On Thu, Jan 20, 2000 at 10:04:16AM -0500, Zhihui Zhang wrote:
> Point 2 seems to be saying that we would rather sacrifice some performance
> to gain a cleaner interface (people are talking about eliminating kernel
> copying for a long time). Consider the physical I/O on a raw device, where
> we map the user data again in the KVA without copying the data. Why do we
> do this double mapping, when we can access the user data directly?
>
Direct I/O to user space should be treated as an optimization. Such I/O
requires wiring down all the user pages before I/O can happen. Hence
it requires special previleges.
Why does it get mapped to KVA ? Because of point 1.
-Arun
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message
