Open Fabrics Enterprise Distribution (OFED)
- Version 1.3.1-rc1
+ Version 1.3.1-rc2
Release Notes
May 2008
- RedHat EL4 up6: 2.6.9-67.ELsmp
- RedHat EL5: 2.6.18-8.el5
- RedHat EL5 up1: 2.6.18-53.el5
+ - RedHat EL5 up2 beta: 2.6.18-84.el5 *
- Fedora C6: 2.6.18-8.fc6 *
- SLES10: 2.6.16.21-0.8-smp
- SLES10 SP1: 2.6.16.46-0.12-smp
- SLES10 SP1 up1: 2.6.16.53-0.16-smp
+ - SLES10 SP2: 2.6.16.60-0.21-smp *
- OpenSuSE 10.3: 2.6.22-*-* *
- kernel.org: 2.6.23 and 2.6.24
- ehca (IBM)
- mlx4 (Mellanox)
- nes (NetEffect)
+- Added backport to support SLES 10 SP2 and RHEL 5 up2 beta
Note: For details regarding the various fixed bugs, please see the
release notes for each component in the docs directory.
datagram mode.
4. For multicast traffic, IPoIB always uses datagram mode.
5. This version adds support for CM mode for HCAs which do not support Infiniband
- Shared Receive Qeueus (SRQ). The detection is done automatically at run time. In
+ Shared Receive Queues (SRQ). The detection is done automatically at run time. In
this case, the driver uses RC connections between the peers for transferring IPOIB
traffic.
Stateless offload features have been introduced:
checksum offload - generate on send and verify on receive of packet checksum
- Large Send Offload - support fragmenting a large quantity of data paseed form the
+ Large Send Offload - support fragmenting a large quantity of data passed form the
TCP/IP stack and fragment it to MSS sized packets.
7. Support for interrupt moderation was introduced. This is done for HCAs which
support this feature (like Mellanox's mlx4). Controlling moderation params is
To increase MTU, enter:
ifconfig ib0 mtu 65520
-5. Switching between CM and UD mode can be done in runtime:
+5. Switching between CM and UD mode can be done in run time:
echo datagram > /sys/class/net/ib0/mode sets the mode of ib0 to UD
echo connected > /sys/class/net/ib0/mode sets the mode ib0 to CM
6. On RedHat EL 4 up4, the IPOIB implementation is not spec-compliant:
- ipoib multicast does not work
- - ipoib cannot interoperate between RHEL4U4 and other hosts. This is due to
+ - ipoib cannot inter-operate between RHEL4U4 and other hosts. This is due to
missing code in the kernel which was available in U3 and U5 but removed in
U4. As a workaround, upgrade to RHEL4U5.
* The bondX name must not contain characters which are disallowed for bash
variable names such as '.' and '-'
* Using /etc/infiniband/openib.conf to create a persistent configuration is
- depracated and not recommended. Use it unless you have no other option and
+ deprecated and not recommended. Use it unless you have no other option and
prefer the standard way. It is not guaranteed that the first method will be
supported in future versions of OFED.
7. Bug Fixes and Enhancements Since OFED 1.3
===============================================================================
- Don't drop multicast sends when they can be queued
-- Bug fix when copying small SKBs
+- IPoIB panics with RHEL5U1, RHEL4U6 and RHEL4U5: Bug fix when copying small
+ SKBs (bug 989)
+- IPoIB failed on stress testing (bug 1004)
+- Kernel Oops during "port up/down test" (bug 1040)
+- Restart the stack during iperf 2.0.4 ver2.0.4 in client side cause to kernel
+ panic (bug 985)
Open Fabrics (www.openfabrics.org) software stack.
NEW SINCE OFED 1.3
- OFED 1.3.1 includes new versions of uDAPL v1 (1.2.6) and v2 (2.0.8)
+ OFED 1.3.1 includes new versions of uDAPL v1 (1.2.7-1) and v2 (2.0.9-1)
Summary of changes since OFED 1.3 release:
v2 - add private data exchange with reject
v1,v2 - support for zero byte operations, iov==NULL
v1,v2 - multi-transport support for inline data and private data differences
v1,v2 - fix memory leaks and other reported bugs since OFED 1.3
+ v1,v2 - dtest,dtestx,dapltest build issues on RHEL5.1
+ v1,v2 - long delay during dat_ia_open when DNS not configured
+ v1,v2 - use rdma_read_in/out from ep_attr per consumer instead of HCA max
NEW SINCE OFED 1.2