Re: Make pipe data structure be a circular list of pages, ratherthan

[Linus Torvalds]

On Fri, 7 Jan 2005, Alan Cox wrote:
> 
> That would break stuff, but holding the last page until it fills 4K
> would work, or just basic sg coalescing when possible. The pipe
> behaviour - particularly size and size of atomic writes is defined by
> SuS and there are people who use pipes two ways between apps and use the
> guarantees to avoid deadlocks

Hmm.. Are there really any other guarantees than the atomicity guarantee 
for a PIPE_BUF transfer? I don't see any in SuS..

That said, existing practice obviously always trumps paper standards, and 
yes, coalescing is possible.

		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/