[GIT PULL] please pull infiniband.git

From: Roland Dreier
Date: Wed Jan 16 2008 - 17:46:46 EST


Linus, please pull from

master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband.git for-linus

This tree is also available from kernel.org mirrors at:

git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git for-linus

This will get the commit below, which moves a little code around to
make a couple of major MPI implementations work on ipath devices:

drivers/infiniband/hw/ipath/ipath_ud.c | 44 ++++++++++++++++----------------
1 files changed, 22 insertions(+), 22 deletions(-)

commit 0a69631b2869093d7306e8f66cca8eb0a05aa919
Author: Ralph Campbell <ralph.campbell@xxxxxxxxxx>
Date: Tue Jan 15 15:58:13 2008 -0800

IB/ipath: Fix receiving UD messages with immediate data

This fixes a small bug in ipath_ud_rcv()'s handling of UD messages
with immediate data. We need to test whether immediate data is
present and update the header size accordingly *before* testing the
packet size from the header against the actual received length.
Otherwise the wrong header size will be used and all messages with
immediate data will be dropped.

This bug keeps MVAPICH-UD and HP MPI from working at all on ipath devices.

Signed-off-by: Ralph Campbell <ralph.campbell@xxxxxxxxxx>
Signed-off-by: Roland Dreier <rolandd@xxxxxxxxx>
---
diff --git a/drivers/infiniband/hw/ipath/ipath_ud.c b/drivers/infiniband/hw/ipath/ipath_ud.c
index 16a2a93..b3df6f3 100644
--- a/drivers/infiniband/hw/ipath/ipath_ud.c
+++ b/drivers/infiniband/hw/ipath/ipath_ud.c
@@ -455,6 +455,28 @@ void ipath_ud_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr,
}
}

+ /*
+ * The opcode is in the low byte when its in network order
+ * (top byte when in host order).
+ */
+ opcode = be32_to_cpu(ohdr->bth[0]) >> 24;
+ if (qp->ibqp.qp_num > 1 &&
+ opcode == IB_OPCODE_UD_SEND_ONLY_WITH_IMMEDIATE) {
+ if (header_in_data) {
+ wc.imm_data = *(__be32 *) data;
+ data += sizeof(__be32);
+ } else
+ wc.imm_data = ohdr->u.ud.imm_data;
+ wc.wc_flags = IB_WC_WITH_IMM;
+ hdrsize += sizeof(u32);
+ } else if (opcode == IB_OPCODE_UD_SEND_ONLY) {
+ wc.imm_data = 0;
+ wc.wc_flags = 0;
+ } else {
+ dev->n_pkt_drops++;
+ goto bail;
+ }
+
/* Get the number of bytes the message was padded by. */
pad = (be32_to_cpu(ohdr->bth[0]) >> 20) & 3;
if (unlikely(tlen < (hdrsize + pad + 4))) {
@@ -482,28 +504,6 @@ void ipath_ud_rcv(struct ipath_ibdev *dev, struct ipath_ib_header *hdr,
wc.byte_len = tlen + sizeof(struct ib_grh);

/*
- * The opcode is in the low byte when its in network order
- * (top byte when in host order).
- */
- opcode = be32_to_cpu(ohdr->bth[0]) >> 24;
- if (qp->ibqp.qp_num > 1 &&
- opcode == IB_OPCODE_UD_SEND_ONLY_WITH_IMMEDIATE) {
- if (header_in_data) {
- wc.imm_data = *(__be32 *) data;
- data += sizeof(__be32);
- } else
- wc.imm_data = ohdr->u.ud.imm_data;
- wc.wc_flags = IB_WC_WITH_IMM;
- hdrsize += sizeof(u32);
- } else if (opcode == IB_OPCODE_UD_SEND_ONLY) {
- wc.imm_data = 0;
- wc.wc_flags = 0;
- } else {
- dev->n_pkt_drops++;
- goto bail;
- }
-
- /*
* Get the next work request entry to find where to put the data.
*/
if (qp->r_reuse_sge)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/