[Bug 216376] New: AMDGPU: display output disables and quickly reenables when switching AVR into/from standby doing HDMI passthrough

[Bug 216376] New: AMDGPU: display output disables and quickly reenables when switching AVR into/from standby doing HDMI passthrough

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=216376

            Bug ID: 216376
           Summary: AMDGPU: display output disables and quickly reenables
                    when switching AVR into/from standby doing HDMI
                    passthrough
           Product: Drivers
           Version: 2.5
    Kernel Version: 5.19.1
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: low
          Priority: P1
         Component: Video(DRI - non Intel)
          Assignee: drivers_video-dri@kernel-bugs.osdl.org
          Reporter: jlp.bugs@gmail.com
        Regression: No

Created attachment 301600
  --> https://bugzilla.kernel.org/attachment.cgi?id=301600&action=edit
dmesg

I have A PC with Radeon RX480 which has HDMI output connected to HDMI input on
Marantz SR5015 AV reciever and there the HDMI output goes to a TV (Panasonic
TX-55CX200E). The DVI output is connected to an LCD monitor. The AVR has the
feature to pass through HDMI signal when put into standby mode. When I switch
AVR from/to standby mode it looks like it for a very short time interrupts the
connection from PC to TV and during this time it appears that the video card
output turns off and monitor layout is re-aranged to one screen (DVI) only and
then almost immediately back two previous dual screen layout. Would it be
possible to delay turning off the HDMI output for a bit longer so that it would
wait long enough that AVR would reconnect entire HDMI path to TV back again and
the HDMI output on video card would not need to be turned off and keep sending
output in the meantime? So that the display layout configuration would keep
being in dual-screen mode for the whole time?

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 216376] AMDGPU: display output disables and quickly reenables when switching AVR into/from standby doing HDMI passthrough

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=216376

--- Comment #1 from Jure Repinc (jlp.bugs@gmail.com) ---
Created attachment 301601
  --> https://bugzilla.kernel.org/attachment.cgi?id=301601&action=edit
Xorg.0.log

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 216376] AMDGPU: display output disables and quickly reenables when switching AVR into/from standby doing HDMI passthrough

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=216376

--- Comment #2 from Jure Repinc (jlp.bugs@gmail.com) ---
Created attachment 301602
  --> https://bugzilla.kernel.org/attachment.cgi?id=301602&action=edit
lspci

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 216376] AMDGPU: display output disables and quickly reenables when switching AVR into/from standby doing HDMI passthrough

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=216376

Alex Deucher (alexdeucher@gmail.com) changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |alexdeucher@gmail.com

--- Comment #3 from Alex Deucher (alexdeucher@gmail.com) ---
I suspect the receiver toggles the hotplug pin when it switches to/from
passthrough mode notify the GPU that the EDID will change (e.g., from receiver
audio to monitor audio) otherwise the GPU won't know that the receiver is no
longer part of the chain.  The driver doesn't actually do anything when it
receives this hotplug interrupt.  It just tells userspace.  Your desktop
environment listens for hotplug events and then acts on them (either enabling
or disabling displays).  Your best bet is probably to change the behavior of
your desktop environment when it receives a hotplug notification. 
Alternatively, you could try and block drm hotplug events in udev.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 216376] AMDGPU: display output disables and quickly reenables when switching AVR into/from standby doing HDMI passthrough

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=216376

--- Comment #4 from Jure Repinc (jlp.bugs@gmail.com) ---
If I got it right, the hotplug events are all posted via udev? If so, is it
correct to use "udevadm monitor --environment --udev" to see these events
posted to userspace? Or is there somethign better? Just trying to make the best
sense out of it to try making a workaround and gather as much relevant
information to report a good bug/feature request with correct info to KDE
developers.


If I use "udevadm monitor --environment --udev" I get the output:

udevadm monitor --environment --udev
monitor will print the received events for:
UDEV - the event which udev sends out after rule processing


< switching AVR from on to standby mode >


UDEV  [35121.304457] change  
/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0 (drm)
ACTION=change
DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
SUBSYSTEM=drm
HOTPLUG=1
CONNECTOR=74
DEVNAME=/dev/dri/card0
DEVTYPE=drm_minor
SEQNUM=12043
USEC_INITIALIZED=4651406
ID_PATH=pci-0000:01:00.0
ID_PATH_TAG=pci-0000_01_00_0
ID_FOR_SEAT=drm-pci-0000_01_00_0
MAJOR=226
MINOR=0
DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
TAGS=:uaccess:master-of-seat:seat:
CURRENT_TAGS=:uaccess:master-of-seat:seat:

UDEV  [35124.216717] change  
/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0 (drm)
ACTION=change
DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
SUBSYSTEM=drm
HOTPLUG=1
CONNECTOR=74
DEVNAME=/dev/dri/card0
DEVTYPE=drm_minor
SEQNUM=12044
USEC_INITIALIZED=4651406
ID_PATH=pci-0000:01:00.0
ID_PATH_TAG=pci-0000_01_00_0
ID_FOR_SEAT=drm-pci-0000_01_00_0
MAJOR=226
MINOR=0
DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
TAGS=:uaccess:master-of-seat:seat:
CURRENT_TAGS=:uaccess:master-of-seat:seat:


< switching AVR from standby to on mode >


UDEV  [35155.373681] change  
/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0 (drm)
ACTION=change
DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
SUBSYSTEM=drm
HOTPLUG=1
CONNECTOR=74
DEVNAME=/dev/dri/card0
DEVTYPE=drm_minor
SEQNUM=12048
USEC_INITIALIZED=4651406
ID_PATH=pci-0000:01:00.0
ID_PATH_TAG=pci-0000_01_00_0
ID_FOR_SEAT=drm-pci-0000_01_00_0
MAJOR=226
MINOR=0
DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
TAGS=:uaccess:master-of-seat:seat:
CURRENT_TAGS=:uaccess:master-of-seat:seat:

UDEV  [35158.196252] change  
/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0 (drm)
ACTION=change
DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
SUBSYSTEM=drm
HOTPLUG=1
CONNECTOR=74
DEVNAME=/dev/dri/card0
DEVTYPE=drm_minor
SEQNUM=12049
USEC_INITIALIZED=4651406
ID_PATH=pci-0000:01:00.0
ID_PATH_TAG=pci-0000_01_00_0
ID_FOR_SEAT=drm-pci-0000_01_00_0
MAJOR=226
MINOR=0
DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
TAGS=:uaccess:master-of-seat:seat:
CURRENT_TAGS=:uaccess:master-of-seat:seat:

UDEV  [35158.953569] change  
/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0 (drm)
ACTION=change
DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
SUBSYSTEM=drm
HOTPLUG=1
CONNECTOR=74
DEVNAME=/dev/dri/card0
DEVTYPE=drm_minor
SEQNUM=12050
USEC_INITIALIZED=4651406
ID_PATH=pci-0000:01:00.0
ID_PATH_TAG=pci-0000_01_00_0
ID_FOR_SEAT=drm-pci-0000_01_00_0
MAJOR=226
MINOR=0
DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
TAGS=:uaccess:master-of-seat:seat:
CURRENT_TAGS=:uaccess:master-of-seat:seat:

UDEV  [35161.036327] change  
/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0 (drm)
ACTION=change
DEVPATH=/devices/pci0000:00/0000:00:01.0/0000:01:00.0/drm/card0
SUBSYSTEM=drm
HOTPLUG=1
CONNECTOR=74
DEVNAME=/dev/dri/card0
DEVTYPE=drm_minor
SEQNUM=12051
USEC_INITIALIZED=4651406
ID_PATH=pci-0000:01:00.0
ID_PATH_TAG=pci-0000_01_00_0
ID_FOR_SEAT=drm-pci-0000_01_00_0
MAJOR=226
MINOR=0
DEVLINKS=/dev/dri/by-path/pci-0000:01:00.0-card
TAGS=:uaccess:master-of-seat:seat:
CURRENT_TAGS=:uaccess:master-of-seat:seat:

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

[Bug 216376] AMDGPU: display output disables and quickly reenables when switching AVR into/from standby doing HDMI passthrough

[bugzilla-daemon]

https://bugzilla.kernel.org/show_bug.cgi?id=216376

--- Comment #5 from Alex Deucher (alexdeucher@gmail.com) ---
Yes, those are the events.  The desktop environment listens for them and
reacts.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.