Not detecting capabilities on the Lenovo G27q

[digitaltrails]

          testing with the new service and I'm not seeing any garbage values anymore, but it is still not detecting capabilities on the G27q, that's the screen 2 in the previous logs.  Though if it has no trouble talking to it if I generate the config file using ddcutil and then switch to the service after that's been saved.

Originally posted by @Extent421 in digitaltrails/vdu_controls#84 (comment)

[digitaltrails]

Screen 2 is only detected fully with ddcutil

ddcutil --display 2 capabilities
Model: Lenovo G27q-20
MCCS version: 2.2
Commands:
   Op Code: 01 (VCP Request)
   Op Code: 02 (VCP Response)
   Op Code: 03 (VCP Set)
   Op Code: 07 (Timing Request)
   Op Code: 0C (Save Settings)
   Op Code: E3 (Capabilities Reply)
   Op Code: F3 (Capabilities Request)
VCP Features:
   Feature: 02 (New control value)
   Feature: 04 (Restore factory defaults)
      Values: 4C 31 2E 30 (interpretation unavailable)
   Feature: 05 (Restore factory brightness/contrast defaults)
   Feature: 08 (Restore color defaults)
   Feature: 10 (Brightness)
   Feature: 12 (Contrast)
   Feature: 14 (Select color preset)
      Values:
         01: sRGB
         05: 6500 K
         06: 7500 K
         08: 9300 K
         12: Unrecognized value
         0b: User 1
   Feature: 16 (Video gain: Red)
   Feature: 18 (Video gain: Green)
   Feature: 1A (Video gain: Blue)
   Feature: 52 (Active control)
   Feature: 60 (Input Source)
      Values:
         11: HDMI-1
         0f: DisplayPort-1
   Feature: 86 (Display Scaling)
      Values:
         01: No scaling
         02: Max image, no aspect ration distortion
   Feature: AC (Horizontal frequency)
   Feature: AE (Vertical frequency)
   Feature: B2 (Flat panel sub-pixel layout)
   Feature: B6 (Display technology type)
   Feature: C6 (Application enable key)
      Values: 02 00 97 00 90 80 00 68 00 8A 80 00 5C 00 54 00 00 (interpretation unavailable)
   Feature: C8 (Display controller type)
   Feature: CA (OSD/Button Control)
   Feature: CC (OSD Language)
      Values:
         02: English
         03: French
         04: German
         05: Italian
         06: Japanese
         09: Russian
         0a: Spanish
         0d: Chinese (simplified / Kantai)
   Feature: D6 (Power mode)
      Values:
         01: DPM: On,  DPMS: Off
         04: DPM: Off, DPMS: Off
         05: Write only value to turn off display
   Feature: DF (VCP Version)
   Feature: E0 (Manufacturer specific feature)
      Values: 00 01 02 (interpretation unavailable)
   Feature: EA (Manufacturer specific feature)
      Values: 00 01 (interpretation unavailable)
   Feature: EB (Manufacturer specific feature)
      Values: 00 01 (interpretation unavailable)
   Feature: EC (Manufacturer specific feature)
      Values: 00 01 FF 00 02 03 (interpretation unavailable)
   Feature: F6 (Manufacturer specific feature)
      Values: 07 (interpretation unavailable)
   Feature: F7 (Manufacturer specific feature)
      Values: 01 02 (interpretation unavailable)
   Feature: F8 (Manufacturer specific feature)
      Values: 01 02 (interpretation unavailable)
   Feature: F9 (Manufacturer specific feature)
      Values: 00 01 02 03 04 05 06 (interpretation unavailable)
   Feature: FA (Manufacturer specific feature)
      Values: 10 12 EA 14 (interpretation unavailable)
   Feature: FD (Manufacturer specific feature)

[digitaltrails]

@Extent421: Can you post the out from the following two ddcutil-service calls by calling them via the busctl command:

busctl --user call com.ddcutil.DdcutilService /com/ddcutil/DdcutilObject com.ddcutil.DdcutilInterface Detect "u" "0"

busctl --user call com.ddcutil.DdcutilService /com/ddcutil/DdcutilObject com.ddcutil.DdcutilInterface GetCapabilitiesMetadata isu "2" ""  "0"

[Extent421]

ya I was just looking at updating it with the new service

busctl --user call com.ddcutil.DdcutilService /com/ddcutil/DdcutilObject com.ddcutil.DdcutilInterface Detect "u" "0"
ia(iiisssqsu)is 2 2 1 0 0 "DEL" "AW3225QF" "66N0YZ3" 41542 "AP///////wAQrEaiU0pFMAUiAQS1Rih4OwMhr04+tSUNUFSlSwBxT4GAqcDRwAEBAQEBAQEBTdAAoPBwPoAwIDUAu4shAAAaAAAA/wA2Nk4wWVozCiAgICAgAAAA/ABBVzMyMjVRRgogICAgAAAA/Qww8P//6gEKICAgICAgAvg=" 809847379 2 0 0 "LEN" "G27q-20" "U63304DD" 26307 "AP///////wAwrsNmREQ0MC4eAQS1PCJ4O3jlrU9GqyYNUFS97wCBgJUAswDRwGFoYXxFaEV8Vl4AoKCgKVAwIDUAVVAhAAAeAAAA/wBVNjMzMDRERAogICAgAAAA/QAwpfr6QQEKICAgICAgAAAA/ABHMjdxLTIwCiAgICAgATw=" 808731716 0 "OK"

busctl --user call com.ddcutil.DdcutilService /com/ddcutil/DdcutilObject com.ddcutil.DdcutilInterface GetCapabilitiesMetadata isu "2" ""  "0"
syya{ys}a{y(ssa{ys})}is "G27q-20" 2 2 7 1 "desc of 1" 2 "desc of 2" 3 "desc of 3" 7 "desc of 7" 12 "desc of 12" 227 "desc of 227" 243 "desc of 243" 33 2 "New control value" "" 0 4 "Restore factory defaults" "" 4 76 "" 49 "" 46 "" 48 "" 5 "Restore factory brightness/contrast defaults" "" 0 8 "Restore color defaults" "" 0 16 "Brightness" "" 0 18 "Contrast" "" 0 20 "Select color preset" "" 11 1 "sRGB" 1 "sRGB" 5 "6500 K" 5 "6500 K" 6 "7500 K" 6 "7500 K" 8 "9300 K" 8 "9300 K" 18 "" 11 "User 1" 11 "User 1" 22 "Video gain: Red" "" 0 24 "Video gain: Green" "" 0 26 "Video gain: Blue" "" 0 82 "Active control" "" 0 96 "Input Source" "" 4 17 "HDMI-1" 17 "HDMI-1" 15 "DisplayPort-1" 15 "DisplayPort-1" 134 "Display Scaling" "" 4 1 "No scaling" 1 "No scaling" 2 "Max image, no aspect ration distortion" 2 "Max image, no aspect ration distortion" 172 "Horizontal frequency" "" 0 174 "Vertical frequency" "" 0 178 "Flat panel sub-pixel layout" "" 0 182 "Display technology type" "" 0 198 "Application enable key" "" 17 2 "" 0 "" 151 "" 0 "" 144 "" 128 "" 0 "" 104 "" 0 "" 138 "" 128 "" 0 "" 92 "" 0 "" 84 "" 0 "" 0 "" 200 "Display controller type" "" 0 202 "OSD/Button Control" "" 0 204 "OSD Language" "" 8 2 "" 3 "" 4 "" 5 "" 6 "" 9 "" 10 "" 13 "" 214 "Power mode" "" 6 1 "DPM: On,  DPMS: Off" 1 "DPM: On,  DPMS: Off" 4 "DPM: Off, DPMS: Off" 4 "DPM: Off, DPMS: Off" 5 "Write only value to turn off display" 5 "Write only value to turn off display" 223 "VCP Version" "" 0 224 "Manufacturer Specific" "" 3 0 "" 1 "" 2 "" 234 "Manufacturer Specific" "" 2 0 "" 1 "" 235 "Manufacturer Specific" "" 2 0 "" 1 "" 236 "Manufacturer Specific" "" 6 0 "" 1 "" 255 "" 0 "" 2 "" 3 "" 246 "Manufacturer Specific" "" 1 7 "" 247 "Manufacturer Specific" "" 2 1 "" 2 "" 248 "Manufacturer Specific" "" 2 1 "" 2 "" 249 "Manufacturer Specific" "" 7 0 "" 1 "" 2 "" 3 "" 4 "" 5 "" 6 "" 250 "Manufacturer Specific" "" 4 16 "" 18 "" 234 "" 20 "" 253 "Manufacturer Specific" "" 0 0 "OK"

I don't know how much it matters, but the service did seem to be really slow in picking up when the screen was connected. I turned the screen on, extended the display, and then for the first couple minutes GetCapabilitiesMetadata was still returning

syya{ys}a{y(ssa{ys})}is "model" 0 0 0 0 -3020 "DDCRC_INVALID_DISPLAY: Error(s) opening ddc devices"

does the Detect need to be manually triggered to pick up changes?

[Extent421]

and just for completeness, the specific error that is getting raised is this, it is picking up the serial number

[digitaltrails]

does the Detect need to be manually triggered to pick up changes?

Yes, kind-of. In theory if a display driver properly supports DRM, then libddcutil should automatically pick up hot-plug events. In practice, depending on GPU/GPU-driver this may not work, or may not work reliably. The service attempts to figure out if libddcutil supports DRM, but it may claim to do so, even if its not 100% working. If the service "thinks" DRM isn't supported, then it can fall back to polling Detect every 30 seconds (prior to libddcutil 2.*, DRM isn't available, polling is always used).

[digitaltrails]

I don't know how much it matters, but the service did seem to be really slow in picking up when the screen was connected. I turned the screen on, extended the display, and then for the first couple minutes GetCapabilitiesMetadata was still returning...

So if you issue a Detect and then a GetCapabilitiesMetadata, and the GetCapabilitiesMetadata is returing something, does vdu_controls then also detect the G27q when you press the refresh button (bottom left of main panel) or the retry button in the above dialog?

Does ddcutil behave the same way - slow/errors at first?

Also, if ddcutil capabilties works, does busctl GetCapabilitiesMetadata also work?

[Extent421]

ddcutil has never given me an error

Initially when I was getting the dbus errors ddcutil was working just fine, and I was beginning to troubleshoot when dbus suddenly began working. I don't know if it's because I ran the dbus detect line first or if it just began working after some time. I'm pretty sure it was going on for longer than 30 seconds, but I'll pay closer attention to it now if it happens again.

That error screenshot was taken immediately after the previous logs with successful detections by restarting vdu.

If I do a refresh it gives a Set value: Failed to communicate with display unknown error D-Bus error org.freedesktop.DBus.Error.NoReply: Message recipient disconnected from message bus without replying and then vdu dies.

If I set vdu to ddcutil mode and do a refresh I get no error.

[digitaltrails]

ddcutil has never given me an error

Initially when I was getting the dbus errors ddcutil was working just fine, and I was beginning to troubleshoot when dbus suddenly began working. I don't know if it's because I ran the dbus detect line first or if it just began working after some time. I'm pretty sure it was going on for longer than 30 seconds, but I'll pay closer attention to it now if it happens again.

Ddcutil and libddcutil share a cache, so if ddcutil has worked, it may be that subsequent service calls will work. The cached timings are dynamically tuned. But the dynamic tuning is a recent feature which has had a few teething issues. The service is much faster than ddcutil, so it could be tripping over timing issues. Hopefully later versions of ddcutil/libddcutil will be better, so it may just pay to wait and see. You could try explicitly setting a sleep multiplier for that specific VDU, something greater than 1.0.

That error screenshot was taken immediately after the previous logs with successful detections by restarting vdu.

If I do a refresh it gives a Set value: Failed to communicate with display unknown error D-Bus error org.freedesktop.DBus.Error.NoReply: Message recipient disconnected from message bus without replying and then vdu dies.

If I set vdu to ddcutil mode and do a refresh I get no error.

So it is likely a timing issue. Maybe it would be worth trying a sleep multiplier. The multiplier is passed to the service, although it's not often used/tested (I'm trying to test it now, but my development environment has an issue at the moment).

With ddcutil/libddcutil 1.4.1, the sleep multiplier will probably affect all operations for all monitors. In later versions of ddcutil/libddcutil libddcutil can set display specific sleep multipliers. In 1.4.1 only the ddcutil command supports display specific sleep multipliers.

You can also force ddcutil/libddcutil to reset its timings cache by removing $HOME/.cache/ddcutil/dsa. Maybe preserve it first, incase it has particularly good values now. (Actually, I think the dsa file might only be for ddcutil/libddcutil >= 2).

[digitaltrails]

I'm trying to test it now, but my development environment has an issue at the moment

Yeah, the sleep multiplier appears to passed through from vdu_controls to the service and applied.

My tumbleweed 20240329 system is now acting a bit strange. After the xz hack dup that updated everything, ddcutil-service --log-info no longer logs info or debug messages. To get that working I've had to changed calls to g_setenv to just setenv and rebuild. Perhaps I've tripped over a glib bug. (Plus the KDE6 and Nvidia 550 updates are causing the odd session lockup/weirdness.)

Unfortunately it means that if you're also on the latest tumbleweed running ddcutil-service 1.0.3 with --log-info might not produce any extra info/debug output, such as showing whether the sleep multiplier was passed correctly. I've updated 1.0.4 to use setenv.

[Extent421]

I am on latest tumbleweed. There is no dsa in ~/.cache/ddcutil/ only capabilities

I had a sleep mult of 4 on the working screen from messing around earlier. I set a mult of 4 on the main settings page too, restarted, then refreshing causes the same error, and then immediately throws a new one that says

Error: RuntimeError: release unlocked lock

Details: Traceback (most recent call last):
  File "/usr/bin/vdu_controls", line 7381, in update_from_vdu
    self.detected_vdu_list = self.ddcutil.detect_monitors()
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/bin/vdu_controls", line 1299, in detect_monitors
    vdu_list = self.ddcutil_impl.detect(1)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/bin/vdu_controls", line 1733, in detect
    vdu_list = [self.DetectedAttributes(*vdu) for vdu in self._validate(result)[1]]
                                                         ^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/bin/vdu_controls", line 1775, in _validate
    raise ValueError(f"D-Bus error {result.errorName()}: {result.errorMessage()}")
ValueError: D-Bus error org.freedesktop.DBus.Error.NoReply: Message recipient disconnected from message bus without replying

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/bin/vdu_controls", line 3711, in run
    self.task_body()
  File "/usr/bin/vdu_controls", line 7386, in update_from_vdu
    self.refresh_lock.release()
RuntimeError: release unlocked lock

and then vdu dies

I'll keep an eye as the service updates

[digitaltrails]

It looks like the ddcutil-service fell over (the dbus daemon automatically starts a new one when next accessed). Possibly it may have logged something in the journal. It would be good to see if it logged anything interesting, but I suspect it may have just segfaulted.

I've never tried a sleep multiplier more than 2, I set 4 in both places, works OK, so it's not the magnitude of the value.

The resultant lock error is interesting. I was clearing up lock issues in the latest vdu_controls code, I had forgotten to remove the unlock line, hopefully that shouldn't happen in the latest code.