On Fri, Jan 26, 2001 at 09:32:14PM +0100, Jaromír Dolecek wrote:

 > Hi,
 > while looking for something else, I also noticed this comment
 > in uvm_vnode.c:
 > 
 > /*
 >  * NOTE: currently we have to use VOP_READ/VOP_WRITE because they go
 >  * through the buffer cache and allow I/O in any size.  These VOPs use
 >  * synchronous i/o.  [vs. VOP_STRATEGY which can be async, but doesn't
 >  * go through the buffer cache or allow I/O sizes larger than a
 >  * block].  we will eventually want to change this.
 >  *
 > 
 > Does it still hold with UBC or the comment should be updated ?

Err, I'm pretty sure that comment is out of date, yes.

-- 
        -- Jason R. Thorpe <thorpej@zembu.com>