Tiny Core Linux
Tiny Core Base => Raspberry Pi => piCore Test Releases => Topic started by: Paul_123 on May 06, 2023, 06:22:39 PM
-
Team Tiny Core is happy to announce the first beta release of piCore-14.0. Both 32 and 64bit images are ready for download/testing.
rpi-kernel 6.1.25
gcc 12.2
binutils 2.39
glibc 2.37
busybox 1.36
micropython 1.19
libffi (v8 library)
piCore 14 also brings openssl to 3.1.0. For applications in the 14.x repo not yet rebuilt, there are compatibility extensions created openssl-1.1.1 and libffi7
32bit - http://www.tinycorelinux.net/14.x/armv7/test_releases/RPi/piCore-14.0-beta1.zip (http://www.tinycorelinux.net/14.x/armv7/test_releases/RPi/piCore-14.0-beta1.zip) (This should run on all rpi boards)
64bit - http://www.tinycorelinux.net/14.x/aarch64/test_releases/RPi/piCore64-14.0-beta1.zip (http://www.tinycorelinux.net/14.x/aarch64/test_releases/RPi/piCore64-14.0-beta1.zip) (rpi zero2W, pi3, pi4)
-
PiCore64:
Tested with my current setup: jwm, pcmanfm, firefox, geany, transmission-gtk, ...
Putting the corresponding new dependencies:
-glib2 from libffi7
-transmission-gtk from openssl-1.1.1
everything seems to work correctly.
Note: the libzstd extension has some spurious files in it.
-
Lots to recompile for OpenSSL 3. But thanks for the info, I’ll check it out.
-
Hi
aarch64
#######
EDIT missing mouse pointer fix post below
anyhow I am using ssh at the moment, I tried to use ssh with cpanel (laughs at knowing little over the years)
but here is a possible issue for you Paul_123
tc@box:~$ tce-load -i inxi
perl5.tcz: OK
libpci.tcz: OK
pciutils.tcz: OK
touch: /usr/local/tce.installed/readline: Permission denied
readline.tcz: OK
bash.tcz: OK
libattr.tcz: OK
libacl.tcz: OK
gmp.tcz: OK
coreutils.tcz: OK
inxi.tcz: OK
tc@box:~$ ls -al /usr/local/tce.installed/ | grep readline
tc@box:~$
I downloaded it here on TC64 and unsquashing it...altho I can not immediately spot the issue....it has Pkgconfig dir in it.
I have a match for http://tinycorelinux.net/14.x/aarch64/tcz/readline.tcz.md5.txt
2) So I then on TC64 downloaded the aarch readline-dev and its missing the pkgconfig file.
Hope you know why I am getting permission error....but inxi works as
inxi -F
System:
Host: box Kernel: 6.1.25-piCore-v8 arch: aarch64 bits: 64 Console: pty pts/0
Distro: piCore 14.0
Machine:
Type: ARM System: Raspberry Pi 4 Model B Rev 1.1 details: BCM2835
rev: c03111 serial: 10000000b0c9e281
CPU:
Info: quad core model: N/A variant: cortex-a72 bits: 64 type: MCP cache:
L2: 1024 KiB
Speed (MHz): avg: 600 min/max: 600/1500 cores: 1: 600 2: 600 3: 600 4: 600
Graphics:
Device-1: bcm2711-hdmi0 driver: N/A
Device-2: bcm2711-hdmi1 driver: N/A
Device-3: bcm2711-vc5 driver: N/A
Display: server: X.org driver: X: loaded: fbdev unloaded: modesetting
gpu: N/A tty: 80x40
API: OpenGL Message: GL data unavailable in console and glxinfo missing.
Audio:
Device-1: bcm2711-hdmi0 driver: N/A
Device-2: bcm2711-hdmi1 driver: N/A
Network:
Device-1: bcm2835-mmc driver: mmc_bcm2835
Device-2: bcm2711-genet-v5 driver: bcmgenet
IF: eth0 state: up speed: 1000 Mbps duplex: full mac: dc:a6:32:4b:4a:65
Drives:
Local Storage: total: 29.72 GiB used: 108.6 MiB (0.4%)
ID-1: /dev/mmcblk0 vendor: SanDisk model: SC32G size: 29.72 GiB
Partition:
Message: No partition data found.
Swap:
ID-1: swap-1 type: zram size: 931.9 MiB used: 0 KiB (0.0%) dev: /dev/zram0
Sensors:
Src: /sys System Temperatures: cpu: 44.8 C mobo: N/A
Fan Speeds (RPM): N/A
Info:
Processes: 89 Uptime: 0h 19m Memory: available: 3.71 GiB
used: 197.1 MiB (5.2%) Init: SysVinit Shell: ash (busybox) inxi: 3.3.27
-
ok I think I have solved my invisible mouse issue.
I have startx with openbox and mouse pointer now appears, so it may be just a gremlin for my mouse in jwm?
its a gecko wired usb mouse, dmesg
[ 2.519245] input: USB OPTICAL MOUSE as /devices/platform/scb/fd500000.pcie/pci0000:00/0000:00:00.0/0000:01:00.0/usb1/1-1/1-1.3/1-1.3:1.0/0003:0000:3825.0001/input/input0
extra info.
mouse pointer was also hidden in fltk-1.3.
I then loaded openbox but not ssh session but for real on RPi and tried startx....startx claimed
/etc/sysconfig/Xserver not found.
Maybe openbox depends on X11....but anyhow I loaded Xorg and the startx works and then mouse pointer appears
-
aarch64
leafpad missing deps appear to be
libXrandr.tcz libXcomposite.tcz
thanks for reading
-
libXrandr, libXcomposite, libXcursor and libXinerama added to gtk2 deps
-
Thanks for the reports. The touch error with readline normally means another extension is misbehaving. What extensions in your tce.installed dir have an actual script? It will be likely be one of those.
-
EDIT
ok so I made onboot.lst only openssh and aterm (if anyone has time to reproduce my error) and I am loading using ssh TC64
tce-load -i inxi
perl5.tcz: OK
libpci.tcz: OK
pciutils.tcz: OK
touch: /usr/local/tce.installed/ncurses: Permission denied
ncurses.tcz: OK
touch: /usr/local/tce.installed/readline: Permission denied
readline.tcz: OK
bash.tcz: OK
libattr.tcz: OK
libacl.tcz: OK
gmp.tcz: OK
coreutils.tcz: OK
inxi.tcz: OK
tc@box:~$ ls -al /usr/local/tce.installed/
total 0
drwxrwsr-x 2 root staff 440 May 8 19:35 ./
drwxr-sr-x 9 root root 180 May 8 19:35 ../
lrwxrwxrwx 1 root root 47 Jan 1 1970 aterm -> /tmp/tcloop/aterm/usr/local/tce.installed/aterm
lrwxrwxrwx 1 root root 45 May 18 12:54 bash -> /tmp/tcloop/bash/usr/local/tce.installed/bash
lrwxrwxrwx 1 root root 67 Jan 1 1970 ca-certificates -> /tmp/tcloop/ca-certificates/usr/local/tce.installed/ca-certificates
-rw-r--r-- 1 tc staff 0 May 18 12:54 coreutils
-rw-r--r-- 1 tc staff 0 May 18 12:54 gmp
lrwxrwxrwx 1 root root 45 May 18 12:54 inxi -> /tmp/tcloop/inxi/usr/local/tce.installed/inxi
-rw-r--r-- 1 tc staff 0 Jan 1 1970 libICE
-rw-r--r-- 1 tc staff 0 Jan 1 1970 libSM
lrwxrwxrwx 1 root root 49 Jan 1 1970 libX11 -> /tmp/tcloop/libX11/usr/local/tce.installed/libX11
-rw-r--r-- 1 tc staff 0 Jan 1 1970 libXau
-rw-r--r-- 1 tc staff 0 Jan 1 1970 libXdmcp
-rw-r--r-- 1 tc staff 0 Jan 1 1970 libXext
-rw-r--r-- 1 tc staff 0 May 18 12:54 libacl
-rw-r--r-- 1 tc staff 0 May 18 12:54 libattr
-rw-r--r-- 1 tc staff 0 May 18 12:54 libpci
-rw-r--r-- 1 tc staff 0 Jan 1 1970 libxcb
-rw-r--r-- 1 tc staff 0 Jan 1 1970 openssh
-rw-r--r-- 1 tc staff 0 Jan 1 1970 openssl
lrwxrwxrwx 1 root root 53 May 18 12:54 pciutils -> /tmp/tcloop/pciutils/usr/local/tce.installed/pciutils
-rw-r--r-- 1 tc staff 0 May 18 12:54 perl5
did not see ncurses before
I know we can do anything in piCore different from TC64 but on TC64 submitqc I think tries to use root:staff for tce.installed dir
-
For applications in the 14.x repo not yet rebuilt, there are compatibility extensions created openssl-1.1.1 and libffi7
thanks found I needed libffi7 otherwise
gdk-pixbuf-query-loaders: error while loading shared libraries: libffi.so.7: cannot open shared object file: No such file or directory
msg appears in manually loading openbox
-
pciutils had the permission problem. - fixed
readline - fixed
-
Hi Paul_123
Thanks for that. in Apps search for tag=terminal AFAIK no terminals in aarch64 depend on vte*.
I plan to build lxterminal which provisionally might depend on vte-2.91
would you prefer TCE name as vte or vte-2.91
vte-2.91 aligns with TC64 TCE name if interested
-
hi mortegai
after updates -> in a terminal does pcmanfm now seg fault?
I have built lxterminal and there is update for pcmanfm.....but I think I will wait for Paul_123 to catch up with openssl and libffi stuff before submitting stuff.
-
I'm fine with aligning with TC64. And if there is anything in the dep chain that you need. Shoot me a PM.
-
hi mortegai
after updates -> in a terminal does pcmanfm now seg fault?
I have built lxterminal and there is update for pcmanfm.....but I think I will wait for Paul_123 to catch up with openssl and libffi stuff before submitting stuff.
I don't see any update for pcmanfm.
For me, in a terminal (urxvt) pcmanfm works correctly, it only shows these warning messages:
root@box:~# pcmanfm
** (pcmanfm:7583): WARNING **: 17:56:46.979: The directory '~/Templates' doesn't exist, ignoring it
** (pcmanfm:7583): WARNING **: 17:56:47.047: modules directory is not accessible
(pcmanfm:7583): Gdk-WARNING **: 17:56:47.106: gdk_window_set_icon_list: icons too large
root@box:~#
-
hi mortegai
my poor writing I meant there is an update upstream for pcmanfm.....and I will build it.
The "updates" I was referrring to was Apps -> check for updates or the command line method if thats what you use.
EEK then the seg fault is something I have done....I will need to ponder my changes. I have made some. As pcmanfm is a GUI I will need a desktop so I will try to see over the weekend what I have done wrong. Or not bother and just do an update of pcmanfm.
-
to anyone who has an invisible mouse pointer, I may have found the solution, from Juanito post 6 for gtk2 dependencies
libXcursor
I made a basic.lst to test out my theory and so I am typing on RPi with firefox-perf loaded after success, visible pointer
cat basic.lst
libXcursor.tcz
libffi7.tcz
jwm.tcz
aterm.tcz
wbar.tcz
I am not expecting anyone to mod their cmdline.txt unless you are weird like me ;)
I append lst=basic.lst after noembed
Naturally my onboot.lst is much larger
-
Sorry Paul_123 I am bit confused on LA files.
I am starting to build pcmanfm update and I removed the la file from libfm-extra-dev and built menu-cache and removed its la file
When I then attemped to make libfm....it complains
libtool: error: '/usr/local/lib/libharfbuzz.la' is not a valid libtool archive
instead of me doing my job correctly ;) I searched apps only for harfbuzz-dev and spotted no la file
so I adjusted my build to download the TC64 12x harfbuzz-dev.tcz because its la file is not in
http://tinycorelinux.net/12.x/x86_64/tcz/src/la/
then my script copies the unpack la file into /usr/local/lib and then I checked and realized we have heaps of LA files.
There is a post by Juanito or Rich for TC64 that says that LA files are no longer loved for TC64 and maybe TC32?
Can you advise what aarch64 intentions are....forgive my apperance that I vaguely know what I am talking about please.
Happy to leave in la files if that is the answer!
-
For *la files to work with autotools they have to either all be there or be totally absent. Since they can’t all be there, they all need removing.
If you delete all *la files from /usr/lib and /usr/local/lib, delete the source and start again things will work.
-
I have been removing .la files as I upgrade extensions.
-
EDIT, me too. Getting late here so I have removed la files in new update attempt but I am still got issues with segmentation fault.
I have eliminated xdg missing $HOME directories
I may have to not strip and atach a gdb as a last resort.
thanks for reading
-
Hi aus9
Juanito had me remove the .la files from the x86 and x86_64 repos.
I did not update any of the piCore repos.
-
Hi aus9
Juanito had me remove the .la files from the x86 and x86_64 repos.
I did not update any of the piCore repos.
You have a script for that? I don’t see why we would not do the same. In the 14.x repos.
-
Hi Paul_123
If you look in my home directory on the repo server there is
a script called CopyRepo.sh. One of the steps it performs is
checking the .list files for filenames ending in .la.
This is the description included in the beginning of the script:
#!/bin/sh
# Written by Richard A. Rost Jan 13, 2023
# Script to copy extensions from one repo version to another.
#
# The .tcz, .tcz.info, .tcz.dep, .tcz.list, and .tcz.md5.txt files get copied.
# The gen_tcz_updates script creates the .tcz.tree and .tcz.zsync files.
#
# Files get copied from source directory to $HOME/staging directory.
#
# The following do not get copied:
# module files.
# files that already exist in the destination repo.
# perl*.tcz files
#
# mpg123*.tcz gets skipped from the following when encountered.
# Extensions that contain .la files:
# get unsquashed
# .la files get removed
# get squashed
# .list file gets updated
# .md5.txt file gets updated
#
#
#-----------------------------------------------------------------------------
-
Hi Paul_123 and Juanito
I have posted about an issue I had for bootcode. It appears our README used NOZSWAP in caps while Rich confirms it is lowercase.
Maybe for beta2 etc we can adjust that term, I followed the remark literally showing my lack of knowledge etc
2) Can someone have a look at the dep file for abiword aarch64 please as
tce-load -w abiword
abiword.tcz.dep OK
wv.tcz.dep OK
Downloading: enchant2.tcz
Connecting to repo.tinycorelinux.net (128.127.66.77:80)
wget: server returned error: HTTP/1.1 404 Not Found
md5sum: enchant2.tcz.md5.txt: No such file or directory
Error on enchant2.tcz
we have enchant.tcz if interested
-
Hi aus9
Change enchant2.tcz to enchant.tcz in your copy of abiword.tcz.dep
and confirm it works.
You might also run ldd on the abiword executable to make sure it
finds the enchant libraries.
If it works, I don't know if Juanito wants to change the .dep file or rename
enchant.tcz to enchant2.tcz to match the other Tinycore repos.
-
Hi Rich
I would have to use a web browser to download the files manually as tce-load -w and Apps GUi fails as per last post.
I will wait if you do not mind.
-
I agree my brain hurts with this one....something funny going on after manual loading of one TCE that causes some TCEs to show root:root against /usr/local/tce.installed
I think this then causes an issue when you manually something else afterwards but I can not spot which is the one that is playing up.
First here is result of my current onboot.lst boot up but as its so long I have removed all but symlinks....and aterm is not showing full symlink but they appear ok
ls -al /usr/local/tce.installed/
total 0
drwxrwxr-x 2 root staff 2120 Jan 1 1970 ./
drwxr-xr-x 9 root root 220 Jan 1 1970 ../
lrwxrwxrwx 1 root staff 47 Jan 1 1970 Xlibs -> /tmp/tcloop/Xl
lrwxrwxrwx 1 root staff 49 Jan 1 1970 Xprogs -> /tmp/tcloop/X
lrwxrwxrwx 1 root staff 47 Jan 1 1970 aterm -> /tmp/tcloop/at
lrwxrwxrwx 1 root staff 67 Jan 1 1970 ca-certificates -> /tmp
lrwxrwxrwx 1 root staff 45 Jan 1 1970 dbus -> /tmp/tcloop/dbu
lrwxrwxrwx 1 root staff 57 Jan 1 1970 /usr/local/tce.installed/fontconfig -> /tmp/tcloop
lrwxrwxrwx 1 root staff 57 Jan 1 1970 gdk-pixbuf -> /tmp/tclo
lrwxrwxrwx 1 root staff 45 Jan 1 1970 gtk3 -> /tmp/tcloop/gtk
lrwxrwxrwx 1 root staff 73 Jan 1 1970 hicolor-icon-theme -> /
lrwxrwxrwx 1 root staff 43 Jan 1 1970 jwm -> /tmp/tcloop/jwm/
lrwxrwxrwx 1 root staff 49 Jan 1 1970 libX11 -> /tmp/tcloop/l
lrwxrwxrwx 1 root staff 53 Jan 1 1970 libinput -> /tmp/tcloop
lrwxrwxrwx 1 root staff 51 Jan 1 1970 librsvg -> /tmp/tcloop/
lrwxrwxrwx 1 root staff 69 Jan 1 1970 shared-mime-info -> /tm
lrwxrwxrwx 1 root staff 45 Jan 1 1970 wbar -> /tmp/tcloop/wba
lrwxrwxrwx 1 root staff 59 Jan 1 1970 xorg-server -> /tmp/tcl
right so load inxi
tc@box:~$ tce-load -i inxi
perl5.tcz: OK
libpci.tcz: OK
pciutils.tcz: OK
readline.tcz: OK
bash.tcz: OK
libattr.tcz: OK
libacl.tcz: OK
gmp.tcz: OK
coreutils.tcz: OK
inxi.tcz: OK
See root:root and I have snipped out no scripts
ls -al /usr/local/tce.installed/
total 0
drwxrwsr-x 2 root staff 2320 May 8 19:35 ./
drwxr-sr-x 9 root root 220 May 8 19:35 ../
lrwxrwxrwx 1 root staff 47 Jan 1 1970 Xlibs -> /tmp/tcloop/Xl
lrwxrwxrwx 1 root staff 49 Jan 1 1970 Xprogs -> /tmp/tcloop/X
lrwxrwxrwx 1 root staff 73 Jan 1 1970 adwaita-icon-theme -> /
lrwxrwxrwx 1 root staff 47 Jan 1 1970 aterm -> /tmp/tcloop/at
lrwxrwxrwx 1 root root 45 May 21 12:19 bash -> /tmp/tcloop/bas
lrwxrwxrwx 1 root staff 67 Jan 1 1970 ca-certificates -> /tmp
lrwxrwxrwx 1 root staff 45 Jan 1 1970 dbus -> /tmp/tcloop/dbu
lrwxrwxrwx 1 root staff 57 Jan 1 1970 fontconfig -> /tmp/tclo
lrwxrwxrwx 1 root staff 57 Jan 1 1970 gdk-pixbuf -> /tmp/tclo
lrwxrwxrwx 1 root staff 45 Jan 1 1970 gtk3 -> /tmp/tcloop/gtk
lrwxrwxrwx 1 root staff 73 Jan 1 1970 hicolor-icon-theme -> /
lrwxrwxrwx 1 root root 45 May 21 12:19 inxi -> /tmp/tcloop/inx
lrwxrwxrwx 1 root staff 43 Jan 1 1970 jwm -> /tmp/tcloop/jwm/
lrwxrwxrwx 1 root staff 49 Jan 1 1970 libX11 -> /tmp/tcloop/l
lrwxrwxrwx 1 root staff 53 Jan 1 1970 libinput -> /tmp/tcloop
lrwxrwxrwx 1 root staff 51 Jan 1 1970 librsvg -> /tmp/tcloop/
lrwxrwxrwx 1 root root 53 May 21 12:19 pciutils -> /tmp/tcloop
lrwxrwxrwx 1 root staff 69 Jan 1 1970 shared-mime-info -> /tm
lrwxrwxrwx 1 root staff 45 Jan 1 1970 wbar -> /tmp/tcloop/wba
lrwxrwxrwx 1 root staff 59 Jan 1 1970 xorg-server -> /tmp/tcl
Lets check the root:root for tcloop
ls -al /tmp/tcloop/bash/usr/local
total 0
drwxr-xr-x 6 root root 71 Mar 31 03:00 ./
drwxr-xr-x 3 root root 28 Mar 31 02:57 ../
drwxr-xr-x 2 root root 42 Mar 31 02:57 bin/
drwxr-xr-x 3 root root 27 Mar 31 03:00 lib/
drwxr-xr-x 3 root root 27 Mar 31 03:00 share/
drwxrwxr-x 2 root staff 27 Mar 31 02:57 tce.installed/
tc@box:~$ ls -al /tmp/tcloop/inxi/usr/local/
total 0
drwxr-sr-x 5 root root 60 May 8 19:35 ./
drwxr-sr-x 3 root root 28 May 8 19:35 ../
drwxr-sr-x 2 root root 27 May 8 19:35 bin/
drwxr-sr-x 3 root root 26 May 8 19:35 share/
drwxrwsr-x 2 root staff 27 May 8 19:35 tce.installed/
tc@box:~$ ls -al /tmp/tcloop/pciutils/usr/local/
total 0
drwxr-xr-x 6 tc staff 72 Mar 21 2021 ./
drwxr-xr-x 3 tc staff 28 Mar 21 2021 ../
drwxr-xr-x 2 tc staff 3 Mar 21 2021 lib/
drwxr-xr-x 2 tc staff 63 Mar 21 2021 sbin/
drwxr-xr-x 4 tc staff 45 Mar 21 2021 share/
drwxrwxr-x 2 root staff 31 Mar 21 2021 tce.installed/
its got me stumped
-
aarch64 GUI for cpanel -> backup
I try to delete an item....as requested by Rich another post
it does not work. I checked my perms and they look ok to me
ls -al /opt/.filetool.lst
-rw-rw-r-- 1 tc staff 76 May 21 00:43 /opt/.filetool.lst
I can use nano to edit the file manually but maybe someone can reproduce,
say click on opt and do dry run or exit if opt disappears?
thanks for reading
-
Pciutils did have a permission problem, did you update the extension?
-
If it works, I don't know if Juanito wants to change the .dep file or rename
enchant.tcz to enchant2.tcz to match the other Tinycore repos.
dep file adjusted - I guess there wasn't an enchant[1] in aarch64..
-
Hi aus9
I don't have a pi but I tried it under TC14 x86_64 and could not
reproduce your results.
Try it again but start it from a terminal like this:
filetool
I added /etc/fstab to my backup:
root
etc/firefox
etc/X11
etc/fstab
When I deleted the entry, this message showed up in the terminal:
sed -i '4d' /opt/.filetool.lst
30
Translation:
The sed command says to remove line number 4 from /opt/.filetool.lst
The sed command had a length of 30 characters.
See if an error message is also displayed.
Maybe busybox sed broke.
Maybe you have sed.tcz installed and it's causing the issue.
-
Hi Paul_123
yes pciutils was updated before posting
Hi Juanito
tce-load -w abiword works as expected
-
Hi Paul_123
I attempted to load flwm and flwm_topside
both fail at console command startx
both complain they are missing /etc/sysconfig/Xserver
Maybe they need xorg-server as a dep?
-
Hi Rich
I am sad to say....your command has revealed that somehow I had changed /opt because the error message in running filetool was
something like sed can not create /opt/.filetool.lstiVSG* permission denied
To anyone having to read my dribble....I am going to take a break....do a fresh install and make notes on every change I make, not what I download as TCEs are pristine state so I will copy that tce dir to a usb stick etc. The fact that no-one else has reported strange permissions etc means its just me.
I accept that so must move on
thanks to Rich I now have a better cmdline.txt with the correct syntax for no swap file and persistent home.
Yippee!
-
There’s a choice between Xorg and Xorg-3d in piCore - for me it makes more sense to say Xorg, flwm, aterm and wbar are required as a minimum for a gui. This is what has been done with the TC meta extension in the armv* repos.
-
EEK then the seg fault is something I have done....I will need to ponder my changes. I have made some. As pcmanfm is a GUI I will need a desktop so I will try to see over the weekend what I have done wrong. Or not bother and just do an update of pcmanfm.
Hi aus9,
I have installed your pcmanfm update (v1.3.2) and everything seems to work correctly
-
Hi mortegai and anyone with the same warning
your post 14, The directory '~/Templates' doesn't exist, ignoring it
The easiest way is just make the directory
mkdir $HOME/Templates
I can't remember if I was that smart of if I loaded xdg-user-dirs and ran
xdg-user-dirs-updatewhich is supposed to look at what dirs you want at /usr/local/etc/xdg/user-dirs.default
YMMV
For me, both pcmanfm and lxterminal open with no warnings launched by terminal
-
I have got a video crash which is repeatable. Initially I was loading libv4l and just before the crash, I saw that alsa-modules-KERNEL had loaded then the crash occurred.
Being a dumbo I rebooted and repeated to get same result. Not being used to RPi I could not get another getty.
2) Checking the deps and knowing the last to load as the alsa-modules on a fresh reboot I only loaded
graphics-KERNEL and that crashed.
3) on next reboot I put graphics-KERNEL.tcz at the bottom of my boot list and rebooted
this time I get to console/getty prompt with warning fail in waitforx
Then at console prompt I typed startx.....and desktop launches, is what I am typing on now (jwm)
4) then tce-load -i libv4l2
v4l-dvb-6.1.25-piCore-v8.tcz: OK
libv4l2.tcz: OK
5) Because I got to the console prompt I have kept the dmesg in case anyone finds it useful
https://www.dropbox.com/s/rsrbtcuoulfjp1m/dmesg.txt
thanks for reading
-
I don’t see any error messages in what you posted. Or are you just getting waitforX failing?
-
Hi at section (3) I do not get a desktop but thewaitforX failing
I too could not see any errors in that dmesg but my skills with Rpi are still low.
Maybe someone else with a desktop, happy to use the power switch to turn off and on to reboot could confirm their desktop fails too in loading
graphics-KERNEL?
2) Is it worthwhile in making beta2 images etc to mod /etc/inittab to allow one more getty?
that way if a member loads something that causes a crash ....black screen output....they can get a second getty and run startx?
cheers
-
Images have ssh enabled by default. Use ssh to see what is going on..
I have a feeling this is a race condition in the audio not loading before the vc4 driver inits. During alpha, we needed to add the cma shared memory driver to the initrd to get this to work reliably, but beta 1 had newer firmware that seemed to work fine without.
If this is the case, when you run dmesg from ssh, it will show there is not a framebuffer created.
-
Hi Paul_123
I removed graphics-KERNEL from my boot list and rebooted and used TC64 to load graphics-KERNEL
in TC64 its shows both alsa and graphics KERNEL TCEs loading but monitor on RPi goes black as prev reported.
2) I am not good as scp and have no plans to learn at this stage so I rebooted.....and for reasons I can not explain.....my persistent home dir was gone. PUFF so much for me thinking I knew what I doing ;D
I have just spotted that partition 1 needs a fsck. I get on to that later and my bootcode I got from Rich in another post....I will have to change that because
[ 0.000000] Unknown kernel command line parameters "nortc noembed nozswap opt=mmcblk0p2 home=mmcblk0p2", will be passed to user space
[ 3.115230] FAT-fs (mmcblk0p1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck..
3) anyhow here is link to dmesg created by ssh command on TC64
https://ln5.sync.com/dl/3b16fd620/r5hrb6fs-cfqi4cfz-ntdicqc3-c7rqk6ch
Altho others do not know.....I chose sync for its 5G free storage but it allows only 3 links to be active any point in time. So link will be deleted when I need a spare link......but you have the piCore64-14.x link which I am keeping.
If you do not mind, as I will be rebuilding my home dir here is dir link (as I lost that link and had to use sync to regain it)
https://ln5.sync.com/dl/1e5a26870/mijx95vj-pfbcqp6p-yevtqeqf-jd95g68b
3) cat dmesg-ssh.txt | grep framebuffer
[ 1.377773] simple-framebuffer 3e4e0000.framebuffer: framebuffer at 0x3e4e0000, 0x6e7000 bytes
[ 1.377795] simple-framebuffer 3e4e0000.framebuffer: format=a8r8g8b8, mode=1824x984x32, linelength=7296
[ 1.385816] simple-framebuffer 3e4e0000.framebuffer: fb0: simplefb registered!
end of dmesg may be useful
[ 13.792413] bcmgenet fd580000.ethernet eth0: Link is Up - 1Gbps/Full - flow control off
[ 127.794271] snd_bcm2835: module is from the staging directory, the quality is unknown, you have been warned.
[ 127.811045] bcm2835_audio bcm2835_audio: card created with 8 channels
(gap by me)
[ 128.709351] vc_sm_cma: module is from the staging directory, the quality is unknown, you have been warned.
[ 128.710512] bcm2835_vc_sm_cma_probe: Videocore shared memory driver
[ 128.710533] [vc_sm_connected_init]: start
[ 128.750170] [vc_sm_connected_init]: installed successfully
[ 128.904083] [drm] Initialized v3d 1.0.0 20180419 for fec00000.v3d on minor 0
[ 128.913311] Console: switching to colour dummy device 80x25
[ 128.942189] vc4-drm gpu: bound fe400000.hvs (ops vc4_hvs_ops [vc4])
[ 128.943842] Registered IR keymap rc-cec
[ 128.943970] rc rc0: vc4-hdmi-0 as /devices/platform/soc/fef00700.hdmi/rc/rc0
[ 128.944087] input: vc4-hdmi-0 as /devices/platform/soc/fef00700.hdmi/rc/rc0/input7
-
I am now thinking the lost of home was because I could not recover within RPi from the blank screen so I used a power off power on to reboot.
I am now thinking I will save /etc/inittab once I resolve a better home bootcode
-
...
I have a feeling this is a race condition in the audio not loading before the vc4 driver inits. During alpha, we needed to add the cma shared memory driver to the initrd to get this to work reliably, but beta 1 had newer firmware that seemed to work fine without.
Perhaps the cause may be the softdep between the vc4 and snd-soc-hdmi-codec modules, as indicated in http://forum.tinycorelinux.net/index.php/topic,26080.msg167351.html
-
Hi Rich
Just tested totem.
It needs at least alsa.tcz as dependency.
Info mentions Xorg-3d.tcz
but the dep file only mentions Xorg.tcz
I made Xorg-3d.tcz one of my higher TCEs in onboot.lst above jwm -> fails for waitforX but then startx works
I am aware that Paul_123 is compiling more Apps to depend on openssl but for the moment
launching totem gives
totem
libssl.so.1.1: cannot open shared object file: No such file or directory
Failed to load module: /usr/local/lib/gio/modules/libgioopenssl.so
thanks for reading
slightly offtopic
Both totem and parole will give numerous gstreamer warnings if launched from a terminal
things like
(gst-plugin-scanner:8894): GStreamer-WARNING **: 11:50:36.412: Failed to load plugin '/usr/local/lib/gstreamer-1.0/libgstwavpack.so': libwavpack.so.1: cannot open shared object file: No such file or directory
-
Hi aus9
If you look at the dependency tree:
http://tinycorelinux.net/14.x/armv6/tcz/totem.tcz.tree
http://tinycorelinux.net/14.x/aarch64/tcz/totem.tcz.tree
you will find Xorg-3d.tcz and alsa-modules-KERNEL.tcz get pulled
in by dependencies further down the chain.
... Info mentions Xorg-3d.tcz ...
If you are not currently running Xorg-3d , and you tce-load -i totem , it probably
won't work because you can't start Xorg-3d while Xorg is running. So totem
in that case should be:
Download only
Exit to prompt
tce-load -i totem
startx
Then add totem to onboot.lst.
Note: Xorg-3d will load graphics-KERNEL.tcz. It hasn't come up in a few years, but loading
graphics-KERNEL.tcz while a GUI is running can/will cause the screen to go black.
... fails for waitforX but then startx works ...
I have a laptop that sometimes liked to do that. I added this to ~/.xsession:
sleep 3
/usr/local/bin/Xorg -nolisten tcp &A short delay prior to running Xorg appears to fix (hide?) the issue.
-
Perhaps the cause may be the softdep between the vc4 and snd-soc-hdmi-codec modules, as indicated in http://forum.tinycorelinux.net/index.php/topic,26080.msg167351.html
Yes, putting the cma shared memory driver in the initrd seems to help change the path of the race track. However the best thing to do is, Load alsa modules (and anything non-graphics related) in onboot, then load all of the graphics and Xorg in bootlocal.sh.
-
Yes, putting the cma shared memory driver in the initrd seems to help change the path of the race track.
Just curious, what is the cma shared memory driver?
However the best thing to do is, Load alsa modules (and anything non-graphics related) in onboot, then load all of the graphics and Xorg in bootlocal.sh.
Seems like a tricky solution, instead of fixing the root problem.
-
Just curious, what is the cma shared memory driver?
It is just the memory driver needed for the GPU. Headless doesn't need it, so that is why its not in the initrd. The kernel will automatically load it if present.....consuming memory.
However the best thing to do is, Load alsa modules (and anything non-graphics related) in onboot, then load all of the graphics and Xorg in bootlocal.sh.
Seems like a tricky solution, instead of fixing the root problem.
The only way to solve soft dependencies would be to switch from busybox to kmod module utilities. Which is alot of extra bloat. Managing the load order, buy how the extensions are loaded should work fine.
-
It is just the memory driver needed for the GPU. Headless doesn't need it, so that is why its not in the initrd. The kernel will automatically load it if present.....consuming memory.
Just to learn something and curiosity,
What is the name of that driver?
-
vc-sm-cma.ko
https://github.com/raspberrypi/linux/tree/rpi-6.1.y/drivers/staging/vc04_services/vc-sm-cma
-
aus9,
As a test. Download the module in a initrd format http://www.tinycorelinux.net/14.x/aarch64/test_releases/RPi/vc-sm-cma.gz
place this file on your boot partition mmcblk0p1 by default
edit your config.txt file. In the [Pi4] section of the config. add the filename to the initramfs line, like this
[PI4]
initramfs rootfs-piCore64-14.0.gz,modules-6.1.25-piCore-v8.gz,vc-sm-cma.gz followkernel
kernel kernel6125v8.img
See if this solves the black screen.
-
Hi Rich
It hasn't come up in a few years, but loading
graphics-KERNEL.tcz while a GUI is running can/will cause the screen to go black.
that explains my crash for post 38. Where I was not loading graphics-KERNEL.tcz but trying to load libv4l in jwm desktop,
and discovered it had graphics-KERNEL.tcz has a recursive dep.
OOPS I forgot to check Apps for libvl2 in piCore64-14.x it has
+ alsa-modules-5.10.39-piCore-v8.tcz 1097728, 1.05 MB
bzip2-lib.tcz 32768, 0.03 MB
gamin.tcz 16384, 0.01 MB
gcc_libs.tcz 1527808, 1.46 MB
glib2.tcz 1495040, 1.43 MB
+ graphics-5.10.39-piCore-v8.tcz 774144, 0.74 MB
libelf.tcz 36864, 0.04 MB
libjpeg-turbo.tcz 262144, 0.25 MB
libudev.tcz 36864, 0.04 MB
libv4l2.tcz 221184, 0.21 MB
pcre.tcz 397312, 0.38 MB
+ v4l-dvb-5.10.39-piCore-v8.tcz 3346432, 3.19 MB
Total size (bytes) 9244672, 8.82 MB
+ Indicates need to download 5218304, 4.98 MB
http://tinycorelinux.net/14.x/aarch64/tcz/libv4l2.tcz.dep says its v4l-dvb-KERNEL.tcz
uname -r
6.1.25-piCore-v8
I am sure this will settle when we are closer to final release
-
Hi Paul_123 before I reboot, I am just posting my kindof normal boot list and reboot with that first before adding in Xorg-3d.tcz
openssh.tcz
libXcursor.tcz
mc.tcz
nano.tcz
jwm.tcz
lxterminal.tcz
leafpad.tcz
firefox-perf.tcz
numlockx.tcz
pcmanfm.tcz
dosfstools.tcz
wbar.tcz
I had removed openssh but leaving it in for now.
-
Hi Paul_123
added Xorg-3d.tcz just above jwm and reboot -> failed in waitforX
I have deleted direct sync link to dmesg-ssh.txt prev posted but dir link still good .
Today dmesg-cma.txt link will be good for a while.
https://ln5.sync.com/dl/35e9843e0/qwwmxdxb-fj6be3ti-hi479r6g-ijpb4u5q
my config.txt reads
[PI4]
initramfs rootfs-piCore64-14.0.gz,modules-6.1.25-piCore-v8.gz,vc-sm-cma.gz followkernel
kernel kernel6125v8.img
arm_64bit=1
in post 42 I posted a snippet that has vc-sm but I can't see anything special?
this dmesg is created at console prompt then I ran startx and startx works
-
I also tried swapping the order to vc-sm-cma.gz,modules-6.1.25-piCore-v8.gz still fail waitforX
-
For piCore32 the libffi6 extension is needed for backward compatibility with: glib2, wayland, ...
-
For piCore32 the libffi6 extension is needed for backward compatibility with: glib2, wayland, ...
lots of old extensions in 32 bit.
-
Hi Paul_123
added Xorg-3d.tcz just above jwm and reboot -> failed in waitforX
this dmesg is created at console prompt then I ran startx and startx works
You need to start from scratch, Start with only openssh in your onboot.lst and make sure openssh is started in bootlocal.sh. Then add one extension to onboot at a time, and reboot, until you cause it to break. If you get a black screen, you will need to ssh in and check things out. Besides dmesg, you also need to look a the xorg log file.
It also might be worth testing between jwm and flwm.
-
My net is down in country town. Using phone with fat fingers.However I am already happy with my current setuo subject to looking at Rich reply
To my nozswap post later.
Startx works. I do not believe we gain anything from changing desktop.
I may be the only person who loads something with a recursive dep of graphics-KERNEL and by me having a backup of etc/inttab to have 2 gettys is already good enough.
I need to move on to other things please. I accept I am not the best tester.
I accept that I lack wifi and openssh and scp skills. I am happy with beta1 in its current form
-
Hi aus9
... still fail waitforX
If you look at my reply #46, you'll see I dealt with that issue on my laptop by
adding a sleep command to the .xsession file in my home directory.
-
Hi Rich and hi Paul_123
sorry for waffle but I have only just setup wifi as ethernet fails as exchange in my town has broken equipement. Luckily I had a wifi dongle for my TC64 and was able to download some missing TCEs, via sim data hotspot. I may do a separate post on that. I ran sudo wifi.sh and full reboot without loading wifi and I have the net.
Very happy with that.
Paul I have removed that vc-sm-cma.gz initrd from my config.txt and done a full reboot.
Hi Rich
3 seconds too quick for RPi model 4 and I know your laptop will be much better spec than that. But 6 sec seems ok
for my .xsessionsleep 6
Xorg -nolisten tcp &
SNIP
I was tempted to say that next image for /etc/skel/.xsession could have that line. But I am NOT tempted because I have already agreed that most users will not be loading bloated software that needs Xorg-3.d.
Due to power switch issue already discussed....I have also removed home=mmcblk0p1 and opt=mmcblk0p1 from my cmd.txt
and am back to being almost normal TC user ;D
Hi Paul_123
I think we can now state this video crash is resolved by Rich sleep suggestion. But I still ask that you please consider next image of /etc/inittab to remove comment from getty2 please.
For any user of a RPi wanting Xorg-3d...I claim the solution is as follows:
1) Add line at top of your ~/.xsession so top 2 lines now reads as
sleep 6
Xorg -nolisten tcp &
2) edit your boot list maybe onboot.lst so somewhere above your preferred desktop you have Xorg-3d.tcz
eg my onboot top section reads
openssh.tcz
libXcursor.tcz
mc.tcz
nano.tcz
Xorg-3d.tcz
jwm.tcz........................change this to your preferred desktop
other TCEs not shown
Good luck
-
Hi aus9
I took a look at /var/log/Xorg.0.log after a failed in waitforX and
saw it reported /dev/dri/card0 did not exist. So I replaced the
sleep command with this:
Timeout=40
while [ $Timeout -gt 0 ]
do
[ -e /dev/dri/card0 ] && break
Timeout=$(($Timeout - 1))
sleep 0.25
doneThis will wait for /dev/dri/card0 to show up for a maximum of
10 seconds before continuing.
[EDIT]: Changed -x to -e in test for /dev/dri/card0. Rich
-
Hi Rich I will test that shortly.
I want to post that I am idiot. I pulled the cable from the RPi and tried to get the net and failed. My last post over reboot.....I must have got ethernet back.
Yes my modem lights are now good. Sorry if above misleads anyone. If you have time can you correct this
and full reboot without loading wifi and I have the net.
to
and full reboot with loading firmware-rpi-wifi and wifi and I have the net.
please....if not thats ok this post should suffice.
-
Hi Rich full reboot.
In the past I have never timed how long my screen is blank (with Xorg-3d in boot list )after saying it is loading extensions etc.
This time I did a rough count.
from when screen goes blank about 16 seconds later, for a second what appears, not sure what you call it,
the ascii text image of tinycore with tux
from when screen goes blank to full desktop appears takes 22 seconds on a RPi model B 4G ram.
-
Hi Rich
retested time to desktop with sleep and its faster 17 seconds, for those reading this I have got Xorg-3d in boot list.
Curiously around 16 seconds there is like a small glitch where I think the ascci tux wanted to appear but a second later my desktop appears.
So I think if the desktop took maybe 18 seconds or more, I would have seen the tux for a second.
I think I will keep it simple and stay with sleep....my impatience is shining thru ;D
-
Hi Rich
this time I remove Xorg-3d from boot list. the accii tux and tinycore stays on the screen for lot longer.
from screen blank to desktop now 10 seconds. Not a huge saving considering we spend time playing with software etc.
I think I will stay with sleep 6
-
I think it's not so simple just remove the # from inittab file in the initrd.
Just because there some differences between the rpi's.
When talking about the serial (uart) interface, the device have lots of different name.
ttyAMA0::respawn:/sbin/getty -L 115200 /dev/ttyAMA0 vt100
And some one have ttyS0 or also Serial0, and i think the device tree change the name, and the bluetooth device shares some.
There are lots of good documenting stuff on the rpi forum.
Here are some links to that.
https://www.raspberrypi.com/documentation/computers/configuration.html#configuring-uarts
https://www.raspberrypi.com/documentation/computers/configuration.html#part5.6
And like neerds like me likes to get all juice logs and things on the serial, you have to do like this to get some more text from the bios(bootcode.bin)
https://www.raspberrypi.com/documentation/computers/raspberry-pi.html#bootcode-bin-uart-enable
-
Paul I have removed that vc-sm-cma.gz initrd from my config.txt and done a full reboot.
Hi Rich
3 seconds too quick for RPi model 4 and I know your laptop will be much better spec than that. But 6 sec seems ok
Does setting something like "cma=384M" in cmdline.txt and/or "dtoverlay=vc4-kms-v3d-pi4,cma-384" cause vc-sm-cma to be loaded first without needing sleep?
-
Hi Juanito
https://github.com/raspberrypi/firmware/blob/master/boot/overlays/README
Name: vc4-fkms-v3d-pi4
Info: Enable Eric Anholt's DRM VC4 V3D driver on top of the dispmanx
display stack.
Load: dtoverlay=vc4-fkms-v3d-pi4,<param>
Params: cma-512 CMA is 512MB (needs 1GB)
cma-448 CMA is 448MB (needs 1GB)
cma-384 CMA is 384MB (needs 1GB)
cma-320 CMA is 320MB (needs 1GB)
cma-256 CMA is 256MB (needs 1GB)
cma-192 CMA is 192MB (needs 1GB)
cma-128 CMA is 128MB
cma-96 CMA is 96MB
cma-64 CMA is 64MB
cma-size CMA size in bytes, 4MB aligned
cma-default Use upstream's default value
I finally wore my goggles and spotted the hypen vs your = ;D
2) I decided to time from my inline power switch to various outcomes as below, with Xorg-3d.tcz in onboot as prev discussed
xsession no sleep 37 seconds to failed in waitfor X
" " config.txt dtoverlay=vc4-kms-v3d-pi4,cma-384 36 seconds to failed in waitforX
xsession sleep 6 config.txt dtoverlay=vc4-kms-v3d-pi4 Desktop appears 37 seconds
I did not test cmdline.txt for a cma config on the grounds I am happy with sleep.
Rich is still the winner :D
3) Because I am weirdo
###################
xsession no sleep and commented out the dtoverlay from pi4 -> desktop appears in 29 seconds.
Its late here so I will digest this tomorrow.
thanks for reading
-
I finally wore my goggles and spotted the hypen vs your =
Indeed - sorry, corrected above
-
Hi aus9
There was a small typo in my reply #63 that has been corrected.
-
I think that by doing modprobe of the snd-soc-hdmi-codec module in tc-config, just before the instructions for udev, no sleep or special order in the loading of extensions (graphics, alsa-modules) would be necessary, since would be loaded, as the vc4 module needs.
To generalize these situations, a bootcode similar to blacklist could be enabled where modules were put to load previously to udev (for whatever reason).
-
I think that by doing modprobe of the snd-soc-hdmi-codec module in tc-config, just before the instructions for udev, no sleep or special order in the loading of extensions (graphics, alsa-modules) would be necessary, since would be loaded, as the vc4 module needs.
The audio kernel modules will not be even mounted by then. Onboot.lst is processed near the end of tc-config. The better place to do what you want is /opt/bootsync.sh. But I'm guessing the driver is long loaded by the time it gets to bootsync.
But I still ask that you please consider next image of /etc/inittab to remove comment from getty2 please.
Aus9,
If you want to play with inittab. Edit your inittab file at /etc/inittab
mount /mnt/mmcblk0p1
cd /
sudo find /etc | grep inittab | sudo cpio -o -H newc | gzip -9 > /mnt/mmcblk0p1/inittab.gz
Then add inittab.gz to your initramfs line in your config. (Like I had you load the cma driver before).
-
piCore14-beta2 posted for 32bit.
- Only change is fixing a libstdc++ issue.
http://www.tinycorelinux.net/14.x/armv7/test_releases/RPi/piCore-14.0-beta2.zip
-
Hi Paul_123
Based on your reply and Patrikg, I withdrawn my ignorant getty request. I already have setup my inittab thanks.
To anyone,
This morning with that crazy setting...I used inxi -Gxxx to prove I was on fbdev with unloaded modesetting.
Xorg log showed no acceleration.
Hi Rich
I will test that edit later today thanks.
-
Hi Rich
Using the editted Timeout xsession settings from power on to failed in waitforX took 36 seconds.
to Paul_123, Rich and interested parties I have editted my sync account to share a dir that contains my past and current dmesg and xorg files
New link takes you to a dir that contains only Xorg-3d files
https://ln5.sync.com/dl/ab5f2e950/im3994fv-bvcpbmst-dxabzv4b-62dehkes
Rich I have in that dir dmesg-timeout.txt and xorg-timeout.txt
this can be compared to a working dmesg-sleep.txt and xorg-sleep.txt
As I have not posted what inxi thinks of a working Xorg-3d in the past here is what it looks like. I like it for summary of what it tried to load and what has loaded.....I also loaded glxinfo to get a bit more detail for those interested.
Output is for sleep config in .xsession.....Raspberry Pi 4 Model B Rev 1.1 4G RAM
tc@box:~$ inxi -Gxxx
Graphics:
Device-1: bcm2711-hdmi0 driver: vc4_hdmi v: N/A bus-ID: N/A
chip-ID: brcm:fef00700 class-ID: hdmi
Device-2: bcm2711-hdmi1 driver: vc4_hdmi v: N/A bus-ID: N/A
chip-ID: brcm:fef05700 class-ID: hdmi
Device-3: bcm2711-vc5 driver: vc4_drm v: N/A bus-ID: N/A chip-ID: brcm:gpu
class-ID: gpu
Display: server: X.org driver: X: loaded: modesetting unloaded: fbdev
dri: vc4
gpu: vc4-drm,vc4_crtc,vc4_dpi,vc4_dsi,vc4_firmware_kms,vc4_hdmi,vc4_hvs,vc4_txp,vc4_v3d,vc4_vec
display-ID: :0.0 note: <missing: xdpyinfo/xrandr>
Monitor-1: HDMI-A-1 model: Lenovo LEN L22e-20 serial: 827215436
res: 1920x1080 dpi: 102 size: 476x268mm (18.74x10.55") diag: 546mm (21.5")
modes: max: 1920x1080 min: 720x400
API: OpenGL v: 2.1 Mesa 22.3.7 renderer: V3D 4.2 direct-render: Yes
past post from Juanito says I should show (and that is real result)
glxinfo | grep Accel
Accelerated: yes
:) :)
-
piCore64-14x issues with TCE lzip
tce-load -i lzip
lzip -h
sh: lzip: not found
tc@box:/tmp$ echo $PATH
/home/tc/.local/bin:/usr/local/sbin:/usr/local/bin:/apps/bin:/usr/sbin:/usr/bin:/sbin:/bin:/etc/sysconfig/tcedir/ondemand
/usr/local/bin/lzip -d icewm-3.3.1.tar.lz
sh: /usr/local/bin/lzip: not found[
file /tmp/tcloop/lzip/usr/local/bin/lzip
/tmp/tcloop/lzip/usr/local/bin/lzip: ELF 32-bit LSB executable, ARM, EABI5 version 1 (GNU/Linux), dynamically linked, interpreter /lib/ld-linux-armhf.so.3, for GNU/Linux 4.9.22, stripped
Maybe I need a ELF 64 bit?
EEK its mine....I will have to rebuild it sorry. Aah the build script suggests it was built for -march=armv6zk
-
The audio kernel modules will not be even mounted by then. Onboot.lst is processed near the end of tc-config. The better place to do what you want is /opt/bootsync.sh. But I'm guessing the driver is long loaded by the time it gets to bootsync.
Yes, you are right, with the configuration/division of modules used by "original" piCore. I am using a configuration in which all the modules (those of RaspiOS upstream, as in your case) are included in the initrd.
Sorry for the possible confusion
-
Hi aus9
Using the editted Timeout xsession settings from power on to failed in waitforX took 36 seconds.
----- Snip -----
Rich I have in that dir dmesg-timeout.txt and xorg-timeout.txt
this can be compared to a working dmesg-sleep.txt and xorg-sleep.txt ...
In the timeout version /dev/dri/card0 is found but not used. It then tries to use /dev/fb0
which does not exist and then gives up.
In the sleep version /dev/dri/card0 is found, but so is /dev/dri/card1 which is what Xorg
winds up using.
Try changing it to this so it tests for both of them:
Timeout=40
while [ $Timeout -gt 0 ]
do
[ -e /dev/dri/card0 ] && [ -e /dev/dri/card1 ] && break
Timeout=$(($Timeout - 1))
sleep 0.25
done
-
Hi Rich
that worked well....acceleration proved using prev commands. Too late for me to upload files but it works. I will test again tomorrow thanks
-
Hi Rich
time from power on to desktop with new .xsession as per last post is about the same. The dmesg timestamps give your TIMEOUT a one second lead to be the new winner
Whats a second between friends :D
If interested I have uploaded dmesg-rich.txt and xorg-rich.txt to same dir
https://ln5.sync.com/dl/ab5f2e950/im3994fv-bvcpbmst-dxabzv4b-62dehkes
Obviously what TCEs I load, what I have in backup will impact on other members boot up time if I choose to load Xorg-3d instead of normal Xorg.
-
Hi aus9
... The dmesg timestamps give your TIMEOUT a one second lead to be the new winner ...
The sleep method will always wait 6 seconds regardless of if or when the /dev/dri/ entries show up.
The Timeout method will only wait as long the /dev/dri/ entries are not present up to a max of 10 seconds.
If you want to see the actual time spent in the Timeout code, you can
add a command to save the elapsed time to a file:
Timeout=40
while [ $Timeout -gt 0 ]
do
[ -e /dev/dri/card0 ] && [ -e /dev/dri/card1 ] && break
Timeout=$(($Timeout - 1))
sleep 0.25
done
calc '(40-'$Timeout')*.25' > Timeout.txtThe Timeout.txt file in your home directory will contain how many seconds it
had to wait for the /dev/dri/ entries to show up.
Remember, the time it takes for the /dev/dri/ entries to show up can vary with
changes made to the contents and order of extensions in your onboot.lst file.
-
thanks Rich
to anyone piCore64-14.x
I think we may have a missing dep for firefox-perf
I have loaded a new desktop not yet ready for submission and I notice
tc@box:~$ firefox-perf
Crash Annotation GraphicsCriticalError: |[0][GFX1-]: glxtest: libpci missing (t=3.57427) [GFX1-]: glxtest: libpci missing
JavaScript error: resource://gre/modules/PromiseWorker.jsm, line 106: Error: Could not get children of file(/home/tc/.cache/mozilla/firefox/0npiijyc.default-default/thumbnails) because it does not exist
Crash Annotation GraphicsCriticalError: |[0][GFX1-]: glxtest: libpci missing (t=3.57427) |[1][GFX1-]: Unrecognized feature VIDEO_OVERLAY (t=11.075) [GFX1-]: Unrecognized feature VIDEO_OVERLAY
^CExiting due to channel error.
tce-load -i libpci
firefox-perf # runs as expected
I may not have noticed before as I had inxi in my boot list which pulls pciutils dep of libpci?
-
fixed the firefox-perf dep for you
-
Hi Paul_123
thanks
-
piCore64-14.x firefox-perf revisitted with less TCEs loaded finds it missing a dep
pcre2
My mistake, I was loading lxterminal that has that as a recursive dep....aterm found it.
also after seeing terminal snippet
firefox-perf
(firefox:6799): Gtk-WARNING **: 12:47:44.943: Could not load a pixbuf from icon theme.
This may indicate that pixbuf loaders or the mime database could not be found.I ran
sudo gdk-pixbuf-query-loaders --update-cache or re-run the install script for gdk-pixbuf
-
Hi Paul_123
mc is missing dep of pcre2 please
next command done by openssh as my new desktop is already loading pcre2. So I made my boot list without a desktop and then
tce-load -i mc
ncurses.tcz: OK
libssh2.tcz: OK
gcc_libs.tcz: OK
bzip2-lib.tcz: OK
pcre.tcz: OK
libelf.tcz: OK
gamin.tcz: OK
glib2.tcz: OK
mc.tcz: OK
tc@box:~$ mc
mc: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
-
Hi Paul_123
EDIT
mc can run in the console....in a later post....I now thinking it might be smarter? to just add pcre2 to Xorg dep file?
I am on a roll with openssh...one of my babies leafpad missing pcre2 found by openssh too
or optionally....make it a dep of gdk-pixbuf?
tce-load -i leafpad
libXinerama.tcz: OK
libXcomposite.tcz: OK
libXrandr.tcz: OK
liblzma.tcz: OK
libxml2.tcz: OK
gcc_libs.tcz: OK
bzip2-lib.tcz: OK
pcre.tcz: OK
libelf.tcz: OK
gamin.tcz: OK
glib2.tcz: OK
shared-mime-info.tcz: OK
liblz4.tcz: OK
libzstd.tcz: OK
zstd.tcz: OK
libjpeg-turbo.tcz: OK
libtiff.tcz: OK
libpng.tcz: OK
gdk-pixbuf-query-loaders: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
gdk-pixbuf.tcz: OK
atk.tcz: OK
fribidi.tcz: OK
graphite2.tcz: OK
harfbuzz.tcz: OK
freetype.tcz: OK
expat2.tcz: OK
fontconfig.tcz: OK
fc-cache: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
libXft.tcz: OK
libudev.tcz: OK
libpciaccess.tcz: OK
libdrm.tcz: OK
ncurses.tcz: OK
llvm15-lib.tcz: OK
libxshmfence.tcz: OK
libXxf86vm.tcz: OK
libXdamage.tcz: OK
mesa.tcz: OK
libGLES.tcz: OK
wayland.tcz: OK
libEGL.tcz: OK
pixman.tcz: OK
cairo.tcz: OK
pango.tcz: OK
gtk-query-immodules-2.0: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
gtk2.tcz: OK
leafpad.tcz: OK
-
Hi Paul_123
Another pcre2 missing dep from pcmanfm but below code, it hangs, that is does return to prompt.
On my desktop pcmanfm works for me and another member on jwm so its not all that bad?
tce-load -i pcmanfm
hicolor-icon-theme.tcz: OK
gdk-pixbuf-query-loaders: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
librsvg.tcz: OK
wayland-protocols.tcz: OK
xkeyboard-config.tcz: OK
libxkbcommon.tcz: OK
libXi.tcz: OK
dbus.tcz: OK
at-spi2-core.tcz: OK
at-spi2-atk.tcz: OK
libepoxy.tcz: OK
gtk-query-immodules-3.0: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
glib-compile-schemas: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
gtk3.tcz: OK
gtk-update-icon-cache: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
adwaita-icon-theme.tcz: OK
desktop-file-utils.tcz: OK
libexif.tcz: OK
menu-cache.tcz: OK
libfm.tcz: OK
pcmanfm.tcz: OK
tc@box:~$ update-desktop-database: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
update-mime-database: error while loading shared libraries: libpcre2-8.so.0: cannot open shared object file: No such file or directory
those last 2 lines are because I have a exe called update-desktops and its called from my install script
-
Hi Paul_123
I am going to stop reporting these missing pcre2 now.
How about we just make it a dep of Xorg as most members will run a desktop with Xorg?
-
..but the default wm, flwm, doesn’t need pcre2.
-
Hi Juanito
EDIT
for me it makes more sense to say Xorg, flwm, aterm and wbar are required as a minimum for a gui.
so umm Xorg can have it as a dep or not?
-
Hi Paul_123
Can you check glib2-python dep. Apps TAB=size names pcre3.
This prevents download of libglade-dev
thanks for reading
-
so umm Xorg can have it as a dep or not?
In the 32bit repos pcre2 is a recursive dep of Xorg
In the 64bit repo Xorg, flwm, aterm and wbar do not depend on pcre2, so I would suggest not.
-
Recent packages have transitioned from pcre to a minimum of pcre2. Lets get the deps in the right spot. The issue is that I messed up changing glib2.tcz.deps from pcre to pcre2 when I updated it. I'm not in a spot where I can get on the server to fix until later today.
-
Hi Paul_123
mc is working well now thanks.
I will need to rebuild/update gtk2 as leafpad and pcmanfm as they now break during loading
[code]SNIP
gtk-query-immodules-2.0: error while loading shared libraries: libpcre.so.1: cannot open shared object file: No such file or directory
gtk2.tcz: OK
leafpad.tcz: OK
SNIP
gtk-query-immodules-2.0: error while loading shared libraries: libpcre.so.1: cannot open shared object file: No such file or directory
gtk2.tcz: OK
libfm.tcz: OK
pcmanfm.tcz: OK
cheers
-
Hi Paul_123
I almost succeeded skipping gtk2 and compiled pcmanfm .
but grep has a dep of pcre and grep is a dep of submitqc and compiletc
I will try changing that dep but can not promise that I will be able to build pcmanfm tonight
-
some parts of gtk2 rebuild are dep/ldd with libpcre2 but at least one does not want to party
root@box:/tmp/gtk2/usr/local/bin# ldd gtk-demo
linux-vdso.so.1 (0x0000007f9a70b000)
libgdk-x11-2.0.so.0 => /usr/local/lib/libgdk-x11-2.0.so.0 (0x0000007f9a5e0000)
SNIP
libpcre.so.1 => not found
I will have to check again...but my brain can't quite understand it when /usr/local/installed has no pcre or dev loaded
I can't spot anything in Makefile but fresh eyes maybe tomorrow
-
at the risk of waffle...tomorrow I think I will build pcmanfm against gtk3 to see if that makes a diff
-
It’s possible to have both a direct dep and a recursive dep on a library.
If you recompiled gtk2 against pcre2, then the problem is probably with the recursive dep.
-
Hi aus9
The ldd command checks for recursive dependencies:
tc@E310:~$ ldd /usr/local/bin/editor
linux-gate.so.1 (0xb7f35000)
libc.so.6 => /lib/libc.so.6 (0xb7df5000)
libfltk.so.1.3 => /usr/local/lib/libfltk.so.1.3 (0xb7d35000)
libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0xb7c08000)
/lib/ld-linux.so.2 (0xb7f36000)
libXcursor.so.1 => /usr/local/lib/libXcursor.so.1 (0xb7bfe000)
libXext.so.6 => /usr/local/lib/libXext.so.6 (0xb7bef000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7bd6000)
libdl.so.2 => /lib/libdl.so.2 (0xb7bd1000)
libX11.so.6 => /usr/local/lib/libX11.so.6 (0xb7ad4000)
libm.so.6 => /lib/libm.so.6 (0xb7a18000)
libgcc_s.so.1 => /usr/lib/libgcc_s.so.1 (0xb7a00000)
libXrender.so.1 => /usr/local/lib/libXrender.so.1 (0xb79f7000)
libXfixes.so.3 => /usr/local/lib/libXfixes.so.3 (0xb79ef000)
libxcb.so.1 => /usr/local/lib/libxcb.so.1 (0xb79cf000)
libXau.so.6 => /usr/local/lib/libXau.so.6 (0xb79ca000)
libXdmcp.so.6 => /usr/local/lib/libXdmcp.so.6 (0xb79c3000)
tc@E310:~$
If you want to find only direct dependencies, install binutils.tcz
and try this command:
tc@E310:~$ readelf -d /usr/local/bin/editor | grep "Shared library" | cut -d ':' -f2 | busybox tr -d [-]
libc.so.6
libfltk.so.1.3
libstdc++.so.6
tc@E310:~$
-
Hi Rich
I have fixed it just now with some new updates I plan to submit shortly. Early testing of pcmanfm starts today and may complete today.
Built against new gtk2
-
rebuilt leafpad. To test it for a printer I will need to update against openssl for ghostscript, qpdf, gutenprint and the cups* packages. That will take me some time. EDIT not all need recompiles but some have updates
-
Hi Paul_123
In attempting to rebuild vte-2.91 to avoid libffi7 I discovered this make error, relates to gnutls TCE
/usr/local/bin/ld: warning: libffi.so.7, needed by /usr/local/lib/libgnutls.soI prefer readelf and it shows
readelf -d /usr/local/lib/libgnutls.so | grep libffi
0x0000000000000001 (NEEDED) Shared library: [libffi.so.7]
the dep file currently reads
libidn2.tcz
unbound.tcz
nettle.tcz
p11-kit.tcz
I know this will already be on your todo list but I just announcing here as the reason why I can not (yet) fix lxterminal
Thanks for reading
-
You can load libffi7 and compile vte without worrying it will link against it - the symlink libffi.so --> libffi.so.7 has been removed.
-
Hi Juanito
have made modest progress but loading of lxterminal now shows a gremlin against glib-networking. That TCE claims dep of openssl but
readelf -d /usr/local/lib/gio/modules/libgioopenssl.so | grep 'NEEDED' # delete TCBs
0x0000000000000001 (NEEDED) Shared library: [libgio-2.0.so.0] glib2
0x0000000000000001 (NEEDED) Shared library: [libgobject-2.0.so.0] glib2
0x0000000000000001 (NEEDED) Shared library: [libglib-2.0.so.0] glib2
0x0000000000000001 (NEEDED) Shared library: [libssl.so.1.1] openssl-1.1.1
0x0000000000000001 (NEEDED) Shared library: [libcrypto.so.1.1] "
-
correction ...goggles on libcrypto.so.1.1 not in openssl-1.1.1 its TCB
-
updated the following
p11-kit (This was your recursive dep to libffi7)
nettle
gnutls
glib-networking
gnutls getting updated hopefully is not too big of an issue, as the abi of the library has not changed. At lease samba4 did not complain about the newer version of the library.
-
Hi Paul_123
Thanks to you lxterminal now launches with only atk-bridge warnings. To see what it would take to eliminate these after searching the net I added
/usr/local/lib/at-spi2-core/at-spi-bus-launcher &
/usr/local/lib/at-spi2-core/at-spi2-registryd &
to /opt/bootlocal.sh
added at-spi2-core.tcz to my onboot.lst but it borks as it claims libXtst missing
2) I am not going to modify my dep list for lxterminal as those atk-bridge warnings do not stop lxterminal from launching and being useful as a terminal. Just FYI
also both readelf and ldd do not give any hits for libXtst as a dep of at-spi2-core
-
correction, in last post I was referring to a rebuilt lxterminal. Just tested existing repo version of lxterminal and it is good as is.Yippee for me and no work for Paul_123
but sorry that re-testing repo edition of gtk2 and repo edition of pcmanfm still shows
pcmanfm
pcmanfm: error while loading shared libraries: libpcre.so.1: cannot open shared object file: No such file or directory
so that is still a submission or 2
-
title: I am trying to trick net-snmp into not looking for pcre.h but pcre2*
what I attempted was
sed 's|pcre|pcre2|g' -i configure
but I get
root@box:/tmp/net-snmp-5.9.4.pre3# cat config.log | grep error | grep pcre
/usr/local/include/pcre2.h:971:2: error: #error PCRE2_CODE_UNIT_WIDTH must be defined before including pcre2.h.
971 | #error PCRE2_CODE_UNIT_WIDTH must be defined before including pcre2.h.
/usr/local/include/pcre2.h:972:2: error: #error Use 8, 16, or 32; or 0 for a multi-width application.
any clues to fix....and there might more once I get past this block.
also I can configure to disable pcre in configure by --without-pcre
I am tempted to do that as so far...without pcre2 effective configure status says
Embedded Perl support: disabled
SNMP Perl modules: disabled
thanks for reading
-
The pkg-config files in pcre-dev are named libpcre*.pc, whereas in pcre2-dev they are named libpcre2*.pc.
If you delete all of the *.la files in /usr/lib and /usr/local/lib, the configure script will probably pick up pcre2.
-
Hi
all my new 14.x builds now delete all *.la files and originally I was loading pcre2-dev without any sed command. error log reported pcre.h not found.
I think I will just continue make and see what happens. net-snmp-dev has a current dep of pcre-dev but net-snmp has no dep of pcre .
I just checked old build for 12.x it too has configure status of Embedded Perl support: disabled & # SNMP Perl modules: disabled
As no-one has complained recently.....lets move on.
2) the 12.x net-snmp claimed "it" was a make dep of sane* TCEs with sane info file in 14.x claiming you may need to use net-snmp but its not a dep of sane* in 14.x so I don't think I will need to rebuild sane.
thanks for suggestion tho
-
Outside of working on the repo and Extensions, Anyone have issues with the base beta images that we need to address?
-
piCore-14.0 and piCore64-14.0 released.