RE: [patch v2 02/37] add opcodes to ib_pack.h

Originally I had xrc support in rxe but there was nothing in Roland's tree
to support it so I stripped it out.
I was experimenting with congestion notification to see if there was a way
to slow down spewing by fast Ethernet devices but that wasn't working, also
all the MAD support was stripped out. In the end the only thing left was the
constants and entries in the opcode table in rxe_opcode.c which as you say
are not used elsewhere. I think it is reasonable to have a complete set of
opcodes somewhere.

I want to eventually complete the IB transport parts of the driver although
they are not used by RoCE. They would be useful for a soft IB implementation
with say an FPGA based MAC/PHY. The long term goal is to have a complete
reference implementation of the IB Architecture.

-----Original Message-----
From: bart.vanassche-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org [mailto:bart.vanassche-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org] On
Behalf Of Bart Van Assche
Sent: Monday, August 15, 2011 4:13 AM
To: rpearson-klaOcWyJdxkshyMvu7JE4pqQE7yCjDx5@public.gmane.org
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: Re: [patch v2 02/37] add opcodes to ib_pack.h

On Sun, Jul 24, 2011 at 9:43 PM, [off-list ref] wrote:

quoted

Bring up to date with the current version of the IBTA spec.
       - add new opcodes for RC and RD
       - add new groups of opcodes for CN and XRC

Signed-off-by: Bob Pearson <redacted>

---
 include/rdma/ib_pack.h |   39

++++++++++++++++++++++++++++++++++++++-

quoted

 1 file changed, 38 insertions(+), 1 deletion(-)

Index: infiniband/include/rdma/ib_pack.h

==========================================================
=========

quoted

--- infiniband.orig/include/rdma/ib_pack.h
+++ infiniband/include/rdma/ib_pack.h

@@ -75,6 +75,8 @@ enum {

       IB_OPCODE_UC                                = 0x20,
       IB_OPCODE_RD                                = 0x40,
       IB_OPCODE_UD                                = 0x60,
+       IB_OPCODE_CN                                = 0x80,
+       IB_OPCODE_XRC                               = 0xA0,

A convention when submitting Linux kernel patches is to add only those
constants that are actually used in the code submitted in the same
patch series. So I'm not sure it's fine to add the XRC constants in
ib_pack.h through the rxe patch series.

Bart.

--
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

On Mon, Aug 15, 2011 at 6:15 PM, Bob Pearson
[off-list ref] wrote:

I want to eventually complete the IB transport parts of the driver although
they are not used by RoCE. They would be useful for a soft IB implementation
with say an FPGA based MAC/PHY. The long term goal is to have a complete
reference implementation of the IB Architecture.

Having a complete reference implementation of the IB architecture
available would be great. IMHO at least the following should be added
to ib_rxe (not necessarily in the first version) before it can called
a reference implementation:
- Support for secondary GIDs. These make it easy to implement
fail-over in a H.A. cluster.
- Support for multicast.

Bart.
--
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