From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eli Cohen Subject: Re: [PATCHv10 08/12] mlx4: Add support for IBoE - address resolution Date: Fri, 22 Oct 2010 14:24:55 +0200 Message-ID: <20101022122455.GA8791@mtldesk30> References: <20100826141851.GI8795@mtldesk30> <20101021230307.GD6367@mtldesk30> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Roland Dreier Cc: RDMA list List-Id: linux-rdma@vger.kernel.org On Thu, Oct 21, 2010 at 09:15:35PM -0700, Roland Dreier wrote: > We don't need ICRC in the header_buf, do we? We're talking about > MLX4_IB_UD_HEADER_SIZE, which is used to hold the packet header we get > from ib_ud_header_pack(). The ICRC is accounted for in a different > segment of the WQE. > > In any case, it shouldn't matter -- I don't see how it could be correct > to increase the buffer size by 4 bytes to account for the change from > LRH to Ethernet header; an LRH is 8 bytes and an Ethernet header > (without .1q etc) is 14 bytes. So we should add 6 bytes, right? > > The math I get is > > IB: LRH + GRH + BTH + DETH + IMM > 8 40 12 8 4 = 72 > > Eth: MAC + GRH + BTH + DETH + IMM > 14 40 12 8 4 = 78 > > I don't see how your change > > > - MLX4_IB_UD_HEADER_SIZE = 72, > > + MLX4_IB_UD_HEADER_SIZE = 76, > > could possibly be correct, although as I said it probably works because > of padding and not hitting the worst case in practice anyway. > Sure, all clear. -- 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