Re: [PATCH 4/4] copyfile: copyfile

[Linus Torvalds]

On Tue, 7 Sep 2004, Jörn Engel wrote:
> 
> > Actually, the 4kB batching one was - if you only max out to using 4kB at a 
> > time, sendfile() is kind of pointless, because then it will never do 
> > multi-page copies in the first place, and all the complexity at a lower 
> > level is worthless..
> 
> Give me a better number.  16k?  1M?  Or would it not be fundamentally
> flawed if the unit was seconds, instead of bytes?  That makes a lot
> more sense, since a floppy and a Ultra320 RAID array differ slightly
> in speed and it's response time the users actually care about.

Well, you can't do it by seconds. What you _can_ do is to just make the 
fundamental page-cache sendfile thing check for interruptible, and just do 
it at a lower level. Maybe.

		Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/