In most case, we use the VPLL directly for HDMI or DP, and the
the frac dclk will bring the big jitter for dclk. So we don't need
to use the dclk_vopx_frac.
Change-Id: I0d27e5fcb8b4c9a28c0102074c1d6da9426386f4
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>
The rk3399 hdmi phy is supplied by the vpll directly and needs to adapt
that frequency depending on the selected resolution on the hdmi output.
For the hdmi-phy the vpll frequency is supplied unchanged without
any dividers being present there.
The vpll also is one of the sources the general display clock of the
visual output processor (vop) and as it is somewhat special for
display operations possibly also the preferred pll source. Here a divider
is available between the pll-mux and the vop clock, so that this part
can adapt the resulting frequency if needed.
So to keep the vop clock in line with the target rate, set the newly
introduced CLK_KEEP_REQ_RATE flag for the dclk_vop clocks on rk3399.
(am from https://patchwork.kernel.org/patch/8993771/)
Change-Id: Iba9a179b764472f22d7531eb0c662dcd982433d4
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>
Given a hirarchy of clk1 -> [div] -> clk2, when the rate of clk1 gets
changed, clk2 changes as well as the divider stays the same. There may
be cases where a user of clk2 needs it at a specific rate, so clk2
needs to be readjusted for the changed rate of clk1.
So if a rate was requested for the clock, and its rate changed during
the underlying rate-change, with this change the clock framework now
tries to readjust the rate back to/near the requested one.
The whole process is protected by a new clock-flag to not force this
behaviour change onto every clock defined in the ccf.
(am from https://patchwork.kernel.org/patch/8993761/)
Change-Id: Ie2636710cb4e66815ee45b28ec86eeaaa47c55c7
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>
The req_rate property seems to be made to hold the rate requested through
clk_set_rate. Currently it gets initialized in clk_init to the clocks
current rate and then adapted in clk_set_rate calls. Orphan clocks and
their children get initialized to a rate of 0 and req_rate never gets
re-set when these lose their orphan-status.
Initializing req_rate to the clocks rate also is unintuitive as it just
copies the value that is already in the rate property and also looses the
information if a component actually requested a specific rate.
So separate the requested rate and only set it in clk_core_set_rate_nolock
when a real rate gets requested. The users of the req_rate __clk_put and
clk_set_rate_range that adjust a clock based on that value use req_rate
at first and fall back to rate if no rate had been requested now.
(am from https://patchwork.kernel.org/patch/8993751/)
Change-Id: I9e345e1e9d0c101d5c3c7064b6c2d2724f9eac4f
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>
As the pervious discussed, disabled the edp node for dts. Otherwise
you connected the mipi panel will show white since the edp will force
to use the extend screen.
Change-Id: I3e716f4a9d92d4faae5d7a12d792c71fe16f489e
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
The gpu can't work with chromeos since the regulator has not enabled.
Okay, the gpu can work for chromeos with rk3399 evb2 on now.
Change-Id: If23affff151a99206b86c2a3f40c21752bdc2d54
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
fix warning:
warning: 'ret' may be used uninitialized in this function
Change-Id: I743ed9788366322beeceddd556fae0c2a7cdd463
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
RK3399 hdmi register layout is similar with rk3288 and rk3368, so most
code can reuse. but hdmi resign is upgrade from 20 to 21 so it has a
little diffrence. and the hdmi phy clk is Independent from dclk too.
Change-Id: I83b30c92d9572fc9ceaf52777d224e5cec1823be
Signed-off-by: xuhuicong <xhc@rock-chips.com>
fix up the warning:
drivers/regulator/lp8752.c: In function 'lp8752_buck_set_mode':
drivers/regulator/lp8752.c:93:2: warning: 'ret' may be used uninitialized in this function [-Wmaybe-uninitialized]
Change-Id: Iee9f69791bbcea2e6b3a16713b76e93cfc0a2b67
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Seperate power-on timer for rkvdec and vpu when they are not
in combo mode (definitly independence).
Before this patch, h/w power off schedule will be interrupted
by the other h/w, and power off will be triggered when h/w
still running a task.
Change-Id: I29124e90afccc727d2e7a04098727aa4a2c3e8bb
Signed-off-by: Alpha Lin <alpha.lin@rock-chips.com>
It will enable the devfreq thermal that's generic devfreq cooling
mechanism through frequency reduction for devices using devfreq.
This will throttle the device by limiting the maximum allowed DVFS
frequency corresponding to the cooling level.
Change-Id: Ia017ecf46599700382b9604e375193135f7d1d24
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
Let's use unlock/lock around phy APIs as them will call
mutex which is sleepable casuing failure of kernel debug
check.
Change-Id: Ic7670bfc9ed763cc9bdec53f85f553bc0be1416c
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
rgb/edp/hdmi/mipi pin_pol is removed after (e5683dd FROMLIST:
drm/rockchip: get rid of rockchip_drm_crtc_mode_config), that
is wrong.
This patch re-add those pin_pol config.
Change-Id: I46f3e32ad405f4b6e2f76110757248e8516693c4
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
enable_time is controlled by capacitance connected to Pin SS,
it described more detailedly in mp8865 datasheet Page 15.
Capacitance is 12nF now.
Change-Id: Ib604a4e109db7ab125104e5cf3067864fefb6fe0
Signed-off-by: Zain Wang <wzz@rock-chips.com>
There is a path that use vskiplines with non-initialize.
That would cause vop abnormal behavior.
Change-Id: I53c6c575d6acc16aeae761dbb4867f3bc8bfe5ce
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
On Async atomic_commit callback, drm_atomic_clean_old_fb will
clean all old fb, but because async, the old fb may be also on
the vop hardware, dma will access the old fb buffer, clean old
fb will cause iommu page fault.
Reference the fb and unreference it when the fb actuall swap out
from vop hardware.
Change-Id: I585786884295060efdaef0a00c3cbd75244399d7
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
The series vop of VOP_FULL framework support area plane, such as
RK3288 and RK3399, one group of area planes share same hardware,
reuse the hardware on different scanout time, this design is
useful to support mulit planes with low hardware consume.
Change-Id: Ie53211ce9ed22d03f7668637efbb7c95d9a8eb5b
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
The plane hardware is used when the display scanout run into plane active
scanout, that means we can reuse the plane hardware resources on plane
non-active scanout.
Because resource share, There are some limit on share plane: one group
of share planes need use same zpos, can't not overlap, etc.
We assume share plane is a universal plane with some limit flags.
people who use the share plane need know the limit, should call the ioctl
DRM_CLIENT_CAP_SHARE_PLANES, and judge the planes limit before use it.
Change-Id: Iecc3d8e7f1ce29d567cdbad689ba4dbad3d594e1
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
We need to take care of the vop status when use
rockchip_drm_crtc_mode_config, if vop is disabled,
the function would failed, that is terrible.
Save output_type and output_mode into rockchip_crtc_state,
it's nice to make them into atomic.
Conflicts:
drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
drivers/gpu/drm/rockchip/dw-mipi-dsi.c
drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c
drivers/gpu/drm/rockchip/inno_hdmi.c
drivers/gpu/drm/rockchip/rockchip_drm_drv.h
drivers/gpu/drm/rockchip/rockchip_drm_vop.c
Change-Id: I43c49a92b2b9df02ce8a055bd16948b400ab0f47
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
(am from https://patchwork.kernel.org/patch/8844321/)
On the RK3368 SoC, support the APB timers for rockchip platform.
Change-Id: I2bee09c4140994d3d2e23f1820663230d82547de
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
(cherry picked from commit c840f28bbf)
Currently rockchip_timer doesn't do some basic cleanup work when
failing to init the timer. Let's add err handle routine to deal
with all the err cases.
Change-Id: I73bbd32592e6fe157a8d166743db3fc130d0004c
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
(cherry picked from commit 522ed95c26)
Increase the compilation test coverage by adding the COMPILE_TEST option.
Due to the dsb() usage in the driver, this driver is only compilable on
ARM and ARM64.
Change-Id: I5f2c1a5353a7b20c80dcfc3cd3900510f56a0729
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Huang, Tao <huangtao@rock-chips.com>
(cherry picked from commit 40ada2aac5)
The dsb() instruction is pointless in this code.
Remove it.
That also fixes the ARM64 compilation issue.
Change-Id: I0c8e33abe0d976714f4df288fe5ac52ffb8ded5b
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Tested-by: Caesar Wang <wxt@rock-chips.com>
Signed-off-by: Huang, Tao <huangtao@rock-chips.com>
(cherry picked from commit 23b8f81f38)
Let's checkstyle to clean up the macros with such trivial details.
Change-Id: I6cf0c7cf5e48bcb4d52a483fdba7c4ce26677f06
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
(cherry picked from commit a0d2216ec0)
Fix the incorrect critial temperture for gpu thermal.
Change-Id: I9ecfc107afcdbb421ae40cc796c40a39d6d68677
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
This patch enables mmc-hs400-enhanced-strobe for rk3399-evb, so
enhanced strobe function will be used if any eMMC 5.1 is probed.
Change-Id: If5f30e0d759f7a9850bec82c3d53d9bb26ba8c3d
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
HS400 enhanced strobe is a new feature introduced by eMMC
spec 5.1, let's implement it and enjoy it!
please note that currently I have no much bandwith to split this
big patch into patchset. So please use, test and applied! Thanks.
Change-Id: I874f18a617a1b69e3ff56f5c134feb817b6985b9
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Let's add some description of mmc-hs400-enhanced-strobe
which can be used to support hs400 enhanced strobe function
introduced by eMMC spec 5.1
Change-Id: I03b8e803071dc7034bddf655892b12eabcaa852a
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
If sock is request_sock then kernel will crash. So use
skb_to_full_sk() and sk_fullsock() helper to make sure
we get full sock.
Change-Id: Iefd548e0591055b1a8031f0835c4dca7b9d42b61
Signed-off-by: Huang, Tao <huangtao@rock-chips.com>
