Only do the i2c configs at the first time or after entering suspend
mode, and add mutex protect during suspend and resume.
Signed-off-by: Wenping Zhang <wenping.zhang@rock-chips.com>
Change-Id: Id3e16da88c526332238682421976a43c5974c936
When clear standby bit, the vop post will start scan out
with the old timing immediately(no need config done bit).
Sometime the old timing is verery different with the
current timing we want config(switch from 4K to 720P),
this may cause problems at some point.
So we clear standby after a new timing config done.
Change-Id: Id621bea6a622c151d3e339b102c6791a6244ac08
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Sometimes isr is disabled, wait for frame start
by poll vcnt is more safe.
Change-Id: I0f3d63e1d51a623d22eedef05920bcff5d63a75d
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
We have move p30x1 and p30x2 reset signal to corresponding
controller drivers. So we need to remove them in PHY driver
in order not to add misleading debug message.
Change-Id: I5f180bf51a449ce2db6c49f937d3898a61d7c4d9
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
If it fails to do dwc3_drd_init() during probe, call the
dwc3_gadget_exit() to release the gadget resource when
the flag dwc->en_runtime is enabled.
Signed-off-by: William Wu <william.wu@rock-chips.com>
Change-Id: I12903841db4d83a7811b7e05d0a950932b4d3085
set Ad0" = Ad0_" to instead of Ad0" = Ad0_" + (Ad0_" >> 7), than VOP
some alpha result is same with GPU output.
Change-Id: If05f561c16b4eb6e46671f4a130fae0fe50191c2
Signed-off-by: Sandy Huang <hjc@rock-chips.com>
Set wb_axi_yrgb_id to 0xd
set wb_axi_uv_id to 0xe
Change-Id: Ie8188c2c307d2167ea237b4a176b03cfdab73792
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Some platforms (such as Eink) during suspend the current
is very low (low than 2Ma), if long time to sleep, it will
cause the dsoc to drain faster.
Signed-off-by: Shunqing Chen <csq@rock-chips.com>
Change-Id: I25330a3a79f9684487bea0f523d97d61b8098e8a
We should not disable it before L2, otherwise all L2 transaction
event would be lost for RC side, leading to unstable L2 link stage.
Change-Id: Ia87f8749d4a0acc9aeb1215682209c853672a84d
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Fixed the error that color fill would cause "req argument is inval".
Signed-off-by: Yu Qiaowei <cerf.yu@rock-chips.com>
Change-Id: I94ebfc2061f16916813a0d85630f8a9862728b0e
Because update_patten_buff_mode is not currently used,
it is removed.
Signed-off-by: Yu Qiaowei <cerf.yu@rock-chips.com>
Change-Id: I48c97e75d27d475eee528ed6c838c67fdaa430f8
We move all the probe stuff to kthread so that it won't block
the system to go on probing other drivers. But that introduced
it bug that PM calls would not be removed by driver core. We
trace back the platform driver to add PM callbacks if we really
detect active devices.
Fixes: 79ac46bdea ("PCI: rockchip: dw: Add kthread to probe PCIe devices")
Change-Id: I93a331fe6c17d97b71010780fba4afbc2dbf6216
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
When get dte_addr for iommu, it must enable reference clock.
Change-Id: Iaec423345463e001b2a7099067ab1e4a2d971f7f
Signed-off-by: Ding Wei <leo.ding@rock-chips.com>
This reverts commit 18a4299e0b.
move to rk356x_eink.config on android which one will overlay kernel configs
Signed-off-by: Zorro Liu <lyx@rock-chips.com>
Change-Id: If43d748cdef541915a218214c274375a2d8e34bd
