Hello List,
Hm... people are very quiet, so I had to keep trying by myself. The good
news is that 3.09, 3.10, and 3.11 do work, but not with a full CD. My
experiments show that somwhere between 400MB and 450MB some limit is reached
and the CDs won't boot no matter what syslinux is on them. When examining
the output of isoinfo -d -i whatever.iso I noticed only a few things
different between images that boot and images that do not.
Volume Size is ...
El Torito VD version 1 found, boot catalog is in ....
Bootoff ....
Values that work for this are:
Volume size is: 228182
El Torito VD version 1 found, boot catalog is in sector 109
Bootoff 6E 110
Values that do not work for this are:
Volume size is: 244596
El Torito VD version 1 found, boot catalog is in sector 110
Bootoff 6F 111
The problem is probably that Bootoff gets too big. Having spent almost a
week and 50 CD blanks, I feel like narrowing it down to exactly where it
fails probably doesn't matter. The more data I add, the larger the boot
catalog start and bootoff get. I don't know if that is the number of
paths on the
CD index or the total number of megabytes on the CD. What I do know is at
least for me if I'm willing to waste about 1/3 of the CD, it boots and works
on the problematic (and obviously incompatible) Award BIOS V6.00PG that
every pentium 4 on campus is using. I didn't get an answer to this query
probably because the answer last year was to upgrade the BIOS. Here, like
many places, BIOS will not be upgraded if it boots windows, so for people
willing to give up space on the first installation CD, there is a workaround
that doesn't require changing BIOS.
Thank you,
JGH
---------- Forwarded message ----------
From: BuraphaLinux Server <buraphalinuxserver at gmail.com>
Date: Apr 8, 2006 3:22 AM
Subject: problems booting isolinux
To: syslinux at zytor.com
Hello List,
I have used isolinux for years without problems. Recently I decided to
upgrade from 3.09 to 3.11. Every CD I create can be read on my Vectra
(Pentium 3)
and all other Pentium 3 boxes, but all my Pentium 4 boxes will not
boot (they ignore
the CD and boot from hard disk). The bochs VM will boot the CDs fine
just like the
Pentium 3 machines. I have tried 3.09 (which worked before) and 3.10, 3.10pre1,
3.10pre3, 3.10pre5, 3.10pre10, 3.11. Each time I verified the isolinux.bin was
identical to the shipped version (I used the DATE= option to make as required).
I was using the simple menu system, but reverted to the old boot: command-line
system to remove the simple menu system code. The only thing I can see that is
different is that I have exceeded the old size for my ISO. Now I am about 470MB
instead of just under 450MB. Is there some requirement to boot from some low
block (like hard disks needed to boot from cylinders<1024). I am beginning
to
think I need to use the almost never used -sort option to mkisofs to force my
/isolinux directory to be earlier on the disk. After burning over 20
CDs with various
combinations of isolinux(-debug).bin and with/without the menu system,
I am at my
wits end. I searched the mailing list archives and found similar
problems mentioned
but no answers. My question is: How do I debug this problem?
I can live without menus, but not without booting. Something about the P4 boxes
(all have award modular bios v6.00PG, but with various date numbers like
12/22/2003-I865P/PE-6A79AD4DC-00 and 09/12/2003-I845PE-6A69YG08C-00)
means I cannot boot. As you can probably guess, most machines here are P4.
I even took the CD-ROM drive out of my vectra (P3) that booted the CD, used the
same drive and CD in my P4 box and the CD failed too boot. That is
why I think it is
some interaction between the BIOS, mkisofs's image generation, and isolinux.
I copied a 3.09 entire /isolinux directory into my image before
burning but nothing
worked on the p4 boxes. This makes me thing that somehow the way mkisofs
generates the image means that isolinux.bin is not loaded (or found).
After booting
from hard drive, the P4 boxes can read all files on the CD and get the
right sha1sums
for every file, so I don't think it is a read failure of the CD drive
or media. While
most CD drives here are Samsung, I did test with two different SONY and one HP
drives The drives boot my older CDs but not the newer ones that are
larger (but still
much less than the 650MB quoted maximum).
I am willing to try paches, provide files, etc. I hope someone on the
list can help.
I tried mkisofs 2.01 and 2.01a7; as you can guess, my head is still
spinning from all
these combinations of boot attempts and the loss of almost an entire
box of CD blanks.
There is no question I am doing something wrong, and probably stupid.
I am ready
to be educated.
I suspect you need some files from me or for me to run some
diagnostics. I am ready;
what must I do?
Thanks - JGH