[ewg] [ANNOUNCE] OFED-RDMAoE-1.5-rc2 release is available

Eli Cohen eli at dev.mellanox.co.il
Thu Nov 5 06:52:35 PST 2009


OFED-RDMAoE-1.5-rc2 is available

The tarball is available on:
http://www.openfabrics.org/downloads/OFED/ofed-rdmaoe-1.5/OFED-RDMAoE-1.5-rc2.tgz


To get BUILD_ID run ofed_info

Please report any issues in bugzilla https://bugs.openfabrics.org/ for
OFED 1.5, component RDMAoE


Release information:
--------------------
Linux Operating Systems:
       - RedHat EL4 up6       2.6.9-67.ELsmp
       - RedHat EL4 up7       2.6.9-78.ELsmp
       - RedHat EL4 up8       2.6.9-89.ELsmp
       - RedHat EL5 up2       2.6.18-92.el5
       - RedHat EL5 up3       2.6.18-128.el5
       - RedHat EL5 up4       2.6.18-164.el5
       - SLES10 SP2           2.6.16.60-0.21-smp
       - SLES10 SP3           2.6.16.60-0.54-smp
       - SLES11               2.6.27.19-5-default
       - OEL 4 up7            2.6.9-78.ELsmp
       - OEL 5 up2            2.6.18-92.el5
       - CentOS5.2            2.6.18-92.el5
       - CentOS5.3            2.6.18-128.el5
       - Fedora Cor           2.6.29              *
       - OpenSuSE 11          2.6.25.5-1.1        *
       - kernel.org           2.6.29 and 2.6.30


     * Minimal QA for these versions

Systems:
     * x86_64
     * x86
     * ia64
     * ppc64

Main changes from 1.5 rc1:
============================
1. Updated packages:
     libibverbs-1.1.2-0.2.g81705cb.tar.gz
     libnes-0.9.0.tar.gz
     libibcm-1.0.5.tar.gz
     librdmacm-1.0.10.tar.gz
     perftest-1.2.2-0.6.g3a2b795.tar.gz
     dapl-2.0.24.tar.gz
     rds-tools-1.5-1.src.rpm

     libibumad-1.3.2_20091029_f342c62.tar.gz
     libibmad-1.3.2_20091029_f342c62.tar.gz
     opensm-3.3.2_20091029_f342c62.tar.gz
     infiniband-diags-1.5.2_20091029_f342c62.tar.gz
     libibverbs-1.1.2-0.7.gaa7b9dd.tar.gz
     libibcm-1.0.5.tar.gz 
     librdmacm-1.0.10.tar.gz
     perftest-1.2.2-0.6.g3a2b795.tar.gz
     dapl-2.0.24.tar.gz
     rds-tools/rds-tools-1.5-1.src.rpm

     - Management packages:
     libibumad-1.3.2_20091029_f342c62.tar.gz
     libibmad-1.3.2_20091029_f342c62.tar.gz
     opensm-3.3.2_20091029_f342c62.tar.gz
     infiniband-diags-1.5.2_20091029_f342c62.tar.gz

     - MPI packages:
     mvapich2-r3510-2.src.rpm
     openmpi-1.3.3-1.src.rpm

2. Bug fixes (attached log file)


Limitations:
============
- SLES10 SP3 on IA64 is not supported yet


Notes:
RDMAoE and mlx4_en capabilities are supported only for x86 and x86_64
architectures.

-------------- next part --------------
commit c35c3bd7b962d3477a28bcec030f28657231d536
Merge: b4ff107 29501ed
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Wed Nov 4 14:37:50 2009 +0200

    Merge branch 'ofed_kernel_1_5' into rdmaoe

commit 29501ed6c478f3e2fd790d092fc00f0a877fc72c
Author: Vladimir Sokolovsky <vlad at mellanox.co.il>
Date:   Wed Nov 4 10:23:47 2009 +0200

    backport/SLES10SP3: Remove redefined variable
    
    Signed-off-by: Vladimir Sokolovsky <vlad at mellanox.co.il>

commit 28ec802a3027d4f03320cf13468a8c933857fc2d
Merge: 90d0131 5255f50
Author: Vladimir Sokolovsky <vlad at mellanox.co.il>
Date:   Tue Nov 3 09:14:25 2009 +0200

    Merge branch 'ofed_kernel_1_5' of ssh://vlad@sofa.openfabrics.org/home/ctung/scm/ofed-1.5 into ofed_kernel_1_5

commit 5255f50880080e3eafbbfeaec5e2b1f64d79816c
Author: Chien Tung <chien.tin.tung at intel.com>
Date:   Mon Nov 2 15:50:19 2009 -0600

    RDMA/nes: close MPA rejected connections
    
    In OpenMPI testing with dynamic connection mode, MPA rejected
    connections are left open thus hanging tests.  Update
    nes_0013_xansation.patch to close MPA rejected connections.
    
    Signed-off-by: Chien Tung <chien.tin.tung at intel.com>

commit b4ff107087aeafeb0e70f58d0e69de2250b5e20c
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Mon Nov 2 14:40:48 2009 +0200

    rdmaoe: fix bug in mlx4_ib_mcg_detach
    
    This bug is related to the one fixed in 87babd9. When dettaching a QP, qp->port
    may still be zero. This patches add a port field to struct gid_entry and uses
    this value if it is available.
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 020ea1876450e06745780af69297cf66c0e95735
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Mon Nov 2 11:42:31 2009 +0200

    mlx4_core: bug fix in trigger_port
    
    Used wrong prototype and relied on non existent variable, attr, causing kernel
    oops.
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 90d0131d56b074cb019dd84f5b8bba83bdee488a
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Mon Nov 2 11:42:31 2009 +0200

    mlx4_core: bug fix in trigger_port
    
    Used wrong prototype and relied on non existent variable, attr, causing kernel
    oops.
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 27547696935c3450c71250b856d05f6b8bdc597a
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Thu Oct 29 16:34:31 2009 +0200

    From 947a02a6076d5f434468d1a5bea446ef3dfbf566 Mon Sep 17 00:00:00 2001
    From: Eli Cohen <eli at mellanox.co.il>
    Date: Thu, 29 Oct 2009 15:31:43 +0200
    Subject: [PATCH] mlx4: Bug fix in mlx4_ib_resolve_grh
    
    The function copied into struct in6_addr a size of struct sockaddr_in6 causing
    overflow and crashes seen on x86 arch.
    
    Note: The fix appears in the patch file mlx4_2030_fix_port_num.patch which
    somehow does not describe the real fix.
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 24e0fbf7c7e22936274022d1aa88391d3c74c1cf
Merge: 87babd9 c124815
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Thu Oct 29 16:10:51 2009 +0200

    Merge branch 'ofed_kernel_1_5' into rdmaoe

commit c12481586c4ba09cb88dc2090c67fdce7c856cde
Author: Jason Gunthorpe <jgunthorpe at obsidianresearch.com>
Date:   Tue Oct 27 18:33:15 2009 +0200

    ipoib/mcast: Fix IPoIB multicast backport
    
    Adjust the dmi_addr directly so ip maddr reports the correct address.
    
    Signed-off-by: Jason Gunthorpe <jgunthorpe at obsidianresearch.com>
    Signed-off-by: Vladimir Sokolovsky <vlad at mellanox.co.il>

commit 87babd9fa0272f16375d50dffdeed15408ff433e
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Thu Oct 29 10:24:23 2009 +0200

    From 360f4f7e93a35200173c0c584ac542f2fbcba617 Mon Sep 17 00:00:00 2001
    From: Eli Cohen <eli at mellanox.co.il>
    Date: Wed, 28 Oct 2009 18:52:46 +0200
    Subject: [PATCH] mlx4: Fix multicast handling
    
    Attaching a QP to a multicast group does not require the QP to be in a state
    different then INIT - this is fine for IB. For RDMAoE however, we need the port
    assigned to the QP in order to call dev_mc_add() for the correct netdevice,
    while port is assgined when moving from INIT to RTR. Hence, we must keep track
    of all the multicast groups attached to a QP and call dev_mc_add() when the
    port becomes available.
    
    Fixes bug 1802.
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 03a761638796dbf4e901a98e6cf31b748191b373
Merge: ab4efd8 282cca8
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Thu Oct 29 10:22:28 2009 +0200

    Merge branch 'ofed_kernel_1_5' into rdmaoe

commit 8562101576249e286431ca4541587bc38119d547
Author: Andy Grover <andy.grover at oracle.com>
Date:   Wed Oct 28 17:05:50 2009 -0700

    backport: fix RDS for rhel4
    
    Signed-off-by: Andy Grover <andy.grover at oracle.com>

commit 282cca8b2478c7f75dbe17557eddc92a3520d956
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Wed Oct 28 12:02:15 2009 +0200

    mlx4: Remove unnecessary backport
    
    Not needed after commit 96e36a81e9e8e5cfe5e1846d03a893682d8a4086
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 5aa4d1291fe7d123265ea1632c8b094927fbc59d
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Wed Oct 28 11:37:55 2009 +0200

    From fa0681d2129732027355d6b7083dd8932b9b799d Mon Sep 17 00:00:00 2001
    From: Roland Dreier <rolandd at cisco.com>
    Date: Sat, 5 Sep 2009 20:24:49 -0700
    Subject: [PATCH] mlx4_core: Allocate and map sufficient ICM memory for EQ context
    
    The current implementation allocates a single host page for EQ context
    memory, which was OK when we only allocated a few EQs.  However, since
    we now allocate an EQ for each CPU core, this patch removes the
    hard-coded limit (which we exceed with 4 KB pages and 128 byte EQ
    context entries with 32 CPUs) and uses the same ICM table code as all
    other context tables, which ends up simplifying the code quite a bit
    while fixing the problem.
    
    This problem was actually hit in practice on a dual-socket Nehalem box
    with 16 real hardware threads and sufficiently odd ACPI tables that it
    shows on boot
    
        SMP: Allowing 32 CPUs, 16 hotplug CPUs
    
    so num_possible_cpus() ends up 32, and mlx4 ends up creating 33 MSI-X
    interrupts and 33 EQs.  This mlx4 bug means that mlx4 can't even
    initialize at all on this quite mainstream system.
    
    Cc: <stable at kernel.org>
    Reported-by: Eli Cohen <eli at mellanox.co.il>
    Tested-by: Christoph Lameter <cl at linux-foundation.org>
    Signed-off-by: Roland Dreier <rolandd at cisco.com>

commit 58124dfcd1d49e31b25819b169417dbea9983a45
Merge: 73d7ecd c379a79
Author: Vladimir Sokolovsky <vlad at mellanox.co.il>
Date:   Wed Oct 28 11:53:54 2009 +0200

    Merge branch 'code-drop/20091027' of git://git.openfabrics.org/~agrover/ofed_1_5/linux-2.6 into ofed_kernel_1_5

commit 73d7ecda345818dc30adf11e8a449cd33c03ef8f
Author: Yevgeny Petrilin <yevgenyp at mellanox.co.il>
Date:   Wed Oct 28 11:31:21 2009 +0200

    mlx4_en: Adjusted moderation backports for RH4
    
    Signed-off-by: Yevgeny Petrilin <yevgenyp at mellanox.co.il>

commit c379a79068d5862562cecfee52decec3b1782521
Author: Andy Grover <andy.grover at oracle.com>
Date:   Wed Sep 23 15:52:10 2009 -0700

    RDS/IB+IW: Move recv processing to a tasklet
    
    Move receive processing from event handler to a tasklet.
    This should help prevent hangcheck timer from going off
    when RDS is under heavy load.
    
    Signed-off-by: Andy Grover <andy.grover at oracle.com>

commit ae20d69d894574a5f379bb38fd6e1de181edc765
Author: Andy Grover <andy.grover at oracle.com>
Date:   Thu Aug 20 14:43:05 2009 -0700

    RDS: Do not send congestion updates to loopback connections
    
    This issue was discovered by HP's Pradeep and fixed in OFED
    1.3, but not fixed in later versions, since the fix's implementation
    was not immediately applyable to the later code. This patch should
    do the trick for 1.4+ codebases.
    
    Signed-off-by: Andy Grover <andy.grover at oracle.com>

commit e4895af26d8cbf9a5a2749808c3895ec5bfe1966
Author: Andy Grover <andy.grover at oracle.com>
Date:   Fri Aug 14 16:34:48 2009 -0700

    RDS: Fix panic on unload
    
    Remove explicit destruction of passive connection when destroying
    active end of the connection. The passive end is also on the
    device's connection list, and will thus be cleaned up properly.
    Panic was caused by trying to clean it up twice.
    
    Signed-off-by: Andy Grover <andy.grover at oracle.com>

commit 2e95bdd703dffacb47529ed512be8bce7fa32ac8
Author: Andy Grover <andy.grover at oracle.com>
Date:   Thu Aug 13 13:30:36 2009 -0700

    RDS: Fix potential race around rds_i[bw]_allocation
    
    From Shin Hong:
    
    "At rds_ib_recv_refill_one(), it first executes atomic_read(&rds_ib_allocation)
    for if-condition checking,
    
    and then executes atomic_inc(&rds_ib_allocation) if the condition was
    not satisfied.
    
    However, if any other code which updates rds_ib_allocation executes
    between these two atomic operation executions,
    it seems that it may result race condition. (especially when
    rds_ib_allocation + 1 == rds_ib_sysctl_max_recv_allocation)"
    
    This patch fixes this by using atomic_inc_unless to eliminate the
    possibility of allocating more than rds_ib_sysctl_max_recv_allocation
    and then decrementing the count if the allocation fails. It also
    makes an identical change to the iwarp transport.
    
    Reported-by: Shin Hong <hongshin at gmail.com>
    Signed-off-by: Andy Grover <andy.grover at oracle.com>

commit 53f7b51485bcd6ba8c06f31c6fb2cc17adc2498a
Author: Yevgeny Petrilin <yevgenyp at mellanox.co.il>
Date:   Tue Oct 27 10:16:13 2009 +0200

    mlx4_core: Rename MSI-X vectors
    
    Signed-off-by: Yevgeny Petrilin <yevgenyp at mellanox.co.il>

commit b2e105af1656326e97784c40a6313de090c340b6
Author: Yevgeny Petrilin <yevgenyp at mellanox.co.il>
Date:   Mon Oct 26 09:36:10 2009 +0200

    mlx4: Added a trigger to change all port types
    
    If user wishes to modify both ports, set the trigger and then modify
    both ports.
    It prevents unneccessary registration iterations.
    
    Signed-off-by: Yevgeny Petrilin <yevgenyp at mellanox.co.il>

commit ab4efd82512a1f19bfa27bf2ac4b8dd96bc2759d
Merge: f6bf582 c6b041d
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Sun Oct 25 15:50:57 2009 +0200

    Merge branch 'ofed_kernel_1_5' into rdmaoe

commit f14435727677366b3b26de3a78122faf910fd577
Author: Yevgeny Petrilin <yevgenyp at mellanox.co.il>
Date:   Sun Oct 25 14:29:25 2009 +0200

    mlx4_en: Using smaller TX queue backports
    
    The backports are applied only to OS with no multiple
    TX queue support.
    
    Signed-off-by: Yevgeny Petrilin <yevgenyp at mellanox.co.il>

commit 2a81d31d45f7a6b7b4403714deb5774c09329f92
Author: Yevgeny Petrilin <yevgenyp at mellanox.co.il>
Date:   Tue Sep 29 11:45:45 2009 +0200

    connectx_port_config: Adjusted script to single port device
    
    Signed-off-by: Yevgeny Petrilin <yevgenyp at mellanox.co.il>

commit 2e6aec781dd6dbe5d8eeeff0a6c315115dfbfe59
Author: Yevgeny Petrilin <yevgenyp at mellanox.co.il>
Date:   Sun Oct 25 14:10:41 2009 +0200

    mlx4_en: Using less resources
    
    We can use less rings while maintaining and even improving
    the throughput.
    In the queue selecting scheme, the "small" packets are the ones
    that are being "moved" to different ring, to avoid large
    "out of order" packets.
    
    Signed-off-by: Yevgeny Petrilin <yevgenyp at mellanox.co.il>

commit 7b810d8b4abf1b719f729dbea6889d0aa992b33a
Author: Yevgeny Petrilin <yevgenyp at mellanox.co.il>
Date:   Sun Oct 25 14:05:09 2009 +0200

    mlx4_en: Use "Latency" mode for ping-pong with small messages
    
    When running in adaptive moderation mode, and in ping-pong mode
    with messages smaller then 256 bytes, enter to latency mode,
    regardless the packet rate.
    
    Signed-off-by: Yevgeny Petrilin <yevgenyp at mellanox.co.il>

commit c6b041d1629feb4b2258e766b0b2fce36abac052
Author: Vladimir Sokolovsky <vlad at mellanox.co.il>
Date:   Thu Oct 22 14:00:09 2009 +0200

    mlx4_en: Enable mlx4_en load on SLES11
    
    Add '--allow-unsupported-modules' to modprobe command on SLES11 to enable
    loading of mlx4_en module.
    
    Signed-off-by: Vladimir Sokolovsky <vlad at mellanox.co.il>

commit f6bf582e987e8f8c09c4986e093d2ed5afdc6688
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Thu Oct 22 10:48:49 2009 +0200

    Fix sles10 sp3 backports
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 449245a4ff0cc034a5dbbce54cfc9961cd7e333a
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Tue Oct 20 16:55:41 2009 +0200

    commit b3c5604d0148aacaac0ffcbe41827b9ec013fc31
    Author: Eli Cohen <eli at mellanox.co.il>
    Date:   Tue Oct 20 15:32:06 2009 +0200
    
        mlx4: fix wrong port used in mlx4_ib_resolve_grh
    
        mlx4_ib_resolve_grh needs to use the correct port depending on whether an
        address vector was provided or not. Failure to do so can cause access to an
        illgal entry in rdmaoe->netdevs and cause access violation.
    
        Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit c89ef56d8ad48d140523bccf203ec2cf6cb70ee1
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Sun Oct 18 11:17:16 2009 +0200

    RDMAOE: update backports for RHEL 4.8
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 9a89c8df3b383f04f5e27c634eebe449f2d403f8
Merge: 5c0ba3c 279476d
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Sun Oct 18 10:46:29 2009 +0200

    Merge branch 'ofed_kernel_1_5' into rdmaoe

commit 279476d4fbf3a794de60b37ba07ad7fe715e776c
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Sun Oct 18 09:46:18 2009 +0200

    From 75eca87d2a5302482af20a77a225ee528c98d365 Mon Sep 17 00:00:00 2001
    From: Eli Cohen <eli at mellanox.co.il>
    Date: Sun, 11 Oct 2009 17:34:16 +0200
    Subject: [PATCH] mlx4: remove unneeded code
    
    There is no such flag DE - the field is reserved and should be zero.
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>
    ---

commit 5c0ba3c8b2e86559576154babd9dfa44d88276be
Merge: 6b79907 72c4789
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Thu Oct 15 15:50:46 2009 +0200

    Merge branch 'ofed_kernel_1_5' into rdmaoe

commit 6b799070483feb58de43ae97e762c4f8304f25c1
Author: Eli Cohen <eli at mellanox.co.il>
Date:   Thu Oct 15 12:48:15 2009 +0200

    mlx4: Add explicit include to linux/notifier.h
    
    Some kernels do not have and the driver failed to build.
    
    Signed-off-by: Eli Cohen <eli at mellanox.co.il>

commit 72c47892e85aa4d8188db492b41a3796b71ffdc6
Author: Don Wood <donald.e.wood at intel.com>
Date:   Wed Oct 14 18:43:01 2009 -0500

    RDMA/nes: Flush queues for a QP that is still in idle state
    
    A QP that is created and destroyed without ever been in RTS needs
    to have its WQEs flushed.
    
    Signed-off-by: Don Wood <donald.e.wood at intel.com>

commit e497f1cfd95e8923451a065dadaca71d3d52d2c9
Author: Chien Tung <chien.tin.tung at intel.com>
Date:   Wed Oct 14 18:41:30 2009 -0500

    RDMA/nes: Implement IB_SIGNAL_ALL_WR as iWARP extension
    
    Add IB_SINGAL_ALL_WR support as an iWARP extension.  If set, make
    sure all WR for the QP is signalled.
    Consolidated flags used in nesqp structure.
    
    Signed-off-by: Chien Tung <chien.tin.tung at intel.com>

commit 727e073a8b5c3751a50e16c4aa4d213950a2ef26
Author: Chien Tung <chien.tin.tung at intel.com>
Date:   Wed Oct 14 18:40:21 2009 -0500

    RDMA/nes: remove unused artifacts
    
    Signed-off-by: Chien Tung <chien.tin.tung at intel.com>

commit a1591e74f35cb520e6e569465bdcb2d5de8dd7a5
Author: Chien Tung <chien.tin.tung at intel.com>
Date:   Wed Oct 14 18:39:13 2009 -0500

    RDMA/nes: Check AMCC PHY after init and reset if necessary
    
    Signed-off-by: Chien Tung <chien.tin.tung at intel.com>

commit 2e089f1875f90d8ffd662ccb85973818411e3851
Author: Chien Tung <chien.tin.tung at intel.com>
Date:   Wed Oct 14 18:38:13 2009 -0500

    RDMA/nes: add send invalidate and FMR
    
    Signed-off-by: Chien Tung <chien.tin.tung at intel.com>

commit 028202e44d718739bac5897fc761cfb9f9be4ca5
Author: Faisal Latif <faisal.latif at intel.com>
Date:   Wed Oct 14 18:35:48 2009 -0500

    RDMA/nes: fixes for reject and MPA handshake from Xansation testing
    
    Signed-off-by: Faisal Latif <faisal.latif at intel.com>

commit 61125cfc80c975dde59adf4930cab153d51c558d
Author: Jack Morgenstein <jackm at dev.mellanox.co.il>
Date:   Tue Oct 13 14:19:32 2009 +0200

    memtrack: adjust kmalloc to avoid spurious inconsistency msgs in kfree for earlier kernels.
    
    Problem was an inconsistency in kmalloc which tested for kernels later than 2.6.9
    to avoid storing 0-sized allocations.  However, in kfree, the condition was
    if ZERO_OR_NULL_PTR was defined (and this came in only at kernel 2.6.23).
    
    Thus, all kernels between 2.6.9 and 2.6.23 had a problem.
    
    Signed-off-by: Jack Morgenstein <jackm at dev.mellanox.co.il>


More information about the ewg mailing list