[Bug]: Unable to compute luma percent: ERROR_INVALID_EXTERNAL_HANDLE #120
Comments
|
Interesting! Sounds like Vulkan is unable to import the sway frame. Have you set |
|
Sure! Also I set |
|
I see, uff this is a tricky one, but let's see if we can find some solution. |
I have just a laptop with single integrated GPU) |
|
Yeah I definitely agree, everything that I know looks good and straightforward, unfortunately I don't have any good ideas 😞 The error is happening in this function, in case you want to try to debug something: Line 286 in c82288e One final place that stood out to me is perhaps this line: Line 337 in c82288e Could you check in the Maybe share the entire While my code is most plausible to have the issue than anything else, we cannot rule out that the issue is in the Vulkan driver. It actually is the case for my current laptop I can't run wluma because of what's currently most likely is the bug in the Vulkan driver for my GPU 😞 |
|
Here is entire vulkaninfo 4 memory types for my integrated GPU as you can see.. I'll be really appreciated to any help from your side |
|
Wow, cannot believe it was (maybe) a lucky hit! Would you be able to test some things for me please?
Line 360 in c82288e Here, just try with |
|
Sure
|
|
Thanks! I guess I was excited too prematurely 😞 It's good to know that the current code (probably) doesn't suffer from the issue with picking among multiple memory types, but with regards to your actual issue we are back to square one, I don't really have any good ideas what might be wrong 😞 Do you do anything special to your sway setup, do you for example run it with some non-default |
|
No, nothing special. I have a basic sway setup |
|
Could you check if you see something interesting in the output, if you install I get some debug logs, but if I understand correctly, if Vulkan decides that some APIs are not used correctly, you would see some error message there, that might indicate issue in the code. |
|
I see this |
|
Very interesting! Could you please try |
|
I tried unfortunatelly no supported formats as I can see |
|
I'm not sure yet, because I also don't get any formats supported, but as I mentioned wluma doesn't work on my laptop as well. I'll try to find someone who can run this branch on a laptop where it definitely works, to see if they get any format that is reported as "supported". I'm still hoping it's a bug in the test code, but I can't see right now what could be wrong - this probably can first happen on Monday. |
|
Sure! Enjoy your weekend |
|
Could you please retry the latest Do still run with validation layers enabled, in case there's now some other validation being thrown to you. Thanks! |
|
OMG this fixed wluma on my hardware too!!! Big thank you for reporting, without this I would have thought that the issue is with Vulkan driver! 🎉 🚀 |
|
Okay maybe "fixed" was a bit too early to say, I am getting "some" output from the driver now, but the colors are wrong 😂 But that's something I can work with, at least 😁 So do not trust the "luma" output just yet, but see if wluma runs at all without crashes and without vulkan validations for you, and we take the next step then. |
|
IT WORKS! Awesome! this is my output |
|
Cool, I'll ping you for a bit more help with extra debugging when I get the colors right, to make sure I do something that works not only on my hardware but on yours too, but glad to see we are making progress! |
|
Sure. Let me know it you need any additional support with debugging from me |
|
Please try the following test: Pull latest First, I'm interested in the line that starts with Second, while it's running, go to this website and hit "Start test": https://lcdtech.info/en/tests/dead.pixel.htm It should open a single color on full-screen. Press If you see other values, please share too, color and corresponding values. Many thanks! |
|
output from the latest lcdtech website is down for me, wont oppen( It was on red, than on green and on blue on the end |
|
Awesome, thank you for the test! I'm still struggling because on my own laptop I'm still getting garbage values for the colors 😭 I'll try to struggle a bit more, but even if I fail for my own laptop, I'll make sure to cut a release with the fix for you very soon 🙏 |
|
It will be great! BTW I don't notice any extra CPU load with latest |
|
In fact I was getting weird colors on lcdtech.info/en/tests/dead.pixel.htm, but on the website you found I am getting correct colors reported! So I think the fix is good after all, I was just confused somewhere during my earlier tests. The final fix for this issue report is very pleasantly short, just 3 lines of code, and actually makes sense :) I'm going to make a release in a moment! Many thanks again for all your help! 🙏 🇺🇦 |
Steps for reproducing the issue
What is the buggy behavior?
App chashes
What is the expected behavior?
Wluma should be fully functional
Logs
RUST_LOG=debug wluma [2024-10-18T19:31:34Z DEBUG wluma] Using Config { als: Iio { path: "/sys/bus/iio/devices", thresholds: { 250: "normal", 0: "night", 800: "outdoors", 80: "dim", 500: "bright", 20: "dark", }, }, output: [ Backlight( BacklightOutput { name: "eDP-1", path: "/sys/class/backlight/intel_backlight", capturer: Wlroots, min_brightness: 1, }, ), Backlight( BacklightOutput { name: "keyboard-dell", path: "/sys/bus/platform/devices/dell-laptop/leds/dell::kbd_backlight", capturer: None, min_brightness: 0, }, ), ], } [2024-10-18T19:31:34Z WARN wluma] Skipping 'keyboard-dell' as it might be disconnected: No such file or directory (os error 2) [2024-10-18T19:31:34Z INFO wluma] Continue adjusting brightness and wluma will learn your preference over time. [2024-10-18T19:31:34Z DEBUG wluma::frame::capturer::wlroots] Using output 'Samsung Display Corp. 0x419D (eDP-1)' for config 'eDP-1' thread 'predictor-eDP-1' panicked at src/frame/capturer/wlroots.rs:128:26: Unable to compute luma percent: ERROR_INVALID_EXTERNAL_HANDLE note: run with `RUST_BACKTRACE=1` environment variable to display a backtraceVersion
4.4.0-1
Environment
The text was updated successfully, but these errors were encountered: