Re: [PATCH v1 1/3] IB/core: Add support for extended query device caps

[Jason Gunthorpe <jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>]

On Thu, Feb 19, 2015 at 02:30:29PM +0200, Matan Barak wrote:

> I think we should all agree about the extension verbs mechanism before
> taking it further.
> IMHO, using comp_mask only for fields that their valid value != 0 is confusing.
> For example, an old kernel should support a command if:
> (a) it knows all its valid bits
> (b) all fields after the size it knows is 0
> 
> In the old schema, the kernel only needs to look at the comp_mask
> bits and execute the command or reject it.

It doesn't really change anything, the kernel still has to go field by
field to check, all that happens is

if ((comp_mask & BIT) && offsetof(field..) < size)
  // field is valid

becomes

if (offsetof(field..) < size && field != 0)
  // field is valid

> Furthermore, ibv_create_flow and ibv_destroy_flow were already
> accepted using the old schema.

IIRC they didn't have values that could meaningfully be 0?

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html