[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Review of Thomas's >2GB ext2fs proposal

From: Thomas Bushnell BSG
Subject: Re: Review of Thomas's >2GB ext2fs proposal
Date: 16 Aug 2004 11:06:32 -0700
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3

So I think that this review is mostly accurate.

My real questions are:

1) Why muck with the well-understood memory object model?

2) Mach won't use more memory under my proposal; you have simply moved
   around where the memory is.  Instead of taking more memory to hold
   a table of physical->virtual address mappings in the memory object,
   you have a table of virtual->backing_store mappings in the pager;
   but it is the same data and it has to be stored either way.

3) I REALLY REALLY prefer keeping the memory object interface clear;
   experience is that this is a huge source of constant bugs.  In
   addition, adopting your strategy is way tougher to debug, because
   it's a big kernel change, rather than a simple user-space one.

4) And finally, what about data caching--which is vastly more
   important than mapping caching?  My version has it that data is
   cached as long as the kernel wants to keep it around, and in a
   fashion decoupled from mappings.

But the moral of Neil's analysis is basically a good one: his proposal
and mine do the same tasks, but in different places.  My proposal is
better because it does these tasks in user space.  


reply via email to

[Prev in Thread] Current Thread [Next in Thread]