Vanilla dpup 9.1.X observations

[Airdale]

Having a strange problem.... when I add new programs there are multiple duplicates present in the menu. Currently running 9.1.10.

There are no duplicates of the pre existing programs that came with Vanilla Dpup.

However, when I install a new program often there are two or even three entries in the menu for it.

My first instinct was that maybe there were extra desktop files. However, I looked in usr/shared/applications, and there was only one desktop file for each application.

I'm using the Synaptic package manager - Could that be the problem? Does Dpup play nicer with the PPM?

Anybody else notice this?

Thanks so much

[TerryH]

snip...
My first instinct was that maybe there were extra desktop files. However, I looked in usr/shared/applications, and there was only one desktop file for each application.

It's not that additional .desktop files exist, it is more than likely that the existing .desktop files contain multiple entries for 'Categories=', this may generate multiple Menu entries.

[dimkr]

9.1.11 will feature https://github.com/puppylinux-woof-CE/woof-CE/pull/2899

[dimkr]

9.1.11 is out and I don't see complaints about new issues. This makes me worried

[williwaw]

updated to 9.1.11 and unwanted suspend? and/or screen blackout is fixed.
I did have to delete my old upper, as first boot after update still lost the screen and locked up the laptop

[Feek]

Observations 1
It seems to me that Xwayland is consuming more resources on my hardware (CPU usage).
lxtask shows CPU usage of approx. 60% when playing video on Youtube using Firefox (although playback is smooth).
After switching to Xorg (and reboot), this value is approx. 30% when playing the same video.

Is this normal with Xwayland?
Or should I switch to Xorg permanently?

Observations 2
In Tray-configuration I can't place application buttons between pager and tasklist (drag and drop).
They either disappear when moving up or remain next to the clock when the Apply button is pressed.

(version 9.1.11)

Thanks for your work.

[dimkr]

Is this normal with Xwayland?

What CPU and GPU do you have?

Technically, Wayland can consume more CPU if the GPU driver is missing, because the fallback is software acceleration (while X doesn't have any kind of acceleration). But this mostly happens in virtual machines.

Also, some browsers disable VAAPI in some setups. The increased CPU usage can happen if it's disabled under Xwayland (because VAAPI makes the GPU handle video decoding, instead of the CPU).

Which process consumes 60% of CPU - is it Firefox?

[Feek]

Thanks for the explanation.

What CPU and GPU do you have?

CPU is Intel core i3-2328M

GPU:
VGA compatible controller
Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller
VendorID=8086 DeviceID=0116 Rev=09
Kernel Driver=i915
Kernel Module=i915

Which process consumes 60% of CPU - is it Firefox?

(I wanted to attach a picture of lxtask window but after pressing "Preview" the picture is invisible)
Captured values:
total CPU usage=62 %
firefox-esr.bin=44 %
Web Content=8 %
RDD Process=5 %
Xwayland=1 %
cage=1 %
WebExtensions=0 %....

[dimkr]

I have a similar laptop (i5, 2nd gen). AFAIK, VAAPI support is very limited, because VAAPI supports only some codecs with this GPU, and sometimes YouTube chooses a different one (which can't be decoded through the GPU) if your internet connection was slower in the moment you clicked the video.

I'd recommend you to try the instructions in https://wiki.archlinux.org/title/Firefo ... celeration, force-enable WebRender or install the h264ify extension.

In another laptop, with a 5th gen i5, I don't have any of these problems.

[retiredt00]

Dear @dimkr thanks for 9.1.11. Works really well.
gdebi integration is very handy (though is missing menu icon and also complains about missing icons when loading from CLI)

I was wondering as the distro matures if you could turn your attention in more trivial puppy things, like being able to see the content of an SFS (BTW how can you see the content of the vanillapup SFSs from a running 9.1.X? i.e. the equivalent of /initrd/pup_* in standard puppy), stop hiding bootmanager/sfs_load etc.

The other issue is the installer and updater. They are designed to work from their own hard drive/USB stick. You can of course manually install them in a single partition but then the update fails. This is too limiting I think.

This issue is relevant to the fact that frugalified vDPup does not play well along other puppies as it needs to have its SFSs and /upper at the root of the partition. I guess you would suggest 9.0.X for this setup but 9.0.X is missing the apt-integration. Is there any reason for omiting this major feature from the initrd-based vanillaDpups?

Thanks again

[dimkr]

i.e. the equivalent of /initrd/pup_* in standard puppy)

It can be done using bind mounts, I'll see what I can do.

stop hiding bootmanager/sfs_load etc.

They're hidden because they're broken - they don't work with overlayfs.

The other issue is the installer and updater. They are designed to work from their own hard drive/USB stick. You can of course manually install them in a single partition but then the update fails. This is too limiting I think.

Installers are indeed a sore spot for Puppy. I tried to build a full-featured installer based on Calamares, but it was huge and I decided to start with something basic, until I find a better solution.

This issue is relevant to the fact that frugalified vDPup does not play well along other puppies as it needs to have its SFSs and /upper at the root of the partition.

Should I add psubdir= support to frugalify?

Is there any reason for omiting this major feature from the initrd-based vanillaDpups?

Consider initrd-based builds to be 100% unsupported. woof-CE's initrd.gz init script is in bad shape, and initrd.gz is built using some ancient prebuilt binaries that cannot be reproduced. At least until I find the time to clean it up and make it fully reproducible by anyone, I want to avoid any use of initrd.gz.

[retiredt00]

Should I add psubdir= support to frugalify?

Please do, if is not a lot of trouble for the (current) installer, updater, efilinux, etc

[Feek]

I'd recommend you to try the instructions in https://wiki.archlinux.org/title/Firefo ... celeration, force-enable WebRender or install the h264ify extension.

@dimkr ,
just FYI (and for anyone else with the same problematic GPU).

I followed the instructions from your link (I hope correctly).
In Firefox in about:config, I changed the following settings:

gfx.webrender.all to true;
media.ffmpeg.vaapi.enabled to true;
media.ffvpx.enabled to false;
media.navigator.mediadatadecoder_vpx_enabled to true;
media.rdd-vpx.enabled to false;

then I restarted Firefox.

Video playback on youtube: CPU usage = approx. 30%

Playback of the same video before changing settings: CPU usage = approx. 60%

This solution seems to have fixed the problem.
Thank you.

[williwaw]

Should I add psubdir= support to frugalify?

yes, thanks you

[pemasu]

Code: Select all

Should I add psubdir= support to frugalify?

In Puppyland people often have kennel of Puppies placed each of them on their own folder inside same partition. Because of that booting from subdir is also normal thing. So the answer is yes.

[retiredt00]

I have noticed that in Xwayland when changing magnification the touchpad is becoming way too sensitive to control although it is OK at native resolution.
in Xorg is behaving OK in all resolutions.
However neither in Xorg nor in Xwayland I could find a utility to control touchpad characteristics (pupx/FlSynclientt and the likes). Is there one? I could not find anything (of reasonable size) in the Debian repo either, though “big distros” have these utilities somehow.
Thanks for any suggestions.

BTW regarding viewing SFSs in 9.1.11, I came across this post https://github.com/puppylinux-woof-CE/w ... 1034106164 and (after strangling a bit to find out what to do with it) it appears to successfully mount SFSs for viewing

[dimkr]

However neither in Xorg nor in Xwayland I could find a utility to control touchpad characteristics (pupx/FlSynclientt and the likes). Is there one?

Yes, you have the input devices wizard (input-wizard):

It's much more powerful under Xwayland.

[Maybe]

Hey!

I really like how apt works! I installed Google Chrome and it works great! Thank you very much @dimkr !

I was unable to use Zoom. I downloaded the .deb package from the official debian site for similar distributions. After installation, the Zoom application starts, but in the form of a single-colored window. Unable to control as there are no buttons.

From Dmitry (Russia)

[retiredt00]

Good day
Is suspend on lid closure activated?
Kernel acpi appears to correctly detect lid state but the screen light stays on.
Same for the rest of the functions (WiFi etc)
THX

[dimkr]

Is suspend on lid closure activated?

Yes.

Does echo mem > /sys/power/state work on this machine? If not, this is probably why suspend on lid close doesn't.

[retiredt00]

Is suspend on lid closure activated?

Yes.

Does echo mem > /sys/power/state work on this machine? If not, this is probably why suspend on lid close doesn't.

Yes echo mem > /sys/power/state suspends (and wakes up) fine.

[pemasu]

Yes echo mem > /sys/power/state suspends (and wakes up) fine.

@retiredt00. Do you have external usb drive or usb flashdrive attached and mounted when you try to suspend ?
From /etc/acpi/actions/suspend.sh which /etc/acpi/LID performs when closing lid:

Code: Select all

# do not suspend if usb media mounted
USBS=$(probedisk2|grep '|usb' | cut -d'|' -f1 )
for USB in $USBS
do
	mount | grep -q "^$USB" && rm -f "$LOCKFILE" && exit 0
done

[retiredt00]

Yes echo mem > /sys/power/state suspends (and wakes up) fine.

@retiredt00. Do you have external usb drive or usb flashdrive attached and mounted when you try to suspend ?

Dear @pemasu indeed is booting from a USB as per vDpup instructions, but it should be more to it
I also put Bionic64 in the same stick and suspends fine on lid closure.
However, it suspends only if it is running with a savefile as a direct savelayer, i.e. pupmode 12, but not with a savefolder in pupmode 12.
In both cases "$USB" has a value but somehow with the savefile it goes through

[pemasu]

@retiredt00. Something in that suspend.sh script prevents the suspend. I had that problem too. I just downloaded acpitool. And I have in that script only this: acpitool -s.
I use acpitool -s also with cbatticon to suspend when battery is low.

[dimkr]

It would be nice to see what sh -x /etc/acpi/actions/suspend.sh has to say.

[retiredt00]

@pemasu acpitool -s works!. Thanks

@dimkr below is the output suspend.sh in 5 different settings.
in vDpup fails both booting from USB and HD
In bionic64 suspends OK from HD, fails from USB with a savefolder but suspends OK with a savefile after a failed call to snapmergepuppy bypassing the USB check.
Hopefully you’ll make sense of it.

Code: Select all

#### vDpup on USB -> NO suspend ####

~$ sh -x /etc/acpi/actions/suspend.sh 
+ ACPI_CONFIG=/etc/acpi/acpi.conf
+ '[' -s /etc/acpi/acpi.conf ']'
+ . /etc/acpi/acpi.conf
+ case "$DISABLE_SUSPEND" in
+ LOCKFILE=/tmp/acpi_suspend-flg
+ '[' -f /tmp/acpi_suspend-flg ']'
+ echo -n 6258
+ sync
++ cat /tmp/acpi_suspend-flg
+ '[' 6258 = 6258 ']'
++ ps
+ PS='PID   USER     TIME  COMMAND
    1 root      0:01 /frugalify
...
...
 6258 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 6261 root      0:00 busybox ps'
+ '[' '!' -f /tmp/suspend ']'
+ grep -qE 'sh[ ].*poweroff'
+ echo 'PID   USER     TIME  COMMAND
    1 root      0:01 /frugalify
...
...
 6258 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 6261 root      0:00 busybox ps'
+ rm -f /tmp/suspend
+ set -x
++ probedisk2
++ grep '|usb'
++ cut '-d|' -f1
+ USBS='/dev/sdb
/dev/sdc'
+ for USB in $USBS
+ mount
+ grep -q '^/dev/sdb'
+ rm -f /tmp/acpi_suspend-flg
+ exit 0


#### vDpup on HD -> NO suspend ####

~$ sh -x /etc/acpi/actions/suspend.sh 
+ ACPI_CONFIG=/etc/acpi/acpi.conf
+ '[' -s /etc/acpi/acpi.conf ']'
+ . /etc/acpi/acpi.conf
+ case "$DISABLE_SUSPEND" in
+ LOCKFILE=/tmp/acpi_suspend-flg
+ '[' -f /tmp/acpi_suspend-flg ']'
+ echo -n 4419
+ sync
++ cat /tmp/acpi_suspend-flg
+ '[' 4419 = 4419 ']'
++ ps
+ PS='PID   USER     TIME  COMMAND
    1 root      0:01 /frugalify
...
...

 4419 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 4422 root      0:00 busybox ps'
+ '[' '!' -f /tmp/suspend ']'
+ grep -qE 'sh[ ].*poweroff'
+ echo 'PID   USER     TIME  COMMAND
    1 root      0:01 /frugalify
...
...
 4419 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 4422 root      0:00 busybox ps'
+ rm -f /tmp/suspend
++ probedisk2
++ cut '-d|' -f1
++ grep '|usb'
+ USBS=/dev/sda
+ for USB in $USBS
+ mount
+ grep -q '^/dev/sda'
+ rm -f /tmp/acpi_suspend-flg
+ exit 0


#### Bionic with savefolder on USB -> NO suspend ####

root# sh -x /etc/acpi/actions/suspend.sh 
+ ACPI_CONFIG=/etc/acpi/acpi.conf
+ '[' -s /etc/acpi/acpi.conf ']'
+ . /etc/acpi/acpi.conf
+ case "$DISABLE_SUSPEND" in
+ LOCKFILE=/tmp/acpi_suspend-flg
+ '[' -f /tmp/acpi_suspend-flg ']'
+ echo -n 5517
+ sync
++ cat /tmp/acpi_suspend-flg
+ '[' 5517 = 5517 ']'
++ ps
+ PS='PID   USER     TIME  COMMAND
    1 root      0:01 /bin/busybox init
...
...
 5517 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 5520 root      0:00 busybox ps'
+ '[' '!' -f /tmp/suspend ']'
+ grep -qE 'sh[ ].*poweroff'
+ echo 'PID   USER     TIME  COMMAND
    1 root      0:01 /bin/busybox init
...
...
 5517 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 5520 root      0:00 busybox ps'
+ rm -f /tmp/suspend
++ probedisk2
++ grep '|usb'
++ cut '-d|' -f1
+ USBS='/dev/sdb
/dev/sdc'
+ for USB in $USBS
+ mount
+ grep -q '^/dev/sdb'
+ rm -f /tmp/acpi_suspend-flg
+ exit 0


#### Bionic with savefolder on HD ->  suspends OK ####

root# sh -x /etc/acpi/actions/suspend.sh 
+ ACPI_CONFIG=/etc/acpi/acpi.conf
+ '[' -s /etc/acpi/acpi.conf ']'
+ . /etc/acpi/acpi.conf
+ case "$DISABLE_SUSPEND" in
+ LOCKFILE=/tmp/acpi_suspend-flg
+ '[' -f /tmp/acpi_suspend-flg ']'
+ echo -n 4251
+ sync
++ cat /tmp/acpi_suspend-flg
+ '[' 4251 = 4251 ']'
++ ps
+ PS='PID   USER     TIME  COMMAND
    1 root      0:00 /bin/busybox init
...
...
 4251 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 4254 root      0:00 busybox ps'
+ '[' '!' -f /tmp/suspend ']'
+ grep -qE 'sh[ ].*poweroff'
+ echo 'PID   USER     TIME  COMMAND
    1 root      0:00 /bin/busybox init
...
...
 4251 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 4254 root      0:00 busybox ps'
+ rm -f /tmp/suspend
++ probedisk2
++ cut '-d|' -f1
++ grep '|usb'
+ USBS=/dev/sda
+ for USB in $USBS
+ mount
+ grep -q '^/dev/sda'
+ sync
+ rmmod ehci_hcd
rmmod: ERROR: Module ehci_hcd is builtin.
+ echo -n mem
+ modprobe ehci_hcd
+ rm -f /tmp/acpi_suspend-flg


#### Bionic with savefile on USB -> suspends OK ####

root# sh -x /etc/acpi/actions/suspend.sh 
+ ACPI_CONFIG=/etc/acpi/acpi.conf
+ '[' -s /etc/acpi/acpi.conf ']'
+ . /etc/acpi/acpi.conf
+ case "$DISABLE_SUSPEND" in
+ LOCKFILE=/tmp/acpi_suspend-flg
+ '[' -f /tmp/acpi_suspend-flg ']'
+ echo -n 6242
+ sync
++ cat /tmp/acpi_suspend-flg
+ '[' 6242 = 6242 ']'
++ ps
+ PS='PID   USER     TIME  COMMAND
    1 root      0:01 /bin/busybox init
...
...
 6242 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 6245 root      0:00 busybox ps'
+ '[' '!' -f /tmp/suspend ']'
+ grep -qE 'sh[ ].*poweroff'
+ echo 'PID   USER     TIME  COMMAND
    1 root      0:01 /bin/busybox init
...
...
 6242 root      0:00 sh -x /etc/acpi/actions/suspend.sh
 6245 root      0:00 busybox ps'
+ rm -f /tmp/suspend
+ snapmergepuppy
/usr/sbin/snapmergepuppy: Wrong PUPMODE (12)
+ sync
+ rmmod ehci_hcd
rmmod: ERROR: Module ehci_hcd is builtin.
+ pm-suspend

+ modprobe ehci_hcd
+ rm -f /tmp/acpi_suspend-flg

[dimkr]

Looks like suspend.sh has changed through the years - and yes, it's exactly as @pemasu says: it refuses to suspend because you have a flash drive plugged in.

[pemasu]

Dpup updater does not update vmlinuz and ucode.cpio in my setup. I have to manually download latest dpup and extract them. All the other files updates fine.
I have small fat partition, which has EFI boot files. It is unused and I cant recall why I have it. I am not sure if that partition is the reason why the updater hangs during EFI handling.

Code: Select all

Updating zdrv_vanilladpup_9.1.12.sfs
+ mv -f zdrv_vanilladpup_9.1.12.sfs ..
+ sync
+ '[' -n /dev/sda10 ']'
+ echo 'Updating boot loader'
Updating boot loader
+ cp -f efilinux.efi /mnt/efi/EFI/BOOT/BOOTX64.EFI
cp: tavallisen tiedoston '/mnt/efi/EFI/BOOT/BOOTX64.EFI' luominen ei onnistu: Tiedostoa tai hakemistoa ei ole
++ cleanup
+ '[' 1 -ne 0 ']'
+ echo 'Failed to move efilinux to EFI partition'
Failed to move efilinux to EFI partition
+ '[' ask = ask ']'
+ read

I suspect that the script should update vmlinuz and ucode.cpio after that step ? And because it does not proceed, the updater does not copy vmlinuz and ucode.cpio.
After reboot obviously the system is unusable.

[dimkr]

The updater is not super smart. It assumes the first FAT32 partition is your ESP, assumes the kernel and initrd are there, and assumes you're using efilinux.efi as the default (bootx64.efi).

If you installed Vanilla Dpup manually, and not using the installer, you have to take the updater's assumptions into consideration if you want it to work.

[pemasu]

If I remove the fat partition, and the hdd does not have any fat partitions left, would the updater update vmlinuz and ucode.cpio where all the .sfs files are ( which are updated fine ) ?