:57:31 +0000 (17:57 +0800)]
rk29: vpu_service: fix bug when __cancel_delayed_work and delay work run at the same time
if in inturrept use async interface else use sync interface

13 years agoarm: Makefile: boot.img use zImage
黄涛 [Tue, 19 Jul 2011 08:50:47 +0000 (16:50 +0800)]
arm: Makefile: boot.img use zImage

13 years agork29: default LZO-compressed kernel for zImage
黄涛 [Tue, 19 Jul 2011 08:24:20 +0000 (16:24 +0800)]
rk29: default LZO-compressed kernel for zImage

13 years agoarm: add support for LZO-compressed kernels
Albin Tonnerre [Fri, 8 Jan 2010 22:42:43 +0000 (14:42 -0800)]
arm: add support for LZO-compressed kernels

- changes to ach/arch/boot/Makefile to make it easier to add new
   compression types
 - new piggy.lzo.S necessary for lzo compression
 - changes in arch/arm/boot/compressed/misc.c to allow the use of lzo or
   gzip, depending on the config
 - Kconfig support

Signed-off-by: Albin Tonnerre <albin.tonnerre@free-electrons.com>
Tested-by: Wu Zhangjin <wuzhangjin@gmail.com>
Acked-by: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Russell King <rmk@arm.linux.org.uk>
Acked-by: Russell King <rmk@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Martin Michlmayr <tbm@cyrius.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agolib: add support for LZO-compressed kernels
Albin Tonnerre [Fri, 8 Jan 2010 22:42:42 +0000 (14:42 -0800)]
lib: add support for LZO-compressed kernels

This patch series adds generic support for creating and extracting
LZO-compressed kernel images, as well as support for using such images on
the x86 and ARM architectures, and support for creating and using
LZO-compressed initrd and initramfs images.

Russell King said:

: Testing on a Cortex A9 model:
: - lzo decompressor is 65% of the time gzip takes to decompress a kernel
: - lzo kernel is 9% larger than a gzip kernel
:
: which I'm happy to say confirms your figures when comparing the two.
:
: However, when comparing your new gzip code to the old gzip code:
: - new is 99% of the size of the old code
: - new takes 42% of the time to decompress than the old code
:
: What this means is that for a proper comparison, the results get even better:
: - lzo is 7.5% larger than the old gzip'd kernel image
: - lzo takes 28% of the time that the old gzip code took
:
: So the expense seems definitely worth the effort.  The only reason I
: can think of ever using gzip would be if you needed the additional
: compression (eg, because you have limited flash to store the image.)
:
: I would argue that the default for ARM should therefore be LZO.

This patch:

The lzo compressor is worse than gzip at compression, but faster at
extraction.  Here are some figures for an ARM board I'm working on:

Uncompressed size: 3.24Mo
gzip  1.61Mo 0.72s
lzo   1.75Mo 0.48s

So for a compression ratio that is still relatively close to gzip, it's
much faster to extract, at least in that case.

This part contains:
 - Makefile routine to support lzo compression
 - Fixes to the existing lzo compressor so that it can be used in
   compressed kernels
 - wrapper around the existing lzo1x_decompress, as it only extracts one
   block at a time, while we need to extract a whole file here
 - config dialog for kernel compression

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: cleanup]
Signed-off-by: Albin Tonnerre <albin.tonnerre@free-electrons.com>
Tested-by: Wu Zhangjin <wuzhangjin@gmail.com>
Acked-by: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Russell King <rmk@arm.linux.org.uk>
Acked-by: Russell King <rmk@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agogpu : fix the gcdOPTIMIZE_HEAP_SAMESIZE for heap alloc.
杜坤明 [Tue, 19 Jul 2011 08:02:03 +0000 (16:02 +0800)]
gpu : fix the gcdOPTIMIZE_HEAP_SAMESIZE for heap alloc.

13 years agork29: i2c: cpu freq notifier only for i2c0
黄涛 [Tue, 19 Jul 2011 07:24:29 +0000 (15:24 +0800)]
rk29: i2c: cpu freq notifier only for i2c0

13 years agoIPP fix bugs:
chenli [Tue, 19 Jul 2011 01:44:37 +0000 (09:44 +0800)]
IPP fix bugs:

1.Disable pre_scaling or deinterlacing when there is no request to do
2.Store clip mode always set to 1 now

13 years agoMerge remote branch 'origin/develop_xbw_update-SDCard' into develop
杨云 [Tue, 19 Jul 2011 01:02:40 +0000 (09:02 +0800)]
Merge remote branch 'origin/develop_xbw_update-SDCard' into develop

13 years agork29: clock: bump version
黄涛 [Mon, 18 Jul 2011 11:43:08 +0000 (19:43 +0800)]
rk29: clock: bump version

13 years agolcd: nt35510: aclk_lcdc max 300MHz
黄涛 [Mon, 18 Jul 2011 11:29:29 +0000 (19:29 +0800)]
lcd: nt35510: aclk_lcdc max 300MHz

13 years agolcd: ls035y8dx02a: aclk_lcdc max 300MHz
黄涛 [Mon, 18 Jul 2011 11:28:48 +0000 (19:28 +0800)]
lcd: ls035y8dx02a: aclk_lcdc max 300MHz

13 years agork29: sdk: more fix for vmac and usb host 1.1 both enabled
黄涛 [Mon, 18 Jul 2011 11:21:12 +0000 (19:21 +0800)]
rk29: sdk: more fix for vmac and usb host 1.1 both enabled

13 years agork29: ddr3sdk: more fix for vmac and usb host 1.1 both enabled
黄涛 [Mon, 18 Jul 2011 11:18:36 +0000 (19:18 +0800)]
rk29: ddr3sdk: more fix for vmac and usb host 1.1 both enabled

13 years agork29: clock: fix aclk_lcdc from ddr pll
黄涛 [Mon, 18 Jul 2011 11:16:33 +0000 (19:16 +0800)]
rk29: clock: fix aclk_lcdc from ddr pll

13 years agoMerge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop
CMY [Mon, 18 Jul 2011 10:02:10 +0000 (18:02 +0800)]
Merge branch 'develop' of 10.10.10.29:/home/rockchip/kernel into develop

13 years ago新的modem驱动
CMY [Mon, 18 Jul 2011 09:59:11 +0000 (17:59 +0800)]
新的modem驱动

13 years ago支持更多的3g dongle
CMY [Mon, 18 Jul 2011 09:58:48 +0000 (17:58 +0800)]
支持更多的3g dongle

13 years agonewton:fix touch screen compile error
蔡枫 [Mon, 18 Jul 2011 09:19:57 +0000 (17:19 +0800)]
newton:fix touch screen compile error

13 years agonewton: modify irda serial device suspend&resume function
lyx [Mon, 18 Jul 2011 06:45:50 +0000 (23:45 -0700)]
newton: modify irda serial device suspend&resume function

13 years agomodify rk29_sdk_defconfig for SDMMC
xbw [Mon, 18 Jul 2011 05:58:43 +0000 (22:58 -0700)]
modify rk29_sdk_defconfig for SDMMC

13 years agoIntegrae the XBW-version of SDMMC-drver into RK29SDK, and the xbw-version is set...
xbw [Mon, 18 Jul 2011 03:45:18 +0000 (20:45 -0700)]
Integrae the XBW-version of SDMMC-drver into RK29SDK, and the xbw-version is set to default(2)

13 years agoIntegrae the XBW-version of SDMMC-drver into RK29SDK, and the xbw-version is set...
xbw [Mon, 18 Jul 2011 03:41:08 +0000 (20:41 -0700)]
Integrae the XBW-version of SDMMC-drver into RK29SDK, and the xbw-version is set to default

13 years agork29_phone: change headset volume
邱建斌 [Mon, 18 Jul 2011 02:40:43 +0000 (10:40 +0800)]
rk29_phone: change headset volume