Re: Patch: add directional argument to socket upcall handler

[christos%astron.com@localhost (Christos Zoulas)]

In article <20090825153236.GA37257%coyotepoint.com@localhost>,
Thor Simon  <tls%coyotepoint.com@localhost> wrote:
>-=-=-=-=-=-
>
>The attached patch adds an argument to socket upcall handlers, so
>they can tell whether they've been called because the socket is
>ready for input or output.  This significantly eases the way for
>in-kernel code that needs to interact with socket flow control,
>such as splice() or things of its ilk.  I think it might also make
>it possible to improve the wakeup mechanism in the NFS code, which
>seems to care whether the socket's come ready for read or write
>when adjusting various timers and handling errors.
>
>Hastily ported forward from netbsd-5 to -current -- compiles, and
>appears to work (works fine in our netbsd-5 tree).
>
>Comments?
>
>-- 
>Thor Lancelot Simon
>Coyote Point Systems, Inc.                     <tls%coyotepoint.com@localhost>
>Millerton, NY, USA

Looks fine to me.

christos