Jordan: Most of this should be machine-independent; the libc malloc is
entirely portable, the implementations of the seg* system calls are too, as
is the current segment table. The size of pointers (and by consequence, if
there is a tree, the depth) would be machine-dependent; if I try to work on
page-table sharing, the page table structures/formats will be too.

I haven't read the Plan 9 source around this in quite some time, though; I
will take a look.

erik: You're absolutely right, I prematurely said 'tree structure'; without
knowing the distribution of segment requests. I said tree structure as every
other kernel I've looked at has some sort of tree structure for this
purpose.

I suspect that working on both the malloc and the VM structures will be
fine; the reason I want to work on this project is that I've started
replacing malloc a few times, but stopped each time since I got frustrated
by the lack of more anonymous segments. The malloc work would provide a
great first consumer for the new segment code, at least, and would satisfy
me more. I'll try to draw up a schedule and see how reasonable it looks...

Thanks,
-- vs

-- 
You received this message because you are subscribed to the Google Groups "Plan 
9 Google Summer of Code" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/plan9-gsoc?hl=en.

Reply via email to