In rk3126, when aclk_vio0\aclk_vio1\hclk_vio were reparented from the
default parent gpll_div2 to gpll in clk_init, the temporary rates
are too high and may lead to failture in VIO idle_request later.
To slove this problem, VIO clocks are modified to auto select parent
and consider the order of reparent and set_div when set_rate.
Signed-off-by: dkl <dkl@rock-chips.com>
If soc is not support feature SUPPORT_TMDS_600M,
or current EDID information is not support the
feature, we should not show the corresponding
option, such as 3840x2160p-60.
Signed-off-by: Zheng Yang <zhengyang@rock-chips.com>
1. fix 3288 hdmi audio source select error.
2. For mid, it need to call rk_fb_get_display_policy
to reset overscan after swichting vop timing.
Signed-off-by: Zheng Yang <zhengyang@rock-chips.com>
1. Rename rk3036 hdmi to rockchip_hdmiv1, support hdmi v1.4;
2. Rename rk3288 hdmi to rockchip_hdmiv2, support hdmi v2.0;
3. Support transmmit 3D picture.
4. 3288 hdmi support HDCP, CEC and TMDSCLK 600M.
5. HDMI video source default is LCDC0, it can be modified in dts by
follow:
rockchip,hdmi_video_source = <DISPLAY_SOURCE_LCDC1>;
HDCP and CEC function default is disabled , it can be enabled in
dts file by follow:
rockchip,hdcp_enable = <1>;
rockchip,cec_enable = <1>;
More information please see
Documentation/devicetree/bindings/video/rockchip_hdmi.txt
disable clock will call disable power domain.
hdmi and lcdc share the power domain, so they must disable
the power domain when system shutdown.
Signed-off-by: zwl <zwl@rockchips.com>
Video stress test failed, may have something relationship to vop mmu stall operation
skip it as a workaround
Reboot stress test failed, error showed a large number of vop/vop mmu interrupt,mask
vop mmu interrupt to avoid reboot stress test failed in rk312x
Signed-off-by: Simon <xxm@rock-chips.com>
This reverts commit 63041eff22. For some
reason this is causing failures before the serial console comes up on
TC2 when booted using UEFI, revert pending investigation.
Signed-off-by: Mark Brown <broonie@kernel.org>
The usage likes following:
$scripts/kconfig/merge_config.sh linaro/configs/linaro-base.conf
linaro/configs/booting-test.conf ...
It was tested on lsk-3.10 pandaboard ES.
Signed-off-by: Alex Shi <alex.shi@linaro.org>
data toggle error without a ChannelHalt status is not a real data
toggle err, it is a transfer error state so just clear the
interrupt bit rather than resubmit this request
Signed-off-by: lyz <lyz@rock-chips.com>
between hcd_reinit() and cil_interrupt handler.
hcd_reinit() should get core_if->lock before modify this
lock, so that can prevent competition between cil_interrupt
handler and hcd_reinit(), hcd_reinit() be scheduled while
cil_interrupt handler holding core_if->lock, hcd_reinit() modify
the lock then previous core_if->lock will never be unlocked.
Signed-off-by: lyz <lyz@rock-chips.com>
changed
When otg stop host mode and before change to device mode hcd_stop
function should do something to double check all usb device is
disconnect because the disconnect interrupt might be lost.
1.kill all active urbs attached to this hcd
2.set hcd->flags to notify usb core to disconnect all usb devices
Signed-off-by: lyz <lyz@rock-chips.com>
