Re: [PATCH v6 00/14] Add some DRM bridge drivers support for i.MX8qm/qxp SoCs
From: Liu Ying
Date: Mon Mar 29 2021 - 04:13:28 EST
Hi Marcel,
On Mon, 2021-03-29 at 00:49 +0000, Marcel Ziswiler wrote:
> Hi Liu
>
> On Tue, 2021-03-23 at 17:09 +0800, Liu Ying wrote:
> > On Tue, 2021-03-23 at 01:03 +0000, Marcel Ziswiler wrote:
> > > Hi Liu
> > >
> > > Some further discrepancy with them binding examples:
> > >
> > > arch/arm64/boot/dts/freescale/imx8qxp.dtsi:335.9-36: Warning (reg_format): /dpu@56180000:reg: property has
> > > invalid length (8 bytes) (#address-cells == 2, #size-cells == 2)
> > > arch/arm64/boot/dts/freescale/imx8qxp.dtsi:508.9-35: Warning (reg_format): /syscon@56221000:reg: property has
> > > invalid length (8 bytes) (#address-cells == 2, #size-cells == 2)
> > > arch/arm64/boot/dts/freescale/imx8qxp.dtsi:601.9-34: Warning (reg_format): /phy@56228300:reg: property has
> > > invalid length (8 bytes) (#address-cells == 2, #size-cells == 2)
> > > arch/arm64/boot/dts/freescale/imx8qxp.dtsi:613.9-36: Warning (reg_format): /pixel-combiner@56020000:reg:
> > > property has invalid length (8 bytes) (#address-cells == 2, #size-cells == 2)
> > >
> > > And with that I am unable to bring it up:
> > >
> > > [ 1.714498] imx8qxp-ldb 5622100000001000.syscon:ldb: [drm:ldb_init_helper] *ERROR* failed to get regmap: -
> > > 12
> > > [ 1.724441] imx8qxp-ldb: probe of 5622100000001000.syscon:ldb failed with error -12
> > > [ 1.734983] imx8qxp-pixel-combiner 5602000000010000.pixel-combiner: invalid resource
> > > [ 1.742830] imx8qxp-pixel-combiner: probe of 5602000000010000.pixel-combiner failed with error -22
> > > [ 1.754040] imx8qxp-display-pixel-link dc0-pixel-link0: [drm:imx8qxp_pixel_link_bridge_probe] *ERROR*
> > > failed
> > > to get pixel link node alias id: -19
> > > [ 1.769626] imx8qxp-pxl2dpi 5622100000001000.syscon:pxl2dpi: [drm:imx8qxp_pxl2dpi_bridge_probe] *ERROR*
> > > failed to get regmap: -12
> > > [ 1.781397] imx8qxp-pxl2dpi: probe of 5622100000001000.syscon:pxl2dpi failed with error -12
> > > [ 1.840547] imx8qxp-lpcg-clk 59580000.clock-controller: deferred probe timeout, ignoring dependency
> > > [ 1.840571] imx8qxp-lpcg-clk: probe of 59580000.clock-controller failed with error -110
> > >
> > > Any suggestions welcome. Thanks!
> >
> > Please reference the patch set I shared in my last reply and see how it
> > goes. Thanks.
>
> Thank you very much. After a little bit of fiddling I can confirm that this also works fine on a Toradex
> Colibri iMX8X [1] with either a Capacitive Touch Display 10.1" LVDS which has a Logic Technologies LT170410-
> 2WHC [2] single-channel panel inside or a dual-channel LG LP156WF1 full HD panel.
Thanks a lot for your testing!
Glad to know that the two LVDS panels work on Toradex Colibri iMX8X.
>
> During boot I noticed quite some clocking/power domain related messages:
>
> [ 0.537965] gpt0_clk: failed to attached the power domain -2
>
> [ 0.562372] dc1_disp0_clk: failed to attached the power domain -2
> [ 0.562800] dc1_disp0_clk: failed to get clock parent -22
> [ 0.562858] dc1_disp0_clk: failed to get clock rate -22
>
> [ 0.563059] dc1_disp1_clk: failed to attached the power domain -2
> [ 0.563463] dc1_disp1_clk: failed to get clock parent -22
> [ 0.563514] dc1_disp1_clk: failed to get clock rate -22
>
> [ 0.563773] dc1_pll0_clk: failed to attached the power domain -2
> [ 0.564174] dc1_pll0_clk: failed to get clock rate -22
>
> [ 0.564413] dc1_pll1_clk: failed to attached the power domain -2
> [ 0.564838] dc1_pll1_clk: failed to get clock rate -22
>
> [ 0.565099] dc1_bypass0_clk: failed to attached the power domain -2
> [ 0.565516] dc1_bypass0_clk: failed to get clock rate -22
>
> [ 0.565755] dc1_bypass1_clk: failed to attached the power domain -2
> [ 0.566159] dc1_bypass1_clk: failed to get clock rate -22
>
> [ 0.574493] lvds0_i2c0_clk: failed to attached the power domain -2
> [ 0.574894] lvds0_i2c0_clk: failed to get clock rate -22
>
> [ 0.575134] lvds0_i2c1_clk: failed to attached the power domain -2
> [ 0.575526] lvds0_i2c1_clk: failed to get clock rate -22
>
> [ 0.575785] lvds0_pwm0_clk: failed to attached the power domain -2
> [ 0.576189] lvds0_pwm0_clk: failed to get clock rate -22
>
> [ 0.576417] lvds1_i2c0_clk: failed to attached the power domain -2
> [ 0.576854] lvds1_i2c0_clk: failed to get clock rate -22
>
> [ 0.577129] lvds1_i2c1_clk: failed to attached the power domain -2
> [ 0.577554] lvds1_i2c1_clk: failed to get clock rate -22
>
> [ 0.577787] lvds1_pwm0_clk: failed to attached the power domain -2
> [ 0.578198] lvds1_pwm0_clk: failed to get clock rate -22
>
> [ 0.578464] mipi_csi0_core_clk: failed to attached the power domain -2
>
> [ 0.579104] mipi_csi0_esc_clk: failed to attached the power domain -2
>
> [ 0.579738] mipi_csi0_i2c0_clk: failed to attached the power domain -2
>
> [ 0.580368] mipi_csi0_pwm0_clk: failed to attached the power domain -2
>
> And the following repeats a couple dozens of times:
>
> [ 4.391495] dc1_disp0_clk: failed to get clock parent -22
> [ 4.398532] dc1_disp1_clk: failed to get clock parent -22
As I mentioned before, there will be logs like 'dc1_disp0_clk: failed
to get clock parent -22' on i.MX8qxp and i.MX8qm/qxp specific clocks
are not split yet. DC1 and LVDS0/1 are i.MX8qm specific. So, once they
are split up, I assume there won't be those logs any more.
If you don't apply the below two patches for i.MX8qm, then dc1 and
lvds0/1 relevant logs won't come. That doesn't impact the i.MX8qxp
displays.
clk: imx: clk-imx8qxp: Add I2C and PWM SCU clocks in LVDS0/1 subsystems
clk: imx: clk-imx8qxp: Add some clocks for i.MX8qm DC1 subsystem
The latest Shawn's for-next branch also generates the gpt and mipi_csi
relevant logs on my i.MX8qxp MEK board like below. So, they are not
related to my patch set.
dmesg | grep clk
[ 1.091534] gpt0_clk: failed to attached the power domain -2
[ 1.133131] mipi_csi0_core_clk: failed to attached the power domain
-2
[ 1.139849] mipi_csi0_esc_clk: failed to attached the power domain
-2
[ 1.146441] mipi_csi0_i2c0_clk: failed to attached the power domain
-2
[ 1.153312] mipi_csi0_pwm0_clk: failed to attached the power domain
-2
On my i.MX8qm MEK board, the latest Shawn's for-next branch behaves
like this:
dmesg | grep clk
[ 0.222517] a35_clk: failed to get clock rate -22
[ 0.225331] gpt0_clk: failed to attached the power domain -2
[ 0.232859] pwm_clk: failed to attached the power domain -2
[ 0.233085] pwm_clk: failed to get clock rate -22
[ 0.233158] lcd_clk: failed to attached the power domain -2
[ 0.233382] lcd_clk: failed to get clock rate -22
[ 0.246576] mipi_csi0_core_clk: failed to attached the power domain
-2
[ 0.246899] mipi_csi0_esc_clk: failed to attached the power domain
-2
[ 0.247218] mipi_csi0_i2c0_clk: failed to attached the power domain
-2
[ 0.247515] mipi_csi0_pwm0_clk: failed to attached the power domain
-2
[ 1.510195] imx8qxp-lpcg-clk 5a4a0000.clock-controller: deferred
probe timeout, ignoring dependency
[ 1.521361] imx8qxp-lpcg-clk: probe of 5a4a0000.clock-controller
failed with error -110
@Aisheng, it looks like we'd better to suppress those warning logs soon
by splitting i.MX8qm/qxp specific clocks up?
> And finally it spits the following:
>
> [ 4.670303] imx8qxp-lpcg-clk 59580000.clock-controller: deferred probe timeout, ignoring dependency
> [ 4.679629] imx8qxp-lpcg-clk: probe of 59580000.clock-controller failed with error -110
I don't see this on my i.MX8qxp MEK board. It looks like it's related
to the 'dsp_lpcg: clock-controller@59580000' node in
imx8-ss-audio.dtsi. Does this reproduce with Shawn's for-next
branch(without my patch set) for you?
>
> Despite those messages the displays do work fine once booted.
>
> I am currently running this with SCFW, SECO, TF-A and U-Boot based off NXP's latest downstream BSP 5.4.70-
> 2.3.0. Not sure whether or not especially the used SCFW version could cause some issues. What SCFW are you
> using?
>
> Full boot logs may be found here [3].
>
> You may add the following to the whole series.
>
> Tested-by: Marcel Ziswiler <marcel.ziswiler@xxxxxxxxxxx> # Colibri iMX8X, LT170410-2WHC, LP156WF1
Thanks for your tag.
>
> Thanks again and just let us know if we may test anything else for you.
Maybe, any Toradex i.MX8qm board with LVDS display, please?
Regards,
Liu Ying
>
> [1] commit ba5a5615d54f ("arm64: dts: freescale: add initial support for colibri imx8x")
> [2] commit 5728fe7fa539 ("drm/panel: simple: add display timings for logic technologies displays")
> [3] https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fshare.toradex.com%2Fs30wwspcr9iwyrg&data=04%7C01%7Cvictor.liu%40nxp.com%7Ca51f827482704468d27208d8f24c75dd%7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C637525757505899800%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=VziEUvn3pnk7QvgIl58CuM8VHMqm6Y5xOuod4ali1Zk%3D&reserved=0
>
> > Liu Ying
>
> Cheers
>
> Marcel