Hyongbin Kim [Thu, 27 Jun 2013 07:07:38 +0000 (16:07 +0900)]
power: max17048: Add export soc get fuction
Remove forced set 100% SOC code when charging done.
Sometimes, charging termination is done under 98% SOC,
when tegratab spend whole power of charger and charging current
drop under ITERM.
Charger driver will check SOC and control charging.
Hyongbin Kim [Wed, 19 Jun 2013 08:32:59 +0000 (17:32 +0900)]
power: max17048: Clear repeated voltage low alert
When battery voltage is under ALRT.VALRTMIN, VL alert occurs.
But irq occur too many times in a row, VL alert need to clear.
And set again when SOC is changed and value is above 1%.
Hyongbin Kim [Thu, 10 Oct 2013 03:05:54 +0000 (12:05 +0900)]
power: bq2419x: Set DCP charging current to 2A.
Tegratab charger is 2A-charger. And BQ24192 has DPM fuction,
it prevent charger voltage falls below input voltage limit
by reducing charging current even with lower-2A-charger is connected.
There is a race condition between hdcp upstream and hdmi disable
during suspend/resume stress with hdmi plugged-in. WAR by making
host1x/hdmi clk enabled and hdmi clk out of reset.
Min-wuk Lee [Wed, 21 Aug 2013 05:32:49 +0000 (14:32 +0900)]
tegra: video: lgd-wxga-7 avdd power sequence fix
This change is fixing LGD WXGA 7 AVDD 3.3V power sequence. This
power rail on needs to be >1000ms after it is off.
Add 1000ms delay after this rail is turned off.
Min-wuk Lee [Thu, 25 Jul 2013 11:47:38 +0000 (20:47 +0900)]
tegra: video: bug fix for sd_brightness usage
sd_brightness should be declared to extern variable since this symbol
is exported and used across kernel source, not only board file,
but also driver side.
Min-wuk Lee [Fri, 19 Jul 2013 11:14:23 +0000 (20:14 +0900)]
ARM: tegra: tegratab: PWFM and pwm based backlight device with DT
For PWFM and pwm based backlight device, platform device register
are not needed in board file since these two devices are registered
by device tree for Tegratab.
Min-wuk Lee [Fri, 19 Jul 2013 10:21:02 +0000 (19:21 +0900)]
video: backlight: pwm_bl: device tree support for pwm_bl
Add device tree support for pwm based backlight driver.
Values for platform data required by pwm backlight driver
are parsed from device tree properties, directly.
For function pointers in platform data, separate devdata
generating driver is prepared. In order to construct functions,
it also parses values from different device tree node.
Backlight output array to make linear backlight response are
parsed in separate devdata generating driver.
Min-wuk Lee [Tue, 2 Jul 2013 07:49:24 +0000 (16:49 +0900)]
ARM: dts: tegratab: pclk change
Before LP0 entry, pclk in tegra_dc_mode struct is referred to.
After LP0 exit, pclk is calculated with dc timing.
Because pclk is set differently, required emc bandwidth is
calculated differently as well beween two cases.
<Without this change>
Before LP0 entry
disp1.emc $ on 1 102000000 (17751000+)
Min-wuk Lee [Mon, 20 May 2013 08:32:03 +0000 (17:32 +0900)]
arm: tegra: tegratab: display controller and panel with DT
Instead of Disp0 and Disp1 platform data delivery by
platform_device_register from the board file, this change
will parse these data from device tree of display controller
node and panel node in tegratab dts.
Separate lcd device data generating driver with device tree
info is prepared.
Min-wuk Lee [Thu, 13 Jun 2013 05:03:53 +0000 (14:03 +0900)]
arm: tegra: tegratab: disable prism in kernel side
Set prism enable node to 0 in kernel side. Brightness change
and color saturation can be visible during kernel dc probe
depending on screen image since splash screen drawn in bootloader
didn't enable prism. Prism on / off will be handled
by power mode selection in user land.
Change-Id: I50a08845d4232e2d462c25bf0f7ae91eb166a141 Signed-off-by: Xiaohui Tao <xtao@nvidia.com>
Reviewed-on: http://git-master/r/280094
(cherry picked from commit 4323e3b03238aafbf52831ae04a0df6e15a46f3f)
Reviewed-on: http://git-master/r/288765
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Change-Id: I4a2917ae2070a4a6500d7e03833218534c117250 Signed-off-by: David Jung <djung@nvidia.com>
Reviewed-on: http://git-master/r/279074
(cherry picked from commit 931518312b66555c862137dfbaaebc8e4265cafe)
Reviewed-on: http://git-master/r/288764 Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Change-Id: Ic6a267e3db268453cb3454e91216a8f97b135ce7 Signed-off-by: David Jung <djung@nvidia.com>
Reviewed-on: http://git-master/r/277469
(cherry picked from commit 5ea5ab3e0910cf77d24a38f520bb03616afb11be)
Reviewed-on: http://git-master/r/288763 Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Change-Id: Icf89f4e3285ba35f92c78552793b9bf5d9f78355 Signed-off-by: Xiaohui Tao <xtao@nvidia.com>
Reviewed-on: http://git-master/r/264601
(cherry picked from commit 06eef70f3568f781da609570d0dfb39c3a869439)
Reviewed-on: http://git-master/r/288762 Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Problems:
If shutdown or remove is called when suspended, it could be stuck in
infinite loop.
Changes:
1. Check kthread_should_stop() in several loops should be helpful to
avoid infinite loop problem.
2. Add __pm_stay_awake() in resume() and __pm_relax() in suspend()
should be helpful to avoid messaging error touch fusion driver and touch
fusion daemon by user space freezing between suspend and resume.
3. shutdown()/remove() stops processing_thread() before killing daemon.
Bo Kim [Wed, 11 Sep 2013 06:01:02 +0000 (15:01 +0900)]
tegratab: use touch_fusion for reversed panel
We don't support DVT1(non reversed panel) device. So we don't need to
support touch_fusion and touch_fusion_rd together. We will use
touch_fusion for reversed panel.
Change-Id: I576f926463c69e1a033b763e5e5c2207e373baa6 Signed-off-by: Bo Kim <bok@nvidia.com>
Reviewed-on: http://git-master/r/272926
(cherry picked from commit 0298d60e90c617b9e2b130e6ce57b478cc5052d1)
Reviewed-on: http://git-master/r/288760 Reviewed-by: Gabby Lee <galee@nvidia.com> Tested-by: Gabby Lee <galee@nvidia.com>
Robert Collins [Sat, 7 Sep 2013 03:27:37 +0000 (20:27 -0700)]
input: touch: maxim_sti: Update to version 1.4.1, v25
Description:
version 1.2.0 TegraTab bringup v25 DVT3+;
Changes:
* Eraser starts as being reported as pen and then switches to eraser without an intervening lift of the pen.
* Touch panel not responsive after standby resume
* Stylus events on virtual keyboard are missing with v24 build.
Change-Id: I4c01e13165074f5c0cbd5a71cf961956a66c07b1 Signed-off-by: Robert Collins <rcollins@nvidia.com>
Reviewed-on: http://git-master/r/271801
(cherry picked from commit e803c70f49318301e41120e140cf466ed8faf025)
Reviewed-on: http://git-master/r/288759
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Bo Kim [Fri, 30 Aug 2013 08:41:04 +0000 (17:41 +0900)]
input: touch: maxim_sti: Shutdown safely
Current shutdown implementation seems to be unsafe. Sometimes it can be
locked up due to interrupt from the chip but the interrupt cannot be
processed and cleared.
Robert Collins [Thu, 15 Aug 2013 16:44:35 +0000 (09:44 -0700)]
input: touch: maxim_sti: Update to version 1.4.1, v22
Changes:
- Water Rejection: Multiple Stylus cursors are getting
displayed when the water drops are there in screen
- Water Rejection: Touch screen is not responding to
stylus or manual touch in water droplet area
Change-Id: If5b403995ca2eeef0106b61553616fc0455351ff Signed-off-by: Robert Collins <rcollins@nvidia.com>
Reviewed-on: http://git-master/r/261949
(cherry picked from commit 3e62d1ff63ef5bfd4cceedfa6cfc0906f9809116)
Reviewed-on: http://git-master/r/288756
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Change-Id: I731da02b5912723b41da553eca19400885e10976 Signed-off-by: Robert Collins <rcollins@nvidia.com>
Reviewed-on: http://git-master/r/260856
(cherry picked from commit ae1911fa1caddd9d1daadffafd53bdf3028a9a97)
Reviewed-on: http://git-master/r/288755
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Youngjin Kim [Thu, 8 Aug 2013 01:55:43 +0000 (10:55 +0900)]
maxim_sti: stylus and finger mutual exclusion
Android frameworks cannot handle input events properly when
there are two or more touch input devices if input events
are coming from two or more devices at the same time.
As a workaround, make stylus and finger events be mutually
exclusive when splitting input devices into two.
Also correct some #if usage and make it report lift event
for stylus and finger respectively when splitting input
devices.
Change-Id: I0dd5da29844433b79db2a00b352ffb50fe4e8a09 Signed-off-by: Youngjin Kim <nkim@nvidia.com>
Reviewed-on: http://git-master/r/259409
(cherry picked from commit eabea86209c26957c7bfb5478806025a0e1db96b)
Reviewed-on: http://git-master/r/288754 Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Robert Collins [Mon, 5 Aug 2013 18:12:32 +0000 (11:12 -0700)]
input: touch: maxim_sti: Update to version 1.3.1, v19
Changes:
- Fixes bug 1322007 We are seeing a halo of no touch/pen activity
around a thumb on the screen.
- Fixes bug 1327520 Touch strokes are missing at intermittent
soft keys on virtual keyboard soft key press.
- Improves bug 1327474 Drawing with finger suddenly starts to
act as eraser if user changes the finger while drawing.
- Improves bug 1296212 Erasing angle is not consistent using
stylus, sometimes draws during erase.
Change-Id: Ic5c909e9d3aa18e8dc5a3fb8fec06ab8ffedc7d6 Signed-off-by: Robert Collins <rcollins@nvidia.com>
Reviewed-on: http://git-master/r/258244
(cherry picked from commit 098cbf25ef437fa32950d81271a752bd8e23b07a)
Reviewed-on: http://git-master/r/288753
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Bo Kim [Thu, 1 Aug 2013 05:49:36 +0000 (14:49 +0900)]
input: touch: maxim_sti: enable device splitting
Many android apps cannot get the touch events correctly
due to they are looking for input device which has only
touchscreen source type instead of using bitmask.
This is to make the apps can get the touch events
correctly without fixing the apps by splitting input
devices for finger touch and stylus.
Change-Id: I3e9a16cc15351c48f0c21cc2cc2065e846237daa Signed-off-by: Bo Kim <bok@nvidia.com>
Reviewed-on: http://git-master/r/252842
(cherry-picked from commit c2a5cdb8a56de8498a8a76068932840d5508eb83)
Reviewed-on: http://git-master/r/288751
GVS: Gerrit_Virtual_Submit Reviewed-by: Gabby Lee <galee@nvidia.com>
Robert Collins [Thu, 18 Jul 2013 04:48:21 +0000 (21:48 -0700)]
input: touch: maxim_sti: Update to version 1.2.0, v16
Description:
Reduce number of SPI interrupts while touch is present. Reduces
from 5000 interrupts per second to 800-1200 interrupts per
second when a touch is present.
Change-Id: I1d88a062d0538526fa24374c5d38d52813857a33 Signed-off-by: Robert Collins <rcollins@nvidia.com>
Reviewed-on: http://git-master/r/250534
(cherry picked from commit 62f6c47cbb70746b5e365fb69a4a7c5b725710db)
Reviewed-on: http://git-master/r/288750
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Change-Id: I78ca8f6119aa948d7d9a57a912bc9946fccba0c6 Signed-off-by: Robert Collins <rcollins@nvidia.com>
Reviewed-on: http://git-master/r/249532
(cherry picked from commit afead9e9e3fd56bc5e94a6fbf4967a1c3b243b1d)
Reviewed-on: http://git-master/r/288749
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Yogish Kulkarni [Wed, 26 Jun 2013 10:24:01 +0000 (15:54 +0530)]
input: touch: maxim: detect and notify out of idle
Detect no spi interrupt for 50ms as idle. On first interrupt after
idle report miscellaneous input event MSC_ACTIVITY. This event will
serve as early notification for actual input event and will allow
cpu frequency governor to boost CPU clk early.
Change-Id: Ibc240c34f1d66ddc924a5ed662c81f323f429eaf Signed-off-by: Yogish Kulkarni <yogishk@nvidia.com>
Reviewed-on: http://git-master/r/242429
(cherry picked from commit 5c82e83a697bdc15db4f7d00f3a6b89e6d16e90a)
Reviewed-on: http://git-master/r/288746
GVS: Gerrit_Virtual_Submit Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Change-Id: I8efa286a02f0953c4887673c8edc9ad8cf8651e3 Signed-off-by: Yogish Kulkarni <yogishk@nvidia.com>
Reviewed-on: http://git-master/r/242427
(cherry picked from commit 93879ea65d4fef06fca3f95dbb1eeb4248d402e0)
Reviewed-on: http://git-master/r/288745 Reviewed-by: Automatic_Commit_Validation_User Reviewed-by: Bo Kim <bok@nvidia.com> Tested-by: Bo Kim <bok@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com>
Change-Id: Ia0462392d6a2ca06b68fcced1b39ec7d68110c74 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/267903
(cherry picked from commit 1f8fa18c03f135d8d12116528971d71cb0d670c6)
Reviewed-on: http://git-master/r/288821 Reviewed-by: Peter Kim <pekim@nvidia.com> Tested-by: Peter Kim <pekim@nvidia.com>
Change-Id: Id8467cadc21270f73154e32711818111b8eb10c0 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/267128
(cherry picked from commit a6477db2ef36ac1662b539492e1f654858ee60ed)
Reviewed-on: http://git-master/r/288820 Reviewed-by: Peter Kim <pekim@nvidia.com> Tested-by: Peter Kim <pekim@nvidia.com>
Change-Id: I3178aca5b09421b5f6c53d28b6327f530d405e9c Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/265925
(cherry picked from commit a69c29fc91cf581fdd678fa05a6b33568ef9ccc7)
Reviewed-on: http://git-master/r/288819 Reviewed-by: Peter Kim <pekim@nvidia.com> Tested-by: Peter Kim <pekim@nvidia.com>
Harry Hong [Mon, 5 Aug 2013 04:35:51 +0000 (13:35 +0900)]
tty: serial: enable irq_wake
enable irq_wake for uart interrupt if wakeup from lp0
is required.
ex) BT host-wake pin is connected to uart3_cts.
so, making uart interrupt as wake irq.
Change-Id: I11f7ea8a627a84972c5d734d256a253cad2bc0af Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/257237
(cherry picked from commit 3c99642533c4d0279c79ef7df502e65623ff7a71)
Reviewed-on: http://git-master/r/288816 Reviewed-by: Peter Kim <pekim@nvidia.com> Tested-by: Peter Kim <pekim@nvidia.com>
Change-Id: I804f1c78b9931279db667d4a40b84af74cf143ef Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/245987
(cherry picked from commit 3b610623648601bf7fc43f40328cc505de569bcc)
Reviewed-on: http://git-master/r/288793 Reviewed-by: Peter Kim <pekim@nvidia.com> Tested-by: Peter Kim <pekim@nvidia.com>
Change-Id: I44581be980c89aa5bcc8080e7473418c3a64ad01 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/247684
(cherry picked from commit 051e83a2f01b803183de316d3d5536e6127c6904)
Reviewed-on: http://git-master/r/288791 Reviewed-by: Peter Kim <pekim@nvidia.com> Tested-by: Peter Kim <pekim@nvidia.com>
Change-Id: Ib570f2c6f51eb8fe40990e0022e788e90f84e889 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/239622
(cherry picked from commit 996d8c3e9a2889f86f701c48775b27fce69a7389)
Reviewed-on: http://git-master/r/288789 Reviewed-by: Peter Kim <pekim@nvidia.com> Tested-by: Peter Kim <pekim@nvidia.com>
Harry Hong [Mon, 26 Aug 2013 05:45:17 +0000 (14:45 +0900)]
arm: tegra: tegratab: set INT_UARTC as st_host_wakup
- add IORESOURCE_IRQ_OPTIONAL in st host wakeup irq resource.
if it's set, don't request irq.
INT_UARTC need to be requested in hsuard driver.
So, don't request it in st host wake drvier. only doing
enable/disable irq in that driver.
- set INT_UARTC as st_host_wakeup irq
Change-Id: Id6ee39b7307dcbbdc812e531ae743a9e14be03b1 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/265923
(cherry picked from commit 2d6b66dda6650ba55ebc6cb8bed55891be41c29e)
Reviewed-on: http://git-master/r/288818 Reviewed-by: Peter Kim <pekim@nvidia.com>
Change-Id: I73fb063219907e11c07166516731cc19d4145fbf Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/264098
(cherry picked from commit 0db751c2fb918edfe383131a63cbecdc8b329a28)
Reviewed-on: http://git-master/r/288817 Reviewed-by: Automatic_Commit_Validation_User Reviewed-by: Peter Kim <pekim@nvidia.com>
Change-Id: I4290bf99942efc2e78a494e25a123de01c73ca91 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/257236
(cherry picked from commit f0fc356449c817a77cdb2f6343ae008ae9c3b696)
Change-Id: I804f1c78b9931279db667d4a40b84af74cf143ef Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/245987
(cherry picked from commit 3b610623648601bf7fc43f40328cc505de569bcc)
Change-Id: I44581be980c89aa5bcc8080e7473418c3a64ad01 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/247684
(cherry picked from commit 051e83a2f01b803183de316d3d5536e6127c6904)
Change-Id: Ib570f2c6f51eb8fe40990e0022e788e90f84e889 Signed-off-by: Harry Hong <hhong@nvidia.com>
Reviewed-on: http://git-master/r/239622
(cherry picked from commit 996d8c3e9a2889f86f701c48775b27fce69a7389)
staging: android: lowmemorykiller: Add config option to support oom_adj values
The conversion to use oom_score_adj instead of the deprecated oom_adj
values breaks existing user-space code. Add a config option to convert
oom_adj values written to oom_score_adj values if they appear to be
valid oom_adj values.
Change-Id: I68308125059b802ee2991feefb07e9703bc48549 Signed-off-by: Arve Hjønnevåg <arve@android.com> Signed-off-by: David Yu <davyu@nvidia.com>
Reviewed-on: http://git-master/r/261325
(cherry picked from commit 8a38dd632283585d882be5973a461b5c4c4de502)
Reviewed-on: http://git-master/r/282183
GVS: Gerrit_Virtual_Submit Reviewed-by: Gabby Lee <galee@nvidia.com>
Simon Je [Fri, 4 Oct 2013 04:54:03 +0000 (13:54 +0900)]
asoc: rt5639: Set Speaker OVCD to default
Set the speaker over-voltage protection value
to the power on reset value.
Set OVCD to low value when audio is not playing.
Make sure codec mclk is enabled when writing the
codec registers.
Change-Id: Ib673d667af54f3b4fd7bf3183cbde1759838ae4a Signed-off-by: Scott Peterson <speterson@nvidia.com> Signed-off-by: Simon Je <sje@nvidia.com>
Reviewed-on: http://git-master/r/274022 Reviewed-by: Automatic_Commit_Validation_User Reviewed-by: Vinod Subbarayalu <vsubbarayalu@nvidia.com> Reviewed-by: Simon Je <sje@nvidia.com>
GVS: Gerrit_Virtual_Submit Reviewed-by: Gabby Lee <galee@nvidia.com> Tested-by: Gabby Lee <galee@nvidia.com>
(cherry picked from commit 2e0cd932ce9c116b5a0e07a3ce5ad865a63121b3)
Reviewed-on: http://git-master/r/281662
Signed-off-by: Scott Peterson <speterson@nvidia.com>
Change-Id: I1e3b216f84f87f36d05c9693f9a97d0444b68718
Reviewed-on: http://git-master/r/273276 Reviewed-by: Simon Je <sje@nvidia.com>
GVS: Gerrit_Virtual_Submit Reviewed-by: Gabby Lee <galee@nvidia.com>
(cherry picked from commit fdaf1ff6e1d21bf00a53ee76b33cd08c3dc9106e)
Reviewed-on: http://git-master/r/281661 Tested-by: Gabby Lee <galee@nvidia.com>
Change-Id: I01ad0ccff952c1f811d24cb22fde70192753d77e Signed-off-by: Simon Je <sje@nvidia.com>
(cherry picked from commit a438685c653c1cd79d1edb057a678276f2dc7cfe)
Reviewed-on: http://git-master/r/281660 Reviewed-by: Automatic_Commit_Validation_User Reviewed-by: Scott Peterson <speterson@nvidia.com> Reviewed-by: Gabby Lee <galee@nvidia.com> Tested-by: Gabby Lee <galee@nvidia.com>
David Yu [Wed, 11 Sep 2013 11:46:51 +0000 (20:46 +0900)]
input: misc: akm8975: fix kernel panic on shutdown
When device is full flashed then reboots, sometimes akm_probe() and
akm_shutdown() races. akm_shutdown() release all the resources thus kernel
panic can occur in akm_probe().
Change-Id: I4ba2d3fc36dd919306637ec550a098e3f174f159 Signed-off-by: David Yu <davyu@nvidia.com> Signed-off-by: Robert Collins <rcollins@nvidia.com>
Reviewed-on: http://git-master/r/273088
(cherry picked from commit a1f0baffc0350c3a94e39f5aece9da97d7baf657)
Reviewed-on: http://git-master/r/280965 Reviewed-by: Automatic_Commit_Validation_User
GVS: Gerrit_Virtual_Submit Reviewed-by: Gabby Lee <galee@nvidia.com>
On Tegra T11x WDT FIQ is not routed to the GIC, so we cannot use Tegra
WDT for soft-hang debugging.
This WAR is to configure GIC group0 to generate the FIQ to CPU, and only
let WDT routed to GIC group0. All other normal IRQs will use GIC group1.
This WAR is only applicable to non-secure OS build; secure OS build will
treat FIQ as secure interrupt for enterring monitor.
This patch is based on previous work of Lucas Dai and Bo Yan.
The FIQ WAR requires all the IRQs using the GIC group1, and let the FIQ
use GIC group0 only. This patch will set all the normal IRQs to group1
if GIC WAR is enabled.
- there is below race condition in cec driver :
-tegra_cec_init() is in progress which is setting many registers
-suspend is called at the same time
-tegra_cec_suspend() disables clock before tegra_cec_init() completes
- above race condition hangs the system with trying to write cec
register while clock is off
- fix this race condition by deferring suspend if init is still in
progress
Need to handle retune request case also in the tuning procedure.
We set retune flag if we remove (un-plug) SD card.
So, tuning should be done next time when SD card is inserted.
Continue tuning procedure for full window calculation,
for the full window size is greater or equal to 50% of
partial window size, else choose partial window.
Change-Id: I7010bb0cfa564bac04add85bd9d99507c7fb547f Signed-off-by: Rick Song <ricks@nvidia.com>
Reviewed-on: http://git-master/r/277462 Reviewed-by: Automatic_Commit_Validation_User Reviewed-by: Laxman Dewangan <ldewangan@nvidia.com> Reviewed-by: Martin Chi <mchi@nvidia.com> Tested-by: Martin Chi <mchi@nvidia.com>
This version fixes the corruption of secure os(TF) stack which is caused by
unexpected hardware cache activities at the time of core #0 wakes up from
per-core power gating.
Caches ON while go to sleep and L1 cache is all cleaned before the sleep,
while L2 cache is selectively cleaned because it's preserved during
the sleep of the cores so L2 cache has some dirty lines from the secure world.
When core #0 wakes up, it turns on i-, d-caches but MMU is still off at this
moment. During this time the wake up process starts synching (flushing) cache
to SDRAM. Until MMU is ON, core #0 uses stacks data from SDRAM, not from the
data cache. It is enough to flush the old line of stack to SDRAM in order to
corrupt call-return linkage if it accidently happens during such stack usage.
To fix this issue, TF cleans the cache lines related to secure os stack
after they have been used last time just before returning to normal OS.
This patch enables the GPIOs and IRQ for the Pike module on
Dalmore. If the pike module is detected, it is enabled instead
of the camera and the motion sensors.
Predict millivolts interface selected pll or dfll mode tables based
on target rate and dynamic dvfs state: rail mode, and current dvfs
clock rate. Since interface clients may not be serialized with run
time rate/voltage scaling (e.g., edp client) it opens possibility for
races. On the other hand there is no need to use current dynamically
changes state to select scaling table - comparing target rate, with
initialized once dfll rate range is sufficient. This commit updated
predict millivolts interface accordingly.
TCTRL/RCTRL values can be calculated either using xusb PADCTL registers
or using SNPS registers. As per latest discussion with HW team, we must
use SNPS registers when any of UTMIP port is owned by SNPS controller and
use XUSB PADCTL registers only when all UTMIP ports are owned by XUSB.
TCTRL/RCTRL is alwasy needed to be programmed on T114 due to a HW bug but
on T124 it needs to be programmed only during ELPG entry when PMC is being
ARMed. Refer table below on this.
T114 T124
1) Init Yes No
2) ELPG entry Yes Yes
3) LP0 exit Yes No
rel-17 branch doesn't have T124 related changes so only T114 relevant changes
are added here.
Xiaohui Tao [Tue, 17 Sep 2013 17:10:54 +0000 (10:10 -0700)]
input: touch: raydium: v60.6 code drop
a. Checkpatch done (Bug ID 1368132)
b. Fix 3.3V control issue (Bug ID 1354253, 1348628) as the attached pictures. The 3.3V is turned off in LP0 mode.
c. Refine suspend/resume timing issue (Bug ID 1359713, 1288233, 1283537)
Change-Id: I468bf04e9eb63ba6b4bb1e646baa73b1fbb55969 Signed-off-by: Xiaohui Tao <xtao@nvidia.com>
Reviewed-on: http://git-master/r/275861 Reviewed-by: Robert Collins <rcollins@nvidia.com> Reviewed-by: Jon Mayo <jmayo@nvidia.com>