> 
> Alex Sofronie wrote:
> > geometry. Also, you must have the swap partition in the first 1024 cyls
> > ENTIRELLY. On my system, RHLinux 62 reported something like <<kernel panic:
> > could not...>> .
> 
> I have several systems with the entire swap space well after cyl 1024
> and have had no problems.  The auto partitioning in Pinstripe and
> Guinness set up swap at the very end of the range of cylinders.  I have
> even booted systems without any swap space at all and ample memory and
> have had no problems.  I have had systems with the swap space on a disk
> other than the "first" hard disk.  I have had swap space on a RAID-0 MD
> device.  AFAIK there is no 1024 cylinder limitation w.r.t. swap space. 
> Possibly it is a limitation if you don't have enough memory, or are
> using an old version of lilo or older kernel rev.  Might the "linear"
> option also make a difference where geometry problems are involved?
> 
> These days, I set up a 24-32MB /boot partition when needed to make sure
> the kernel is located within the first 1024 cyls.  I allow more space in
> /boot for SMP systems or systems which need special drivers where I
> would tend to have multiple kernel versions that I want to boot.  I have
> even had a case where "/boot" was on a partition of /dev/hda and "/" was
> on partition of /dev/hdc and it works fine.
> 

Access to swap is entirely a Linux not BIOS matter so it conforms to
Linux not BIOS restrictions.  In other worsds there is no 1024 limit
for swap.

There are there three things who must conformd to BIOS restrictions:
the boot loader, the kernel and initrd.  However the GRUB boot loader
seems to be abe to can access kernels and in initrds who are beyond
the 1024 cylinder or in the second IDE controller.  Perhaps modern
BIOSSes don't have those stringent limits.

-- 
                        Jean Francois Martinez

Project Independence: Linux for the Masses
http://www.independence.seul.org



_______________________________________________
Redhat-devel-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-devel-list

Reply via email to