I'm tracking two bugs in the database about how lilo isn't sizing memory properly when it loads ramdisks. This appears to affect a certain class of machine. I don't have a machine with < 64 MB of memory, so I cannot test it directly. I can tell that the latest LILO includes a new codepath that detects machines with more than 64MB of memory and puts the ramdisk at that boundary. One user reported that his k6-2 wouldn't boot at all, but it isn't clear that his problem is with ramdisks since he didn't state it as such. He did claim that one of the interrupt calls used in LILO returns an incorrect memory limit.
Has anyone on this list encountered this problem? I'm thinking that one solution would be to disable this new memory check and fallback to the older code that seems to work. Opinions? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
