Some USB host controller seems to have problems with
autosuspend. For example, Rockchip rk3328 SoC USB 3.0
wouldn't handle remote wakeup correctly with external
hub after entered autosuspend, caused to resume SS
port fail.
This patch introduces a new quirk flag for hub that
should remain disabled for autosuspend.
Change-Id: I6d14222b2c5025583fea811a6afd6abd22f41cb9
Signed-off-by: William Wu <wulf@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch adds superspeed descriptors in device
applications to support USB 3.0 ffs gadget.
Change-Id: I5a364c935b1d30e2e929791ff16a34cf0d1c87e1
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
The spi which's version is higher than ver 2 will automatically
enable this feature.
If the length of master transmission is uncertain, the RK spi slave
is better to automatically stop after cs inactive instead of waiting
for xfer_completion forever.
Change-Id: If99e51d35391b824f48e31a3e4508db036593c8a
Signed-off-by: Jon Lin <jon.lin@rock-chips.com>
1.Add standard spi-cs-high support
2.Refer to spi-controller.yaml for details
Change-Id: I899bce8d9418ee99c784726bb56534aaed27c00b
Signed-off-by: Jon Lin <jon.lin@rock-chips.com>
1.Add standard cs-gpio support
2.Refer to spi-controller.yaml for details
Change-Id: I8f839189038afd77d534d767d938c845aa54fedb
Signed-off-by: Jon Lin <jon.lin@rock-chips.com>
Add compatible string for rv1126 for potential applications.
Signed-off-by: Jon Lin <jon.lin@rock-chips.com>
Change-Id: I2ad2de0c715b064870b35214f25e9705412fd62f
Fixes: f27d8c9975 ("spi: rockchip: get pinctrl for lookup pinctrl state")
Fixes: 87dbea63d5 ("spi: rockchip: set higher io driver when sclk higher than 24MHz")
Change-Id: I963c92eab7f7bff0b32e2ac262aa79f0667f39ee
Signed-off-by: Huibin Hong <huibin.hong@rock-chips.com>
Signed-off-by: Jianqun Xu <jay.xu@rock-chips.com>
Signed-off-by: Jon Lin <jon.lin@rock-chips.com>
RK3568 u2phy used shared interrupt and do not used id irq.
Change-Id: I341cc0edb0f74996f159c095545465673cc2a990
Signed-off-by: Bin Yang <yangbin@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Test on RK3568S EVB1, the otg device data eye test fail with
far end template when use 1.2 meter long cable. So tuning
the hs eye height from 400mv(default) to 437.5mv. And we
test on RK3568 EVB1, it can also benefit from this patch.
Change-Id: Ie2342aba5546990838fdd6faf27a007a8843fd0d
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
When do USB 3.0 Receiver Jitter Tolerance Test, it fails at
Sj Frequency 2.0/4.9/10.0 [MHz]. This patch adjusts the PLL
parameters for USB to pass the Receiver Jitter Tolerance Test,
and it's helpful to improve the USB 3.0 signal compatibility.
Change-Id: I58eb687a4677fe22cf5bc324578b033526310859
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
If the Type-C PHY works as DP 4 lanes, the Tx/Rx lanes
of Type-C USB 3.0 PHY is used for DP lanes, so it needs
to force the USB 3.0 to USB 2.0 only, and make sure USB
3.0 xHCI controller doesn't support USB 3.0 port.
With this patch, an ASUS Type-C to HDMI dongle which
supports DP 4 lanes and USB 2.0 simultaneously can
suspend/resume successfully.
Change-Id: I77049702c768bd56d638d11c29aae07eeb608282
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Bind / unbind stress testing of the USB controller on rk3399 found
that we'd often end up with lots of failures that looked like this:
phy phy-ff800000.phy.9: phy poweron failed --> -110
dwc3 fe900000.dwc3: failed to initialize core
dwc3: probe of fe900000.dwc3 failed with error -110
Those errors were sometimes seen at bootup too, in which case USB
peripherals wouldn't work until unplugged and re-plugged in.
I spent some time trying to figure out why the PHY was failing to
power on but I wasn't able to. Possibly this has to do with the fact
that the PHY docs say that the USB controller "needs to be held in
reset to hold pipe power state in P2 before initializing the Type C
PHY" but that doesn't appear to be easy to do with the dwc3 driver
today. Messing around with the ordering of the reset vs. the PHY
initialization in the dwc3 driver didn't seem to fix things.
I did, however, find that if I simply retry the power on it seems to
have a good chance of working. So let's add some retries. I ran a
pretty tight bind/unbind loop overnight. When I did so, I found that
I need to retry between 1% and 2% of the time. Overnight I found only
a small handful of times where I needed 2 retries. I never found a
case where I needed 3 retries.
I'm completely aware of the fact that this is quite an ugly hack and I
wish I didn't have to resort to it, but I have no other real idea how
to make this hardware reliable. If Rockchip in the future can come up
with a solution we can always revert this hack. Until then, let's at
least have something that works.
This patch is tested atop Enric's latest dwc3 patch series ending at:
https://patchwork.kernel.org/patch/10095527/
...but it could be applied independently of that series without any
bad effects.
For some more details on this bug, you can refer to:
https://bugs.chromium.org/p/chromium/issues/detail?id=783464
Change-Id: I7909731247739694f56bf89ab3064889f2b34d3c
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
(am from https://patchwork.kernel.org/patch/10105833/)
the phy config values used to fix in dp firmware, but some boards
need change these values to do training and get the better eye diagram
result. So support that in phy driver.
Change-Id: I195727b2a81130606e66ffc4471df74e5782a7fa
Signed-off-by: Chris Zhong <zyw@rock-chips.com>
Signed-off-by: Lin Huang <hl@rock-chips.com>
Signed-off-by: Wyon Bi <bivvy.bi@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
I test on RK3399 IND EVB running Android R system, and
set the USB gadget as UVC function via configfs, both
the USB 2.0 and 3.0 UVC have preview abnormal problems.
If the CPU enter deeper sleep, the USB DMA transfer gets
corrupted, and the UVC ISOC transmission lost data.
This patch puts in a "pm_qos_latency" requirement which
will keep the CPU out of the deeper sleep states when
UVC stream on. And allow the CPU to enter deeper sleep
state after UVC stream off.
Change-Id: I808290f124c6a32da3888819348093a205bfad61
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch adds uvc_function_suspend and uvc_function_resume
for uvc function. With this patch, if usb bus enter suspend
or resume state, the usb controller driver will call the
uvc_function_suspend or uvc_function_resume to send event
to uvc app.
Change-Id: I9c584aae25298747c5a287243cb3efd71c8adfe6
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch add new uvc trace for setup request processing.
1. enable trace
echo 4 > /sys/module/usb_f_uvc/parameters/trace
2. dump trace
dmesg
Test on RV1109 EVB with UVC function when streaming on:
[ 426.441316] uvcvideo: setup request a1 81 value 0100 index 0003 001a
[ 426.450280] uvcvideo: uvc_send_response: req len 26
[ 426.450444] uvcvideo: event_setup_out 0, data len 26
[ 426.451545] uvcvideo: setup request 21 01 value 0100 index 0003 001a
[ 426.460862] uvcvideo: uvc_send_response: req len 26
[ 426.461014] uvcvideo: event_setup_out 1, data len 26
[ 426.461173] uvcvideo: setup request a1 81 value 0100 index 0003 001a
[ 426.485791] uvcvideo: uvc_send_response: req len 26
[ 426.486672] uvcvideo: event_setup_out 0, data len 26
[ 426.487145] uvcvideo: setup request a1 83 value 0100 index 0003 001a
[ 426.496529] uvcvideo: uvc_send_response: req len 26
[ 426.496655] uvcvideo: event_setup_out 0, data len 26
[ 426.497485] uvcvideo: setup request a1 82 value 0100 index 0003 001a
[ 426.506737] uvcvideo: uvc_send_response: req len 26
[ 426.507195] uvcvideo: event_setup_out 0, data len 26
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Change-Id: If882728e891d033c249e6c466b40d3e2a28bccb6
According to USB Audio Device 1.0 Spec, Ch4.6.1.1:
The wMaxPacketSize of endpoint is defined as follows:
Maximum packet size of endpoint is capable of sending
or receiving when this configuration is selected. This
is determined by the audio bandwidth constraints of
the endpoint.
In current code, the wMaxPacketSize is limited to 200,
and the bInterval is set to 4 (1ms). That is, the maximum
bandwidth over USB bus is 200 * 1000 = 200000 bytes.
We find an issue about bandwidth limitation when we try to
support UAC1 with 8ch * 16bit * 16KHz on RK3308 platform,
which needs more bandwidth than it can support.
This patch sets the wMaxPacketSize dynamically according
to the parameters of UAC1. It is similar to the same thing
done earlier for f_uac2.
Change-Id: I9af3fd7665a80b5eb0cfb5dc91ebe2c20df1dd46
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch add uevent to notify the application layer how much ppm
is different between USB clk and AUDIO clk.
The event include two parts USB_STATE and PPM. For example:
g_audio_work: sent uac uevent USB_STATE=SET_AUDIO_CLK PPM=12
g_audio_work: sent uac uevent USB_STATE=SET_AUDIO_CLK PPM=-1
Note: The ppm compensation depends on the method implement of
clk drift and compensation in the rockchip_pdm.c driver. So if
you want the ppm compensation to take effect, please make sure
the commit "ASoC: rockchip: pdm: Add support for clk compensation"
is merged.
Change-Id: Id25411397fe376342c773c11f1989ed5854f8ad9
Signed-off-by: Ren Jianing <jianing.ren@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch add direction index for eps_revmap, which can avoid matching
to the requests for other functions.
For example, if we config ep1in for uac and ep1out for adb, the requests
for ep1in will be matching to ffs without this patch.
Change-Id: Ic7a52fca35503656bbc306c9f0014dfdea61392b
Signed-off-by: Ren Jianing <jianing.ren@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
The usb gadget core set the chosen endpoint descriptor for
each endpoints in config_ep_by_speed(), however, we want
to get the transfer type of the endpoints earlier on the
rockchip platforms for usb controller initialization
(e.g. do tx fifos resize for rockchip usb dwc3 controller),
so this patch add transfer_type in the struct usb_ep, and
set the transfer_type in the usb_ep_autoconfig_ss().
Change-Id: Ia2added218e180dda7a7ca5da09ee18d63be1ff0
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Without this patch, we only calculate usb request packet_size when
start playback. If PC set sample rate after opening stream, the value
of packet_size will not be updated to uac driver and the gadget will
send error size of pcm data.
Signed-off-by: Ren Jianing <jianing.ren@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Change-Id: Iaa54ed740c98e5f12bf9cf680b50cf4be931cfbd
This patch adds feature unit descriptor for f_uac1 and f_uac2,
and supports volume control and mute control for capture and
playback.
By default, the volume and mute feature unit descriptors are NOT
add in the UAC descriptor. The user can define whether the UAC1/
UAC2 shall support volume and mute functionality via the attributes
c_feature_unit/p_feature_unit in the UAC function directory.
For example, user can add the volume and mute feature unit descriptors
for UAC1 capture and playback:
echo 1 > /config/usb-gadget/gadget/functions/uac1.name/c_feature_unit
echo 1 > /config/usb-gadget/gadget/functions/uac1.name/p_feature_unit
This patch also adds uevents for volume and mute functionality. The
user can complete the real volume and mute control functionality in
the user space depends on these new uevents (like the uac_app in the
RV1126/RV1109 SDK).
Change-Id: I76d447a19fd69e038851040cd73e6c7d420f467d
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Ren Jianing <jianing.ren@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch handles the class specific control request
"USB_CDC_SET_ETHERNET_PACKET_FILTER" for rndis. Actually,
the cdc_filter is not rndis specific, so we don't really
set the cdc_filter to the wValue of setup request for
rndis, and just return value 0 for the setup request to
avoid usb controllers stall ep0.
Change-Id: I42f9835a9812b8dfa640562b8555bc5f9b022256
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch add uevent to notify the application layer when it should
playback or capture and setting the accuration sample rate.
The event consists of three parts, namely USB_STATE, STREAM_DIRECTION
and STREAM_STATE. For example:
g_audio_work: sent uac uevent USB_STATE=SET_INTERFACE STREAM_DIRECTION=IN STREAM_STATE=OFF
g_audio_work: sent uac uevent USB_STATE=SET_INTERFACE STREAM_DIRECTION=OUT STREAM_STATE=ON
g_audio_work: sent uac uevent USB_STATE=SET_SAMPLE_RATE STREAM_DIRECTION=OUT SAMPLE_RATE=44100
Signed-off-by: Ren Jianing <jianing.ren@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Change-Id: Ia01572d92d0c3ae40b19e772571282b8806da9e5
Implement support for multiple sampling rates in the USB Audio gadgets.
A list of sampling rates can be specified via configfs. All enabled
sampling rates are sent to the USB host on request. When the host
selects a sampling rate the internal active rate is updated. The
currently configured rates are exposed through amixer controls. Also on
pcm open from userspace the requested rated is checked against the
currently configured rate of the host.
Link: https://lore.kernel.org/patchwork/patch/805286/
Signed-off-by: Julian Scheel <julian@jusst.de>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Change-Id: I43dc28462062c9aa179c2b9c8e7e50c307137874
This replaces the dedicated headers for uac1 and uac2 functions with a
shared header for both of them. Apart from unifying the struct names,
further duplicated code for configfs setup is moved out of the function
files into the shared header.
Link: https://lore.kernel.org/patchwork/patch/805285/
Signed-off-by: Julian Scheel <julian@jusst.de>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Change-Id: Ie866bd02f174dc91168fe0bc49b571a77d849a42
For rockchip platforms, the userspace uvc application use
bytesused == 0 as a way to indicate that the data is all
zero and unused. In that case, the uvc driver should use
the allow_zero_bytesused flag to keep bytesused 0 rather
than use the buf actual size instead. This can help to
preview 1080P@MJPG faster.
Change-Id: Ic4d561f18e85d759b5a2405db4f6fb48533e1781
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch adds a new uvc attribute "uvc_num_request" to
set the number of uvc requests. And the maximum limit is
set to UVC_MAX_NUM_REQUESTS (8), the default value of
uvc_num_request is set to UVC_NUM_REQUESTS (4).
Change-Id: I69d6a4d6547a81b67ce447f271c7a5103671767e
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch fixes the following panic when use uvc function and
do reboot during uvc preview.
Unable to handle kernel NULL pointer dereference at virtual address 000001fd
pgd = 85dd55c1
[000001fd] *pgd=00000000
Internal error: Oops: 17 [#1] PREEMPT SMP ARM
Modules linked in: galcore(O)
CPU: 0 PID: 716 Comm: xc:RkAiqCoreThr Tainted: G W O 4.19.111 #18
Hardware name: Generic DT based system
PC is at usb_gadget_deactivate+0x0/0x6c
LR is at usb_function_deactivate+0x54/0x74
It's because that do reboot operation will call configfs_composite_unbind()
to set cdev->gadget to NULL.
Change-Id: I6fbfe9b58f865113d04ca7ce0b74b00f8d89227c
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
As we know, The framebased format can be used to support a lot of
formats other than YUV and MJPEG, for example: H264 or H265.
And Nowadays, the H264 and H265 is used more and more compared to the
MJPEG, so there is a need to support such usecase, although the new UVC
1.5 and the UVC1.1 extensions already support H264, but they are much
more expensive and most of the time, what we need is just a channel to
stream the video frames.
In this patch, we implement the framebased format and set it to H264 by
default, and the users could change the video format as they need through
the guid configfs attribute node.
Change-Id: Ie1c62f5fb4fa57c24a2878f9992c982e9deb302e
Signed-off-by: Zeng Tao <prime.zeng@hisilicon.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
(am from https://patchwork.kernel.org/patch/10722957/)
If a configs function sets bind_deactivated flag, upon removal it will be
left with an unbalanced deactivation. Let's make sure that we conditionally
call usb_function_activate() from purge_configs_funcs() and make sure
purge_configs_funcs() is called from configfs_composite_unbind().
Change-Id: Ibe91837fdb7754f8af343a18138556901a0e089f
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
According to the spec "USB Interface Association Descriptor Device
Class Code and Use Model"[1], the Interface Association Descriptor
(IAD) is needed for composite device which has multiple interfaces
controlled independently of each other. Without the IAD, the device
may not work as expected because the USB system software will not
properly bind the interfaces with drivers (e.g UAC1 && UVC composite
device not recognized in Win10).
[1] https://www.usb.org/sites/default/files/iadclasscode_r10.pdf
Change-Id: I446c771fe182bd556a054d9ae0ada1374442d1d0
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This patch adds bulk endpoint for uvc to be used as
video streaming transfer. By default, the uvc gadget
still uses isoc endpoint for video streaming.
The important difference between the bulk and isoc
method is that, alt-settings in a video streaming
interface are supported only for isoc endpoints as
there are different alt-settings for zero-bandwidth
and full-bandwidth use-cases, but the same is not
true for bulk endpoints, as they support only a single
alt-setting.
How to use:
1. Enable the bulk transfer:
echo 1 > /config/usb-gadget/g1/functions/uvc.name/streaming_bulk
2. Disable the bulk transfer:
echo 0 > /config/usb-gadget/g1/functions/uvc.name/streaming_bulk
3. Testing the uvc bulk function transfer:
device: run the uvc-gadget program
# uvc-gadget -b -u /dev/video<uvc video node #> -v /dev/video<vivid video node #>
where uvc-gadget is this program:
http://git.ideasonboard.org/uvc-gadget.git
with these patches:
http://www.spinics.net/lists/linux-usb/msg99220.html
Change-Id: Ifedfe3f5c4354dd2bdf07382290107e9bcc89f59
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
We find that the UAC2 gadget can't be recognized on Windows 10.
It's because that the descriptors of UAC2 doesn't meet the
requirements of Windows.
According to the USB Audio 2.0 Drivers of Windows [1], if the
bmAttributes of OUT-EP in the audio stream interface is set
to asynchronous (use USB_ENDPOINT_SYNC_ASYNC), then a feedback
endpoint must be implemented in the respective alternate setting
of the AS interface. The Windows driver does not support implicit
feedback.
However, it's difficult to implement the feedback endpoint now.
So the patch changes the bmAttributes of OUT-EP to adaptive, and
changes the bmAttributes of IN-EP to synchronous at the same time.
This patch also sets the wTerminalType of terminal descriptor to
microphone and speaker by default.
With this patch, we also fix the wTotalLength of the ac_hdr_desc.
[1] https://docs.microsoft.com/en-us/windows-hardware/drivers/audio/usb-2-0-audio-drivers
Change-Id: I3597d5f321235fcbce56dbfbfe95172d02e58892
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
This allows 32 bit owners of USB Class to make ioctls
into a 64 bit kernel.
Change-Id: Ia31b26147ab619f0673f94b6662eaf181a9eb5dd
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
Add extension unit descriptor for uvc. Support 3
controls in this extension unit. The rockchip IQ
tool use it to transfer vendor specific control
data.
Change-Id: I219e12616629bc75548b30ce63d46136aeac6561
Signed-off-by: William Wu <william.wu@rock-chips.com>
The default period size is only 64 frames, this
will cause usb audio playback with noise via
internal audio codec. This patch sets the period
size to (snd->rate / 10), and also sets the buffer
size to snd->rate.
Change-Id: I4a4eb1b4dd79aec65f5c44eacd8a2fa101dfbd1b
Signed-off-by: William Wu <william.wu@rock-chips.com>
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
The f_audio_disable() doesn't disable usb ep, and
this cause usb enumeration fail. So add usb ep
disable operation.
This patch also reinitializes the opts->bound flag
to false in f_audio_free(), and then it can setup
ALSA audio device again in f_audio_bind().
Change-Id: I7b10630f5085b1a03792bc4b9e7eabb02d2bd5a2
Signed-off-by: William Wu <william.wu@rock-chips.com>
In current code, the connect detective event will be send before the
connect state of controller change. The event is sent when receive
the first setup packet and the connect state of controller is changed
when receive set address command. So if disconnect interrupt trigger
between the first setup packet and set address command, the disconnect
process will not be done, while the application state has been changed
and keep in connected. As a result, the UI may still show the USB connect
option event if the device has been disconnect from PC. This patch send
the connect uevent when get descriptor of configuration which follow
set address command to make sure the application state change after
controller.
Change-Id: Icf2124327db93687b4b644672edb0dbecc8f127a
Signed-off-by: Meng Dongyang <daniel.meng@rock-chips.com>
The RNDIS gadget function has USB class of 2 and subclass
of 2, which matches "USB\Class_02&SubClass_02" in the
usbser.inf file in the Windows system, so the device is
initially detected as a COM port instead of RNDIS. This
is why we need to install RNDIS manually.
Refer to Defined Class Codes [1] and rndis_host of Linux [2],
this patch sets the RNDIS gadget with base class of 0xE0h
(Wireless Controller) and subclass of 0x01h and protocol
of 0x03h.
[1] https://www.usb.org/defined-class-codes
[2] http://www.embedded-os.de/en/proto-rndis_host.shtml
Change-Id: Ida366749f378a0ce770d707b4ba56b87f9e188cf
Signed-off-by: William Wu <william.wu@rock-chips.com>
