On Thu, 2012-03-15 at 08:03 +0000, Jan Beulich wrote:

quoted

quoted

quoted

On 14.03.12 at 18:01, Justin Gibbs [off-list ref] wrote:

While we're talking about fixing ring data structures, can RING_IDX
be defined as a "uint32_t" instead of "unsigned int".  The structure
padding in the ring macros assumes RING_IDX is exactly 4 bytes,
so this should be made explicit.  ILP64 machines may still be a way
out, but the use of non-fixed sized types in places where size really
matters just isn't clean.

Yes, if we're going to rev the interface, then any such flaws should be
corrected.

There has been talk of doing something similar for netif too. IIRC the
netchannel2 work included a new generic ring scheme with support for
variable sized req/rsp elements and such.

If we are going to rev the rings then should we try and use a common
ring mechanism? I think so. If so then we could do worse than to start
from the netchannel2 ring stuff and/or concepts?

Looks like that is
http://xenbits.xen.org/ext/netchannel2/linux-2.6.18/log/075f6677a290/include/xen/interface/io/uring.h
still a bit nc2 specific though.

Ian.