From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Weiny, Ira" Subject: RE: [RFC PATCH 03/16] ib/mad: Add check for jumbo MADs support on a device Date: Mon, 8 Dec 2014 00:23:48 +0000 Message-ID: <2807E5FD2F6FDA4886F6618EAC48510E0CBD4F23@CRSMSX101.amr.corp.intel.com> References: <1415908465-24392-1-git-send-email-ira.weiny@intel.com> <1415908465-24392-4-git-send-email-ira.weiny@intel.com> <54770F44.2090909@mellanox.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 8BIT Return-path: In-Reply-To: <54770F44.2090909-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> Content-Language: en-US Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: 'Or Gerlitz' , "roland-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org" Cc: "linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" List-Id: linux-rdma@vger.kernel.org > > On 11/13/2014 9:54 PM, ira.weiny-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org wrote: > > > Add a device capability flag for OPA devices to signal their support of "jumbo" > > MADs. > > > > Check for IB_DEVICE_JUMBO_MAD_SUPPORT in the device capabilities and > > if supported mark the special QPs created. > > Ira, > > Few comments here, the device capability is OK, specifically if it helps for the > mad layer logic to be simpler and/or more robust. > > You should add device attribute telling what is the size of MADs supported by > the device, I think the IBTA mandated 256Band in the OPA case, the driver will > fill there 2k. I don't think this is a bad idea, however, the complication is determining the kmem_cache object size in that case. How about we limit this value to < 2K until such time as there is a need for larger support? > > You can't just state that (jumbo == 2k) and go to complete design/changes > which use this hard-coded assumption. You need to see how to re-spin this > series w.o this assumption. I'm looking into it. I believe I will still need to have a capability bit, but it may end up having a different meaning. > > I find it very annoying that upper level drivers replicate in different ways > elements from the IB device attributes returned by ib_query_device. I met that > in multiple drivers and upcoming designs for which I do code review. Are you > up to come up with a patch that caches the device attributes on the device > structure? I don't follow what you are asking for. Could you give more details? > if not, > I can do that.. and have your code to see it. > > Also, I would go and unify (squash) this patch and the preceding one. > I'll keep this in mind as I rework the patches. Ira -- 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