[syslinux] Re: mboot.c32, weird e820 map on HP blade machine, possible memory corruption
H. Peter Anvin
hpa at zytor.com
Fri Mar 17 09:16:46 PST 2006
Tim Deegan wrote:
> On Wed, Mar 15, 2006 at 01:23:20PM -0800, H. Peter Anvin wrote:
>>One thing I think in particular might be important is to set eflags,
>>which mboot.c currently doesn't do. eflags set to a random value is
>>usually bad news.
> Good point. I'm surprised that didn't cause trouble before now.
Well, BIOSes are *supposed* to not depend on it, and although it might
not look like it, for any particular feature, odds are most BIOSes
actually do it correctly. It's just across all BIOSes and all features
that it's a minefield.
Don't expect EFI to be any better, either. EFI is a 1500 page spec.
More information about the Syslinux