Subject: Re: Per-thread user <-> kernel communication area
To: Andrew Doran <firstname.lastname@example.org>
From: Bill Stouder-Studenmund <email@example.com>
Date: 11/08/2007 16:33:08
Content-Type: text/plain; charset=us-ascii
On Thu, Nov 08, 2007 at 10:41:37PM +0000, Andrew Doran wrote:
> Hi all,
> The below diff provides a wired-down, per-LWP area that is shared between=
> user process and the kernel. It's not implemented as a commpage, because I
> want to be able to send information in both directions, meaning the user
> part has to be mapped read/write. Also, one advantage to having the kernel
> allocate single blocks on behalf of the user process is that we can add to
> the block later without backward compatibility concerns.
1) kern_lpw.c seems to be losing a lot of DPRINTF() stuff. I think that=20
should be a separate change.
2) What thoughts do you have for versioning this stuff? Specifically, I=20
can see a utility in eventually having the app communicate some things to=
the kernel via this page. How do we avoid having either side end up=20
waiting for something the other doesn't know to do?
A sub-comment of the above, once we've added more stuff, how would we=20
handle obsoleting stuff? Say we add something then, after having actually=
tried it, realize that it's unworkable. How do we get rid of it?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (NetBSD)
-----END PGP SIGNATURE-----