]> git.karo-electronics.de Git - karo-tx-linux.git/log
karo-tx-linux.git
11 years agoENGR00225981-8 csi/imx6sl: add the platform-related setting
Robby Cai [Thu, 27 Sep 2012 15:50:08 +0000 (23:50 +0800)]
ENGR00225981-8 csi/imx6sl: add the platform-related setting

- add platform data for csi driver
- change the regulator name to reflect the voltage really used
- select OSC as csi parent clock to get 24MHz
- add an boot option to use csi feature while filter out the EPDC/SPDC, since
  there are pin conflicts with xPDC.
- both ov5640 and ov5642 are verified okay, ov5640 is used by default.
- remove IPU from update_defconfig

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225981-7 csi: Unify the api set_mclk_rate
Robby Cai [Thu, 27 Sep 2012 14:47:20 +0000 (22:47 +0800)]
ENGR00225981-7 csi: Unify the api set_mclk_rate

For CSI on MX6SL, there's only 1 CSI and hence no need to change csi souce.
We add this parameter just for compatibility.

Adjust the called api in related camera drivers.

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225981-6 csi: Add PxP post-processing for CSC
Robby Cai [Thu, 27 Sep 2012 13:50:59 +0000 (21:50 +0800)]
ENGR00225981-6 csi: Add PxP post-processing for CSC

The default output format from camera is UYVY. For preview purpose, need
convert the format to RGB565. Use PxP CSC as post processing.

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225981-5 imx6sl: update the defconfig
Robby Cai [Fri, 28 Sep 2012 10:41:47 +0000 (18:41 +0800)]
ENGR00225981-5 imx6sl: update the defconfig

- remove IPU define
- add csi camera define

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225981-4 IPU: Remove IPU fb build for MX6SL
Robby Cai [Thu, 27 Sep 2012 13:02:01 +0000 (21:02 +0800)]
ENGR00225981-4 IPU: Remove IPU fb build for MX6SL

Filter out IPU framebuffer for MX6SL by adding conditional control.

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225981-3 csi: Add csi driver support on MX6SL
Robby Cai [Thu, 27 Sep 2012 13:02:01 +0000 (21:02 +0800)]
ENGR00225981-3 csi: Add csi driver support on MX6SL

- re-use csi driver from mx25
- set csi irq/baseaddr as platform resource

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225981-2 pxp: fix the coefficient for the YCbCr-RGB CSC
Robby Cai [Fri, 28 Sep 2012 09:25:25 +0000 (17:25 +0800)]
ENGR00225981-2 pxp: fix the coefficient for the YCbCr-RGB CSC

Fix the coefficient for the YCbCr-RGB Color space conversion.

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225981-1 pxp: add UYVY input pixel format
Robby Cai [Thu, 27 Sep 2012 08:29:07 +0000 (16:29 +0800)]
ENGR00225981-1 pxp: add UYVY input pixel format

The default camera output format is UYVY, PxP is used to convert UYVY to RGB565
for displaying on LCD.

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00225960-04 mx6sl evk: Add sii902x GPIO/IOMux setting
Sandor Yu [Fri, 28 Sep 2012 05:16:19 +0000 (13:16 +0800)]
ENGR00225960-04 mx6sl evk: Add sii902x GPIO/IOMux setting

Add sii902x GPIO and IOMUX setting to mx6sl evk board.

Signed-off-by: Sandor Yu <R01008@freescale.com>
11 years agoENGR00225960-03 sii902x: Added sii902x HDMI driver with ELCDIF FB
Sandor Yu [Fri, 28 Sep 2012 05:15:27 +0000 (13:15 +0800)]
ENGR00225960-03 sii902x: Added sii902x HDMI driver with ELCDIF FB

Enable sii902x HDMI driver with ELCDIF framebuffer interface.
HDMI cable Hotplug support.
Video mode switch support.
Blank/Unblank HDMI display support.
Added config variable CONFIG_FB_MXC_SII902X_ELCDIF for sii902x_elcdif module

Signed-off-by: Sandor Yu <R01008@freescale.com>
11 years agoENGR00225960-02 FB: Support sii902x HDMI driver in ELCDIF FB
Sandor Yu [Fri, 28 Sep 2012 05:12:52 +0000 (13:12 +0800)]
ENGR00225960-02 FB: Support sii902x HDMI driver in ELCDIF FB

- Added mxcfb_elcdif_register_mode function.
- Create video mode list, and check default video mode
with video mode list before setting.
- Adjust elcdif pixel clock setting, reconfig elcdif pixel
parent clock video pll, get more accurate pixel clock
according video mode.
- Added video mode dump function for debug

Signed-off-by: Sandor Yu <R01008@freescale.com>
11 years agoENGR00225960-01 Sii902x HDMI: added SII902x Config variable
Sandor Yu [Fri, 28 Sep 2012 03:17:18 +0000 (11:17 +0800)]
ENGR00225960-01 Sii902x HDMI: added SII902x Config variable

-Add CONFIG_FB_MXC_SII902x_ELCDIF variable to mx6s config files.
-Default enable CONFIG_FB_MXC_ELCDIF_FB,
CONFIG_FB_MXC_SEIKO_WVGA_SYNC_PANEL
and CONFIG_FB_MXC_SII902x_ELCDIF.

Signed-off-by: Sandor Yu <R01008@freescale.com>
11 years agoENGR00225735-3 GPU: add gpu regulator disable/enable in gpu driver
Robin Gong [Tue, 25 Sep 2012 08:47:48 +0000 (16:47 +0800)]
ENGR00225735-3 GPU: add gpu regulator disable/enable in gpu driver

add gpu regulator management in gpu driver
Signed-off-by: Robin Gong <b38343@freescale.com>
Acked-by: Lily Zhang
Signed-off-by: Robin Gong <b38343@freescale.com>
11 years agoENGR00225735-2 VPU: add vpu regulator enable/disable in driver
Robin Gong [Tue, 25 Sep 2012 08:43:57 +0000 (16:43 +0800)]
ENGR00225735-2 VPU: add vpu regulator enable/disable in driver

add vpu regulator management including enable/disable in vpu driver. Mainly
in open/release and suspend/resume.
Signed-off-by: Robin Gong <b38343@freescale.com>
11 years agoENGR00225735-1 PU anatop: PU regulator can be disabled/enabled by drivers
Robin Gong [Tue, 25 Sep 2012 08:33:37 +0000 (16:33 +0800)]
ENGR00225735-1 PU anatop: PU regulator can be disabled/enabled by drivers

Before, PU regulator only be turned off in DSM, which means it kept on in
system normal mode even GPU/VPU driver didn't run. To decrease power number,
PU regulator can be disabled/enabled by GPU/VPU driver, and the voltage value
is tracked by VDDARM which change in cpufreq driver.The patch including:
1.implement PU regulator enable/disable interface in anatop regulator driver
2.remove gpu_power_down and gpu_power_up in system suspend/resume flow.
3.skip change pu regulator set if it has been disabled.

Note: There is three power supply on  VDDPU:
a).Use internal anatop PU regulator, VDDDPU_IN is fixed. In this case,VDDPU_CAP
can be turned off or dynamic change by internal anatop(track with VDDSOCi_CAP).
In other words, it use "cpu_vddvpu" regulator as PU regulator,not only in GPU/
VPU driver, but also in cpufreq driver.Sabresd,Sololite EVK is in this case if
disable CONFIG_MX6_INTER_LDO_BYPASS.

b).Use external pmic regulator as PU regulator, it means in LDO bypass way
(CONFIG_MX6_INTER_LDO_BYPASS).But VDDPU_IN is connected with VDDSOC_IN. Because
VDDSOC can't be turned off for ever, and VDDPU track with VDDSOC always, we
remove  "pu_id", so that cpufreq driver will never touch it.But GPU/VPU driver
will turn off/on VDDPU_CAP by disabling/enabling "cpu_vddvpu".In this case,
although VDDPU_IN is supplied by external pmic in hardware level, but we can't
turn off external pmic directly, because it connect with VDDSOC_IN.So only we
can do is turn off VDDPU_CAP by internal anatop regulator("cpu_vddvpu").Sabresd
,Sololite EVK is in this case if enable CONFIG_MX6_INTER_LDO_BYPASS.

c).Use external pmic regulator as PU regulator, it means in LDO bypass way
(CONFIG_MX6_INTER_LDO_BYPASS). And VDDPU_IN is in dependent with VDDSOC_IN(
whether short in VDDPU_IN with VDDPU_CAP directly or not).In this case,the only
thing we need to do is that add the right external vddpu regulator name to
"pu_id" in their boad file, and nothing else.
Just like below (arch/arm/mach-mx6/board-mx6sl_evk.c):

static struct mxc_dvfs_platform_data mx6sl_evk_dvfscore_data = {
.reg_id = "VDDCORE",
.soc_id = "VDDSOC",
*******************************************************************
.pu_id = "what you used VDDPU_IN regulator name from external pmic"
*******************************************************************
.reg_id = "cpu_vddgp",
.soc_id = "cpu_vddsoc",
.pu_id = "cpu_vddvpu",
......

Signed-off-by: Robin Gong <b38343@freescale.com>
11 years agoENGR00225894 MX6SL-Improve system IDLE power numbers
Ranjani Vaidyanathan [Mon, 17 Sep 2012 09:16:51 +0000 (04:16 -0500)]
ENGR00225894 MX6SL-Improve system IDLE power numbers

Add the following power optimizations when all PLLs
are either disabled or in bypass:
1. Disable 2P5 in system IDLE mode and enable weak 2P5.
2. Set OSC bias current to -37.5% just before the WFI instruction
and set it back to 0% after WFI.
3. Enable the low power bandgap and power down the regular
bandgap.

Also lower AHB and AXI to 3MHz in this mode.

Signed-off-by: Ranjani Vaidyanathan <ra5478@freescale.com>
11 years agoENGR00226095-2 SPDIF: add HDMI AUDIO in device name in mx6_sl evk
Gary Zhang [Thu, 27 Sep 2012 07:49:08 +0000 (15:49 +0800)]
ENGR00226095-2 SPDIF: add HDMI AUDIO in device name in mx6_sl evk

because HDMI AUDIO in mx6sl evk comes from SPDIF TX, change to
HDMI AUDIO in device name.

Signed-off-by: Gary Zhang <b13634@freescale.com>
11 years agoENGR00226095-1 MX6 SL: add SPDIF support in EVK
Gary Zhang [Thu, 27 Sep 2012 07:45:41 +0000 (15:45 +0800)]
ENGR00226095-1 MX6 SL: add SPDIF support in EVK

HDMI AUDIO in evk comes from SPDIF TX, and add SPDIF TX support
to implement HDMI AUDIO feature

Signed-off-by: Gary Zhang <b13634@freescale.com>
11 years agoENGR00226080 Update GC2000/GC320 AXI bus cache attribute
Loren Huang [Thu, 27 Sep 2012 08:43:10 +0000 (16:43 +0800)]
ENGR00226080 Update GC2000/GC320 AXI bus cache attribute

This patch changed the GPU AXI bus cache attribute
to improve the performance of fillrate bound case.

Here is some test result: BEFORE:AFTER (larger better)
mm06 samruai fps:    117.74:137.64
mm06 proxycon fps:   117.90:135.00
Fillrate with 1 tex: 251.3M:331.1M
Fillrate with 2 tex: 402.6M:475.4M

Signed-off-by: Loren Huang <b02279@freescale.com>
Acked-by: Lily Zhang
11 years agoENGR00226096 mx6sl: remove high set point for usdhc
Ryan QIAN [Thu, 27 Sep 2012 08:12:56 +0000 (16:12 +0800)]
ENGR00226096 mx6sl: remove high set point for usdhc

- Due to usdhc ADMA mode has issue working with low bus freq, in
driver usdhc has been configured to SDMA mode already. So it is
not necessary to keep high set point flag any more.

Signed-off-by: Ryan QIAN <b32804@freescale.com>
Acked-by: Dong Aisheng <b29396@freescale.com>
Acked-by: Robby CAI <r63905@freescale.com>
11 years agoENGR00223797-10 ARM:MX6 Kconfig:Support MX6_CLK_FOR_BOOTUI_TRANS
Liu Ying [Thu, 13 Sep 2012 09:30:10 +0000 (17:30 +0800)]
ENGR00223797-10 ARM:MX6 Kconfig:Support MX6_CLK_FOR_BOOTUI_TRANS

This patch adds CONFIG_MX6_CLK_FOR_BOOTUI_TRANS support. Setting
this Kconfig may keep enable IPU related clocks and PWM clocks
and avoid setting IPU related clocks' parents when initializing
clock tree so that bootloader splashimage can transition to
kernel smoothly.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit 21b5e4d436666967585313948a03421da1f8f68b)

11 years agoENGR00223797-7 MX6 SabreSD:Fix ipuv3 fb platform driver register
Liu Ying [Tue, 11 Sep 2012 04:49:49 +0000 (12:49 +0800)]
ENGR00223797-7 MX6 SabreSD:Fix ipuv3 fb platform driver register

As there is only one IPU embedded in MX6DL and two IPUs embedded
in MX6DQ. The max ipuv3 fb platform driver number is two for
MX6DL and four for MX6DQ.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
11 years agoENGR00223797-6 MX6 SabreSD:Initialize late init for IPUv3 fb pdata
Liu Ying [Wed, 12 Sep 2012 01:43:12 +0000 (09:43 +0800)]
ENGR00223797-6 MX6 SabreSD:Initialize late init for IPUv3 fb pdata

This patch initializes late init field to false for IPUv3 fb pdata,
so we don't support late init by default unless we change them to
true in other places specificly.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit cc9baf8aa84f4dd0120aa85960d254646a3aa523)

11 years agoENGR00223797-5 MX6 SabreSD:Initialize bypass reset for IPUv3 pdata
Liu Ying [Wed, 12 Sep 2012 01:36:12 +0000 (09:36 +0800)]
ENGR00223797-5 MX6 SabreSD:Initialize bypass reset for IPUv3 pdata

This patch initializes bypass reset field to false for IPUv3 pdata,
so we don't support bypass reset by default unless we change them
to true in other places specificly.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit 6c294588a1987810ed4d45bcb5e1d3416d66d0be)

11 years agoENGR00223797-4 IPUv3 fb:Support framebuffer late init
Liu Ying [Tue, 11 Sep 2012 10:10:11 +0000 (18:10 +0800)]
ENGR00223797-4 IPUv3 fb:Support framebuffer late init

This patch adds late init support in IPUv3 fb driver
to avoid IPUv3 fb being re-initialized during probe if
a platform supports smooth transition from bootloader
splashimage to system UI. The re-initialization will
be delayed to the first time the user triggers
mxcfb_set_par() and unblank the framebuffer.
The following items are done to support this:
1) Move global alpha and color key setting in probe
   after framebuffer is registered(before registering
   we enable IPU hsp clock), because the 2 APIs enable
   and disable IPU hsp clock which may cause IPU stops
   running in ipuv3 fb probe function.
2) Do not clear framebuffer content in probe function
   if late init is set. This is to avoid bootloader
   splashimage content is cleared.
3) If late init is set, do not re-initialize and
   unblank framebuffer in probe function, but initialize
   and enable ipu display channel instead to enable
   the ipu hsp clock. Refer to the code comment for
   detail.
4) Delay register IPU interrupts used by framebuffer
   until IPU hsp clock is enabled by 3). As the APIs
   to register IPU interrupts may enable and disable
   IPU hsp clock as well.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit 926a6187d89afb5a37dc3b1f3c7bc5225742d09b)

11 years agoENGR00223797-3 ARM:IPUv3 fb:Add late init in pdata
Liu Ying [Tue, 11 Sep 2012 10:20:39 +0000 (18:20 +0800)]
ENGR00223797-3 ARM:IPUv3 fb:Add late init in pdata

This patch adds late init field support in ipuv3 fb
platform data, so that a platform may choose to not
to initialize framebuffer until the user triggers
set_par(), which may support smooth transition from
bootloader splashimage to system UI.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit 3a2547ac167fb2c06d9f63c7a925b9ebdc896d96)

11 years agoENGR00223797-2 IPUv3:Support bypass reset for probe
Liu Ying [Tue, 11 Sep 2012 10:18:39 +0000 (18:18 +0800)]
ENGR00223797-2 IPUv3:Support bypass reset for probe

This patch adds bypass reset support in IPUv3 driver
probe function to avoid IPUv3 being reset if a plat-
form supports smooth transition from bootloader
splashimage to system UI.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit d8c12ced755491d8337fb7456fa78004701d5408)

11 years agoENGR00223797-1 ARM:IPUv3:Add bypass reset in pdata
Liu Ying [Tue, 11 Sep 2012 10:16:12 +0000 (18:16 +0800)]
ENGR00223797-1 ARM:IPUv3:Add bypass reset in pdata

This patch adds bypass reset field support in ipuv3
platform data, so that a platform may choose not to
reset ipuv3 when doing probe, which may support
smooth transition from bootloader splashimage to
system UI.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit e558b2b6877a2b230c564a1c621bbba3d546f5f7)

11 years agoENGR00220747 V4L2 out: Fix mosaic and flash issue if use VDOA input crop mode
Wayne Zou [Thu, 30 Aug 2012 09:20:06 +0000 (17:20 +0800)]
ENGR00220747 V4L2 out: Fix mosaic and flash issue if use VDOA input crop mode

Fix bug:
A lot of mosaic and flash issue if use VDOA mode when doing video playback
on stream H264_BP13_fc_320x136_29.97_444_AAC_48_130_2_vbr.mp4.

For this special stream, the frame size is 320x144 and the stream includes
input cropping information:320x136. It needs to use 320x144 frame size to
do tiled format conversion instead of 320x128 after alignment.

Signed-off-by: Wayne Zou <b36644@freescale.com>
11 years agoENGR00225912 ARM: etm: Add trace buffer support
Frank Li [Wed, 26 Sep 2012 06:05:57 +0000 (14:05 +0800)]
ENGR00225912 ARM: etm: Add trace buffer support

start trace:
echo 1 > echo 1 >/sys/devices/etm.0/trace_running

Notes: The other cores ptm also enabled by above command.

dump trace buffer:
echo v >/proc/sysrq-trigger

Decode trace buffer:
/unit_test/etm --pft-1.1 --sourceid-match 0 < /dev/tracebuf

Notes: this version need connect JTAG to make etm work.

Signed-off-by: Frank Li <Frank.Li@freescale.com>
11 years agoARM: etm: Add sysfs entry to enable return stack if supported
Arve Hjønnevåg [Wed, 4 Apr 2012 01:01:03 +0000 (18:01 -0700)]
ARM: etm: Add sysfs entry to enable return stack if supported

Change-Id: Icb73d60324ad0ddfc3e8a450a28bb3d90c702788
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Add sysfs entry to disable branch_output flag
Arve Hjønnevåg [Tue, 3 Apr 2012 23:15:36 +0000 (16:15 -0700)]
ARM: etm: Add sysfs entry to disable branch_output flag

Change-Id: Ib91208a2c33621aa2d7bd9aa72bfbc670d9d5f1d
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Add sysfs entry to set context-id-size
Arve Hjønnevåg [Tue, 3 Apr 2012 00:20:32 +0000 (17:20 -0700)]
ARM: etm: Add sysfs entry to set context-id-size

Change-Id: I520dfb6e593dac131de8b9b1db77f1c734f18c24
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Add sysfs entry to enable timestamps if supported
Arve Hjønnevåg [Tue, 3 Apr 2012 00:20:32 +0000 (17:20 -0700)]
ARM: etm: Add sysfs entry to enable timestamps if supported

Change-Id: Iff964ba2f6236ed81863e02ec7b3ec9fbc48044a
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Check arch version and disable data tracing for ptm
Arve Hjønnevåg [Tue, 3 Apr 2012 03:18:56 +0000 (20:18 -0700)]
ARM: etm: Check arch version and disable data tracing for ptm

Change-Id: If2cb7928d0711f48348443d882a12416be9c5910
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Wait for etm/ptm(s) to stop before requesting PowerDown
Arve Hjønnevåg [Thu, 29 Mar 2012 04:03:13 +0000 (21:03 -0700)]
ARM: etm: Wait for etm/ptm(s) to stop before requesting PowerDown

When PowerDown was requested at the same time as ProgBit, the
formatter flush command that follows could get stuck.

Change-Id: Iafb665f61f055819e64ca1dcb60398c656f593e4
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Power down etm(s) when tracing is not enabled
Arve Hjønnevåg [Thu, 24 Feb 2011 00:51:58 +0000 (16:51 -0800)]
ARM: etm: Power down etm(s) when tracing is not enabled

Without this change a saw an 18% increase in idle power consumption
on one deivce when trace support is compiled into the kernel. Now
I see the same increase only when tracing.

Change-Id: I21bb5ecf1b7d29ce3790ceeb5323409cc22d5a3b
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Support multiple ETMs/PTMs.
Arve Hjønnevåg [Sat, 5 Feb 2011 06:38:14 +0000 (22:38 -0800)]
ARM: etm: Support multiple ETMs/PTMs.

If more than one ETM or PTM are present, configure all of them
and enable the formatter in the ETB. This allows tracing on dual
core systems (e.g. omap4).

Change-Id: I028657d5cf2bee1b23f193d4387b607953b35888
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Return the entire trace buffer if it is empty after reset
Arve Hjønnevåg [Sat, 5 Feb 2011 06:38:14 +0000 (22:38 -0800)]
ARM: etm: Return the entire trace buffer if it is empty after reset

On some SOCs the read and write pointer are reset when the chip
resets, but the trace buffer content is preserved. If the status
bits indicates that the buffer is empty and we have never started
tracing, assume the buffer is full instead. This can be useful
if the system rebooted from a watchdog reset.

Change-Id: Iaf21c2c329c6059004ee1d38e3dfff66d7d28029
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Add some missing locks and error checks
Arve Hjønnevåg [Tue, 15 Feb 2011 05:09:51 +0000 (21:09 -0800)]
ARM: etm: Add some missing locks and error checks

It is not safe to call etm_lock or etb_lock without holding the
mutex since another thread may also have unlocked the registers.

Also add some missing checks for valid etb_regs in the etm sysfs
entries.

Change-Id: I939f76a6ea7546a8fc0d4ddafa2fd2b6f38103bb
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Configure data tracing
Arve Hjønnevåg [Tue, 1 Feb 2011 02:33:55 +0000 (18:33 -0800)]
ARM: etm: Configure data tracing

The old code enabled data tracing, but did not configure the
range. We now configure it to trace all data addresses by default,
and add a trace_data_range attribute to change the range or disable
data tracing.

Change-Id: I9d04e3e1ea0d0b4d4d5bcb93b1b042938ad738b2
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Allow range selection
Arve Hjønnevåg [Sat, 29 Jan 2011 07:44:43 +0000 (23:44 -0800)]
ARM: etm: Allow range selection

Trace kernel text segment by default as before, allow tracing of other
ranges by writing a range to /sys/devices/etm/trace_range, or to trace
everything by writing 0 0.

Change-Id: Ibb734ca820fedf79560b20536247f1e1700cdc71
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Don't try to clear the buffer full status after reading the buffer
Arve Hjønnevåg [Tue, 1 Feb 2011 05:34:47 +0000 (21:34 -0800)]
ARM: etm: Don't try to clear the buffer full status after reading the buffer

If the write address was at the end of the buffer, toggling the trace
capture bit would set the RAM-full status instead of clearing it, and
if any of the stop bits in the formatter is set toggling the trace
capture bit may not do anything.

Instead use the read position to find out if the data has already
been returned.

This also fixes the read function so it works when the trace buffer is
larger than the buffer passed in from user space. The old version
would reset the trace buffer pointers after every read, so the second
call to read would always return 0.

Change-Id: I75256abe2556adfd66fd5963e46f9e84ae4645e1
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Don't limit tracing to only non-secure code.
Arve Hjønnevåg [Sat, 29 Jan 2011 07:33:11 +0000 (23:33 -0800)]
ARM: etm: Don't limit tracing to only non-secure code.

On some systems kernel code is considered secure, and this code
already limits tracing to the kernel text segment which results
in no trace data.

Change-Id: I098a0753e874859446d098e1ee209f67fc13cd5d
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoARM: etm: Don't require clock control
Arve Hjønnevåg [Sat, 29 Jan 2011 07:12:32 +0000 (23:12 -0800)]
ARM: etm: Don't require clock control

If clk_get fail, assume the etb does not need a separate clock.

Change-Id: Ia0bf3f5391e94a60ea45876aa7afc8a88a7ec3bf
Signed-off-by: Arve Hjønnevåg <arve@android.com>
11 years agoENGR00225131-06 MX6 PM: clear stop_mode_config after system resume
make shi [Thu, 20 Sep 2012 09:20:14 +0000 (17:20 +0800)]
ENGR00225131-06 MX6 PM: clear stop_mode_config after system resume

- If stop_mode_config is set as 1, the USB otg vbus wakeup system can be
  supported
- After system resume, need clear stop_mode_config bit to keep the 1P1
  default off

Signed-off-by: make shi <b15407@freescale.com>
11 years agoENGR00225131-05 MX6 USB: set stop_mode_config bit if wake up is enabled
make shi [Thu, 20 Sep 2012 09:15:26 +0000 (17:15 +0800)]
ENGR00225131-05 MX6 USB: set stop_mode_config bit if wake up is enabled

IC designer had clarified that 1P1 can be turned off if we do not need support
remote wakeup. So If there is no requirement for USB remote wake up, the 1P1
can be turn off. USB driver will support dynamically turn on(off) 1P1 during
system suspend. 1P1 will be turn on depend on USB wakeup is enabled.

- Set stop_mode_config bit if USB host need support USB remote wake up
- Set stop_mode_config bit if USB device need support USB DP/DM wake
  up system

Signed-off-by: make shi <b15407@freescale.com>
11 years agoENGR00225131-04 MX6 USB: implement platform_phy_power_on for USB otg and h1
make shi [Thu, 20 Sep 2012 09:12:47 +0000 (17:12 +0800)]
ENGR00225131-04 MX6 USB: implement platform_phy_power_on for USB otg and h1

IC designer had clarified that 1P1 can be turned off if we do not need support
remote wakeup. So If there is no requirement for USB remote wake up, the 1P1
can be turn off. USB driver will support dynamically turn on(off) 1P1 during
system suspend. 1P1 will be turn on depend on USB wakeup is enabled.

Set stop_mode_config bit to turn on 1P1 during system suspend to support USB
host remote wake up or USB device DP/DM wake up.

Signed-off-by: make shi <b15407@freescale.com>
11 years agoENGR00225131-03 MX6 USB: add platform_phy_power_on platform data in head file
make shi [Thu, 20 Sep 2012 09:10:20 +0000 (17:10 +0800)]
ENGR00225131-03 MX6 USB: add platform_phy_power_on platform data in head file

IC designer had clarified that 1P1 can be turned off if we do not need support
remote wakeup. So If there is no requirement for USB remote wake up, the 1P1
can be turn off. USB driver will support dynamically turn on(off) 1P1 during
system suspend. 1P1 will be turn on depend on USB wakeup is enabled.

add platform_1p1_on platform data in head file.

Signed-off-by: make shi <b15407@freescale.com>
11 years agoENGR00225131-02 USB: revert ENGR00221317-02 stop_mode_config patch head file
make shi [Thu, 20 Sep 2012 07:14:58 +0000 (15:14 +0800)]
ENGR00225131-02 USB: revert ENGR00221317-02 stop_mode_config patch head file

IC designer had clarified that 1P1 can be turned off if we do not need support
remote wakeup. So If there is no requirement for USB remote wake up, the 1P1
can be turn off. USB driver will support dynamically turn on(off) 1P1 during
system suspend. 1P1 will be turn on depend on USB wakeup is enabled.

This reverts commit cc2e51368f2620a37370746eaf35f0b70674b31e.

Signed-off-by: make shi <b15407@freescale.com>
11 years agoENGR00225131-01 USB: revert ENGR00221317-02 about set stop_mode_config patch
make shi [Thu, 20 Sep 2012 07:11:46 +0000 (15:11 +0800)]
ENGR00225131-01 USB: revert ENGR00221317-02 about set stop_mode_config patch

IC designer had clarified that 1P1 can be turned off if we do not need support
remote wakeup. So If there is no requirement for USB remote wake up, the 1P1
can be turn off. USB driver will support dynamically turn on(off) 1P1 during
system suspend. 1P1 will be turn on depend on USB wakeup is enabled.

This reverts commit c6f69925f9d1f39ccc8c6fcc583122fcf03c105c.

Signed-off-by: make shi <b15407@freescale.com>
11 years agoENGR00225594 mmc: sdhci: remove calling enable clk in interrupt context
Ryan QIAN [Mon, 24 Sep 2012 07:25:36 +0000 (15:25 +0800)]
ENGR00225594 mmc: sdhci: remove calling enable clk in interrupt context

There's kernel warning without when insmod WiFi module as follows:

------------[ cut here ]------------
WARNING: at kernel/irq/handle.c:130 handle_irq_event_percpu+0x164/0x180()
irq 55 handler sdhci_irq+0x0/0x988 enabled interrupts
Modules linked in: ar6000
[<800471dc>] (unwind_backtrace+0x0/0xfc) from
[<80072fd0>] (warn_slowpath_common+0x4c/0x64)
[<80072fd0>] (warn_slowpath_common+0x4c/0x64) from
[<8007307c>] (warn_slowpath_fmt+0x30/0x40)
[<8007307c>] (warn_slowpath_fmt+0x30/0x40) from
[<800a9f54>] (handle_irq_event_percpu+0x164/0x180)
[<800a9f54>] (handle_irq_event_percpu+0x164/0x180) from
[<800a9fac>] (handle_irq_event+0x3c/0x5c)
[<800a9fac>] (handle_irq_event+0x3c/0x5c) from
[<800ac2a8>] (handle_fasteoi_irq+0x98/0x148)
[<800ac2a8>] (handle_fasteoi_irq+0x98/0x148) from
[<800a9a2c>] (generic_handle_irq+0x2c/0x38)
[<800a9a2c>] (generic_handle_irq+0x2c/0x38) from
[<80041f60>] (handle_IRQ+0x4c/0xb8)
[<80041f60>] (handle_IRQ+0x4c/0xb8) from
[<80040f8c>] (__irq_svc+0x4c/0xe8)
[<80040f8c>] (__irq_svc+0x4c/0xe8) from
[<800420e4>] (default_idle+0x24/0x28)
[<800420e4>] (default_idle+0x24/0x28) from
[<80042730>] (cpu_idle+0xbc/0xfc)
[<80042730>] (cpu_idle+0xbc/0xfc) from
[<80008904>] (start_kernel+0x248/0x288)
[<80008904>] (start_kernel+0x248/0x288) from
[<10008040>] (0x10008040)
---[ end trace 95ab51b95e0e8e5f ]---

it is caused by calling cancel_delayed_work_sync in sdhci_enable_clk, since
it will call spin_unlock_irq which will possibly enable irq in interrupt
context. if sdhci_enable_clk is called in interrupt context this warning
will be shown.

sdhci_enable_clk will be called in interrupt context in the following path:
sdhci_irq->mmc_signal_sdio_irq->mmc_signal_sdio_irq->sdhci_enable_sdio_irq->...
     |____________________interrupt context____________________________________

fix:
Remove calling sdhci_enable_clk in sdhci_enable_sdio_irq. For sdio cards, sdhci
clk will not be disabled, so it is save to remove it.

Signed-off-by: Ryan QIAN <b32804@freescale.com>
Acked-by: Dong Aisheng <b29396@freescale.com>
11 years agoENGR00225534-2 mmc: sdhci: using cancel_work_sync instread of cancel_delayed_work
Dong Aisheng [Fri, 21 Sep 2012 10:33:06 +0000 (18:33 +0800)]
ENGR00225534-2 mmc: sdhci: using cancel_work_sync instread of cancel_delayed_work

We recently met an rarely happened sychronization issue which can cause
mmc timeout during transfer as follows:
[ OK ]onfiguring network interfaces...
[ OK ]ctivating swap...
mmc0: Timeout waiting for hardware interrupt.
mmc0: Timeout waiting for hardware interrupt.
mmc0: Timeout waiting for hardware interrupt.
mmcblk0: error -110 sending status command, retrying
mmcblk0: error -110 sending stop command, original cmd response 0x900, card status 0xd00
mmcblk0: error -110 transferring data, sector 9981952, nr 16, cmd response 0x900, card status 0xc00
end_request: I/O error, dev mmcblk0, sector 9981952

The root cause is that host uses cancel_delayed_work to cancel the delayed
clock disable work to avoid clock to be disabled if a new cmd
transfer request happens.
However, the work callback may already be running during the excution
of cancel_delayed_work which can not be cancelled and causes
the clock probably still to be disabled during cmd transfer,
then cmd timeout happens.

Using cancel_work_sync instead to wait for the completion of clock disable
first then we can make sure the clock can not be disabled during the cmd
transfer.

BTW, although the original code checks if in interrupt context, however,
it's still not interrupt context safe due to the unsafe platform_clk_ctrl,
so it's ok to directly use cancel_work_sync here and sdhci_enable_clk is
simply not allowed to be called in irq context.

Acked-by: Ryan QIAN <b32804@freescale.com>
Signed-off-by: Dong Aisheng <b29396@freescale.com>
11 years agoENGR00225534-1 mmc: core: move ->request() call from atomic context
Adrian Hunter [Fri, 21 Sep 2012 10:28:42 +0000 (18:28 +0800)]
ENGR00225534-1 mmc: core: move ->request() call from atomic context

mmc_request_done() is sometimes called from interrupt or other atomic
context.  Mostly all mmc_request_done() does is complete(), however it
contains code to retry on error, which uses ->request().  As the error
path is certainly not performance critical, this may be moved to the
waiting function mmc_wait_for_req_done().

This allows ->request() to use runtime PM get_sync() and guarantee it
is never in an atomic context.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Acked-by: Ulf Hansson <ulf.hansson@stericsson.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Acked-by: Ryan QIAN <b32804@freescale.com>
Signed-off-by: Dong Aisheng <b29396@freescale.com>
11 years agoENGR00220342 mmc: sdhci:reform fix for dat1 glitch causes system panic
Jianzheng Zhou [Mon, 24 Sep 2012 01:50:59 +0000 (09:50 +0800)]
ENGR00220342 mmc: sdhci:reform fix for dat1 glitch causes system panic

refix this bug, sdhci_init should be done before request_irq to
clear those wild interrupts.
Former fix commit is 4e3a2b40a16b8d5065b680ab86df288515cdea9c.

Signed-off-by: Jianzheng Zhou <jianzheng.zhou@freescale.com>
11 years agoENGR00224476 pgc: fix display power gating causes PxP processing timeout
Robby Cai [Wed, 12 Sep 2012 06:32:34 +0000 (14:32 +0800)]
ENGR00224476 pgc: fix display power gating causes PxP processing timeout

The root-cause is PxP need the clock to be on for synchronous reset.
This patch turned on PXP axi clock, and EPDC/LCDIF pix clock, and
EPDC/LCDIF axi clock before power up. The driver codes can guarantee
the clock setting be restored to the one before suspend.

Signed-off-by: Robby Cai <R63905@freescale.com>
11 years agoENGR00223533 - FEC : fix net watchdog timeout due not refreshing TX timers
Fugang Duan [Tue, 18 Sep 2012 08:27:43 +0000 (16:27 +0800)]
ENGR00223533 - FEC : fix net watchdog timeout due not refreshing TX timers

When do overnight gpu stress test throught nfs, kernel dump as below:

------------[ cut here ]------------
WARNING: at net/sched/sch_generic.c:255 dev_watchdog+0x284/0x2a8()
NETDEV WATCHDOG: eth0 (fec): transmit queue 0 timed out
Modules linked in: galcore imx2_wd_test [last unloaded: galcore]
[<80045854>] (unwind_backtrace+0x0/0xf8) from [<80070c2c>]
(warn_slowpath_common+0x4c/0x64)
<snip>

In sometime, ethernet cannot recover after watchdog timeout which
results in nfs no responding and system hang on.

The patch fixes the way that ->trans_start is refreshed to avoid watchdog
timeout during ethernet status change (such as speed, duplex), and re-init
fec to recover ethernet after watchdog timeout.

Signed-off-by: Fugang Duan <B38611@freescale.com>
11 years agoENGR00224700 IPUv3 FB:Support RGB24/RGB32/ABGR32 pixel formats
Liu Ying [Sat, 15 Sep 2012 16:06:43 +0000 (00:06 +0800)]
ENGR00224700 IPUv3 FB:Support RGB24/RGB32/ABGR32 pixel formats

Users may call FBIOPUT_VSCREENINFO framebuffer ioctrl to set
a framebuffer's pixel format by specifying bits_per_pixel field
and red/green/blue/transp fields of struct fb_var_screeninfo.
Users may also know a framebuffer's pixel format at which it is
working by calling FBIOGET_VSCREENINFO framebuffer ioctrl, red/
green/blue/transp bitfields of struct fb_var_screeninfo tell the
exact offset and length of each color component in a pixel. This
patch supports RGB24/RGB32/ABGR32 pixel formats with these 2 ioctrls.
To change the default pixel format when initializeing a framebuffer
at the first time, users may add 'fbpix=' option in framebuffer
kernel command line option, for example, 'video=mxcfb0:fbpix=ABGR32'
makes fb0's default pixel format be ABGR32. 'bpp=' option cannot
overwrite 'fbpix=' option, for example, 'video=mxfb0:fbpix=RGB24,
bpp=16' still makes fb0 work at RGB24 by default. To be back
compatible, this patch doesn't change the default pixel formats
(BGR24/BGR32/RGB565) for each bits_per_pixel, if users don't
provide valid 'fbpix=' option.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
11 years agoENGR00224910 IPUv3 fb:Correct logic to check BGR24 for if= option
Liu Ying [Wed, 19 Sep 2012 02:26:25 +0000 (10:26 +0800)]
ENGR00224910 IPUv3 fb:Correct logic to check BGR24 for if= option

This patch corrects the logic to check BGR24 pixel format for
'if=' option in kernel bootup command line of framebuffer.
After applying the patch, users may use 'if=BGR24' to specify
the display data format to be BGR24. For example, 'video=mxcfb0:
dev=lcd,LCD_VIDEO_NAME,if=BGR24' makes IPUv3 output BGR24 data
to fb0's lcd panel.

Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
(cherry picked from commit 7ab85674f63eb1388e44a61e17891f0e283c409a)

11 years agoENGR00224696 mx6:Improve DSM power stability
Anson Huang [Tue, 18 Sep 2012 04:24:29 +0000 (12:24 +0800)]
ENGR00224696 mx6:Improve DSM power stability

1. If weak 2P5 is not enabled in DSM, need to enable pull down
function to make sure its voltage is 0V;

2. WB_COUNT need to set to a higher value instead of the reset
value to make sure it function normally.

Signed-off-by: Anson Huang <b20788@freescale.com>
11 years agoENGR00221594 Add Dithering algorism support into EPDC driver
Daiyu Ko [Wed, 19 Sep 2012 16:15:43 +0000 (11:15 -0500)]
ENGR00221594 Add Dithering algorism support into EPDC driver

Adding Atkinson's dithering alorism implementation into our EPDC
Driver with Y8->Y1 and Y8->Y4 supported. Two EPDC flags have been
added to support the features. EPDC_FLAG_USE_DITHERING_Y1 and
EPDC_FLAG_USE_DITHERING_Y4.

Signed-off-by: Daiyu Ko <dko@freescale.com>
11 years agoENGR00219882 wm8962: avoid the digital volume changed by user
Gary Zhang [Thu, 13 Sep 2012 09:33:41 +0000 (17:33 +0800)]
ENGR00219882 wm8962: avoid the digital volume changed by user

Wolfson suggests that digital volume should not be changed by user
to avoid affecting SNR and dynamic range, and user should adjust
analog volume.

Signed-off-by: Gary Zhang <b13634@freescale.com>
11 years agoENGR00224938 HDMI audio: clear HDMI dma done bit before start
Chen Liangjun [Wed, 19 Sep 2012 05:30:27 +0000 (13:30 +0800)]
ENGR00224938 HDMI audio: clear HDMI dma done bit before start

HDMI hardware fix: signal of HDMI DMA DONE is hard connected to SDMA
event line. SDMA event is triggered by edge. If the HDMI DMA done is
already 1 before start, there would be no SDMA event being trigged after
HDMI generates another HDMI DONE signal.

In this patch, clear HDMI DONE bit before start HDMI audio DMA.

Signed-off-by: Chen Liangjun <b36089@freescale.com>
11 years agoENGR00224513 V4L2 output: Avoid fb blank/unblank when open/close video device
Wayne Zou [Mon, 17 Sep 2012 07:48:15 +0000 (15:48 +0800)]
ENGR00224513 V4L2 output: Avoid fb blank/unblank when open/close video device

Initialize release variable to avoid fb blank/unblank
when open/close video device only.

Signed-off-by: Wayne Zou <b36644@freescale.com>
11 years agoENGR00224886 board-mx6q-sabreauto: remove camera control lines
Adrian Alonso [Thu, 13 Sep 2012 21:43:10 +0000 (16:43 -0500)]
ENGR00224886 board-mx6q-sabreauto: remove camera control lines

* Remove camera control lines, this gpio control lines are used by
  other periherals
* GPIO_4_5 correspond to RGMII_INT
* GPIO_3_24 correspond to CSI0_DAT5 parallel tv-in

Signed-off-by: Adrian Alonso <aalonso@freescale.com>
11 years agoENGR00224467 GPU driver debug version build failure fix
Loren Huang [Mon, 17 Sep 2012 09:31:28 +0000 (17:31 +0800)]
ENGR00224467 GPU driver debug version build failure fix

-Fix a value name typo in gc_hal_kernel_hardware.c

Signed-off-by: Loren Huang <b02279@freescale.com>
11 years agoENGR00224275-2 board-mx6q-sabreauto: remove unsupported ov5640 camera
Adrian Alonso [Thu, 13 Sep 2012 16:26:56 +0000 (11:26 -0500)]
ENGR00224275-2 board-mx6q-sabreauto: remove unsupported ov5640 camera

* Remove unsupported camera device ov5640, there is no
  hardware module available for sabreauto target board.
  No mechanical connector compability with existing capture
  sensor.

Signed-off-by: Adrian Alonso <aalonso@freescale.com>
11 years agoENGR00224275-1 board-mx6q-sabreauto: remove unsupported ov3640 camera
Adrian Alonso [Tue, 11 Sep 2012 23:25:42 +0000 (18:25 -0500)]
ENGR00224275-1 board-mx6q-sabreauto: remove unsupported ov3640 camera

* Remove unsupported camera device ov3640, there is no
  hardware module available for sabreauto target board.
* No mechanical connector compability with existing capture
  sensors.

Signed-off-by: Adrian Alonso <aalonso@freescale.com>
11 years agoperf tools: Fix build against newer glibc
Josh Boyer [Thu, 18 Aug 2011 11:37:21 +0000 (07:37 -0400)]
perf tools: Fix build against newer glibc

Upstream glibc commit 295e904 added a definition for __attribute_const__
to cdefs.h.  This causes the following error when building perf:

util/include/linux/compiler.h:8:0: error: "__attribute_const__"
redefined [-Werror] /usr/include/sys/cdefs.h:226:0: note: this is the
location of the previous definition

Wrap __attribute_const__ in #ifndef as we do for __always_inline.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20110818113720.GL2227@zod.bos.redhat.com
Signed-off-by: Josh Boyer <jwboyer@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
11 years agoENGR00224338 mx6:gpu:Make gpu as buildin driver by default
Loren Huang [Fri, 14 Sep 2012 08:08:20 +0000 (16:08 +0800)]
ENGR00224338 mx6:gpu:Make gpu as buildin driver by default

Make gpu as build-in driver to simplify gpu test and usage.

Signed-off-by: Loren Huang <b02279@freescale.com>
11 years agoENGR00224137 mx6 : restore VDD_PU in resume with the old value before suspend
Robin Gong [Thu, 13 Sep 2012 06:19:41 +0000 (14:19 +0800)]
ENGR00224137 mx6 : restore VDD_PU in resume with the old value before suspend

Before, VDD_PU regulator will be set 1.1V for ever, but now VDD_PU track with
VDD_ARM, so we can not restore with 1.1V when system resume back.We need save
the old value set when power off PU in suspend flow, and then restore it in
resume flow.
Signed-off-by: Robin Gong <B38343@freescale.com>
11 years agoENGR00224245 HDMI AUDIO: stop/start PCM while unplug,blank/plug, unblank
Chen Liangjun [Thu, 13 Sep 2012 13:28:11 +0000 (21:28 +0800)]
ENGR00224245 HDMI AUDIO: stop/start PCM while unplug,blank/plug, unblank

When unplug, blank happens, HDMI audio can't play properly. So in
driver, audio pcm would be disconnected when event above happens.
However, pulse audio can't process disconnect event properly and if an
blank or unplug event happens, HDMI sink would lost and can't be back
again.

In this patch, instead of disconnecting audio PCM stream, triggering
stop audio pcm while unplug and blank, triggering start again while plug
and unblank if the audio pcm is triggerd stop in the unplug/blank event.
Signed-off-by: Chen Liangjun <b36089@freescale.com>
11 years agoENGR00223349-4 gpmi: enable the EDO support for mx6q
Huang Shijie [Tue, 28 Aug 2012 05:53:49 +0000 (13:53 +0800)]
ENGR00223349-4 gpmi: enable the EDO support for mx6q

Enable the EDO mode for mx6q.

The following is the test result with the same nand chip (Micron MT29F32G08QAA)
in mode 4:

The test result without enable the EDO mode:
=================================================
mtd_speedtest: MTD device: 2
mtd_speedtest: MTD device size 209715200, eraseblock size 524288,
page size 4096, count of eraseblocks 400,
pages per eraseblock 128, OOB size 218
mtd_speedtest: scanned 400 eraseblocks, 6 are bad
mtd_speedtest: testing eraseblock write speed
mtd_speedtest: eraseblock write speed is 1945 KiB/s
mtd_speedtest: testing eraseblock read speed
mtd_speedtest: eraseblock read speed is 3384 KiB/s
mtd_speedtest: testing page write speed
mtd_speedtest: page write speed is 1841 KiB/s
mtd_speedtest: testing page read speed
mtd_speedtest: page read speed is 3136 KiB/s
mtd_speedtest: testing 2 page write speed
mtd_speedtest: 2 page write speed is 1853 KiB/s
mtd_speedtest: testing 2 page read speed
mtd_speedtest: 2 page read speed is 3164 KiB/s
mtd_speedtest: Testing erase speed
mtd_speedtest: erase speed is 145441 KiB/s
mtd_speedtest: Testing 2x multi-block erase speed
mtd_speedtest: 2x multi-block erase speed is 146711 KiB/s
mtd_speedtest: Testing 4x multi-block erase speed
mtd_speedtest: 4x multi-block erase speed is 147139 KiB/s
mtd_speedtest: Testing 8x multi-block erase speed
mtd_speedtest: 8x multi-block erase speed is 147786 KiB/s
mtd_speedtest: Testing 16x multi-block erase speed
mtd_speedtest: 16x multi-block erase speed is 147569 KiB/s
mtd_speedtest: Testing 32x multi-block erase speed
mtd_speedtest: 32x multi-block erase speed is 147677 KiB/s
mtd_speedtest: Testing 64x multi-block erase speed
mtd_speedtest: 64x multi-block erase speed is 147677 KiB/s
mtd_speedtest: finished
=================================================

The test result enable the EDO mode:
=================================================
mtd_speedtest: MTD device: 2
mtd_speedtest: MTD device size 209715200, eraseblock size 524288,
page size 4096, count of eraseblocks 400,
pages per eraseblock 128, OOB size 218
mtd_speedtest: scanned 400 eraseblocks, 6 are bad
mtd_speedtest: testing eraseblock write speed
mtd_speedtest: eraseblock write speed is 3733 KiB/s
mtd_speedtest: testing eraseblock read speed
mtd_speedtest: eraseblock read speed is 20413 KiB/s
mtd_speedtest: testing page write speed
mtd_speedtest: page write speed is 3603 KiB/s
mtd_speedtest: testing page read speed
mtd_speedtest: page read speed is 18966 KiB/s
mtd_speedtest: testing 2 page write speed
mtd_speedtest: 2 page write speed is 3668 KiB/s
mtd_speedtest: testing 2 page read speed
mtd_speedtest: 2 page read speed is 19686 KiB/s
mtd_speedtest: Testing erase speed
mtd_speedtest: erase speed is 146604 KiB/s
mtd_speedtest: Testing 2x multi-block erase speed
mtd_speedtest: 2x multi-block erase speed is 147354 KiB/s
mtd_speedtest: Testing 4x multi-block erase speed
mtd_speedtest: 4x multi-block erase speed is 147677 KiB/s
mtd_speedtest: Testing 8x multi-block erase speed
mtd_speedtest: 8x multi-block erase speed is 148002 KiB/s
mtd_speedtest: Testing 16x multi-block erase speed
mtd_speedtest: 16x multi-block erase speed is 147894 KiB/s
mtd_speedtest: Testing 32x multi-block erase speed
mtd_speedtest: 32x multi-block erase speed is 148329 KiB/s
mtd_speedtest: Testing 64x multi-block erase speed
mtd_speedtest: 64x multi-block erase speed is 148220 KiB/s
mtd_speedtest: finished
=================================================

We can see that there is 6 times performance improvement for reading
when we enable the EDO mode.

Signed-off-by: Huang Shijie <b32955@freescale.com>
11 years agoENGR00223349-3 gpmi: add a new field for HW_GPMI_CTRL1
Huang Shijie [Fri, 7 Sep 2012 06:37:47 +0000 (14:37 +0800)]
ENGR00223349-3 gpmi: add a new field for HW_GPMI_CTRL1

add the WRN_DLY_SEL field for HW_GPMI_CTRL1.
This field is used as delay for gpmi write strobe.

Signed-off-by: Huang Shijie <b32955@freescale.com>
11 years agoENGR00223349-2 gpmi: do not get the clock frequency in gpmi_begin()
Huang Shijie [Fri, 7 Sep 2012 05:00:58 +0000 (13:00 +0800)]
ENGR00223349-2 gpmi: do not get the clock frequency in gpmi_begin()

The current code will gets the clock frequency which is used by
gpmi_nfc_compute_hardware_timing(). It makes the code a little mess.

So move the `get clock frequency` code to the
gpmi_nfc_compute_hardware_timing() itself. This makes the code tidy
and clean.

Signed-off-by: Huang Shijie <b32955@freescale.com>
11 years agoENGR00223349-1 gpmi: add a new field for HW_GPMI_TIMING1
Huang Shijie [Fri, 7 Sep 2012 02:51:01 +0000 (10:51 +0800)]
ENGR00223349-1 gpmi: add a new field for HW_GPMI_TIMING1

The gpmi_nfc_compute_hardware_timing{} should contains all the
fields setting for gpmi timing registers. It already contains the fields
for HW_GPMI_TIMING0 and HW_GPMI_CTRL1.

So it is better to add a new field setting for HW_GPMI_TIMING1 in
this data structure. This makes the code more clear in logic.

This patch also changs some comments to make the code more readable.

Signed-off-by: Huang Shijie <b32955@freescale.com>
11 years agoENGR00223816 HDMI AUDIO: fix kernel panic cause by accessing unavailable memory
Chen Liangjun [Wed, 12 Sep 2012 10:34:28 +0000 (18:34 +0800)]
ENGR00223816 HDMI AUDIO: fix kernel panic cause by accessing unavailable memory

HDMI audio driver is responsible for add IEC header into audio sample.
In HDMI audio driver, a variable named rtd->appl_bytes is maintained to
stand for how many audio sample have already processed. appl_bytes
stands for how many audio sample the user space have already feed into
kernel driver. So we use the connt = appl_bytes - rtd->appl_bytes to
decide how many data need to be processed. And the processed data would
be write into an preallocated buffer called hw_buf in driver.

When doing seek operation, the appl_bytes changes in an wide range. So
it is possible that the count value is far larger than the size of
hw_buf and the memory access un-existed address error would happens.

In this patch, Add check operation for count to avoid kernel panic.

Kernel panic log:
seeking: 0:00:18.000000000/0:03:0Unable to handle kernel paging request
at virtual address ffdf0000
pgd = 80004000
[ffdf0000] *pgd=71e35811, *pte=00000000, *ppte=00000000
Internal error: Oops: 7 [#1] PREEMPT SMP
Modules linked in: vivante drm galcore
CPU: 0    Not tainted  (3.0.35-2014-g7a9337b #1)
PC is at hdmi_dma_mmap_copy+0x134/0x190
LR is at hdmi_dma_mmap_copy+0x5c/0x190
pc : [<803e1e4c>]    lr : [<803e1d74>]    psr: 800f0193
sp : 80a61e98  ip : ffdf0000  fp : ffdeffc0
r10: 00000055  r9 : ffdeff80  r8 : 0029b450
r7 : 00000060  r6 : ffdf0200  r5 : 00000240  r4 : 00000120
r3 : 00000000  r2 : ffdf0000  r1 : 00000000  r0 : 00000090
Flags: Nzcv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
Control: 10c53c7d  Table: 6d28804a  DAC: 00000015
Process swapper (pid: 0, stack limit = 0x80a602f0)
Stack: (0x80a61e98 to 0x80a62000)
1e80:                                                       00000002
000037f1
1ea0: 00006000 e1d3a7e4 a00f0193 e1d3a780 e1dcab00 00a83100 00a83100
00006000
1ec0: 00006000 803e24f4 413187b9 00000000 0073001a e1dd1e00 00000001
00000001
1ee0: 00000080 00000093 80ac7070 80a66a80 00000001 800a5ca8 8c80e568
efb3e7b9
1f00: 00000055 80a66a80 80a66acc e1ea9bc0 00000093 00000000 80a60000
00000000
1f20: 00000000 800a5e14 80a66a80 80a66acc 0000107f 800a8198 80a71cc0
80038c00
1f40: 80a60000 800a5610 000001f0 80040830 ffffffff f2a00100 00000093
00000002
1f60: 00000001 8003f9cc 80ac5f60 800f0093 00000001 00000000 80a60000
80abeb64
1f80: 804e1a54 80a74e7c 1000406a 412fc09a 00000000 00000000 00000000
80a61fb0
1fa0: 8004d52c 80040ac4 400f0013 ffffffff 80040aa0 80040cbc 00000001
80a71b3c
1fc0: 80abeac0 8002e3c4 8c80b140 80008868 800082f8 00000000 00000000
8002e3c4
1fe0: 00000000 10c53c7d 80a71a6c 8002e3c0 80a74e74 10008040 00000000
00000000
[<803e1e4c>] (hdmi_dma_mmap_copy+0x134/0x190) from [<803e24f4>]
(hdmi_dma_isr+0x17c/0x1a0)
[<803e24f4>] (hdmi_dma_isr+0x17c/0x1a0) from [<800a5ca8>]
(handle_irq_event_percpu+0x50/0x180)
[<800a5ca8>] (handle_irq_event_percpu+0x50/0x180) from [<800a5e14>]
(handle_irq_event+0x3c/0x5c)
[<800a5e14>] (handle_irq_event+0x3c/0x5c) from [<800a8198>]
(handle_fasteoi_irq+0xbc/0x154)
[<800a8198>] (handle_fasteoi_irq+0xbc/0x154) from [<800a5610>]
(generic_handle_irq+0x28/0x3c)
[<800a5610>] (generic_handle_irq+0x28/0x3c) from [<80040830>]
(handle_IRQ+0x4c/0xac)
[<80040830>] (handle_IRQ+0x4c/0xac) from [<8003f9cc>]
(__irq_svc+0x4c/0xe8)
[<8003f9cc>] (__irq_svc+0x4c/0xe8) from [<80040ac4>]
(default_idle+0x24/0x28)
[<80040ac4>] (default_idle+0x24/0x28) from [<80040cbc>]
(cpu_idle+0xbc/0xfc)
[<80040cbc>] (cpu_idle+0xbc/0xfc) from [<80008868>]
(start_kernel+0x248/0x288)
[<80008868>] (start_kernel+0x248/0x288) from [<10008040>] (0x10008040)
Code: c1a0c009 c08b6005 c1a0200b da00000a (e0d230b2)

Signed-off-by: Chen Liangjun <b36089@freescale.com>
11 years agoENGR00224112 wm8962: correct lrclk mismatch
Gary Zhang [Wed, 12 Sep 2012 12:29:01 +0000 (20:29 +0800)]
ENGR00224112 wm8962: correct lrclk mismatch

lrclk trigger settings between ssi and wm8962 are different,
which causes L/R channel swap.

Signed-off-by: Gary Zhang <b13634@freescale.com>
11 years agoENGR00222253 Merge vivante 4.6.9_p7 kernel part code
Loren Huang [Sun, 9 Sep 2012 16:00:13 +0000 (00:00 +0800)]
ENGR00222253 Merge vivante 4.6.9_p7 kernel part code

Merge vivante 4.6.9 p7 kernel part code.

Signed-off-by: Loren Huang <b02279@freescale.com>
Acked-by: Lily Zhang
11 years agoENGR00223964: CAAM: enable ahash feature in config
Terry Lv [Wed, 12 Sep 2012 06:06:15 +0000 (14:06 +0800)]
ENGR00223964: CAAM: enable ahash feature in config

Enable ahash feature in mx6q config.

Signed-off-by: Terry Lv <r65388@freescale.com>
11 years agoENGR00223679-3 battery: Fix codes formatting
Rong Dian [Wed, 12 Sep 2012 02:12:01 +0000 (10:12 +0800)]
ENGR00223679-3 battery: Fix codes formatting

Just fix codes formatting,there's no functionality
change in codes

Signed-off-by: Rong Dian <b38775@freescale.com>
11 years agoENGR00223679-2 battery: Add battery driver support for mx6sl_evk
Rong Dian [Mon, 10 Sep 2012 10:07:21 +0000 (18:07 +0800)]
ENGR00223679-2 battery: Add battery driver support for mx6sl_evk

Add battery driver support for mx6sl_evk.

Signed-off-by: Rong Dian <b38775@freescale.com>
11 years agoENGR00223679-1 mx6sl_evk: Add pin config for MAX8903
Rong Dian [Mon, 10 Sep 2012 09:55:11 +0000 (17:55 +0800)]
ENGR00223679-1 mx6sl_evk: Add pin config for MAX8903

Configure PINMUX for max8903 driver on mx6sl_evk board.

Signed-off-by: Rong Dian <b38775@freescale.com>
11 years agoENGR00223911 wm8962: correct typo
Gary Zhang [Wed, 12 Sep 2012 02:03:11 +0000 (10:03 +0800)]
ENGR00223911 wm8962: correct typo

correct typo for wm8962 description in Kconfig

Signed-off-by: Gary Zhang <b13634@freescale.com>
11 years agoENGR00223450 [MX6]Fix secondary cores BogoMIPs error
Anson Huang [Mon, 10 Sep 2012 18:41:54 +0000 (02:41 +0800)]
ENGR00223450 [MX6]Fix secondary cores BogoMIPs error

After doing some suspend/resuem test, secondary cores BogoMIPs
will be wrong, the root cause is that when cpufreq is changed,
we only update the online cpus' loops_per_jiffy, and when secondary
cores back to online, we skip the loops_per_jiffy calibration. During
suspend/resume, the cpufreq can be changed during disabling/enabling
secondary cores, which will make secondary cores loops_per_jiffy
wrong, so here we need to update all possible cpus' loops_per_jiffy
when cpufreq is changed.

Signed-off-by: Anson Huang <b20788@freescale.com>
11 years agoENGR00221671 - i2c :imx : fix some i2c devices can not suspend
Fugang Duan [Mon, 10 Sep 2012 05:05:26 +0000 (13:05 +0800)]
ENGR00221671 - i2c :imx : fix some i2c devices can not suspend

i2c device (isl29023) can not suspend once during hdmi
audio suspend/resume test, and print log:

pm_op(): i2c_device_pm_suspend+0x0/0x38 returns -4 PM: Device
2-0044 failed to suspend: error -4 PM: Some devices failed to
suspend PM: resume of devices complete after 40.936 msecs
Restarting tasks ... done.

Because suspend function in isl29023 driver requires i2c bus
to write isl29023 device. I2C apdater driver process any signal
as exception during waiting the bus idle, so once user space sent
out signal during suspend, I2C device cannot request bus.

Using "fatal_signal_pending()" instead of "signal_pending()"
to avoid the waiting of bus idle to be terminated by general
signals except SIGKILL. After the change, i2c adapter can be
terminated by kill signal from user space with "CTRL+C" or
kill command operation.

Signed-off-by: Fugang Duan <B38611@freescale.com>
11 years agoENGR00223365 [MX6DL]Update cpu setpoint
Anson Huang [Fri, 7 Sep 2012 14:14:11 +0000 (22:14 +0800)]
ENGR00223365 [MX6DL]Update cpu setpoint

Update CPU setpoint according to latest datasheet.

Signed-off-by: Anson Huang <b20788@freescale.com>
11 years agoENGR00221895 IPU Device: release IPU device resources when process is killed
Wayne Zou [Wed, 29 Aug 2012 09:34:51 +0000 (17:34 +0800)]
ENGR00221895 IPU Device: release IPU device resources when process is killed

Release IPU device resources when process is killed

Signed-off-by: Wayne Zou <b36644@freescale.com>
11 years agoENGR00223249 : fix Ethernet performance downgrade issue.
Fugang Duan [Thu, 6 Sep 2012 06:13:00 +0000 (14:13 +0800)]
ENGR00223249 : fix Ethernet performance downgrade issue.

Ethernet performance is downgraded when wait mode on in
100Mbps mode.
wait mode off:
100Mbps mode: tx bandwidth is 94Mbps
      rx bandwidth is 94Mbps
wait mode on:
100Mbps mode: tx bandwidth is 30Mbps
      rx bandwidth is 94Mbps
After apply the patch:
wait mode on:
100Mbps mode: tx bandwidth is 94Mbps
      rx bandwidth is 94Mbps

Wait mode on cause enet interrupt has long latency, which
results in BD entries are full and stop tx queue, so cpus
have more chance to enter wait mode.

Incresing TX BD entries can properly accommodate the blance
between BD request before tx packets and BD release after tx
completion in interrupt process.

Signed-off-by: Fugang Duan <B38611@freescale.com>
11 years agomtd: gpmi-nand: fix read page when reading to vmalloced area
Sascha Hauer [Tue, 26 Jun 2012 15:26:16 +0000 (17:26 +0200)]
mtd: gpmi-nand: fix read page when reading to vmalloced area

The gpmi-nand driver uses virt_addr_valid() to check whether a buffer
is suitable for dma. If it's not, a driver allocated buffer is used
instead. Then after a page read the driver allocated buffer must be
copied to the user supplied buffer. This does not happen since commit
7725cc85932bd02dd12c23108e0ef748c551ccba.

This patch fixes the issue. The bug is encountered with UBI which uses a
vmalloced buffer for the volume table.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Tested-by: snijsure@grid-net.com
Acked-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Huang Shijie <b32955@freescale.com>
11 years agomtd: gpmi-nand: utilize oob_requested parameter
Brian Norris [Wed, 2 May 2012 17:15:02 +0000 (10:15 -0700)]
mtd: gpmi-nand: utilize oob_requested parameter

Don't read OOB if the caller didn't request it.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Acked-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Huang Shijie <b32955@freescale.com>
11 years agomtd: nand: pass proper 'oob_required' parameter
Huang Shijie [Wed, 5 Sep 2012 08:47:02 +0000 (16:47 +0800)]
mtd: nand: pass proper 'oob_required' parameter

We now have an interface for notifying the nand_ecc_ctrl functions when OOB
data must be returned to the upper layers and when it may be left untouched.
This patch fills in the 'oob_required' parameter properly from
nand_do_{read,write}_ops. When utilized properly in the lower layers, this
parameter can improve performance and/or reduce complexity for NAND HW and SW
that can simply avoid transferring the OOB data.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Acked-by: Jiandong Zheng <jdzheng@broadcom.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
11 years agomtd: nand: add 'oob_required' argument to NAND {read,write}_page interfaces
Huang Shijie [Wed, 5 Sep 2012 08:30:53 +0000 (16:30 +0800)]
mtd: nand: add 'oob_required' argument to NAND {read,write}_page interfaces

New NAND controllers can perform read/write via HW engines which don't expose
OOB data in their DMA mode. To reflect this, we should rework the nand_chip /
nand_ecc_ctrl interfaces that assume that drivers will always read/write OOB
data in the nand_chip.oob_poi buffer. A better interface includes a boolean
argument that explicitly tells the callee when OOB data is requested by the
calling layer (for reading/writing to/from nand_chip.oob_poi).

This patch adds the 'oob_required' parameter to each relevant {read,write}_page
interface; all 'oob_required' parameters are left unused for now. The next
patch will set the parameter properly in the nand_base.c callers, and follow-up
patches will make use of 'oob_required' in some of the callee functions.

Note that currently, there is no harm in ignoring the 'oob_required' parameter
and *always* utilizing nand_chip.oob_poi, but there can be
performance/complexity/design benefits from avoiding filling oob_poi in the
common case. I will try to implement this for some drivers which can be ported
easily.

Note: I couldn't compile-test all of these easily, as some had ARCH
dependencies.

[Huang Shijie: I remove the unused code for the other drivers.]

[dwmw2: Merge later 1/0 vs. true/false cleanup]
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Acked-by: Jiandong Zheng <jdzheng@broadcom.com>
Acked-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Huang Shijie <b32955@freescale.com>
11 years agoENGR00223344 [Thermal]Fix clk enable flow bug
Anson Huang [Fri, 7 Sep 2012 01:07:56 +0000 (09:07 +0800)]
ENGR00223344 [Thermal]Fix clk enable flow bug

We should make sure clk_enable is called after clk_get.

Signed-off-by: Anson Huang <b20788@freescale.com>
11 years agoENGR00223236-2 [MX6]Change the temperature range to -40C ~ 125C
Anson Huang [Thu, 6 Sep 2012 07:38:07 +0000 (15:38 +0800)]
ENGR00223236-2 [MX6]Change the temperature range to -40C ~ 125C

Previous temperature range is -25C ~ 125C, according to latest
datasheet, change it to -40C to 125C.

Signed-off-by: Anson Huang <b20788@freescale.com>
11 years agoENGR00222954 Enable PLL3 in GPU power up flow
Loren Huang [Thu, 6 Sep 2012 07:37:37 +0000 (15:37 +0800)]
ENGR00222954 Enable PLL3 in GPU power up flow

If GPU2D used PLL3 as root, we need enable PLL
during GPU power up flow so that we can power up
GPU2D properly.
Till now, this issue can only be duplicated on
Android.

Signed-off-by: Loren Huang <b02279@freescale.com>
11 years agoENGR00223103 ESAI ASRC: clean build warning
Chen Liangjun [Wed, 5 Sep 2012 08:45:18 +0000 (16:45 +0800)]
ENGR00223103 ESAI ASRC: clean build warning

BUILD WARNING:
WARNING: arch/arm/mach-mx6/built-in.o(.data+0x7e44): Section mismatch in
reference from the variable sab_audio_data to the (unknown reference)
.init.rodata:(unknown) The variable sab_audio_data references the
(unknown reference) __initconst (unknown) If the reference is valid then
annotate the variable with __init* or __refdata (see linux/init.h) or
name the variable:
*_template, *_timer, *_sht, *_ops, *_probe, *_probe_one, *_console

In this patch, remove esai_p2p struct with init attribute.

Signed-off-by: Chen Liangjun <b36089@freescale.com>
11 years agoENGR00223056 Fix HDMI build warning
Sandor Yu [Wed, 5 Sep 2012 05:26:50 +0000 (13:26 +0800)]
ENGR00223056 Fix HDMI build warning

Fix HDMI build warning.
drivers/video/mxc_hdmi.c: In function 'mxc_hdmi_set_mode':
drivers/video/mxc_hdmi.c:1659: warning: assignment discards
 qualifiers from pointer target type
drivers/video/mxc_hdmi.c: At top level:
driver/video/mxc_hdmi.c:1398: warning: 'mxc_hdmi_enable_pins'
 defined but not used

Remove unused function mxc_hdmi_enable_pins() and mxc_hdmi_disable_pins()
from code. Fix defined but unused function build warning.
Added pointer conversion from const poniter to non-const pointer.

Signed-off-by: Sandor Yu <R01008@freescale.com>
11 years agoENGR00222078 power battery:fix charger attach detect missing after resume
Rong Dian [Mon, 3 Sep 2012 10:37:43 +0000 (18:37 +0800)]
ENGR00222078 power battery:fix charger attach detect missing after resume

1.config gpio dok for AC charger as wake up irq, config gpio uok
for USB charger as wake up irq.
2.add AC/USB charger detect in resume,fix charger detect status update
missing after attach AC/USB charger and resume system

Signed-off-by: Rong Dian <b38775@freescale.com>
11 years agoENGR00223097 sabresd pfuze:increase VDDARM_IN&VDDSOC_IN to 1.475V for 1.2G
Robin Gong [Wed, 5 Sep 2012 08:17:32 +0000 (16:17 +0800)]
ENGR00223097 sabresd pfuze:increase VDDARM_IN&VDDSOC_IN to 1.475V for 1.2G

consider ripple, IR drop and pfuze tolerance, we need incrase VDDARM_IN and
VDDSOC_IN to 1.475V.

Signed-off-by: Robin Gong <B38343@freescale.com>