[ofa-general] Re: [PATCH for-2.6.22] ipoib/cm: fix interoperability when mtu don't match

[Roland Dreier]

OK, I applied this for 2.6.22 since it looks quite safe (I even took
the risk of replacing the "4" in the warning string with a "%d" and
printing IPOIB_ENCAP_LEN, because it seemed funny to test against a
named constant and then print a raw number).  But I'm really going to
be disappointed if this breaks something...

BTW, any objection to merging the patch below for 2.6.22 too?  It's
compile-tested only but it looks *REALLY* safe.

commit f667e4b9c4d7b2772105d2872becffbe9e65ecea
Author: Roland Dreier <rolandd at cisco.com>
Date:   Thu Jun 21 13:37:05 2007 -0700

    IPoIB/cm: Remove dead definition of struct ipoib_cm_id
    
    It's completely unused.
    
    Signed-off-by: Roland Dreier <rolandd at cisco.com>

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
index 1fe7f66..b3d0a31 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
@@ -56,13 +56,6 @@ MODULE_PARM_DESC(cm_data_debug_level,
 #define IPOIB_CM_RX_DELAY       (3 * 256 * HZ)
 #define IPOIB_CM_RX_UPDATE_MASK (0x3)
 
-struct ipoib_cm_id {
-	struct ib_cm_id *id;
-	int flags;
-	u32 remote_qpn;
-	u32 remote_mtu;
-};
-
 static struct ib_qp_attr ipoib_cm_err_attr = {
 	.qp_state = IB_QPS_ERR
 };