On Sat, Apr 1, 2023 at 3:55 PM Shawn Lindberg <shawn.lindberg@xxxxxxxxx> wrote:
>
> On Sat, Apr 1, 2023 at 3:17 AM Hans Verkuil <hverkuil@xxxxxxxxx> wrote:
> >
> > On 31/03/2023 22:23, Shawn Lindberg wrote:
> > > On Wed, Mar 29, 2023 at 2:56 PM Shawn Lindberg <shawn.lindberg@xxxxxxxxx> wrote:
> > >>
> > >> On Tue, Mar 28, 2023 at 2:58 AM Hans Verkuil <hverkuil@xxxxxxxxx> wrote:
> > >>>
> > >>>> cec-ctl -d0 --tv --cec-version-1.4
> > >>>
> > >>> That's wrong, the RPi is a Playback device, not a TV. So use --playback instead.
> > >>>
> > >>> You should also add this line to the config.txt:
> > >>>
> > >>> hdmi_ignore_cec=1
> > >>>
> > >>> otherwise the RPi's firmware tries to process CEC messages as well.
> > >>
> > >> Oh, I thought that the TV/playback command was indicating what sort of
> > >> device the connected device is. This wasn't clear from the man page,
> > >> either. Thank you for that. I made the change to config.txt and
> > >> strangely when the RPi rebooted (I have it set to do this
> > >> automatically once a day) the projector automatically turned on. I
> > >> have never experienced this before.
> > >
> > > Further update on this. I continue to see the projector automatically
> > > power on every time the RPi does its daily reboot, so I think I may
> > > have to remove the hdmi_ignore_cec from the config.txt. Especially
> > > since I can't figure out how to reliably shut the projector back off
> > > again.
> >
> > From what I can tell, the Raspberry Pi doesn't transmit anything over CEC
> > at boot time, regardless of whether hdmi_ignore_cec is present or not.
> > That's with a Raspberry Pi 4B. It might be different for an RPi 3.
>
> That is strange. I am also using a Raspberry Pi 4B. I do have the
> following changes to my config.txt:
>
> # uncomment if hdmi display is not detected and composite is being output
> hdmi_force_hotplug=1
>
> # uncomment to force a specific HDMI mode (this will force VGA)
> hdmi_group=1
> hdmi_mode=16
>
> # Additional line added to prevent firmware from processing CEC messages.
> hdmi_ignore_cec=1
>
>
> > >>>> During this time, if I try to poll the projector, it will succeed.
> > >>>> However, if I monitor events, after a significant amount of time
> > >>>> (appears to be greater than 20 minutes, although this is difficult to
> > >>>> verify because of how long it takes) I go will eventually see the
> > >>>> following:
> > >>>>
> > >>>> Event: State Change: PA: 1.0.0.0, LA mask: 0x0000, Conn Info: yes
> > >>>> Timestamp: 30981.428s
> > >>>
> > >>> Now it appears to be able to read the EDID again and it has a valid
> > >>> physical address.
> > >>>
> > >>>> Transmitted by Specific to Specific (14 to 14): POLL
> > >>>> Tx, Not Acknowledged (4), Max Retries
> > >>>> Sequence: 21 Tx Timestamp: 30981.561s Tx, Not Acknowledged (4), Max Retries
> > >>>>
> > >>>> Event: State Change: PA: 1.0.0.0, LA mask: 0x4000, Conn Info: yes
> > >>>> Timestamp: 30981.561s
> > >>>> Transmitted by Specific to all (14 to 15): REPORT_PHYSICAL_ADDR (0x84):
> > >>>> phys-addr: 1.0.0.0
> > >>>> prim-devtype: tv (0x00)
> > >>>> Sequence: 22 Tx Timestamp: 30981.696s
> > >>>> Transmitted by Specific to all (14 to 15): DEVICE_VENDOR_ID (0x87):
> > >>>> vendor-id: 3075 (0x00000c03)
> > >>>> Sequence: 23 Tx Timestamp: 30981.835s
> > >>>> Received from TV to Specific (0 to 14): FEATURE_ABORT (0x00):
> > >>>> abort-msg: 132 (0x84, REPORT_PHYSICAL_ADDR)
> > >>>> reason: invalid-op (0x03)
> > >>>> Sequence: 0 Rx Timestamp: 30981.949s
> > >>>> Received from TV to Specific (0 to 14): GIVE_OSD_NAME (0x46)
> > >>>> Sequence: 0 Rx Timestamp: 30982.026s
> > >>>> Transmitted by Specific to TV (14 to 0): SET_OSD_NAME (0x47):
> > >>>> name: TV
> > >>>> Sequence: 24 Tx Timestamp: 30982.137s
> > >>>>
> > >>>> After this point in time the standby command will succeed and the
> > >>>> projector will turn off. It's quite inconvenient to have to wait over
> > >>>> 20 minutes to turn the projector back off again. Any idea how I can
> > >>>> shorten this delay?
> > >>>
> > >>> There is something weird about your setup and EDID. I can't really tell
> > >>> what it is.
> > >>
> > >> After making the above changes and retesting, the behavior didn't
> > >> change. I still get the device not connected message and the invalid
> > >> physical address when I try to do standby. I should also note that one
> > >> way around this issue is to reboot the RPi. For some reason that seems
> > >> to get around the long delay in getting the physical address.
> > >>
> > >> I don't know what would be strange about my set up other than the
> > >> projector itself and a couple of lines I uncommented in the config.txt
> > >> to set the RPi to use HDMI even if the projector is not on at the time
> > >> of booting. Is there more information I can provide that would allow
> > >> us to figure out what's going on? If you are correct that for some
> > >> reason it is just not reading the EDID, is there a way to manually
> > >> provide that? I don't know much about it, but it's a static property
> > >> of the device (the projector in this case), right?
> > >
> > > Since I noticed that the physical address is populated properly when
> > > the RPi is booted while the projector is turned on, I did that and
> > > then tried using the get-edid utility to see if I could read the EDID
> > > block and save it to a file. Unfortunately, this didn't work, as the
> > > utility reports that there was no EDID available on any of the buses.
> > > So once again I am out of ideas.
> >
> > The EDID also appears in /sys:
> >
> > /sys/devices/platform/gpu/drm/card1/card1-HDMI-A-1/edid
> > /sys/devices/platform/gpu/drm/card1/card1-HDMI-A-2/edid
> >
> > get-edid works fine on my RPi 4B, so if that doesn't work, then it really
> > looks like there is something weird going on with your projector.
>
> Thank you for that. Based on this information with the projector on
> during boot, I was able to get the following output from parse-edid. I
> did not copy the entire output, just a snippet to indicate that it is
> able to get a valid EDID at boot time.
>
> cat /sys/devices/platform/gpu/drm/card1/card1-HDMI-A-1/edid | parse-edid
> Checksum Correct
>
> Section "Monitor"
> Identifier "LG PROJECTOR"
> ModelName "LG PROJECTOR"
> VendorName "GSM"
> # Monitor Manufactured week 33 of 2017
> # EDID version 1.3
> # Digital Display
> DisplaySize 1600 900
> Gamma 2.20
> Option "DPMS" "false"
> Horizsync 30-83
> VertRefresh 58-62
> # Maximum pixel clock is 160MHz
> #Not giving standard mode: 640x480, 60Hz
>
> In summary, it seems that at boot time and after waiting an extended
> amount of time after turning on the projector using the image-view-on
> command, the EDID file is available and valid. So although my
> projector may be strange, I am guessing that there must be a way to
> manually provide the EDID file, force the physical address, or some
> other workaround/solution. However, I'm certainly no expert in this
> area so I would defer to those of you who are.
I finally found some more time to experiment with this, and I tried
saving a copy of the EDID while it was available after the machine
rebooted and the projector automatically turned on. Unsurprisingly,
after I turned the projector off using standby and then turned it back
on using image-view-on, the EDID was not available from
/sys/devices/platform/gpu/drm/card1/card1-HDMI-A-1/edid.
So I tried the following command to see if I could get the physical
address to update from the saved EDID.
cec-ctl -d0 -t0 -T --phys-addr-from-edid projector_edid
CEC_ADAP_G_CAPS returned 0 (Success)
The CEC adapter doesn't allow setting the physical address manually,
ignore this option.
CEC_ADAP_S_PHYS_ADDR returned -1 (Inappropriate ioctl for device)
CEC_ADAP_G_PHYS_ADDR returned 0 (Success)
CEC_ADAP_G_LOG_ADDRS returned 0 (Success)
CEC_ADAP_G_CONNECTOR_INFO returned 0 (Success)
I'm not really sure why this would be the case.
