RC7 (STABLE) WeeDogLinux Arch 64 now released

A home for all kinds of Puppy related projects
Post Reply
Message
Author
User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#121 Post by fredx181 »

rockedge wrote:Hello Fred! to always make it eth0 add this line to the kernel command line ->
Code:
net.ifnames=0
Yes, that should work ok.
How is running X going for you now ?

Fred

User avatar
rockedge
Posts: 1864
Joined: Wed 11 Apr 2012, 13:32
Location: Connecticut, United States
Contact:

#122 Post by rockedge »

not quite.... strange....now I am either getting a blank screen with a hard reboot to get out of it or X starts xclock and xterm start with JWM...but no keyboard or mouse....tried many things...

my 32 bit version will not boot ... even the simplest version...could be that my partition is sdb1 and is an external USB drive? but it starts booting and goes into an endless loop with can't chroot and use setsid sh -c 'exec sh

not sure whats going on. the 64 bit I am about to add some of the mounts you suggested in my script...I noticed that as well when I could run firefox or palemoon with installing base-system and using Xnest and a Bionic64 host.

User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#123 Post by fredx181 »

rockedge wrote:my 32 bit version will not boot ... even the simplest version...could be that my partition is sdb1 and is an external USB drive? but it starts booting and goes into an endless loop with can't chroot and use setsid sh -c 'exec sh
First time test for me booting from external USB drive (before only from internal) and doesn't work for me either (endless loop, setsid not found).

Fred

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#124 Post by wiak »

fredx181 wrote:
rockedge wrote:my 32 bit version will not boot ... even the simplest version...could be that my partition is sdb1 and is an external USB drive? but it starts booting and goes into an endless loop with can't chroot and use setsid sh -c 'exec sh
First time test for me booting from external USB drive (before only from internal) and doesn't work for me either (endless loop, setsid not found).

Fred
@rockledge: Are you installing to external usb or to internal drive?

Anyway, I tried with bionicpup32 kernel and zdrv. Worked fine for me with internal drive.

I have never tried using FirstRib with any external usb sticks or somesuch. There is no wait for usb code in FirstRib, so maybe that is the problem. I tried installing onto usb flash stick (specified partition sdb1 when creating the initramfs.gz), but no good, got the endless loop.

I believe, from previous tests I now forget, that the endless loop indicates overlay.ko module couldn't be found. Since that is including in the zdr from BionicPup32, the implication is the init inside initramfs.gz cannot find firstrib_firmware_module.sfs (for my install it would be tying to find it at /mnt/sdb1/firstrib/firstrib_firmware_module.sfs). I can only presume at the moment that it can't find it cos usb not ready. Perhaps if I add a sleep before trying to mount the sfs... I'll try that in build initramfs script. Otherwise, I have no idea and can only suggest, at this stage, always use internal harddrive installs for now at least (as you know, FirstRib is not all-singing all dancing - it's minimalistic and thus with ease of script understanding but also with the limitations that result from its lack of 'intelligence').

As I say, using BionicPup32 vmlinuz and renamed zdrv works for me if installed to internal disc (such as sda5) but not when installed to usb flash disk (such as sdb1) which failes with reported endless loop.

wiak

User avatar
rockedge
Posts: 1864
Joined: Wed 11 Apr 2012, 13:32
Location: Connecticut, United States
Contact:

#125 Post by rockedge »

I finally have a X server working! with JWM, xterm, xclock. 64 bit on internal HDD sda1

firstrib is using 86 megs of RAM (top)

I will try to see if a sleep will let the externel USB drive (500 gigs) be seen using the 32 bit version

I have another 250 gig hard drive with a usb adaptor that is recognized much faster at boot than the 500 gig usb hdd which I will test out to see how that works....the 2 drives behave differently. I have not tried a usb stick yet at all.

the 64 bit firstrib with X and JWM going is very fast on this DELL Optiplex 990 quad core 8 megs of RAM

I can connect to the network and ping the router or another subnet machine but can NOT access internet so xbps-install will not work

persistence storage seems to be working

I installed base-minimal utils-linux xorg ncurses-base xterm jwm using firstrib with the mount script....umount script then built iniframfs.gz
added to the kernel command line : net.ifnames=0

User avatar
rockedge
Posts: 1864
Joined: Wed 11 Apr 2012, 13:32
Location: Connecticut, United States
Contact:

#126 Post by rockedge »

for a moment I could reach google.com but now I can't again.

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#127 Post by wiak »

rockedge wrote:for a moment I could reach google.com but now I can't again.
Sounds close. If you can ping your router, udhcpc -i <interface_name>
should have automatically set up /etc/resolv.conf for DNS working (e.g. ping google.com should work; can't imagine why it isn't - is that with eth?)

ip link

or

ip address

will show the interface names.

Yeah, not sure about how to get it working if installed to usb stick. Haven't tried again as yet.

wiak

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#128 Post by wiak »

Just put in a 10 second sleep immediately after udev & in the initramfs build script and with a 32bit firstrib_rootfs did:

Code: Select all

./build_firstrib_initramfs01ver007pre.sh sdb1
Then copied the resultant initramfs.gz to /mnt/sdb1/firstrib along with vmlinuz and firstrib_firmware_modules.sfs from BionicPup32 and (actually to my surprise) it booted successfully. I couldn't connect to wifi yet because I need to manually put my wifi firmware into that firstrib_firmware_modules.sfs but otherwise it did, with that sleep 10, manage to find the firstrib_firmware_modules.sfs and thus worked without the endless loop issue.

I'll produce build_firstrib_initramfs01ver008pre.sh now allowing optional entry of second argument on the commandline, usbwait=XXX.

My only concern is that I imagine it won't always work because udev can change names of partitions so sdb1 might not end up being correct for usb stick. But it will be a for testing release only (hence pre.sh). Will think about more complex (but not too complex) processing of arguments and udev matters later.

wiak

User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#129 Post by fredx181 »

wiak wrote:Just put in a 10 second sleep immediately after udev & in the initramfs build script and with a 32bit firstrib_rootfs
That works, it does boot fine now for me from sdb1 (external hdd).

Fred

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#130 Post by wiak »

Currently testing build_firstrib_initramfs01ver008pre

This one contains kernel boot params:

usbwait=number_of_seconds for slow devices,

and

bootfrom="<bootdir_location>"

e.g. bootfrom=/mnt/sda5/myfirstrib1 or whatever

Should be able to copy the resultant initramfs.gz to any such bootdir_location thereafter without having to rebuild it for particular partition (all going well...). Will save a lot of mucking around with very little extra code complexity at all (all done inside the init).

Have finished first draft of it, but not yet tested it at all (so likely won't work at all yet!)...

wiak

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#131 Post by wiak »

Per my above post. I was right... it didn't work... but, almost amazingly to me, it seems to now. I just have one extra to add (making /lib/modules and /lib/firmware mount binds rather than symlinks so they don't need to be removed from firstrib_rootfs if firmware from Void Linux happens to be loaded by some apps/templates, such as base-system or whatever).

So will likely post in 12 hours or so (after I now sleep).

wiak

User avatar
rockedge
Posts: 1864
Joined: Wed 11 Apr 2012, 13:32
Location: Connecticut, United States
Contact:

#132 Post by rockedge »

I also added a

Code: Select all

sleep 12
immediately after udev & in the initramfs build script and that worked and booted
now I am looking to find what modprobe to run on this particular machine for the eth0.

Update: success booting and connecting to network via eth0 on the 32 bit version using sdb1 which is an external usb hdd
also ping is successfull......modprobe b44 did the trick
Attachments
2019-07-04-154700_1280x1024_scrot.png
(65.83 KiB) Downloaded 504 times
2019-07-04-154107_600x400_scrot.png
(103.08 KiB) Downloaded 561 times

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#133 Post by wiak »

rockedge wrote: Update: success booting and connecting to network via eth0 on the 32 bit version using sdb1 which is an external usb hdd
also ping is successfull......modprobe b44 did the trick
Nice! Which browser is that in your screenshot (oh, now I see it is firefox)? I tried xbps-install firefox and also firefox-esr but had some issues, which I haven't addressed yet. How did you install it?

I'll be uploading build initramfs ver008pre later today.

Later I'll be adding option for changes/save only on shutdown with no_save option too. Such additions are simple enough. Actually, I wasn't really intending working to make FirstRib rootfs build bootable directly (since mainly for chroot on Linux host), but since it works, why not...

wiak

User avatar
rockedge
Posts: 1864
Joined: Wed 11 Apr 2012, 13:32
Location: Connecticut, United States
Contact:

#134 Post by rockedge »

the screenshots are from firstrib 64 bit .....the 32 bit version looks basically the same.

I used the latest firefox from the Void repos and I mounted /tmp to /dev/shm
as Fred suggested. Then firefox 67.04 starts very quickly

I built the firstrib file system and used mount_chrootver007.shmount.sh to chroot and start the system terminal.
then used

Code: Select all

xbps-install -Sy base-minimal xorg xterm xclock util-linux ncurses-base jwm mc
and installed these and the exited and ran umount.sh
ran build_firstrib_initramfs01_ver007pre.sh sdb1

Booted the system and modprobe e1000e (other system b44) connected and ping'ed the network successfully with google.com
I then installed rox, firefox, geany, htop which lands in /upper_changes

important was Fred's tip about mounting a /dev/shm for FF to work...also I did not modify /etc/X11/xinit/xinitrc so I started X server manually and
then in in an open terminal started jwm and rox -p default

the system is very responsive running with a 64 bit Bionic64-v8 kernel and zdrv.sfs
so far very fast and extremely low CPU loads on a quad core quad thread 8 gigs RAM
Attachments
2019-07-04-230440_600x480_scrot.png
(243.8 KiB) Downloaded 478 times

User avatar
fredx181
Posts: 4448
Joined: Wed 11 Dec 2013, 12:37
Location: holland

#135 Post by fredx181 »

rockedge wrote:important was Fred's tip about mounting a /dev/shm for FF to work
Actually that tip came from wiak.
Btw,I've found that it's only required for firefox to work that /dev/shm exists.
So only "mkdir /dev/shm" just before "udev &" in build_firstrib_initramfs01_ver007pre.sh should be enough.
(but eventually I think it will be good to have /dev/shm or /run/shm as separate tmpfs filesystem).
Looks very nice, rockedge !

Fred

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#136 Post by wiak »

Yes, your desktop is looking very good rockedge. Nice seeing jwm in FirstRib/WeeDog.
fredx181 wrote: (but eventually I think it will be good to have /dev/shm or /run/shm as separate tmpfs filesystem).
Actually, I'm about to upload build initramfs01 ver008pre, which includes tmptfs for /dev/shm and /run/shm (actually, I followed a recommendation I found elsewhere, and on the booted system have /dev/shm as a symlink to /run/shm.

Upload will be there in a few minutes...

As I said earlier, this build initramfs script contains support for slow boot devices and also for specifying boot partition/dir you wish to use. Does that via grub config (menu.lst for grub4dos) kernel-line arguments usbwait=duration (which is optional) and bootfrom=/mnt/partition/directory (which is required). No commandline parameter is used now in running the initramfs build script since the grub menu.lst line supplies the information instead. Big advantage is that the resulting initramfs.gz can be moved around between partitions (and into different dir names) without having to be re-built each time (same one will work everywhere - the init script inside the initramfs.gz sorts everything out based on what grub cmdline produces).

wiak
Last edited by wiak on Fri 05 Jul 2019, 09:46, edited 1 time in total.

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

build initramfs script ver008pre

#137 Post by wiak »

I've tested WeeDog with boot from my internal hd and also from usbstick (using different boot directory names) and with startx. All went fine (posting this right now, via wifi, using latest Firefox in WeeDog; I keep an eye on things with top and free command, but also lxtask is a small install).

EDIT: Also got sound working with youtube in Firefox now. For my system, I needed (did in this order I think...). Note that I found which snd module by looking at lsmod whist on my running XenialDog system and checking dependencies in lib/modules/kernel_version/modules.dep:

Code: Select all

xbps-install alsa-utils
xbps-install pulseaudio
modprobe snd_hda_intel
alsactl init
pulseaudio --start
(ignore warning about running as root)
Used alsamixer to turn up volumes, then
firefox &
[EDIT] NOTE: rufwoof has since added some extra information regarding getting sound system to work:

http://www.murga-linux.com/puppy/viewto ... 99#1033499

http://www.murga-linux.com/puppy/viewto ... 25#1033525

wiak
Last edited by wiak on Thu 15 Aug 2019, 08:35, edited 5 times in total.

User avatar
rockedge
Posts: 1864
Joined: Wed 11 Apr 2012, 13:32
Location: Connecticut, United States
Contact:

#138 Post by rockedge »

Hello wiak...I am about to boot ver 008

don't forget in your documentation for booting to add net.ifnames=0

Code: Select all

title firstrib (Void Linux Flavour)
  root (hd0,0)
  kernel /firstrib/vmlinuz  usbwait=12 bootfrom=/mnt/sda1/firstrib net.ifnames=0 
  initrd /firstrib/initramfs.gz

wiak
Posts: 2040
Joined: Tue 11 Dec 2007, 05:12
Location: not Bulgaria

#139 Post by wiak »

rockedge wrote:Hello wiak...I am about to boot ver 008

don't forget in your documentation for booting to add net.ifnames=0

Code: Select all

title firstrib (Void Linux Flavour)
  root (hd0,0)
  kernel /firstrib/vmlinuz  usbwait=12 bootfrom=/mnt/sda1/firstrib net.ifnames=0 
  initrd /firstrib/initramfs.gz
Good idea. That's an actual kernel option for keeping traditional interface names. I don't do that myself; I use the new form from udev, and check names with ip link command, but still a good alternative to use the likes of eth0 or wlan0 format, which grub menu.lst kernel-line option net.ifnames=0 forces; on some systems you may, rarely (only if biosdevname package installed), also need parameter biosdevname=0 if you choose to disable the so-called 'consistent network device naming' and use eth0, wlan0 and so on instead (but seems biosdevname pkg not used by Void Linux anyway):

https://access.redhat.com/discussions/916973
https://access.redhat.com/documentation ... ice_naming
https://wiki.archlinux.org/index.php/Ne ... rface_name

I'll add that in to docs somewhere. Thanks.

EDIT: For the moment, I've now included that suggestion (and link to this post) in the post/quick-howto I detailed setting up network connection:

http://www.murga-linux.com/puppy/viewto ... 22#1031022

wiak

User avatar
rockedge
Posts: 1864
Joined: Wed 11 Apr 2012, 13:32
Location: Connecticut, United States
Contact:

#140 Post by rockedge »

I completed a working system using the 008 version. the build went smoothly.

I used the kernel and zdrv.sfs for Bionic64-v8 4.19.23

once the file system was built I used the mount script and then ->

Code: Select all

xbps-install -Sy base-minimal bash xorg xterm xclock util-linux ncurses-base jwm mc
next run I will be using the firstrib.plug to perform the above commands during the initial build of the file system.

after the installation is completed I needed to replace the /usr/bin/sh symlink that points to dash with that symlink now pointing at /usr/bin/bash.
seems that using udhcpc -i eth0 needs that to work correctly

also modified /etc/X11/xinit/xinitrc to start JWM and rox -p default and a xterm terminal

Code: Select all

#twm &
#xclock -geometry 50x50-1+1 &
#xterm -geometry 80x50+494+51 &
#xterm -geometry 80x20+494-0 &
#exec xterm -geometry 80x66+0+0 -name login
jwm &
rox -p default &
exec xterm -geometry 80x66+0+0 -name login
then umount script is run followed by the build initramfs script.

boot code doesn't need the usbwait= in this case->

Code: Select all

title firstrib (Void Linux Flavour)
  root (hd0,0)
  kernel /firstrib/vmlinuz  bootfrom=/mnt/sda1/firstrib net.ifnames=0 
  initrd /firstrib/initramfs.gz
the system boots well and at the shell command line used xbps-install -Su and xbps-install -S to put rox, firefox,geany on the system.
I had to un-comment a line to set locales en_US.UTF-8 using

Code: Select all

xbps-reconfigure  -f glibc-locales
after all the little details typed startx in the shell and the desktop started right up.

did some cosmetic additions added scrot and NetworkManager ( which I have not really used at all yet and may remove it..don't know yet)

free- h returns ->

Code: Select all

sh-5.0# free -h
              total        used        free      shared  buff/cache   available
Mem:          7.7Gi       323Mi       6.4Gi       812Mi       993Mi       6.3Gi
Swap:            0B          0B          0B
_
Attachments
2019-07-06-093112_1280x1024_scrot.png
(247.49 KiB) Downloaded 731 times

Post Reply