Hi all,

I decided to play with a custom 2.6.18.

I only have two (and I hope) minor issues. Wondering if someone may be able to shed light on this for me.

Issue 1: Only with the custom 2.6.18 kernel, at 'init3' run level, when I log in as a normal user, I get this immediately echoed on the screen under my name.

"Warning: add_groups: Invalid Argument"

This does not occur on the installed 2.4.33 or the 2.6.17/huge26.s kernel that are in the pc too.

Issue 2: There are only two modules that I need to force my slackware to load, maybe I should just build them in for now on. They are for the temperature sensors of the CPU. For now, I'm just modprobing them thru /etc/rc.d/rc.local. However, are we supposed to create a new (as for this example) /etc/rc.d/rc.modules-2.6.18 and maybe remove the current symlink'd rc.modules then create a symlink called "rc.modules" pointing to /etc/rc.d/rc.modules-2.6.18?

so for example:

rm /etc/rc.d/rc.modules
cp /etc/rc.d/rc.modules-2.6.17.3 /etc/rc.d/rc.modules-2.6.18
ln -s /etc/rc.d/rc.modules-2.6.18 /etc/rc.d/rc.modules

The way I made my kernel is I pulled out the config-generic-2.6.18 from my slackware 11.0 cd's. I renamed it .config. I placed it in a clean freshly extracted linux-2.6.18 source folder, and pretty much just added in preempt kernel, 1000khz timer, and removed stuff I know that I don't use to trim it down a little, mostly dallas one wire, stuff like that.

Thank you in advance.

OK,

I took myself out of 'man pages' group and now I dont get the error.

No biggie, I didn't like man pages anyway LOL.

Problem solved for that.

Also, /etc/rc.d/rc.S explains what's going on with the modules.

so yeah it's differnt that slack 10.2 but that explains it. i needed a new rc.modules-2.6.18 made it executable.

it was just a duplicate of my 2.6.17 that I had already tweaked.

bye for now all.

I am trying to figure out why your membership in the man group in /etc/group (if I have understood you correctly) would have anyting to do with the login. Also, I just checked my box, and I am not in the man group, but I can read man pages fine. Have I misunderstood something? I don't mean to hash over something you've already solved, but this sounds odd and has made me curious.

Did you upgrade that box to Slackware 11.0 as opposed to installing from scratch? Is the correct "shadow" package installed? Are all the new rc scripts in place (most notably /etc/rc.d/rc.udev and /etc/udev/rules.d/udev.rules).What is the difference between the kernel config of the Slackware 2.6.18 kernel and your custom config?

Eric

Hi guys,

I read "gileads" post here on the forums:
http://www.linuxquestions.org/questi...d.php?t=426383

The originator of the thread had the same issue on slackware current that I appeared to be having. Why google all of the sudden found this for me after hours of searching and initiating this thread I don't know.

I figured I would give that a shot, remove myself from some of the lesser used groups, one at a time and see what happens. So I just chose to remove myself from man pages first, and that solved my issue. I can live with that, I tend to use online web page - man pages anyway.

PKGTOOL - > view reports that "shadow-4.0.3-i486-13" is installed.

The installation was:
1. clean partition /dev/hda6 with slackware 10.2 on /dev/hda5
2. full nooob install from gpg verified and md5sum'd cd's, i used sata.i to install, then at kernel selection I chose huge 2.6s
2a. installed lilo to the boot sector of the hda6 and my slack 11 is booted up from the lilo in the bootrecord that is owned by slack 10.2 on /dev/hda5, make sense?
3. rebooted, went to extra installed my modules
4. tweaked slack 11, insalled video, printers, and packages that I already had made or downloaded from Alien bob's packages or linuxpackages.
5. saw a thread where someone said they couldn't roll their own 2.6.18 so figured I'd make my own kernel to see if anything different.
6. i do have the files that you ask about Alien Bob.
7. the only udev rules I have are the ones that I made for my mouse, the rule that I have in my guide on xbindkeys, xvkbd and the networking udev rule so that my eth0 and eth1 stay in sync with the respective mac addresses. the exact way and what that I did is here:http://www.linuxquestions.org/questi...d.php?t=489285
8. I am using the dazuko module for security as I describe in the end of this post here:
http://www.linuxquestions.org/questi...d.php?t=439994
9. I have made the /etc/rc.d/rc.modules-2.6.18 so that slackware knows to modprobe the dazuko for me at boot time, before capabilities, and brings up the temperature sensors for my cpu. tho I really could use xmbmon as it directly accesses the chips and doesn't even need linux to do it, but a really bad gui so I prefer to use the modules from the kernel for the kicker applet 'cpuinfo'
10. I'm trying to figure out how to work 'kompare' to just give me the text differences of my kernel config.
basically:
changed to :1000 khz, changed so kernel is for an amd cpu, force module unload, removed dallas one wire, built in xfs and reiser and ext3. removed video cards and chipsets in character devices for hardware i dont have. added preempt kernel, hpet, removed bluetooth, removed appletalk &novell internet protocols,using highmem (tho I can't choose 3G/1G split? or any of them),stayed with flat memory, removed generic i386 under cpu, added ACPI compliant motherboard (experimental yes, but fixes issues for mapping devices on my nforce2 board); added 'apic' as well as I do have this device, it's listed in my bios settings too. Did not change any of the security modules that are in Mr. Pat V's config for testing 2.6.18 kernel generic.

I must say that this slackware 11 runs really nice...but...I kind of feel like I'm using SuSe or Ubuntu in a way, this udev makes the pc have a mind of it's own.

Here's my 1 week perspective, obviously subject to change as I get smarter or so I hope. These aren't the plithera of things that I'm really happy with that's a whole new thread.

This is the only machine I'm testing 11.0 on, and I was able to get 11.0 into vmware too http://www.linuxquestions.org/questi...d.php?t=490483.
So I do have "Pam" on here. issue? not sure.

But this computer here running native 11.0, well the boot processes are not 'consistent'. Sometimes booting either the huge26 or my custom kernel, I get pauses at different points along the boot process once udev detection kicks in.

Sometimes, as I boot up the pc, I get messages spammed on my console during the boot process for my external USB 2.0 hubs, then after about 20 lines or so, I guess the udev finds some way to map that hub and then it assigns the nodes and continues booting. But on some boots, the udev just rips thru the USB hardware detection lightening fast. Overall the udev is still much faster than the hotplug, but sometimes it appears to be fighting to figure out what it wants to do.

Another issue, due to udev for me was that I was using a usb keyboard, slack 11 crashed, and I had no keyboad to run e2fsck, had to find a ps/2 keyboard and boot up w/that to enter my root pass to do administration.

One last issue I have is for my ati video card. I'm not sure, but I may need to use the xorg patch. My video card has reset on me while playing a game. Constant no mtrr found then re-initialization of the card is spammed in dmesg. this is on huge26 & my 2.6.18 custom.

the above is spammed for about 1000 lines! then I get DRI.

lastly: i get a message at boot for some kind of sub-tree check when I try and nfs mount a share from a different pc during the boot process. I have not had a chance to even read up on this but it does not happen in 10.2, i basically just copied the text of /etc/host.allow..deny...and exports over and the text of the nfs mounts to my fstab.

and I'm still trying to figure out this problem here in 11.0 too, this issue never happened in 10.2 either;
http://www.linuxquestions.org/questi...d.php?t=489929