IB/ehca: Prevent RDMA-related connection failures on some eHCA2 hardware

Browse source

Some HW revisions of eHCA2 may cause an RC connection to break if they
received RDMA Reads over that connection before.  This can be
prevented by assuring that, after the first RDMA Read, the QP receives
a new RDMA Read every few million link packets.

Include code into the driver that inserts an empty (size 0) RDMA Read
into the message stream every now and then if the consumer doesn't
post them frequently enough.

Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

...

This commit is contained in:

Joachim Fenkes 2008-01-17 15:07:24 +01:00 • committed by Roland Dreier

parent bbdd267ef2

commit 2ec8e66241

3 changed files with 95 additions and 36 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

5

drivers/infiniband/hw/ehca/ehca_classes.h

View file

14

drivers/infiniband/hw/ehca/ehca_qp.c

View file

112

drivers/infiniband/hw/ehca/ehca_reqs.c

View file