HDMI Display does not work after update to JetPack 5.0.2

nkpyq · August 31, 2022, 6:28am

step1. No HDMI display after startup.
step2. Unplug HDMI connector, execute xrandr, and then connect HDMI, the monitor displays with a resolution of 1024 * 768.
step3. Execute xrandr again, then the monitor displays a black screen.
The following is the relevant information. Why does HDMI fail to display normally?
Thanks.

log
boot.log (93.0 KB)
dmesg.log (68.5 KB)
Xorg.0.log (20.9 KB)
$ lsmod | grep nv

nvidia_modeset       1093632  3
nvgpu                2793472  20
nvadsp                110592  1 snd_soc_tegra210_adsp
nvidia               1339392  7 nvidia_modeset
nvmap                 192512  76 nvgpu

3.hdmi dcb

./dcb_tool -r tegra234-dcb-p3701-0000-a02-p3737-0000-a01_hdmi.dtsi 
=== Reading DCB blob ===

########## Tegra DCB BLOB ###############
########### Display Devices #############
Display Devices::
	Display Devices : [0]
		Type               : [TMDS]
		CCB                : [0]
		Heads              : 0:[Y] 1:[Y]
		Sor                : [0 ]
		HDMI capable       : [1]
		Connector          : [0]
		Bus                : [0]
############### CCB Entries ###############
CCB::
*CCB entries that have both I2C and AUX ports unused (value = 31) are not displayed
	CCB Index : 0
		I2C Port           : [6]
		AUX Port           : [0]
########### Connector entries #############
Connectors::
	Connector Index : 0x0
		Type               : [HDMI]
		Hotplug            : A:[Y]

################# ******* #################

WayneWWW · September 1, 2022, 3:02am

There is no software change in this part. You better checking the full procedure to enable the HDMI port again.

nkpyq · September 19, 2022, 3:38am

hi, WayneWWW

I tried the following two methods, but the problem still exists with JetPack 5.0.2GA.

https://docs.nvidia.com/jetson/archives/r34.1/DeveloperGuide/text/SD/Kernel/DisplayConfigurationAndBringUp.html
tegra234-dcb-p3701-0000-a02-p3737-0000-a01_hdmi.dtsi (27.5 KB)
This works well with JetPack 5.0.1.
https://docs.nvidia.com/jetson/archives/r35.1/DeveloperGuide/text/SD/Kernel/DisplayConfigurationAndBringUp/OrinDisplayconfig.html
Refer to the <kernel-source-path>/nvidia/Documentation/devicetree/bindings/video/nvidia,tegra234-display.txt, i regenerate the following hdmi.dtsi (differ with method 1).
tegra234-dcb-p3701-0000-a02-p3737-0000-a01.dtsi-modified.dtsi (26.2 KB)

Both methods can be displayed by the following steps

But i found the color displayed was wrong, just like this post.

WayneWWW · September 19, 2022, 3:51am

Hi,

在這邊先澄清一下, 如果你的hdmi 是畫面已經出來了. 但是顏色不對或是有雜訊, 那就已經不是device tree相關的問題. 所以麻煩澄清一下現在到底是怎樣的問題？

Just want to clarify this first. If your HDMI monitor already render the desktop but with wrong color or noise, then it may not be the device tree issue.

So, please clarify what is your exact question.

nkpyq · September 19, 2022, 6:27am

Sorry, let me summarize the whole process.

The hdmi from our custom carrier board work normally, when using Jetpack 5.0.1 with the patch below. The monitor displays boot logo, ubuntu desktop.
https://docs.nvidia.com/jetson/archives/r34.1/DeveloperGuide/text/SD/Kernel/DisplayConfigurationAndBringUp.html
Then, the hdmi does not work after update to Jeptpack 5.0.2 with the patch as above. The monitor displays “No video input”.
The monitor displays ubuntu desktop only when I perform the following actions. If you execute the xrandr command again after “step2”, the monitor turns to black screen, and displays “No video input” again.

Later, I find that the color display is abnormal when the monitor displays ubuntu desktop, just like this post.

I reflash Jetpack 5.0.1, the monitor displays boot logo, ubuntu desktop.
I follow the instruction below(use dcbtool), and reflash Jetpack 5.0.2, the monitor displays “No video input”.

WayneWWW · September 19, 2022, 6:34am

Could you check if your jetpack5.0.1 and jetpack5.0.2 are using the same dcb setting in the device tree?

nkpyq · September 19, 2022, 6:46am

I’ve already confirmed jetpack5.0.1 and jetpack5.0.2 are using the same dcb setting.

I will check again, upload boot log, and hexdump /proc/device-tree/display@13800000/nvidia,dcb-image.

WayneWWW · September 19, 2022, 7:20am

Hi,

想再跟您確認一下, 請問目前 jp5.0.2到底在哪個狀況或步驟下, 你能看到ubuntu桌面？雜訊或是顏色錯誤先不討論.

Just want to double confirm with you. What is the exact step or situation that you could see ubuntu desktop under jp5.0.2? The color issue or noise is not under discussion for now.

nkpyq · September 19, 2022, 7:28am

Step1. 連接HDMI線啓動，此時無任何顯示；
Step2. 拔掉HDMI線，執行如下命令後，再連接上HDMI線。
export DISPLAY=:0
xrandr

nkpyq · September 19, 2022, 7:34am

按照以上步驟可顯示ubuntu桌面，但是分辨率不是1080p，而是1024 * 768。

WayneWWW · September 19, 2022, 8:10am

Are you saying that the desktop immediately show up just after you run “xrandr” command?
We have a new dcb setting here. Could you check if this is same as what you are using?
Custom board HDMI hotplug detect(HPD) did't work - #13 by WayneWWW

nkpyq · September 19, 2022, 9:17am

Yes
The logs of Step1 & Step2 are as follows
dmesg:

[  686.886374] snd_hda_codec_hdmi hdaudioC0D0: HDMI: invalid ELD data byte 1
[  686.886390] snd_hda_codec_hdmi hdaudioC0D0: HDMI: out of range MNL 13
[  984.579941] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
[  984.579965] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device Philips PHL 271E1 (HDMI-0)
[  984.656689] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
[  984.656700] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0
[  985.201024] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
[  985.201042] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0
[  985.265995] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
[  985.266017] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0

Xorg.log:

[   984.061] (--) NVIDIA(GPU-0): DFP-0: connected
[   984.061] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[   984.061] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[   984.061] (--) NVIDIA(GPU-0): 
[   984.101] (--) NVIDIA(GPU-0): DFP-0: connected
[   984.101] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[   984.101] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[   984.101] (--) NVIDIA(GPU-0): 
[   984.221] (II) NVIDIA(0): Setting mode "NULL"
[   984.321] (II) NVIDIA(0): Setting mode "HDMI-0: nvidia-auto-select @1024x768 +0+0 {ViewPortIn=1024x768, ViewPortOut=1024x768+0+0}"
[   984.645] (--) NVIDIA(GPU-0): DFP-0: connected
[   984.645] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[   984.646] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[   984.646] (--) NVIDIA(GPU-0): 
[   984.710] (--) NVIDIA(GPU-0): DFP-0: connected
[   984.711] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[   984.711] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[   984.711] (--) NVIDIA(GPU-0): 
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1920x1080_75" could not be re-validated
[  1033.649] (WW) NVIDIA(0):     against the current hardware configuration; removing.
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1920x1080_60_0" could not be re-validated
[  1033.649] (WW) NVIDIA(0):     against the current hardware configuration; removing.
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1920x1080_50" could not be re-validated
[  1033.649] (WW) NVIDIA(0):     against the current hardware configuration; removing.
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1680x1050" could not be re-validated against
[  1033.649] (WW) NVIDIA(0):     the current hardware configuration; removing.
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1440x900" could not be re-validated against
[  1033.649] (WW) NVIDIA(0):     the current hardware configuration; removing.
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1280x1024" could not be re-validated against
[  1033.649] (WW) NVIDIA(0):     the current hardware configuration; removing.
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1280x1024_60" could not be re-validated
[  1033.649] (WW) NVIDIA(0):     against the current hardware configuration; removing.
[  1033.649] (WW) NVIDIA(0): MetaMode "HDMI-0:1280x960" could not be re-validated against
[  1033.649] (WW) NVIDIA(0):     the current hardware configuration; removing.
[  1033.687] (WW) NVIDIA(0): MetaMode "HDMI-0:1280x720_50" could not be re-validated
[  1033.687] (WW) NVIDIA(0):     against the current hardware configuration; removing.
[  1033.717] (WW) NVIDIA(0): MetaMode "HDMI-0:1024x768_70" could not be re-validated
[  1033.717] (WW) NVIDIA(0):     against the current hardware configuration; removing.
[  1033.840] (WW) NVIDIA(0): MetaMode "HDMI-0:720x576" could not be re-validated against
[  1033.840] (WW) NVIDIA(0):     the current hardware configuration; removing.
[  1033.840] (WW) NVIDIA(0): MetaMode "HDMI-0:720x480" could not be re-validated against
[  1033.840] (WW) NVIDIA(0):     the current hardware configuration; removing.
[  1033.862] (WW) NVIDIA(0): MetaMode "HDMI-0:640x480_72" could not be re-validated against
[  1033.862] (WW) NVIDIA(0):     the current hardware configuration; removing.
[  1033.884] (WW) NVIDIA(0): MetaMode "HDMI-0:640x480_60_0" could not be re-validated
[  1033.884] (WW) NVIDIA(0):     against the current hardware configuration; removing.
[  1033.884] (WW) NVIDIA(0): custom viewPortOut, 1920 x 1079 +0 +0, does not fit within
[  1033.884] (WW) NVIDIA(0):     rasterSize, 1024x768; ignoring custom viewPortOut
[  1033.936] (WW) NVIDIA(0): custom viewPortOut, 1728 x 1080 +96 +0, does not fit within
[  1033.936] (WW) NVIDIA(0):     rasterSize, 1024x768; ignoring custom viewPortOut

It’s different from the one I’m using, but the problem is the same.

The following are logs of jp5.0.1/jp5.0.2, or different dcb setting.
(1) jp5.0.2 with new dcb setting from below.

boot_jp5_0_2_new.log (93.1 KB)
Xorg_jp5_0_2_new.0.log (19.2 KB)
dcb-image_jp5_0_2_new.txt (38.1 KB)

(2) jp5.0.2 with dcb setting from below.

boot_jp5_0_2.log (93.0 KB)
Xorg_jp5_0_2.0.log (19.2 KB)
dcb-image_jp5_0_2.txt (38.6 KB)

(3) jp5.0.1, dcb setting is the same as (2).
boot_jp5_0_1.log (88.2 KB)
Xorg_jp5_0_1.0.log (19.2 KB)
dcb-image_jp5_0_2.txt (38.6 KB)

WayneWWW · September 19, 2022, 9:23am

Hi,

Are you sure your EDID of this monitor is able to get read when it is jetpack5.0.1?
If you only see a hardcoded mode, it means EDID fails to read so we give you a hardcoded value.

[  984.579965] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device Philips PHL 271E1 (HDMI-0)
[  984.656689] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
[  984.656700] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0
[  985.201024] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
[  985.201042] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0
[  985.265995] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
[  985.266017] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0

nkpyq · September 19, 2022, 10:18am

Hi, WayneWWW

The logs of Step1 & Step2 are as follows

dmesg:

Sep 19 10:10:09 nvidia-agx-orin kernel: [ 1510.530119] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:10:40 nvidia-agx-orin kernel: [ 1541.303535] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:10:40 nvidia-agx-orin kernel: [ 1541.331774] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:10:40 nvidia-agx-orin kernel: [ 1541.332714] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:10:41 nvidia-agx-orin kernel: [ 1542.879639] snd_hda_codec_hdmi hdaudioC0D0: HDMI: invalid ELD data byte 5

Sep 19 10:10:41 nvidia-agx-orin kernel: [ 1542.879649] snd_hda_codec_hdmi hdaudioC0D0: HDMI: out of range MNL 13

Sep 19 10:12:39 nvidia-agx-orin kernel: [ 1661.074561] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call result 0xffff:

Sep 19 10:12:39 nvidia-agx-orin kernel: [ 1661.074574] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device Philips PHL 271E1 (HDMI-0)

Sep 19 10:12:39 nvidia-agx-orin kernel: [ 1661.118284] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call result 0xffff:

Sep 19 10:12:39 nvidia-agx-orin kernel: [ 1661.118297] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0

Sep 19 10:12:39 nvidia-agx-orin kernel: [ 1661.249057] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.434242] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.437292] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.484443] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.502388] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call result 0xffff:

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.502396] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.583558] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call result 0xffff:

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.583569] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.794888] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Sep 19 10:12:40 nvidia-agx-orin kernel: [ 1661.795660] NVRM nvAssertFailed: Assertion failed: 0 @ g_mem_mgr_nvoc.h:1175

Xorg.log:

[  1661.107] (--) NVIDIA(GPU-0): DFP-0: connected
[  1661.107] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[  1661.107] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[  1661.107] (--) NVIDIA(GPU-0): 
[  1661.126] (--) NVIDIA(GPU-0): DFP-0: connected
[  1661.127] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[  1661.127] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[  1661.127] (--) NVIDIA(GPU-0): 
[  1661.193] (II) NVIDIA(0): Setting mode "NULL"
[  1661.277] (II) NVIDIA(0): Setting mode "HDMI-0: nvidia-auto-select @1024x768 +0+0 {ViewPortIn=1024x768, ViewPortOut=1024x768+0+0}"
[  1661.511] (--) NVIDIA(GPU-0): DFP-0: connected
[  1661.511] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[  1661.511] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[  1661.511] (--) NVIDIA(GPU-0): 
[  1661.592] (--) NVIDIA(GPU-0): DFP-0: connected
[  1661.592] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[  1661.592] (--) NVIDIA(GPU-0): DFP-0: 165.0 MHz maximum pixel clock
[  1661.592] (--) NVIDIA(GPU-0):

WayneWWW · September 19, 2022, 10:46am

抱歉請問你現在是在說你在jetpack5.0.1也有一樣的EDID read failure 問題嗎？

Sorry, are you telling that you have same EDID read failure issue in jetpack5.0.1?

nkpyq · September 19, 2022, 12:33pm

我确认了" EDID read failure issue" 发生的时间点，jp5.0.1和jp5.0.2都是发生在拔掉hdmi线再执行xrandr后，我想这是正常现象。

回到jp5.0.2，使用new dcb setting，如下信息可能有助于帮助分析问题。

1.启动后，hdmi无画面显示，kernel如下报错信息

Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084379] nvidia-modeset: Loading NVIDIA UNIX Open Kernel Mode Setting Driver for aarch64 35.1.0 Release Build (buildbrain@mobile-u64-5273-d7000) Wed Aug 10 20:32:41 PDT 2022
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084496] NVRM gpumgrGetSomeGpu: Failed to retrieve pGpu - Too early call!.
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084502] NVRM nvAssertFailedNoLog: Assertion failed: NV_FALSE @ gpu_mgr.c:296
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084513] CPU: 7 PID: 1795 Comm: modprobe Tainted: G OE 5.10.104-tegra #28
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084516] Hardware name: /, BIOS 1.0-d7fb19b 08/10/2022
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084519] Call trace:
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084541] dump_backtrace+0x0/0x1d0
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084545] show_stack+0x2c/0x40
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084563] dump_stack+0xd8/0x138
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084640] os_dump_stack+0x18/0x20 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084693] tlsEntryGet+0x130/0x138 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084739] gpumgrGetSomeGpu+0x7c/0x90 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084788] threadPriorityStateFree+0xa8/0x2a8 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084833] rmclientConstruct_IMPL+0x170/0x240 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084886] __nvoc_ctor_RmClient+0x5c/0x88 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084930] __nvoc_objCreate_RmClient+0x78/0x110 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.084974] resservClientFactory+0x28/0x90 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085018] serverAllocClient+0x19c/0x318 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085063] serverAllocResource+0x140/0x2e0 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085111] rmapiAllocWithSecInfo+0x180/0x2e0 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085157] rmapiAllocWithSecInfoTls+0x74/0xa8 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085203] rmapiControlWithSecInfoTls+0x4a8/0x520 [nvidia]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085265] nvkms_call_rm+0x60/0x98 [nvidia_modeset]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085313] nvRmApiAlloc+0x30/0x40 [nvidia_modeset]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085320] do_one_initcall+0x4c/0x2d0
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085337] do_init_module+0x4c/0x200
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085339] load_module+0x1ff8/0x22f0
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085342] __do_sys_finit_module+0xb4/0xf0
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085345] __arm64_sys_finit_module+0x28/0x40
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085356] el0_svc_common.constprop.0+0x7c/0x1c0
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085359] do_el0_svc+0x34/0xa0
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085369] el0_svc+0x1c/0x30
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085374] el0_sync_handler+0xa8/0xb0
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085376] el0_sync+0x16c/0x180
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085387] NVRM gpumgrGetSomeGpu: Failed to retrieve pGpu - Too early call!.
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.085392] NVRM nvAssertFailedNoLog: Assertion failed: NV_FALSE @ gpu_mgr.c:296
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.326033] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x2080013f result 0x56:
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.327565] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x2080017e result 0x56:
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.335630] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x2080014a result 0x56:
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.411854] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730190 result 0x56:
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.857233] dce: dce_ipc_read_message:678 Error getting next free buf to read
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.857239] dce: dce_client_process_event_ipc:319 Error in reading DCE msg for ch_type [3]
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.857242] dce: dce_ipc_read_message:678 Error getting next free buf to read
Sep 19 12:10:34 nvidia-agx-orin kernel: [ 31.857243] dce: dce_client_process_event_ipc:319 Error in reading DCE msg for ch_type [3]
Sep 19 12:10:47 nvidia-agx-orin kernel: [ 44.568527] vdd-12v-sys: disabling
Sep 19 12:10:47 nvidia-agx-orin kernel: [ 44.568547] vdd-3v3-ftdi: disabling

2.run xrandr后，hdmi无画面显示，kenel 无新增打印信息

3.拔掉hdmi线再run xrandr后，出现EDID read failure issue，hdmi无画面显示

Sep 19 12:17:49 nvidia-agx-orin kernel: [ 283.346956] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
Sep 19 12:17:49 nvidia-agx-orin kernel: [ 283.346976] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device Philips PHL 271E1 (HDMI-0)
Sep 19 12:17:50 nvidia-agx-orin kernel: [ 283.416949] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
Sep 19 12:17:50 nvidia-agx-orin kernel: [ 283.416963] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0
Sep 19 12:17:50 nvidia-agx-orin kernel: [ 283.989460] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
Sep 19 12:17:50 nvidia-agx-orin kernel: [ 283.989476] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0
Sep 19 12:17:50 nvidia-agx-orin kernel: [ 284.070555] NVRM rpcRmApiControl_dce: NVRM_RPC_DCE: Failed RM ctrl call cmd:0x730245 result 0xffff:
Sep 19 12:17:50 nvidia-agx-orin kernel: [ 284.070577] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0

4.连接上hdmi线后，此时可显示ubuntu桌面，kernel 无新增打印信息

5.再次执行xrandr后，hdmi显示变为黑屏，kernel如下报错信息

Sep 19 12:19:09 nvidia-agx-orin kernel: [ 362.781485] arm-smmu 10000000.iommu: Unhandled context fault: fsr=0x80000402, iova=0x7ffca01000, fsynr=0x280003, cbfrsynra=0x401, cb=2
Sep 19 12:19:09 nvidia-agx-orin kernel: [ 362.794008] mc-err: unknown mcerr fault, int_status=0x00000000, ch_int_status=0x00000000, hubc_int_status=0x00000000 sbs_int_status=0x00000000, hub_int_status=0x00000000
Sep 19 12:19:10 nvidia-agx-orin kernel: [ 364.053877] snd_hda_codec_hdmi hdaudioC0D0: HDMI: invalid ELD data byte 1
Sep 19 12:19:10 nvidia-agx-orin kernel: [ 364.053895] snd_hda_codec_hdmi hdaudioC0D0: HDMI: out of range MNL 13

WayneWWW · September 20, 2022, 3:03am

Hi,

可以麻煩您之後的敘述都用中文回覆就好. 我發現你中文回覆所提供的資訊比起用英文回覆的時候多了不少…

有幾點問題再請問你

請問你的 “拔掉hdmi线再run xrandr后，出现EDID read failure issue，hdmi无画面显示” 是要讓畫面出現的必要步驟嗎? 比方說今天如果你都不跑xrandr, 只有插拔HDMI, 請問畫面會出現嗎?
請問你有嘗試過其他款的HDMI螢幕嗎？

另外, 基本上你可以參考上面我貼那個新DCB的討論串. 那個用戶目前是hotplug沒辦法使用. 但是畫面都是正常.
你可以確認一下我們那個討論串上面測試過的東西.

---------Eng version

Is your “plug out HDMI cable and run xrandr, and it will show EDID read failure” a must-have step to make the display show the desktop? For example, if you don’t run xrandr but only hotplug HDMI, will the screen show up?
Have you ever tried other kind of monitor?

Actually you can just check the post I shared for the new DCB. The user from that post already had normal HDMI working. But just hotplug has some issue. You can go through the check we have done there.

WayneWWW · September 20, 2022, 3:11am

另外完整dump log的方法如下基本上就是重新載入數個kernel module + debug print. 然後請貼上抓下來的dmesg.

1. sudo systemctl stop gdm
2. echo 7 > /proc/sys/kernel/printk
3. rmmod nvidia-modeset
4. rmmod nvidia

5. insmod /lib/modules/$(uname -r)/extra/opensrc-disp/nvidia.ko NVreg_RmMsg=":" rm_firmware_active="all"
6. insmod /lib/modules/$(uname -r)/extra/opensrc-disp/nvidia-modeset.ko
7. dmesg > dmesg.0.log

8. sudo systemctl start gdm
9. dmesg > dmesg.1.log

— Eng ver----

Above steps are the full method to dump the display log. Basically, it reloads multiple drivers with debug print enabled. Please attach new dmesg again.

nkpyq · September 20, 2022, 4:01am

Hi，感谢耐心解答。

1.“拔掉hdmi线再run xrandr后，出现EDID read failure issue，hdmi无画面显示” 是让画面出現的必要步驟。
如果上电启动后从未执行过xrandr，插拔HDMI，画面都不会出现；
上电启动后执行过上述“必要步骤”，插拔HDMI，画面都可以显示（但分辨率和颜色不对）。

2.试过3种屏幕：Pilips、Voc冠捷、（小尺寸）调试屏幕
只有调试屏幕上电后即可显示，不需要上述“必要步骤”，插拔HDMI也可以显示（但是桌面右侧有细彩条缝隙，jp5.0.1使用该调试屏幕无细彩条，相应图片上传不了，一直卡在Uploading 0%）。

nkpyq · September 20, 2022, 4:04am

如下是 full method dump log
dmesg.0.log (74.0 KB)
dmesg.1.log (133.5 KB)
kern.log (769.5 KB)

如下是相应图片效果：