Discussion, talk and tips
https://forum.puppylinux.com/
Same Machine Preprealpha4 still same problem with shutting down, Do I need to change, modify my grub4dos.
title Puppy Cagepup (sda1/cagepup)
root (hd0,0)
kernel /cagepup/vmlinuz psubdir=/cagepup pmedia=atahd pfix=fsck
initrd /cagepup/initrd.gz
Machine Comapaq Presario CQ61
Solved by adding kernel /cagepup/vmlinuz modprobe.blacklist=snd_hda_codec_idt
I have managed to cagepup to shutdown by pressing briefly power button. Tried acpi=off option in grub4dos but it affects my wireless.
Cagepup64 8.0 Alpha 4
Frugal install on a USB drive.
Computer is a home built desktop.
Intel i7 9700 3.0GHz,
Nvidia Geforce RTX 2060 graphics card.
Boots to working desktop.
Seems to be using the nouveau graphics driver.
Glxgears works OK.
Seems like the processor is being throttled to running at a very low speed.
Programs I have tried so far work, but the scrolling in them seems sluggish.
Tried a reboot and making a save. Worked OK.
A normal shutdown worked OK.
@darry19662018 your system make be different: Here's my situation on one of my alpha4 tests.
Scenario: Clicking Shutdown does nothing. Soooo, to shutdown, I am required to exit desktop to the prompt and enter "poweroff now". Then follow the normal Puppy shutdown questions.
@dimkr
Please fix the loopback.cfg requirement as soon as you can. Too much time is spent on moving from download ISO to media to testing. This is what makes the prior PUPs so advantagious, as, I merely-only download the ISO and boot directly for tests with NO other requirements.
Thanks for any helpful consideration to this minor boot need which mirrors the prior PUPs since 2019..
I got my copy of Vivaldi working on Cagepup64 and I'm using it to post this.
I adopted the brute force approach to the dependencies: I downloaded the latest Debian Bullseye basic live ISO and made an SFS of it's /usr/lib/x86_64-linux-gnu contents minus all of the subdirectories. It's 109MB.
With that loaded Vivaldi doesn't have any missing dependencies.
I've just tried VA-API on my Athlon 3000G box and found some discrepancies:
If I run vainfo I get:
libva error: vaGetDriverNameByIndex() failed with unknown libva error, driver_name = (null)
I can fix this error with:
export LIBVA_DRIVER_NAME=radeonsi
But this workaround has not been necessary with X11 for sometime, since libva can work out which driver to use. Maybe it can't determine the correct GPU type under Wayland.
While I was checking this, I also noticed that i965_drv_video.so was missing from /usr/lib/x86_64-linux-gnu/dri but the video drivers from mesa-va-drivers were present. The Intel driver is in a separate package.
Looks like it's not just VA-API related stuff that's missing, but also gstreamer/ffmpeg. Firefox refuses to play video in many sites.
I'll finish the NVIDIA bug hunting round on Dell laptops, then take a look at video decoding.
Almost done preparing for testing on my newest laptop - https://github.com/dimkr/woof-CE/compar ... i?expand=1.
It has no support (= no support at all, not disabled by default) for BIOS boot, so I had to make dpup OS capable of UEFI boot (using the kernel's "EFI stub" feature) before I can test the Xwayland solution with it.
How about fixing the shutdown issue - Never had this issue with other Pups.
darry19662018 wrote:How about fixing the shutdown issue - Never had this issue with other Pups.
I hope it's a race condition between Cage's death and xwin that happens only on fast or slow machines, so I prefer to test on this new Dell first and hope I'll reproduce the shutdown issue too.
Hi all, newbie here.
Am I the only one have this problem because all my partitions are in NTFS? Never had problems with other puppys (savefiles)!
It says "No suitable partitions to save to. Shutting down without saving."
BUG: Running a DVD AAP4 boot ... on shutdown, failure in writing the session to the booted DVD.
@NightKnight2, @Clarity: I'm intentionally ignoring any issue related to save files, PPM, etc', and focus only on stuff related to the Xwayland switch.
The goal is not to stabilize Cagepup64 and release a stable version, but to stabilize support for using Xwayland instead of Xorg in woof-CE, so any Puppy can do this.
Amazing stuff, now that UEFI boot works, I got the empty tray issue reproduced on this Dell Vostro 14 5490 (16 GB, 10510U, MX250):
Hi @dimkr, thanks for letting me know 
Awesome Wayland. A curated list of Wayland code and resources.
It seems that something is being done 
This is interesting, from the first release on May 17th, my installation was done using UEFI boot. I've had no issues with panels or any issues booting.
This is the boot stanza I have used:
Code: Select all
menuentry 'Cagepup64 8.0 (by dimkr on /dev/sda5 17/05/21)' {
search --set -f /cagepup/vmlinuz
linux /cagepup/vmlinuz root=UUID=d3271454-716f-4fde-b98b-85691d4b4c83 psubdir=/cagepup pmedia=ataflash
initrd /cagepup/initrd.gz
} UEFI is not causing this issue. Maybe it's a nouveau bug, so it affects only NVIDIA GPUs, and UEFI is related only because the only NVIDIA GPU I have at home is in a new laptop that supports only UEFI boot and doesn't have the option to turn on BIOS boot 
In the UEFI setup. See if it offers enable CSM.
That is the new setting to make the UEFI work as legacy bios.
Also in the UEFI setup look for disable secure boot.
What about disable secure boot?
May have to do that before enable CSM is active.
This UEFI could be new enough to just auto adjust to what it is seeing on the selected boot device.
Could try and use the nomodeset option added to the linux or kernel line of the boot loader entry.
The newest kernels have moved the video mode setting into the kernel.
So all the programming of the hardware specific clock rates and registers on the video card happen in the kernel rather than in the X driver when the X server starts..
This makes it possible to have high resolution nice looking splash (boot) screens and flicker free transitions from boot splash to login screen.
Unfortunately, on some cards this doesn't work properly and you end up with a black screen.
Adding the nomodeset parameter, instructs the kernel to not load video drivers and use BIOS modes instead, until X is loaded.
Shutdown issue is solved turns out to be a kernel issue in this case the 5.10 series of kernels and my machine. Anyway shutdown and savefile working now. Booted to desktop. Will test more later.
I'll try to switch to the woof-CE build of 5.4.x to see if the black screen issue is gone.
Maybe 5.10 is the source of more than one problem.
Connects to internet with Compaq CQ61.
Firefox general browsing ok but trying to view youtube videos crashes Firefox. Restarting Xserver leads to black screen,
Savefile works with boot parameter I used from Ozsouth.
Looks like downgrading the kernel to latest 5.4.x doesn't change anything on the Dell Vostro 14 5490. Still the same black screen issue @01micko is seeing on Dell Latitude E7270, and the errors are identical.
Just a quicknote as i've opnly just had a quick look at this topic.
If your having problems at least have a look at the notes on RebeccsblackOS, one of the first wayland builds, they have come across most of the problems by now.
Current pointer is https://sourceforge.net/projects/rebecc ... 021-05-26/
https://github.com/puppylinux-woof-CE/woof-CE/pull/2291 should fix issues with HTML 5 video playback in Firefox. I hope that once Firefox can play video, we won't see crazy regressions like graphical artifacts, blinking video output and CPU usage spikes.
Chrome seems to work just fine, and I haven't seen anything unusual.
https://github.com/puppylinux-woof-CE/w ... 2294/files adds VA API drivers, to make video decoding faster, at least in on some systems.
Once this one is merged into woof-CE, I'll publish another Cagepup64.
I have updated to A5 with a new savefile with Vivaldi working.
I think the problem with vainfo is caused by an omission in XWayland. There is a current issue on github vaapi about it, but the people reporting the problem and the people responding don't seem to understand each other, because it's not actually a vaapi problem.
The solution is to use:
vainfo --display drm
Then vainfo works correctly without needing to set an environment variable.
I assume that vainfo defaults to --display X11 but doesn't get an answer from XWayland.
So whether vaapi will work depends on whether the application has a workaround for it.
