> > Agree with all your points above , except for point# 2. That IMHO would be > more of a code refactoring/cleanup patch > that can be taken up later once the baseline version is checked in as it's > not really related to this patch ? > Also, if we do that , then we also have to take care of the IB case as well > where GID is something that is obtained from the port/card(query_gid) > as opposed to something that is pushed down from the stack as you are now > proposing for IP-based GIDs.. i.e implications are much more than > what this patch intends to address. > So I'd rather do this at a later point in time and not club it with V2, sound > good ? > > Thanks > Som
Hi Som I think that it can be done in 2 stages. Moving GID table management to IB/core is indeed not mandatory for RoCEv2. However, you might want to consider it anyway since RoCEv2 support is going to add changes to GID table management anyhow. So, instead of modifying each device driver sperately, do it once centrally and add a plugin function in struct ib_device (e.g. ib_write_gid(union ib_gid, struct ib_gid_info *info)). We can work in parallel and implement the mlx4 function while you concentrate in IB/core and ocrdma. I leave it for you to decide how to proceed from here. -- 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
