Just tried a reinstall off the cooking lowram-cd. After upgrade, and confirming that I am running the 2.6.36 kernel, still having the same problems.

FWIW, udev is failing.

See http://forum.slitaz.org/index.php/discussion/comment/10013/#Comment_10013

Yea, I saw that discussion. As it advises, I upgraded. I now have the 2.6.36 kernel installed, and my computer went from completely functional aside from X to bailing before mount completes and having no modules load. I may be having a different problem here.

I'm running an old laptop, Micron Transport Trek2, Pentium II 300 with 64MB RAM. Probably not common hadware these days - this may be the first time somone has run this cooking or kernel on this type of Hardware. Could be the 2.6.36 kernel has some problems with my hardware, could be that the new Cooking's configuration broke my working install somehow. Its hard to troubleshoot since boot bails before any drivers are loaded, and its not entirely clear to me yet how to init SliTaz manually. I can say that uname reveals 2.6.36, so that much I did "right".

Just my 2cents on the issue - a lot of people have had trouble with the Cooking iso as the current X drivers are compiled against a newer kernel. They have been advised to upgrade, and that seems reasonable - if it works. But in my case the upgrade broke my machine. There are two big variables at play here - the kernel and the Cooking release. As someone else said, it may be helpful to have an iso that works rather than having to upgrade. At least then we have a known configuration to start from when testing inevitably reveals a problem.

The Stable SliTaz is wonderful, and exactly what I'm looking for. I went to Cooking for the Xorg R128 driver. Looks like I'll just go back and compile R128 against stable.

I have the same problems. SliTaz 3.0 is running perfectly on my Pentium III machine, but never managed to successfully install any cooking iso. The latest cooking, I was only able to login as root, otherwise it "failed to execute login command". Then as root I upgraded to the latest 2.36 kernel and packages. Now I have "Unable to locate IOAPIC for GSI..." and "mounting devpts failed" and no X/slim login screen, just the console. startx doesn't help. I even tried to build my own iso with tazlito and the latest cooking packages, but no change. And as it is mentioned, the Live-CD works well, only booting from hd fails, so I guess there must be a bug in the installer...

If you run into the 'unable to mount devpts' issue, copying vmlinuz-*.slitaz from a working setup usually resolves it. That is what I was trying to point out in my earlier post.

You can generate a working cooking iso by doing "# tazlito clean-distro ; tazlito get-flavor core ; tazlito gen-distro". Then create a live-usb using that iso, install to hdd. This should already have the upgraded packages. If for whatever reason you run a tazpkg upgrade, that will most likely break the install, and if that happens, copying vmlinuz-* from a working setup should 'unbreak' it. Atleast, it did for me.

to login as a user different than root you need .config/slitaz and .xinitrc in the user home directory.
i don't know if files are present in tux home, but i changed my name from tux to fabio and i have done the following to login as fabio:

cp -a /root/.config /home/fabio
chown -R fabio:fabio /home/fabio/.config
cp /root/.xinitrc /home/fabio
chown fabio:fabio /home/fabio/.xinitrc 

do these commands with you username instead of fabio and you will be able to login.
http://mirror.slitaz.org/iso/cooking/flavors/slitaz-cooking-4in1.iso already has 2.6.36, if you install from it you don't need to upgrade the kernel.
i had the same issues from upgrading the core flavor, but i think (i had no time to investigate) that they are caused by busybox upgrade, i had some grep error from the post install script

thanks ms3811, your solution works and busybox is not involved at all.
i've upgrade from the 4+1 iso to resolve some issues with my intel 945GM VGA adapter and the 'unable to mount devpts' issue has come up again, overwriting /boot/vmlinuz-2.6.36-slitaz with the one in the 4+1 iso solved the problem.
again, thanks :-)

I gave up and these were my findings. Installing 2.6.34 cooking created the file .xinitrc and the .config folder in /home instead of /home/user, but with owner 'root'. So moving these to /home/user and "#chown -R user:user /home/user" fixes that problem and it's not a big deal. The problem for me was that 2.6.34 works and 2.6.36 doesn't. When I try to install the "ati" driver I get the kernel modules for 2.6.36 and they don't fit the 2.6.34 kernel. It's also unclear how the two different kernel module folders can work together. So I'll be back for the next stable release.

I gave up and these were my findings. Installing 2.6.34 cooking created the file .xinitrc and the .config folder in /home instead of /home/user, but with owner 'root'. So moving these to /home/user and "#chown -R user:user /home/user" fixes that problem and it's not a big deal. The problem for me was that 2.6.34 works and 2.6.36 doesn't. When I try to install the "ati" driver I get the kernel modules for 2.6.36 and they don't fit the 2.6.34 kernel. It's also unclear how the two different kernel module folders can work together. So I'll be back for the next stable release.

me too...

@ernia, to give credit where its due, that was actually @kultex's idea. But you are welcome anyway :).

@ms3811
:-)
@hymie
what if you tazpkg recharge and tazpkg upgrade and reboot with the 2.6.36 kernel?

@ernia
That's what I wrote in my first post:
"Now I have "Unable to locate IOAPIC for GSI..." and "mounting devpts failed" and no X/slim login screen."
But oddly the Live-CD I managed to create with tazlito and 2.6.36 kernel "works well, only booting from hd fails" same way.

Off topic:
When using chromium browser and posting comments I get:
Gdn_ErrorException thrown within the exception handler.
Message: Cannot modify header information - headers already sent by (output started at /home/slitaz/www/forum/library/core/class.email.php:113) in /home/slitaz/www/forum/library/core/functions.error.php on line 62
But the comment appears correctly - that's the reason why the previous post is double - I sent it again due to the error which then reappeared again.

the init scripts from livecd to hd install are different, so it could be possible to have different behaviors.
did you try to overwrite /boot/vmlinuz-2.6.36-slitaz in your hd install with the vmlinuz-2.6.36-slitaz in the 4 in 1 iso image? it worked for me. you can boot with the 4in1 iso http://mirror.slitaz.org/iso/cooking/flavors/slitaz-cooking-4in1.iso and use it to fix your hd install

I'm not an expert in that, but are the init-scripts part of the vmlinuz-...? I'll try to install the 4in1 cooking flavor on the Pentium III 512 MiB RAM system, just for curiosity. Still have the Slitaz 3.0 stable in another partition working well, but was simply trying to have more up to date programs - not a big deal...

So I had the 4in1 cooking successfully installed, but my radeon 9200 card is not working propperly after "tazhw setup ati". I have only a "software rasterizer". In boot time I can see a message "GPU initialization failed (errorcode -2)" or something like that. I tried a few things in the boot options (nomodeset and noapic) and manipulating "xorg.conf" and now the system is broken again ;-). The 3.0 stable version that I'm using right now everything is perfect, glxinfo: "OpenGL renderer string: Mesa DRI R200 20060602 x86/MMX/SSE TCL".
By the way ".xinitrc" and ".config/.." are still misplaced in the "/home" instead of "/home/user" folder and have to be moved manually...

My very last feedback on this ;-) The GPU problem seems to be due to missing firmware, because there is another message before that says "failed to load firmware radeon/R200_cp.bin". I don't remember that I had to provide any special firmware when installing the stable release, except for my p54usb wlan adapter. So long...

"tazpkg get-install linux-firmware" did the trick.

OpenGL renderer string: Mesa DRI OpenGL renderer string: Mesa DRI R200 (RV280 5960) 20090101 x86/MMX/SSE TCL DRI2R200 (RV280 5960) 20090101

But the frame rate is only at 300 in 5 sec, while in 3.0 stable I have 2400 ??

I decided to give slitaz another try after a six month hiatus and got this same problem. /dev/pts fails to mount...

mount: mounting devpts on /dev/pts failed: No such file or directory
mount: mounting tmpfs on /dev/shm failed: No such file or directory

So I wipe the installation, do a clean reinstall, don't install any restricted drivers, do a tazpkg recharge tazpkg upgrade and voila, the same exact problem. Upon booting I get dumped at tty1. When I ran startx, I was unable to interact with the computer using mouse or keyboard. I think it froze because the cpu graph in the bottom right wasn't changing.

This is using the latest slitaz 3.0 cooking iso.

I decided to give slitaz another try after a six month hiatus and got this same problem. /dev/pts fails to mount...

mount: mounting devpts on /dev/pts failed: No such file or directory
mount: mounting tmpfs on /dev/shm failed: No such file or directory

So I wipe the installation, do a clean reinstall, don't install any restricted drivers, do a tazpkg recharge tazpkg upgrade and voila, the same exact problem. Upon booting I get dumped at tty1. When I ran startx, I was unable to interact with the computer using mouse or keyboard. I think it froze because the cpu graph in the bottom right wasn't changing.

This is using the latest slitaz 3.0 cooking iso.

sorry for double posting, when I push post comment I get a php error from the slitaz website.

I have it perfectly running now. Even copied to an ext4 partition. glxgears says that now it's "Running synchronized to the vertical refresh." So I cannot compare that, but the feel is that graphics at least is not slower than before and running from ext4 makes the hd feel faster. From what I've tried with Linux I think SliTaz is the fastest Linux distro around, although these few issues are a pain (when someone has no clue of what the problem is) and I don't think that it should be a big deal fixing them.

mmm, did you made a standard cooking install in an ext4 partition and upgraded without overwriting the kernel?
if it's so and it works out of the box in an ext4 partition, could the issues be due to the fact that direct ext3 support it's disabled in the kernel and demandend to ext4 support?

root@q1s:~# zcat /proc/config.gz | grep EXT3
# CONFIG_EXT3_FS is not set
root@q1s:~# zcat /proc/config.gz | grep EXT4
CONFIG_EXT4_FS=y
CONFIG_EXT4_USE_FOR_EXT23=y
CONFIG_EXT4_FS_XATTR=y
CONFIG_EXT4_FS_POSIX_ACL=y
# CONFIG_EXT4_FS_SECURITY is not set
# CONFIG_EXT4_DEBUG is not set

@ernia
What is the

uname -r

of the kernel where
# CONFIG_EXT3_FS is not set
So everyone is on the same page.
Thanks ;-)

it's 2.6.36, but my suppositions were wrong, i've tried to install in an ext4 partition in a spare disk i have and nothing has changed after the upgrade.
the working vmlinuz-2.6.36-slitaz has the same ext3 setting so it is not the point, sorry
i will zcat the non working vs workin kernel but maybe the right way to check difference would be to check the wok for changes

ok, this is the edited diff (to remove some row)

--- workingkernel
+++ nonworkingkernel

-# Sat Nov 13 07:20:38 2010
+# Tue Nov 23 20:01:35 2010

-# CONFIG_APM is not set
+CONFIG_X86_APM_BOOT=y
+CONFIG_APM=y
+# CONFIG_APM_IGNORE_USER_SUSPEND is not set
+CONFIG_APM_DO_ENABLE=y
+# CONFIG_APM_CPU_IDLE is not set
+# CONFIG_APM_DISPLAY_BLANK is not set
+# CONFIG_APM_ALLOW_INTS is not set
 

-# CONFIG_DEVTMPFS is not set
+CONFIG_DEVTMPFS=y
+CONFIG_DEVTMPFS_MOUNT=y

-# CONFIG_XFS_QUOTA is not set
+CONFIG_XFS_QUOTA=y

-# CONFIG_AUTOFS4_FS is not set
+CONFIG_AUTOFS4_FS=m

-CONFIG_AUFS_FS=m
-# CONFIG_AUFS_BRANCH_MAX_127 is not set
-# CONFIG_AUFS_BRANCH_MAX_511 is not set
-CONFIG_AUFS_BRANCH_MAX_1023=y
-# CONFIG_AUFS_BRANCH_MAX_32767 is not set
-CONFIG_AUFS_HNOTIFY=y
-CONFIG_AUFS_HFSNOTIFY=y
-# CONFIG_AUFS_EXPORT is not set
-# CONFIG_AUFS_RDU is not set
-# CONFIG_AUFS_SP_IATTR is not set
-CONFIG_AUFS_SHWH=y
-CONFIG_AUFS_BR_RAMFS=y
-# CONFIG_AUFS_BR_FUSE is not set
-# CONFIG_AUFS_BR_HFSPLUS is not set
-CONFIG_AUFS_BDEV_LOOP=y
-# CONFIG_AUFS_DEBUG is not set

-# CONFIG_NFSD is not set
+CONFIG_NFSD=m
+CONFIG_NFSD_V2_ACL=y
+CONFIG_NFSD_V3=y
+CONFIG_NFSD_V3_ACL=y
+# CONFIG_NFSD_V4 is not set

-# CONFIG_RPCSEC_GSS_KRB5 is not set
+CONFIG_SUNRPC_GSS=m
+CONFIG_RPCSEC_GSS_KRB5=m

full version: http://pastebin.com/7BmJY2iE
i don't have any idea right now, i'm going bed and i leave you the guessing game, but the first thing that i note is that CONFIG_AUFS_FS=m disappear without being replaced from something else, could it be something wrong with some patch?

What I did was just installing the 4in1 on an ext3 partition. And when I had everything working, I made a full backup of the filesystem to a usb-hd. Then I wanted to try out if it works on ext4, changed the partition to ext4 and restored from the backup. And it still works, that's all...

@ernia
aufs is now its own package. I did this so source tarball will be saved and not purge when running tazdev purge on tank.

@Hymie
The speed of glxgears on 3.0 stable and cooking may vary in big different. Mostly cause 3.0 stable uses xorg-server 1.5.2 and mesa 7.4.4.

so the correct solution would be to have all working after the upgrade would be to install the aufs package?
if yes, should not the kernel depends on aufs?

@ernia
The kernel doesn't depend on aufs. Just explaning that aufs is not build with linux package any more. Thats all.

sorry, i misunderstood

the problem seems to be due to the

CONFIG_DEVTMPFS=y

option:
http://cateee.net/lkddb/web-lkddb/DEVTMPFS.html
i've rebuild the kernel without this option and it works, if i reinclude the option it stops working. Maybe it interfere with the init script?

@ernia
That option is added so /dev is writable when booting from hard drive. Otherwise i think /dev/ will end up being read only and it will fail.

Check to see if its CONFIG_DEVTMPFS_MOUNT=y option. Disable that and keep CONFIG_DEVTMPFS=y to see if that CONFIG_DEVTMPFS_MOUNT is causing the problem.

fabio@cooking:~$ zcat /proc/config.gz | grep CONFIG_DEVTMPFS
CONFIG_DEVTMPFS=y
# CONFIG_DEVTMPFS_MOUNT is not set

works!
reading http://cateee.net/lkddb/web-lkddb/DEVTMPFS_MOUNT.html (where is stated that the option does not affect initramfs based booting, which would explain why it does not cause trouble in live mode) i've tried to boot the repo kernel with

devtmpfs.mount=0

and it works, so it is possible to correctly boot the repo kernel with this kernel commandline parameter, until the problem get fixed.

i like wok logs :-)
http://hg.slitaz.org/wok/rev/b6f6e0af98c6
thanks

I made a full upgrade on a machine that I'm only using for tests, a very old Pentium II with Riva TNT (nv driver). I managed to upgrade everything but exactly "linux" which I had to block. And it seems not even the vmlinuz... itself but something is wrong with the kernel modules that get installed/replaced along, because when I copy the working vmlinuz... it's still the same problem. I have a black screen and monitor being switched to standby, only Ctrl-Alt-Del still working ;-) Could it be that "nv" moved to linux-firmware or something like that?

It works on Pentium III / ATI-Radeon 9200 and linux-firmware installed. Writing here after full upgrade.

Now the "nv" Riva TNT also works. Had to add the "nomodeset" parameter. Everything perfect...