linux-toradex.git/drivers/phy/rockchip, branch v6.7

phy: rockchip-inno-usb2: Drop unnecessary DT includes

2023-10-13T10:16:09+00:00

Neither of_address.h or of_platform.h are used in the driver, so drop them.

Signed-off-by: Rob Herring 
Reviewed-by: Heiko Stuebner 
Link: https://lore.kernel.org/r/20231009172923.2457844-17-robh@kernel.org
Signed-off-by: Vinod Koul

phy: Use device_get_match_data()

2023-10-13T10:15:40+00:00

Use preferred device_get_match_data() instead of of_match_device() to
get the driver match data. With this, adjust the includes to explicitly
include the correct headers.

Signed-off-by: Rob Herring 
Reviewed-by: Heiko Stuebner 
Link: https://lore.kernel.org/r/20231009172923.2457844-15-robh@kernel.org
Signed-off-by: Vinod Koul

phy: rockchip: inno-dsidphy: Add rv1126 support

2023-08-22T13:58:11+00:00

Add support for Rockchip RV1126 DSI-DPHY.

The existing 2.5GHz phy timing table added for RK3568 is working
as it is for RV1126 as well.

Signed-off-by: Jagan Teki 
Reviewed-by: Heiko Stuebner 
Link: https://lore.kernel.org/r/20230731110012.2913742-5-jagan@edgeble.ai
Signed-off-by: Vinod Koul

phy: Explicitly include correct DT includes

2023-07-17T06:22:56+00:00

The DT of_device.h and of_platform.h date back to the separate
of_platform_bus_type before it as merged into the regular platform bus.
As part of that merge prepping Arm DT support 13 years ago, they
"temporarily" include each other. They also include platform_device.h
and of.h. As a result, there's a pretty much random mix of those include
files used throughout the tree. In order to detangle these headers and
replace the implicit includes with struct declarations, users need to
explicitly include the correct includes.

Signed-off-by: Rob Herring 
Acked-by: Marc Kleine-Budde  # for drivers/phy/phy-can-transceiver.c
Acked-by: Heiko Stuebner 
Acked-by: Sergio Paracuellos 
Link: https://lore.kernel.org/r/20230714174841.4061919-1-robh@kernel.org
Signed-off-by: Vinod Koul

phy/rockchip: inno-hdmi: add more supported pre-pll rates

2023-07-12T16:57:43+00:00

This adds a bunch of new pixel clock- and tmds rates to the pre-pll
table which are required to get more VESA and some DMT rates working.

It has been completely re-calculated to match the min- and max-vco of
(750 MHz - 3.2 GHz) requirements. If more than one configuration would
have been possible the lowest fbdiv and refdiv (and therefore lowest
vco rate) has been preferred.

It's important to note, that RK3228 version of the phy does not support
fractional dividers. To support the most possible rates for this version
also in both 8-bit and 10-bit variant, some rates are not exact. The
maximum deviation of the pixel clock is 0.26, which perfectly fits into
VESA DMT recommendation of 0.5%.

I tested all possible rates on several screens from different
manufacturers with both RK3228 and RK3328. Both pre- and post-PLL
locking are slighlty faster now.

Signed-off-by: Alex Bee 
Signed-off-by: Jonas Karlman 
Link: https://lore.kernel.org/r/20230615171005.2251032-7-jonas@kwiboo.se
Signed-off-by: Vinod Koul

phy/rockchip: inno-hdmi: force set_rate on power_on

2023-07-12T16:57:43+00:00

Regular 8-bit and Deep Color video formats mainly differ in TMDS rate and
not in pixel clock rate.
When the hdmiphy clock is configured with the same pixel clock rate using
clk_set_rate() the clock framework do not signal the hdmi phy driver
to set_rate when switching between 8-bit and Deep Color.
This result in pre/post pll not being re-configured when switching between
regular 8-bit and Deep Color video formats.

Fix this by calling set_rate in power_on to force pre pll re-configuration.

Signed-off-by: Huicong Xu 
Signed-off-by: Jonas Karlman 
Link: https://lore.kernel.org/r/20230615171005.2251032-6-jonas@kwiboo.se
Signed-off-by: Vinod Koul

phy/rockchip: inno-hdmi: do not power on rk3328 post pll on reg write

2023-07-12T16:57:43+00:00

inno_write is used to configure 0xaa reg, that also hold the
POST_PLL_POWER_DOWN bit.
When POST_PLL_REFCLK_SEL_TMDS is configured the power down bit is not
taken into consideration.

Fix this by keeping the power down bit until configuration is complete.
Also reorder the reg write order for consistency.

Fixes: 53706a116863 ("phy: add Rockchip Innosilicon hdmi phy")
Signed-off-by: Jonas Karlman 
Link: https://lore.kernel.org/r/20230615171005.2251032-5-jonas@kwiboo.se
Signed-off-by: Vinod Koul

phy/rockchip: inno-hdmi: remove unused no_c from rk3328 recalc_rate

2023-07-12T16:57:43+00:00

no_c is not used in any calculation, lets remove it.

Signed-off-by: Jonas Karlman 
Link: https://lore.kernel.org/r/20230615171005.2251032-4-jonas@kwiboo.se
Signed-off-by: Vinod Koul

phy/rockchip: inno-hdmi: round fractal pixclock in rk3328 recalc_rate

2023-07-12T16:57:43+00:00

inno_hdmi_phy_rk3328_clk_recalc_rate() is returning a rate not found
in the pre pll config table when the fractal divider is used.
This can prevent proper power_on because a tmdsclock for the new rate
is not found in the pre pll config table.

Fix this by saving and returning a rounded pixel rate that exist
in the pre pll config table.

Fixes: 53706a116863 ("phy: add Rockchip Innosilicon hdmi phy")
Signed-off-by: Zheng Yang 
Signed-off-by: Jonas Karlman 
Link: https://lore.kernel.org/r/20230615171005.2251032-3-jonas@kwiboo.se
Signed-off-by: Vinod Koul

phy/rockchip: inno-hdmi: use correct vco_div_5 macro on rk3328

2023-07-12T16:57:43+00:00

inno_hdmi_phy_rk3328_clk_set_rate() is using the RK3228 macro
when configuring vco_div_5 on RK3328.

Fix this by using correct vco_div_5 macro for RK3328.

Fixes: 53706a116863 ("phy: add Rockchip Innosilicon hdmi phy")
Signed-off-by: Jonas Karlman 
Link: https://lore.kernel.org/r/20230615171005.2251032-2-jonas@kwiboo.se
Signed-off-by: Vinod Koul