| Age | Commit message (Collapse) | Author |
|---|
|
add header files to make cherry-pick compile on 2.6.35 kernel
change wireless name from eth to wlan
Signed-off-by: Tony Lin <tony.lin@freescale.com>
|
|
AR6003 is a single stream, SDIO based 802.11 chipset from
Atheros optimized for mobile and embedded devices. ath6kl is a
cfg80211 driver for AR6003 and supports both the station and
AP mode of operation.
Station mode supports 802.11 a/b/g/n with HT20 on 2.4/5GHz and
HT40 only on 5GHz. Some of the other features include WPA/WPA2,
WPS, WMM, WMM-PS, and BT coexistence. AP mode can be operated
only in b/g mode with support for a subset of features mentioned
above.
The driver supports cfg80211 but comes with its own set of
wext ioctls which have historically supported some of our
customers with features like BT 3.0 and AP mode of operation.
For further details, please refer to:
http://wireless.kernel.org/en/users/Drivers/ath6kl
The driver requires firmware that runs on the chip's network
processor. The majority of it is stored in ROM. The binaries
that are downloaded and executed from RAM are as follows:
1) Patch against the code in ROM for bug fixes and feature
enhancements.
2) Code to copy the data from the OTP region of the memory
into RAM.
3) Calibration file carrying board specific data.
The above files need to be present in the directory
'/lib/firmware/ath6k/AR6003/hw2.0/' for the driver to initialize
the chip upon enumeration. The files can be downloaded from the
link specified at the following location:
http://wireless.kernel.org/en/users/Drivers/ath6kl#Download
This driver is only provided in the interim while we work on
the mac80211 replacement, ath6k. Once the mac80211 driver
achieves feature parity with the ath6kl driver, the ath6kl will
be deprecated and removed from staging.
Signed-off-by: Vipin Mehta <vmehta@atheros.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
Customer provided this patch, the SET_BLOCKLEN shouldn't be sent when
enable the DDR mode required by customer's iNand.
Tested by customer on it's own board.
BTW, this patch is not needed on 2.6.38 kernel, should be discarded
after upgrade the kernel version.
Signed-off-by: Richard Zhu <r65037@freescale.com>
|
|
Fix the errors after implement the new soc rev APIs
Refine the SOC revision APIs in the esdhc driver.
Signed-off-by: Richard Zhu <r65037@freescale.com>
|
|
This patch adds MXC V4L2 capture maximum buffer number
from 3 to 10. This makes users be able to get rid of
losing frames in streaming mode by requiring more buffers.
Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
|
|
We used to set the interrupt thresholds to the measure range
of the light sensor in the interrupt handler, this potentially
makes user poll on the input event and never be waken up.
This patch keeps the interrupt thresholds in the interrupt
handler and is tested on MX53 SMD.
Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
|
|
Since Ubuntu needs report single pointer, I add BTN_TOUCH
event in single pointer event report.
It will not affact Android, also made ubuntu touchscreen works.
Signed-off-by: Zhang Jiejing <jiejing.zhang@freescale.com>
|
|
- Virtual address parameter for alternate buffer feature removed. Virtual
address now computed based on physical address parameter.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
- Create new per-update list of markers instead of just one marker
per update.
- Allow updates with markers to merge and compile all markers onto a
per-update list.
- Changed full marker list member variable to a list_head to reduce
memory allocations and wasted memory.
- Also redefined buffer queue objects as list_head objects
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
The GPU hang when run two cubes together with one video playback.
According to the suggestion from AMD, we'd better not read register
when GPU active, especially for CP block.
Signed-off-by: Zhou, Jie <b30303@freescale.com>
|
|
1. Set initial max poll delay time in mma8450 and mma8451
accelerometer sensor driver. And remove one build warning.
2. Remove dev parent set.
Signed-off-by: Sammy He <r62914@freescale.com>
|
|
For better alignment with the upstream i.MX kernel, switch from
cpu_is_mx5x_rev() to mx5x_revision().
This is the drivers/ part of the patch.
Signed-off-by: Dinh Nguyen <Dinh.Nguyen@freescale.com>
|
|
Created a list for pending update and a list for update markers.
Memory for list object for both is allocated dynamically. Accordingly,
code was modified to work with these new object structures.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
This patch adds ISL29023 light sensor support.
Users may control the light sensor work at
different modes/ranges/resolutions or read
the lux value via sysfs.
A power state node is also created in sysfs.
If the light sensor works at ALS once mode,
users need to power down/power up the sensor
before read the lux value again.
Currently, IR(Infra Red) mode is not supported.
Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
|
|
Remove interlaced mode from fb modelist before it can be supported.
Should revert this patch after interlaced mode support on normal display port.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
Refresh rate nearness is not calculated or reset when nearest resolution
changes.
This patch resets the refresh rate differential measurement whenever a
new nearest resolution is discovered. This fixes two error cases;
first, wherein the first mode's refresh rate differential is never
calculated and second, when the closest refresh rate from a previous
nearest resolution is erroneously preserved.
Signed-off-by: Andrew Kephart <andrew.kephart@alereon.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
(cherry picked from commit e504b84805c574cf18d705c610c9e4a8f6e61017)
|
|
Add name sysnode to mxc_ddc and sii902x, which can be check under:
/sys/devices/platform/mxc_ddc.0/fb_name
or
/sys/devices/platform/sii902x.0/fb_name
It's the name of fb fix id which it associated.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
Need enable analog regulator.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
add battery charger change event.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
Add console output for iim value.
Signed-off-by: Terry Lv <r65388@freescale.com>
|
|
Add max17085 battery charger support
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
Added code to detect the potential for a "missed collision", where
an update would cause a collision with an update sitting on the collision
list, but does not do so with currently active updates. For this case,
we conservatively treat the update as if it has generated a collision.
Refined merging algorithm by blocking the merge process upon
detecting certain conditions where no further merges should be done.
Also added check for a special case where update regions are identical.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
After this patch, you can check edid info under
/sys/devices/platform/sii902x.0/edid for sii902x HDMI.
And under /sys/devices/platform/mxc_ddc.0/edid
for common DDC.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
fix build break for imx25
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
Add YV12 format support.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
Add common ddc driver support.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
This is HannStar p1003 touch screen driver with Ilitek firmware
multitouch(two point at max) touch screen driver.
Signed-off-by: Xinyu Chen <xinyu.chen@freescale.com>
Signed-off-by: Zhang Jiejing <jiejing.zhang@freescale.com>
|
|
during multi-block transfer and MX50 TO1.1/MX53 TO2
SDHCI_VERDOR_SPEC register will be overwritten
Signed-off-by: Tony Lin <tony.lin@freescale.com>
|
|
This patch changes video mode setting method for LDB
related video modes.
Users need to add 'ldb=dix' for default video mode
support, otherwise, need to specify the ldb mode
in Kernel bootup command line.
Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
|
|
fsl_ep_disable may be called while unregister gadget driver. Gedget speed
was set to UNKNOWN. So we should disable endpoint whatever link speed.
Correct uninitialized variables in spin lock.
Signed-off-by: Zhang Yan <b34916@freescale.com>
|
|
HDMI Sii902x support for changing the video mode setting method.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
TVE & related ipu driver support for changing the video mode setting method.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
FB driver support for the video mode setting method.
Signed-off-by: Jason Chen <b02280@freescale.com>
|
|
Some functions at setup_received_irq will cause spin lock recursion problem,
unlock spin locks before calling these functions, and re-lock spin locks
after executing them.
Signed-off-by: Peter Chen <peter.chen@freescale.com>
|
|
Instead of dropping updates with a different panning offset than the
current framebuffer panning offset, we just process those updates
using the current framebuffer panning offset. This ensures that
all requested update regions are processed from the latest buffer.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
PxP mutex was being released and retaken while EPDC power turned on.
This provides opportunity for a race condition where another thread
attempts to being a PxP operation and corrupts the PxP data
structures of the EPDC. PxP mutex now held through the EPDC
powerup sequence.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
- When update is processed, we check against latest
panning offset and drop the update if they don't match. This
ensures that we always update using the most current
buffer location.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
Statically allocate buffers for each update object, rather
than dynamically allocating/freeing buffers per use. These buffers
are used when PxP limitations require update content first be copied
into a temporary buffer.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
Markers now cleared once the associated LUT completes. This
ensures that markers will always be cleared. The marker does not
have to be waited on to be cleared.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
- Corrected the computation of PxP input/output buffer address offsets
- Corrected the computation of EPDC input buffer address offset
- Expanded cases requiring SW of update contents pre-PxP processing,
now covering:
- Case where input address offset is not 32-bit aligned
and auto-waveform mode selected
- Case where input address is not 32-bit aligned and
total pixels added per-line to PxP processing task
is 8 or more.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
- Software workaround for PxP alignment/histogram limitation implemented
For updates using auto-waveform selection and using update region
dimensions that are not 8x8 multiples, the update region is copied
into a temporary buffer and the margins filled with 0s. This buffer
becomes the input to the PxP.
- PxP/EPDC buffer offsets adjusted to properly handle all rotation cases.
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
1) Fix the double locking IPU spinlock issue.
2) Protect g_channel_enable_mask with IPU spinlock.
Signed-off-by: Liu Ying <Ying.Liu@freescale.com>
|
|
Disable usb interrupt in host wakeup to avoid bad irq. USB host
irq was re-enabled in resume. BTW, set host-to-gadget discharge 20ms.
Signed-off-by: Zhang Yan <b34916@freescale.com>
|
|
- Ensure that no updates are active when changing the FB info
- Fix criteria for evaluating whether any updates are active
Signed-off-by: Danny Nold <dannynold@freescale.com>
|
|
1. Add discharge vbus when set vbus lower;
2. Due to unplug from HOST bring up suspend irq. add discharge in suspend irq.
3. Update port speed when port connect changed.
4. Add port speed verify in ep operation code.
Signed-off-by: Zhang Yan <b34916@freescale.com>
|
|
When open mxc_iim twice, mxc_iim can't be close twice.
Signed-off-by: Terry Lv <r65388@freescale.com>
|
|
mxc_iim: let mxc_iim work with dd and echo.
Signed-off-by: Terry Lv <r65388@freescale.com>
|
|
Error check of clk pointer should use IS_ERR instead of
ERR_PTR(ENOENT). Updated SCC2 and SAHARA drivers to use
this method for error checks as well as to propagate the
error code where applicable.
Signed-off-by: Anish Trivedi <anish@freescale.com>
|
|
mxc_iim can't be open twice.
The reason is that request_irq failed when open mxc_iim again.
So put request_irq function to probe function.
Signed-off-by: Terry Lv <r65388@freescale.com>
|
|
Open usb clock before view usb debug file. avoid system hang when view
usb debug information in suspend state.
Signed-off-by: Zhang Yan <b34916@freescale.com>
|
