-
-
Notifications
You must be signed in to change notification settings - Fork 492
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Chromium Kiosk Mode no longer working #4175
Comments
Many thanks for your report. That directory is not present at all?
Can you try the following:
If that does not work:
And if you're in mood to compare how well the alternative behaves:
|
Can confirm. Its broken on a fresh installation and existing setups are broken after Update 😢. |
Hi, I did a test on RPi4B 64bit and there Chromium is still working on kiosk mode after updating to 7.0.2. Will test 32bit now EDIT:
|
Actually the DRI device should be always available, if I'm not mistaken, so I think it's a kernel error. But with last release, we slimmed down the X server install so that no additional GPU driver is installed automatically, while before we installed most of them, including AMD, legacy ATI, VMware and fallback VESA drivers on all SBCs, which does not harm but is pretty non-required. On Odroids, dedicated fbdev or fbturbo drivers are installed, on many other SBCs we ship an xorg.conf that explicitly uses the regular modesetting driver, which is always installed as part of the X.org core package, and, compared to fallback drivers, provides a basic GPU acceleration. On RPi, I wonder what might be the difference between 32-bit and 64-bit. @Joulinar since there is still not 64-bit/arm64 Chromium package available, does it mean the 32-bit/armhf package just installed and runs fine on RPi 64-bit (now)? There are two binaries, in |
Nope, I changed to |
Okay, I guess the RPi Chromium invokes the X server or GPU features internally differently. Dies Since we add armhf as foreign DPkg architecture on the 64-bit image, it could be actually possible to install I'll do some tests the next days, also to compare the different X/GPU drivers. |
not sure why, but out of curiosity I did another test on an older 64bit backup and this time it's failing as well. Not sure what is the difference to my first test now. Anyway same fix is working on 64bit also. |
Does starting a desktop work (without the fix)? |
starting the desktop is quite a challenge. If I install Same as this EDIT: REMOVED OBSOLETE URL Will need to play with it to get additional details... EDIT1
Combination of both issues is resulting in an empty screen with a blinking cursor.
|
Okay, to not force the use of the KMS driver, I'll re-add the The autostart setting is strange, as nothing is changed when cancelling the dialog, and if it's not cancelled, the autostart menu is shown but nothing automatically selected 🤔. |
nice behaviour of |
Verified failing X server without KMS overlay enabled. I can swear I tested it before, but either I did something wrong or something changed kernel/firmware-wise. However it sounds pretty logic that fbdev is for /dev/fb0 and modesetting/KMS is for /dev/dri/card0 and that KMS X display driver requires the KMS firmware driver enabled. Fixed both issues with: 7e8046c I'm currently running tests about which driver combination works best. Sadly with full KMS, probably the best performing, the monitor blanks after a few minutes and it's not possible to wake it up again. X server seems to have crashed 😢. glxgears FPS test fullscreen 1920x1200
It somehow makes all sense as KMS enables some 3D acceleration, while fbdev/fbturbo don't do it. But for desktop 2D AFAIK this is not responsible. With fbdev/fbturbo e.g. scolling in Firefox feels snappier. Important: Once the tbturbo package is installed, it is forcefully loaded, not as fallback, so that KMS is not used anymore by the X server. Another downside of fbturbo is tearing in videos, if I remember right. So I think installing fbdev by default and letting the user decide whether to enable OpenGL/3D acceleration or not is best. Chromium btw benefits from 3D acceleration, so probably we should enable the more compatible fake KMS on it's install by default, or recommend it at the end of the install. Also the DispmanX API stays available, which is used by some projects, like Amiberry and Kodi on RPi 4. |
Could this also be used for a Pi Zero? Thanks |
Please do instead:
|
there was a similar issue on a Odroid N+ running DietPi v7.3.2 EDIT: REMOVED OBSOLETE LINK running |
Makes sense, fixed with: e4ca471 |
I run into the issue again with updates to 7.6.2 but I could not fix it again.
|
You used wrong package name |
Creating a bug report/issue
Required Information
G_DIETPI_VERSION_CORE=7 G_DIETPI_VERSION_SUB=0 G_DIETPI_VERSION_RC=2' G_GITBRANCH='master' G_GITOWNER='MichaIng'
10.8
Linux DietPi 5.10.17-v7l+ #1403 SMP Mon Feb 22 11:33:35 GMT 2021 armv7l GNU/Linux
RPi 4 Model B (armv7l)
Additional Information (if applicable)
Steps to reproduce
Expected behaviour
Actual behaviour
Extra details
X.Org X Server logs:
The text was updated successfully, but these errors were encountered: