Re: it works!!! (almost)

[Jesper S. <js...@re...>]

>>>>> "Roman" == Roman Zippel <zi...@fh...> writes:

Roman> Hi,
>> > Yesterday I got 2.3.99pre8 running, I could start an init shell,
>> do a ls > and go to bed. :-)
>> 
>> Now, tell us, what was the showstopper?

Roman> A wrong entry in the hash table, the usual assumption that all
Roman> memory starts at physical address zero. :( The problem was,
Roman> that it had such a strange effect, the cpu did basically
Roman> nothing (visibly) for about a minute after the hash table entry
Roman> was updated. There was no further page fault for that, so I did
Roman> not expect that the hash table entry was wrong. And as soon the
Roman> first page fault from user space happened, I saw that the entry
Roman> was updated, it returned and after that I got machine check
Roman> exception from the interrupt handler...

I suspected as much, but since I didn't see any further exceptions
after hash table updating, I figured it had to be something else. I've
never had the patience to let the fscker sit for a full minute and
cook in its own fat :)

Very nice!

Jesper