Very Computer

I am trying to install Linux using the SLS 1.03 disk suite on my
ThinkPad 350. The kernal is 99.pl12. Lilo is 0.10.
Because the ThinkPad does not store hard disk info in CMOS I use the
following method to install Linux:
1) Boot with a1.3, select RAMDISK & HD=cyl,hd,sec
2) Log in as root
3) FDISK to create a primary Linux partition about 62M (n). Toggle
   as bootable (a). Write table to disk (w).
4) Reboot as in (1) log in as install
5) Select 1 from SLS and mke2fs creates file system on /dev/hda1
6) Select 7 from SLS to install a minimal system from 3.5 A drive
   without asking me questions.
7) Creates floppy disk and adds Linux to HDD
8) Reboot with a1.3 specify HARDDISK ROOT=/dev/hda1 hd=cyl,hd,sec
9) rdev /zImage to use /dev/hda1 as root
10) go to /etc/lilo to modify disktab. Set the only uncommented
    disktab entry to '0x300 0x80 sec hd cyl 0' (no quotes :-) )
11) config has boot as /dev/hda1 & root as /dev/hda1
12) run install -v -v -v and boot sector was added.
13) Boot DOS 5.02 diskette and do FDISK /MBR to get master boot record.

Now depending on the geometry of the disk I get varied results.
Size    Cyl     HD     Sec      Result
85M     981     10     17       Works Linux boots to login prompt
85M     989     10     17       Works Linux boots to login prompt
128M   1024     14     17       Won't boot from any disk. Floppy boot
                                dies with errors uncompressing Linux with
                                CRC errors or corrupt input. File system
                                is hosed e2fsck gives all sorts of errors.
                                Programs won't run. Segmentation violations.
252M    569     16     54       Will only boot from a1.3 w/ harddisk root= hd=
                                Gives LI on hd boot. Floppy can't find root.
252M    989     16     34       Will only boot from a1.3 w/ harddisk root= hd=
                                Floopy & HD boot can't find root.

I am a newbie, but I have read LILO FAQ & Linux FAQ.
- Is there newer SLS that has a newer LILO or newer kernal
  that would fix this?
- Am I missing something?
- Should I ask this question elsewhere?

-----------------------------------------------------------------

More info.
1) Reinstalled following same exact procedure on 85M again and found that
   they could no longer find root. I must have done something different for
   them originally.
2) If at the LILO prompt I specified 'linux HD=cyl,hd,sec' the systems do
   find the root and boot properly (85 & 252 that gave LILO prompt)
3) I looked at the map files and saw that there was no HD=cyl,hd,sec
   phrase after the linux text. Somehow I got this in to the map file
   for the 85M & it booted properly again. Still trying to figure what
   I need to run to get this in. I know it is the -R "HD=cyl,hd,sec"
   but install & lilo keep flaking out when I run with this parameter.
   I ran bootargs since it looked like it should do the correct thing
   which worked on the 85 but not on the 252. I am sure I did something
   different between the two.
4) The suggested fix for the 252 that booted to LI in both the
   LILO FAQ & README did not fix the problem.

-----------------------------------------------------------------

I was able to get it to work properly once I had the proper disk
parameters. I queried the drive using the 'Identify' command and
found my numbers were wrong. Now all I need to do is

Quote:>1) Boot with a1.3, select RAMDISK & HD=cyl,hd,sec
>2) Log in as root
>3) FDISK to create a primary Linux partition about 62M (n). Toggle
>   Create remaining part of disk as swap Toggle 1st partition
>   as bootable (a). Write table to disk (w).
>4) Reboot as in (1) log in as install
>5) Select 1 from SLS and mke2fs creates file system on /dev/hda1
>6) Select 2 from SLS and mkswap creates swap partition on /dev/hda2
>6) Select 7 from SLS to install a minimal system from 3.5 A drive
>   without asking me questions.
>7) Creates floppy disk and adds Linux to HDD
>8) Boot DOS 5.02 diskette and do FDISK /MBR to get master boot record.

Can't the kernal be modified to query the parameters from the drive
first and if that fails then go to CMOS? It would have saved me
a lot of trouble.
-----------------------------------------------------------------