On 12/23/2013 8:21 PM, Jason Gunthorpe wrote:
On Mon, Dec 23, 2013 at 09:35:46AM -0800, Roland Dreier wrote:
On Thu, Aug 1, 2013 at 8:05 AM, Yishai Hadas <[email protected]> wrote:
#ifdef HAVE_IBV_REGISTER_DRIVER
static __attribute__((constructor)) void mlx4_register_driver(void)
{
- ibv_register_driver("mlx4", mlx4_driver_init);
+ verbs_register_driver("mlx4", mlx4_driver_init);
}
#else
So just to be clear, after this change libmlx4 will no longer build or
work with libibverbs <= 1.1.7 (1.1.8 is unreleased but is the one with
the verbs extension changes)?
Right, but is it more than just that hunk..
If so does it make sense to have this autoconf stuff
(HAVE_IBV_REGISTER_DRIVER) for compatibility with libibverbs 1.0.x?
None, it should be dropped, and the kind thing would be to have
configure bail if verbs_register isn't found.
How hard would it be to add an autoconf check for
verbs_register_driver() and keep compatibility with libibverbs 1.1.7?
Not sure that is useful. I can't think of any reason someone would
actually want to do that, and doing the autoconf test sort of implies
that someone actually tested it, it works, and it is a sane thing to
do..
Jason
Right, the idea was that libibverbs with verbs extension will
support both types of providers, w/w.o verbs extension but provider with
verbs extension
(e.g. libmlx4) will require the extended libibverbs.
Re libibverbs, I have reviewed your GIT with the extensions - seems
to be fine.
Roland - please move forward and take libmlx4 as well, there are
some distros who are waiting for.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
