Re: [PATCH 2/4] IB/uverbs: Support for XRC

[Roland Dreier <rdreier-FYB4Gu1CFyUAvxtiuMwx3w@public.gmane.org>]

 > > -	IB_USER_VERBS_CMD_POST_SRQ_RECV
 > > +	IB_USER_VERBS_CMD_POST_SRQ_RECV,
 > > +	IB_USER_VERBS_CMD_CREATE_XRC_SRQ,
 > > +	IB_USER_VERBS_CMD_OPEN_XRCD,
 > > +	IB_USER_VERBS_CMD_CLOSE_XRCD
 > 
 > Jack,Roland I believe that adding entries to this enum means ABI change towards libibverbs or eventually cause ABI change between libibverbs and its consumers, (e.g change in size/content of the verbs context structure, etc), correct? 

This is a backwards-compatible ABI extension -- we are adding new
commands.  And I believe OFED has been shipping this ABI for quite a
while now (since 1.4?).

 > I'd like to use this ABI change and introduce/expose also the ib_modify_cq verb to user space. Basically, the related patches should be quite simple (uverbs, libibverbs, libmlx4), still, I'll be glad if you tell/coach me against what code base to do that. 

OK, but we should coordinate this with all the other ABI extensions that
OFED has already made.

 > One more thing I'd like to add to libibverbs and some device driver libraries is send-with-invalidate, but this is only user space patch set without any kernel/user, lib/app ABI change, correct?

It is a library ABI change of course to add the send with invalidate
opcode.  But that is OK.

 - R.
--
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