Tiny Core Linux
Tiny Core Base => TCB Bugs => Topic started by: Misalf on May 15, 2015, 09:01:08 AM
-
Since 6.0 I can't see the current size / usage of rootfs "/" .
I noticed this isn't working anymore since Conky under TC5.x and below was able to display that the usage increase if I copy lots of data to /tmp . Not so under 6.x.
I didn't gave it much attention, sorry. I thought I'd have to fix my conky config but never did.
Recently I prepared a USB flash drive to boot Tiny Core for use on other computers with the exact same setup, extensions, even the same backup file.
On those computers I tested the USB flash drive on, Conky and df show the usage of rootfs correctly.
It's just my box that's acting strange. Kinda old MSI Wind U100 netbook.
This is also true If I boot with base norestore boot codes.
By the way, the pc speaker doesn't work either.
Can anyone tell me where I should start digging?
Might I need to recompile kernel and/or modules with some special netbook settings (if something like that exists).
busybox df -a:
Filesystem 1K-blocks Used Available Use% Mounted on
rootfs 0 0 0 0% /
proc 0 0 0 0% /proc
sysfs 0 0 0 0% /sys
devpts 0 0 0 0% /dev/pts
tmpfs 1031380 0 1031380 0% /dev/shm
fusectl 0 0 0 0% /sys/fs/fuse/connections
/dev/sda3 17869132 16300376 637992 96% /mnt/sda3
...
df -a with coreutils.tcz installed:
Filesystem Size Used Avail Use% Mounted on
rootfs 0 0 0 - /
proc 0 0 0 - /proc
sysfs 0 0 0 - /sys
devpts 0 0 0 - /dev/pts
tmpfs 1008M 0 1008M 0% /dev/shm
fusectl 0 0 0 - /sys/fs/fuse/connections
/dev/sda3 18G 16G 624M 97% /mnt/sda3
...
-
Show us the output of 'free' and 'showbootcodes'.
-
free
total used free shared buffers
Mem: 2062760 758076 1304684 0 101004
-/+ buffers: 657072 1405688
Swap: 472900 0 472900
showbootcodes
root=/dev/sda3 initrd=/tce-6.x/boot/core62.gz,/tce-6.x/boot/more-test-62.gz,/tce-6.x/boot/fonts-test.gz,/tce-6.x/boot/xorg.gz,/tce-6.x/boot/intel.gz vga=788 video=vesafb:ywrap,mtrr:3 logo.nologo quiet loglevel=4 mydata=myxorg kmap=qwertz/de-latin1 noutc tz=CET-1CEST,M3.5.0,M10.5.0/3 tce=UUID="cb6f8b98-91fd-4c96-b115-3d2bb9cb3e57"/tce-6.x laptop nodhcp desktop=openbox blacklist=pcmcia,ssb,b43,bcma,rt2860sta,bluetooth,btusb,pcspkr BOOT_IMAGE=/tce-6.x/boot/vmlinuz
-
You are not running a standard core system. You are loading multiple initrds, and now for some reason all your RAM has been allocated to your tmpfs.
Boot a stock core and see if the problem persists.
-
Results from stock TinyCore 6.2 iso:
(http://i.imgur.com/DQONiMr.png)
Only busybox df is expected to work. Try making sure there is only one busybox in the initrds, and it is the current version.
edit: It's not clear if you tested the working USB stick on the netbook, or if the netbook results are from the hd/ssd install?
-
Hi Misalf
By the way, the pc speaker doesn't work either.
blacklist=pcmcia,ssb,b43,bcma,rt2860sta,bluetooth,btusb,pcspkr
You blacklisted pcspkr.
-
You are not running a standard core system. You are loading multiple initrds
True, yet it works on other computers.
Boot a stock core and see if the problem persists.
The problem does persist. I freshly downloaded the 6.2 versions of core.gz and vmlinuz booted from /tce/boot/ without any extensions using this boot stanza via grub4dos
LABEL core62def
MENU LABEL Tiny Core 6.2 Default
KERNEL /tce/boot/vmlinuz
APPEND root=/dev/sda3 initrd=/tce/boot/core.gz noutc tz=CET-1CEST,M3.5.0,M10.5.0/3 tce=UUID="cb6f8b98-91fd-4c96-b115-3d2bb9cb3e57"/tce laptop nodhcp
edit: It's not clear if you tested the working USB stick on the netbook, or if the netbook results are from the hd/ssd install?
The results are from HDD install.
Strangely, booting from USB stick and df shows correct values!
You blacklisted pcspkr.
Yep, because it's not working anyway.
-
The problem does persist. I freshly downloaded the 6.2 versions of core.gz and vmlinuz booted from /tce/boot/ without any extensions using this boot stanza via grub4dos
Correction: I'm using extlinux on this netbook. Not grub4dos.
On the USB stick I'm using grub2.
I copied the USB install back to HDD into a seperate tce-dir to make sure I havn't done any modifications I forgot about that would make things work. Still the same result.
-
So you're thinking it's the bootloader? Use your grub2 stick, in the grub menu open the console, and boot from HD using the HD bootcodes.
-
That would be strange if this could actually be the case, right? I will try another bootloader later.
Currently compiling kernel and modules. I remember someone on the forum reported it would fix the not working pc speaker.
When browsing through make menuconfig I saw I can build the kernel for Atom CPU instead of 486.
Apart from the fact that it wouldn't boot on other systems, would this break compatibility with software build for 486 CPUs or just modules and firmware?
-
Are you sure /proc and /sys are mounted?
-
23:49:35 tc;box:/$> \ls -la
total 4
drwxr-xr-x 17 tc staff 0 May 3 14:38 .
drwxr-xr-x 17 tc staff 0 May 3 14:38 ..
drwxr-xr-x 2 root root 0 May 16 16:21 bin
drwxrwxr-x 16 root staff 0 May 16 20:38 dev
drwxr-xr-x 12 root root 0 May 16 20:38 etc
drwxrwxr-x 3 root staff 0 Feb 24 19:48 home
-rwxr-xr-x 1 root root 496 May 3 18:38 init
drwxr-xr-x 4 root root 0 Feb 9 17:42 lib
lrwxrwxrwx 1 root root 11 May 3 14:38 linuxrc -> bin/busybox
drwxrwxr-x 8 root staff 0 May 16 15:48 mnt
drwxrwsr-x 4 root staff 0 May 16 15:46 opt
dr-xr-xr-x 217 root root 0 May 16 17:45 proc
drwxrwxr-x 3 root staff 0 May 16 16:21 root
drwxrwxr-x 4 root staff 0 May 16 17:45 run
drwxr-xr-x 2 root root 0 May 3 14:38 sbin
dr-xr-xr-x 12 root root 0 May 16 17:45 sys
drwxrwxr-t 6 root staff 0 May 16 23:45 tmp
drwxr-xr-x 9 tc staff 0 Feb 21 01:45 usr
drwxr-xr-x 8 root root 0 May 3 14:38 var
-
Apart from the fact that it wouldn't boot on other systems, would this break compatibility with software build for 486 CPUs or just modules and firmware?
It would only need new modules. Firmware and userspace unaffected.
-
Sounds good. I may try an Atom optimized kernel later then. Well, the rebuilt 486 kernel fixed nothing.
But guess what, booting my Tiny Core HDD install via grub2 on USB stick fixes it.
I guess I should update or change my bootloader.
-
But guess what, booting my Tiny Core HDD install via grub2 on USB stick fixes it.
I guess I should update or change my bootloader.
Glad you're making progress. Just wanted to mention 'df' correctly displays rootfs usage on my TC 6.2, utilizing Grub2 for simple boot. showbootcodes output: BOOT_IMAGE=/tinycore/boot/vmlinuz nozswap showapps tce=sdb3
-
Linux does not know what bootloader was used, nor what it booted off of.
Unless you are using the exact same kernel, initrd, tce directory, home, opt, and boot codes the two bootups are not the same.
I suspect you are loading something that causing your problem.
-
Well, what could be a little weird boot method from my side, is that I have syslinux on sda1 which loads Windows or chainloads to extlinux on sda3 where I have my Linux stuff.
I wasn't able to boot Linux directly from sda1 so I have this weird setup.
I now chainload from Syslinux on sda1 to Grub4DOS on sda1 which in turn boot Linux on sda3.
O.o
I can assure you that "kernel, initrd, tce directory, home, opt" are in fact the exact same.
"and boot codes" may have slight differences, though. However, even with only the necessary boot codes I got the same result.
LABEL core62test
MENU LABEL Tiny Core 6.2
KERNEL /tce-6.x/boot/vmlinuz
APPEND root=/dev/sda3 initrd=/tce-6.x/boot/core62.gz,/tce-6.x/boot/more-test-62.gz,/tce-6.x/boot/fonts-test.gz,/tce-6.x/boot/xorg.gz,/tce-6.x/boot/intel.gz vga=788 video=vesafb:ywrap,mtrr:3 logo.nologo quiet loglevel=4 mydata=myxorg kmap=qwertz/de-latin1 noutc tz=CET-1CEST,M3.5.0,M10.5.0/3 tce=UUID="cb6f8b98-91fd-4c96-b115-3d2bb9cb3e57"/tce-6.x laptop nodhcp desktop=openbox blacklist=pcmcia,ssb,b43,bcma,rt2860sta,bluetooth,btusb
timeout 0
default 0
title Tiny Core Linux
find --set-root /tce-6.x/boot/vmlinuz
kernel /tce-6.x/boot/vmlinuz logo.nologo quiet loglevel=3 vga=788 video=vesafb:ywrap,mtrr:3 blacklist=pcmcia,ssb,b43,bcma,rt2860sta,bluetooth,btusb kmap=qwertz/de-latin1 noutc tz=CET-1CEST,M3.5.0,M10.5.0/3 tce=UUID="cb6f8b98-91fd-4c96-b115-3d2bb9cb3e57"/tce-6.x mydata=myxorg laptop nodhcp desktop=openbox
initrd /tce-6.x/boot/core62.gz /tce-6.x/boot/more-test-62.gz /tce-6.x/boot/fonts-test.gz /tce-6.x/boot/xorg.gz /tce-6.x/boot/intel.gz
Also, if changing back and fourth from syslinux to grub4dos and back to syslinux (I didn't plan this - was an accident (: ), then booting Linux, Linux thinks I have only 1GB of RAM instead of 2GB.
So, even if Linux doesn't know about any of this, maybe the bootloaders do some memory hacks which don't work as expected or getting wrong information from the BIOS?
It's not the most expensive netbook. Plop Boot Manager is known to have problems booting off of USB on some laptops apparently because the BIOSes don't always stick to the standards.
-
You have something messed up in one of your initrds.
As I stated before. Boot STOCK initrd only. We can't troubleshoot your custom initrds.
-
You probably missed my reply?
You are not running a standard core system. You are loading multiple initrds
True, yet it works on other computers.
Boot a stock core and see if the problem persists.
The problem does persist. I freshly downloaded the 6.2 versions of core.gz and vmlinuz booted from /tce/boot/ without any extensions using this boot stanza via grub4dos
LABEL core62def
MENU LABEL Tiny Core 6.2 Default
KERNEL /tce/boot/vmlinuz
APPEND root=/dev/sda3 initrd=/tce/boot/core.gz noutc tz=CET-1CEST,M3.5.0,M10.5.0/3 tce=UUID="cb6f8b98-91fd-4c96-b115-3d2bb9cb3e57"/tce laptop nodhcp
Anyway, the issue can be worked around quite easily. As long as nobody else suffers from this, whether if it's a configuration incompatibility or a bug, I consider it solved for myself. Even when I boot stock Core, my boot setup might be unusual.
-
I have no idea what you have in your tce directory, or what is in your onboot.lst.
Add "base norestore" to your boot options. That will limit it it "Tiny Core Base" issues. Otherwise it belongs in the extension bugs forum, and you can start checking extensions one-by-one.
-
Another thing ... make sure you do not have any .gz files in your tce directory, or they will be loaded over the top of the root file system.
-
Thanks Gerald, I know you're trying to help and also trying to make sure it's not Core's fault.
But really, I have already made sure it's not because of any modifications I did, nor any extensions, by booting stock Core, just core.gz and vmlinuz, which gives same results.
The problem does persist. I freshly downloaded the 6.2 versions of core.gz and vmlinuz booted from /tce/boot/ without any extensions
[...]
It's just my box that's acting strange. Kinda old MSI Wind U100 netbook.
This is also true If I boot with base norestore boot codes.
[...]
It's 100% reproducible:
Syslinux -> Extlinux -> Core Linux = size of rootfs displayed incorrectly.
Grub2 -> Core Linux = size of rootfs displayed OK.
Syslinux -> Grub4DOS -> Core Linux = size of rootfs displayed OK.
Without having much knowledge about this topic, I'd guess it's either this particular netbook which gives values unexpected by Syslinux, or my Syslinux config needs to be changed, since everything works as expected when I use another bootloader.
-
I have also fsck'ed the partition that holds Extlinux. Seems to be alright.
-
I have no idea what you have in your tce directory
While testing stock Core, nothing was stored in my tce directory.
Except one folder /boot which contains core.gz and vmlinuz .
/tce/
/boot/
core.gz
vmlinuz
Should be enough, right?
-
I've always placed boot alongside tce at the drive level, can't say I've ever seen boot inside of tce before,
tce is for extensions..
Normally I install like this
sda1
/boot
/tce
-
TC-install creates /tce/boot/ on the boot drive.
-
OK thanks
-
If you could test the latest extlinux, I'm sure the syslinux project would like to hear if the bug still applies.