On 5/26/06, Eric W. Biederman <[EMAIL PROTECTED]> wrote:
> "Magnus Damm" <[EMAIL PROTECTED]> writes:
> >
> > Also, I feel that my approach with a valid idt and gdt is more robust.
>
> One of my biggest concerns with the current code is that it is not
> sufficiently robust, in the kdump case.  So I am all in favor things
> that improve that situation.  At the same time just moving code from C
> to assembly doesn't make it more robust, especially when the comments
> explaining what the code does don't come along.

I agree that just moving the code does not help. But my code actually
loads a new set of gdts and idts and I'm hoping that it will improve
the robustness.

Regarding more comments I totally agree with you.

> >> The big problem was you did several things with a single patch,
> >> and that made the review much more difficult than it had to be.
> >>
> >> Having to check if you correctly modified the page tables, while also
> >> having to check for segmentation, and the interrupt descriptor
> >> transformations was distracting.
> >
> > Let me know which parts you think are good and we will implement and
> > review them bit by bit instead then.
>
> Skip the infrastructure changes, and the rest looks like real
> possibilities.

But I need to store my page tables somewhere, and there is no good
place to store them now. With good reasoning I can be convinced that
storing things on the control page is a good thing, and I'd like to
agree on something, but without good reasoning I will continue to
think that the control page solution is overly complex.

Thanks,

/ magnus

_______________________________________________
fastboot mailing list
[email protected]
https://lists.osdl.org/mailman/listinfo/fastboot

Reply via email to