[syslinux] Patch sensible callback framework

[H. Peter Anvin]

On 04/29/2010 07:42 AM, Steffen Winterfeldt wrote:
>>
>> Keep in mind that when we had 64K steps the maximum kernel size was
>> 512K... no more than 8 steps.
> 
> Ok, 64k is maybe a bit extreme. All I'm saying is that it would be nice to
> have the chunk size configurable.
> 

Actually, for a lot of applications the lower layer should be able to
recognize a timeout, which might be more appropriate.  In the case of
PXE, for example, there is a lower layer packet-reading loop, which
could check for a timeout.  This should make it possible to invoke a
callback back on time units rather than on data flow.

As far as kernel/initrd size... they vary all over the place.  At the
same time, your point about progress bars is also true: one may not even
want to bother drawing it until the boot takes more than say four
seconds (in fact, "menu hidden" is there in no small part for the people
who do *not* want to show any kind of progress info.)

	-hpa

-- 
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.