Prev: Acer Nplify
Next: Slackware 12.2 audio
From: Henrik Carlqvist on 7 Apr 2010 14:32 Longfellow <not(a)this.address> wrote: > On 2010-04-05, Henrik Carlqvist <Henrik.Carlqvist(a)deadspam.com> wrote: >> So what is your next step? > Well, I'm slowly but surely regaining a hint of *NIX chops. It's been > so long since I've done any of this, I have forgotten more than I > thought. Is that your way of saying "trying to fix my current installation?" > Thing is, 'lsmode' says nothing is inserted. It seems to me as if your current installation really is screwed up. Maybe this has something to do with the fact that X tried to load modules from an old an now non-existant directory. Did you do a standard Slackware install or did you try to reinvent everything like udev, initrd and kernel compilation from scratch? Unfortunately I haven't run Slackware 13 myself, so I can't answer how the boot process is supposed to be, maybe it should use an initial ramdisk, maybe it will do with a standalone kernel. regards Henrik -- The address in the header is only to prevent spam. My real address is: hc3(at)poolhem.se Examples of addresses which go to spammers: root(a)localhost postmaster(a)localhost
From: Martin on 7 Apr 2010 14:40 Longfellow wrote: > Slackware 12.0 didn't require an initrd and I don't recall having to > have had one before. Does this version need one, and if so, is there > anything I should know before trying to make one. > > Or what else am I missing here? > > Thanks, > > Longfellow If you didn't need one then you won't need one now. Anyway, since you said your problem is starting X, you can simply boot into runlevel 3 and take all the time in the world to troubleshoot your problem. No initrd required. Martin
From: Douglas Mayne on 7 Apr 2010 16:54 On Tue, 06 Apr 2010 18:52:58 -0500, Longfellow wrote: > On 2010-04-05, Henrik Carlqvist <Henrik.Carlqvist(a)deadspam.com> wrote: >> Longfellow <not(a)this.address> wrote: >>> I'll have to think about the next step; I definitely want an >>> up-to-date OS. Suddenly I hear an echo in the distance: >>> U-b-u-n-t-u!!! Nahh, I'm not that far gone... >> >> So what is your next step? Trying to find out what is wrong with your >> current installation? Redoing the install from scratch and hoping for >> better luck next time? Switching distribution or version as indicated >> above? Switching graphics hardware as indicated in a post where nVidia >> was mentioned? >> >> regards Henrik > > Well, I'm slowly but surely regaining a hint of *NIX chops. It's been > so long since I've done any of this, I have forgotten more than I > thought. > > It finally dawned on me that /boot contains the current configuration > file for the kernel, and slogging through the kernel build was entirely > unnecessary. The configuration file lists almost everything as a > module, which is just fine. Thing is, 'lsmode' says nothing is > inserted. Eventually, I recalled that modules need to be loaded early, > before the root fs is mounted. Which requires an 'initrd', I think. > > So, as there is no initrd in /boot, (or is that the > /etc/rc.d/initrd/functions file?), it looks like I need to make one. I > can't read shell scripts like I once could, and don't particularly > relish having to relearn everything again. In addition, there is no > instruction for mounting an initrd in the lilo.conf. > > Slackware 12.0 didn't require an initrd and I don't recall having to > have had one before. Does this version need one, and if so, is there > anything I should know before trying to make one. > > Or what else am I missing here? > > Thanks, > > Longfellow > AFAIK, it is highly recommended to use the modular kernel which requires an initrd. A lot of people either prefer compiling their own kernel, or using the non-modular kernel, but if you are using the standard modular kernel, then you will need an initrd. IIRC, Slackware 12.1 was the first slackware version to require the initrd. It includes no support for any disk filesystem (ext2, xfs). IMO, using a modular kernel makes sense. There is this help file in Slackware 13.0's base directory: README.initrd The command to create an initrd is mkinitrd. This creates a basic initrd at /boot/initrd-tree. The result is not made into a loopback file system as had been done earlier. Instead, it is created via cpio by the script. The initrd can also be tweaked and recreated using something similar to this: # (cd /boot/initrd-tree && find . | cpio -o -H newc ) | gzip >initrd.gz Hint: Include kernel modules for any devices and filesystems you'll need. Also, check the file modules.dep to verify all dependancies are present. -- Douglas Mayne
From: Longfellow on 7 Apr 2010 19:42 On 2010-04-07, Henrik Carlqvist <Henrik.Carlqvist(a)deadspam.com> wrote: > Longfellow <not(a)this.address> wrote: > >> On 2010-04-05, Henrik Carlqvist <Henrik.Carlqvist(a)deadspam.com> wrote: >>> So what is your next step? > >> Well, I'm slowly but surely regaining a hint of *NIX chops. It's been >> so long since I've done any of this, I have forgotten more than I >> thought. > > Is that your way of saying "trying to fix my current installation?" I guess so. I don't ever remember all this being so complicated! Something is missing that is needed to run X Window, or so say the complaints. /boot/config has everything compiled as modules, which then need to be loaded for them to work. 'lsmod' says they're not loaded. What I'm saying is that an initrd loads them before mounting the root fs. Slackware uses /etc/rc.d/rcX to call for desktop manager (without Gnome, Kde comes up), but that doesn't load the modules. Actually, I can't find anything that does load the modules, so I can't figure out (yet) how it works without an initrd. >> Thing is, 'lsmode' says nothing is inserted. > > It seems to me as if your current installation really is screwed up. Maybe > this has something to do with the fact that X tried to load modules from > an old an now non-existant directory. This installation is in a brand new partition, formatted with bad block checking and everything. I cannot imagine how it gets 'uname -r' to come up with the 12.0 kernel version, but it does. Which is why I just reinstalled the entire distro again. It still gives 'uname -r' as 2.6.21.5-smp, the 12.0 kernel. The only thing these two versions share is the swap partition, and there isn't supposed to be anything there after a cold reboot (shutdown, wait 5 minutes, boot up). > Did you do a standard Slackware install or did you try to reinvent > everything like udev, initrd and kernel compilation from scratch? > Unfortunately I haven't run Slackware 13 myself, so I can't answer how the > boot process is supposed to be, maybe it should use an initial ramdisk, > maybe it will do with a standalone kernel. No, the installation was virtually all default, except for time zone and a few rc.d's on boot up (like Sendmail and cups, etc). I'm stumped. Slackware has always been as smooth as silk, but not this time. Maybe I should see if Pat V. is about, and ask him what he thinks. Longfellow > regards Henrik
From: Chick Tower on 7 Apr 2010 22:52
On 2010-04-06, Longfellow <not(a)this.address> wrote: > Slackware 12.0 didn't require an initrd and I don't recall having to > have had one before. Does this version need one, and if so, is there > anything I should know before trying to make one. > > Or what else am I missing here? If you use the huge kernel or compile your own kernel, you won't need an initrd. If you want to use the generic kernel, see /boot/README.initrd. -- Chick Tower For e-mail: aols2 DOT sent DOT towerboy AT xoxy DOT net |