Re: [PATCH rdma-core 4/5] verbs: Add an option to provide vendor private data when creating a CQ

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

On Tue, Mar 14, 2017 at 05:14:55PM +0200, Yishai Hadas wrote:
> It's not an issue of 'basic' and 'complex' API but a question of code
> sharing and resource managing.

This can always be acheived with some code restructuring.

eg we can put the tail of __lib_ibv_create_cq_ex into a helper, or
have an internal create_cq that accepts the void *provider_data

There is a lot less code inside libibverbs that outside, I'd rather
have an obvious clean user API than make our lives a tiny bit easier.

> >>Once the correct structure is used there should be *no* safety
> >>issue, please note that the vendor data defined to be as some
> >>extended structure which can only be grown based on comp_mask so
> >>that compatibility will be saved in any case.
> >
> >No, you still have to make sure that your ib_device is somehow a mlx5.
> 
> We are talking on an application that previously read some vendor
> capabilities using a direct API (e.g. mlx5dv_query_device) to know whether
> the vendor functionality is supported.

With these patches if the app passes the wrong thing into the
vendor_data it crashes, with a mlx5_create_cq it would get an error
code. Error codes are better.

Yishai, even if we go this way, stop using the word 'vendor'. It is
not vendor data.

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