Re: First Test Cycle starts today
>>>>> "Richard" == Richard Braakman <dark@xs4all.nl> writes:
Richard> On Tue, May 02, 2000 at 02:22:32PM -0700, Karl M. Hegbloom wrote:
>>
>> Please tell me the new Lilo made it into the distribution in time.
Richard> Yes.
Hey! Great! Thank you, Richard.
Richard> You know, the reason I kept it out originally is that I didn't want
Richard> to spring such a change on the boot-floppies people. You could have
Richard> told me you were waiting for it.
Well, I did tell you, I thought... just not as soon as I should
have. 's ok, we're in time for release, and for this test cycle too,
even. :-)
`dbootstrap' and the Installation Manual both tell of the new BIOS's
large disk access extensions that are supported by the new Lilo boot
sector and by MBR. It explains that older BIOS don't have that, so
in the case of old BIOS, you must partition a large disk with a boot
partition, mounted on "/boot". There is some support for doing that;
"/boot" is offered as a mount point, and a dialog repeats the
information about BIOS large disk extensions etc. prior to running
`cfdisk'. I started trying to make it so it would verify that the
"/boot" partition is all below the 1024 cylinder boundary (hacking in
libfdisk), but was unsure of the correctness of my code, and backed
that part out. Hopefully people will follow directions and not try
to make a huge "/boot" partition. I recommend, in the docs and
dialog, a 5-10Mb partition, which is well below the limit.
I've read at least one email (I think it was in the local Linux
user's group list) from a person who had just upgraded the kernel and
Lilo wouldn't work, when it had worked fine before that... the disk
filled up, and the newly compiled kernel was no longer all below the
1024 cylinder boundary. The person did not have a "/boot" partition.
It worked fine until the disk filled up more. Hopefully, future
Debian users will not have that problem on older equipment ever.
Reply to: