Olga Shern (Voltaire) wrote:
- 27 bugs fixed (see attached for details)
Hi Vlad,
I don't see the attached file.
Olga
Hi Olga,
Here are the missing files:
The list of the fixed bugs and ofed kernel changes between rc2 and rc3.
Thanks,
Regards,
Vladimir
bug_id,bug_severity,priority,op_sys,assigned_to,bug_status,resolution,short_short_desc
1244,blocker,P3,All,[EMAIL PROTECTED],RESOLVED,FIXED,XRC finalization code is broken
1278,blocker,P1,RHEL 5,[EMAIL PROTECTED],RESOLVED,FIXED,Mvapich2 fails to compile when uDAPL option is specified
1117,critical,P2,SLES 10,[EMAIL PROTECTED],RESOLVED,FIXED,ib_ipath module hangs on unload
1186,critical,P2,RHEL 5,[EMAIL PROTECTED],RESOLVED,FIXED,lockup in IPoIB during module unload
1188,critical,P3,RHEL 5,[EMAIL PROTECTED],RESOLVED,FIXED,kernel panic in neigh_destroy during ib_ipoib unload
1191,critical,P3,SLES 10,[EMAIL PROTECTED],RESOLVED,FIXED,"Failed to load ib_ipoib on sles10 sp2, ia64"
1192,critical,P3,RHEL 4,[EMAIL PROTECTED],RESOLVED,FIXED,"Failed to comple ofa_kernel RPM on RH4 up7, ia64 arch'"
1194,critical,P3,RHEL 5,[EMAIL PROTECTED],CLOSED,FIXED,ipoib_ud_test causes Kernel panic over ofed 1.4 caused by deadlock
1196,critical,P3,SLES 10,[EMAIL PROTECTED],RESOLVED,FIXED,rds module is in use right after installation and thus openibd stop is not responds on sles10sp2
1199,critical,P1,RHEL 5,[EMAIL PROTECTED],RESOLVED,FIXED,tvflash and tgt-generic fail to build if --prefix is used
1250,critical,P3,RHEL 4,[EMAIL PROTECTED],RESOLVED,FIXED,Bonding: ib1 not become the active slave after ib0 down on RH4 up6
1277,critical,P1,Other,[EMAIL PROTECTED],RESOLVED,FIXED,OpenSM: IPv6 SNM consolidation is broken
1236,major,P3,RHEL 4,[EMAIL PROTECTED],RESOLVED,FIXED,ib-bonding fails to install on RHEL4.4
1247,major,P3,RHEL 4,[EMAIL PROTECTED],RESOLVED,FIXED,ipoib_ud_test caused kernel oops on ofed_1_4 (sw083/084)
1160,normal,P3,Other,[EMAIL PROTECTED],RESOLVED,FIXED,mvapich2 build on ppc64 fails (in rh5up2 and sles10sp2)
1178,normal,P3,Other,[EMAIL PROTECTED],RESOLVED,FIXED,OFED-1.4-rc1 - mvapich-1.1.0-2960.src.rpm: rpmbuild fails on Fedora Core 9
1179,normal,P3,SLES 10,[EMAIL PROTECTED],RESOLVED,FIXED,OFED-1.4-rc1 compilation failed on SLES10SP2 PPC64 platform
1203,normal,P3,SLES 10,[EMAIL PROTECTED],VERIFIED,FIXED,RDS crashes when running qperf on SLES10 sp2/OFED1.4 RC1
1225,normal,P3,All,[EMAIL PROTECTED],RESOLVED,FIXED,Number of warnings found in 1.4 RC2
1228,normal,P3,RHEL 5,[EMAIL PROTECTED],RESOLVED,FIXED,DAPL v2 UD extensions fails to resolve ah on multiple endpoints
1229,normal,P3,RHEL 5,[EMAIL PROTECTED],RESOLVED,FIXED,install/packaging issues with hand written hooks for dat.conf
1232,normal,P3,All,[EMAIL PROTECTED],RESOLVED,FIXED,Still some dapl-related warnings in 1.4 RC2
1234,normal,P3,All,[EMAIL PROTECTED],RESOLVED,FIXED,SDP related warnings in RC2
1235,normal,P3,SLES 10,[EMAIL PROTECTED],RESOLVED,FIXED,Installer doesn't check for kernel-ib dependency on nfs-utils >= 1.1.1
1238,normal,P3,Other,[EMAIL PROTECTED],RESOLVED,FIXED,Installation freezes because sdpnetstat is going interactive
1240,normal,P3,RedHat Fedora Core 6,[EMAIL PROTECTED],RESOLVED,FIXED,OFED1.4-rc2 install fails on Fedora 9 PPC64
1241,normal,P3,All,[EMAIL PROTECTED],RESOLVED,FIXED,Warnings in OFED 1.4 RC2
commit 5d325de3d5f3963a40074bdccdc751e49eb150f2
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Oct 16 10:50:56 2008 +0200
mlx4: Merged multiple port support patches
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit d7dc2499811045f2ba082b4eacdf36aae969f9c7
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Wed Oct 15 15:31:51 2008 +0200
mlx4_core: Port configuration bug fix.
When port type is not set for one of the ports, set the possible
port (could be auto) and not the actual port.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 8a8a5d0a3e5913e43b5558720e2b74fee036a48b
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 18:05:37 2008 +0200
mlx4: Added DEFFERABLE task backport to 2.6.2* kernels
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit bd42011a08c2244d1ee68b314666cdd0581aef38
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 16:34:03 2008 +0200
mlx4: Fixed sysfs backport patches format for RH4 kernels
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 3f0aaaf73dd474af60fedda007a5d9382d0f27ac
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 11:15:55 2008 +0200
Added 'INIT_DELAYED_WORK_DEFERRABLE' definition to backports
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit df969b09d72b02cb4da24f80ac71f179fea5d447
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 10:27:57 2008 +0200
mlx4: Using Defferable task for sensing
Since the task is not time critical, it will run every 3 round seconds
Added "resched" flag to enable canselation of the task from its context.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit b19ca2b8d318307d0c33bc2af260116110ab2ed2
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 10:21:32 2008 +0200
mlx4: Doubled default number of QPs for FEXCH
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 758de3b8e4b149cb6264ce07f44bdf3d003e6f6a
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 10:18:04 2008 +0200
mlx4/IB: Fixed 'Least_attached_vector' documentation typos
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 26332ec5f6970c47660d21e4d06079e5deb21042
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 10:12:06 2008 +0200
mlx4: Removed port_type module parameter.
Ports type is set to IB by default, and changed to Ethernet if IB is not
supported. Port type can still be changed through sysfs interface.
"check_port_params" function is no longer set to check 2 ports,
It is more generic, to handle the actual number of device ports.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 1b09e388bdf1f3f72d9ad357e6478f244bdd9c71
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 10:03:29 2008 +0200
mlx4: Fixed VLAN table default indexes
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 189e44c321e35af426c63b34cefb2c650c3c50b1
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Oct 12 10:00:04 2008 +0200
mlx4: Removed "mlx4_get_ports_of_type" inline function.
Instead mlx4 device holds ports mask in mlx4_caps struct.
Each user uses this mask to go over its ports with "mlx4_foreach_port"
macro, giving the required port type as parameter.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 71db9e0ea4ef3005f825e249868a5133f1520bb7
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Tue Oct 7 18:10:06 2008 +0200
mlx4_en: Removed unused code
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 4a39df88f5b50d582ea092892cc09fa4f39db7df
Author: Amir Vadai <[EMAIL PROTECTED]>
Date: Tue Oct 7 13:29:05 2008 +0200
SDP: print socket tcp-state in /proc/net/sdp
Signed-off-by: Amir Vadai <[EMAIL PROTECTED]>
commit d342aca80a2617ca9c7e944b38c3135aa0440498
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Tue Oct 7 15:34:01 2008 +0200
IB/mlx4: Set RLKEY bit in QP context
Set RLKEY bit in QP context so that QP can use Reserved L_Key for
memory reference.
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
commit 350f0247433445e7fbefc4bd9386ec5d934a6a9a
Author: Jack Morgenstein <[EMAIL PROTECTED]>
Date: Mon Oct 6 14:07:49 2008 +0200
srp: fix DEVICE_ATTR show-method procedures, and all associated backports
As of kernel 2.6.26, there is no struct class_device. As a result,
the show method prototype for srp_target_infabric and srp_target_oofabric
was incorrect. As of kernel 2.6.26, we must use DEVICE_ATTR, and its show
method requires struct device, and const char *buf.
This also impacted all the srp backports.
Signed-off-by: Jack Morgenstein <[EMAIL PROTECTED]>
commit e3b4fd4af18e77724f285ccf23bcda95a642c1c0
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Mon Oct 6 16:04:19 2008 +0200
mlx4_en: Fixed compilation warnings
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 81a898b9bb2361508037c285d9a5534c17b8015f
Merge: 5de32b2... 501bd88...
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Mon Oct 6 14:15:04 2008 +0200
Merge branch 'ofed_kernel' of git://git.openfabrics.org/~ctung/ofed-1.4
into ofed_kernel
commit 5de32b27b81cf1573ee558b38420d3c6ff75858d
Author: Jack Morgenstein <[EMAIL PROTECTED]>
Date: Mon Oct 6 11:36:06 2008 +0200
mlx4_core: fixed small bug in commit
405ee8d84f8573e6d91de55a7de077933facfe12
Signed-off-by: Jack Morgenstein <[EMAIL PROTECTED]>
commit 501bd88e800adc5e19f2ae2b637cc7f4fb7addf1
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Mon Oct 6 02:09:25 2008 -0700
RDMA/nes: Updated ofed1.4 backports
Updated backports to take care of strict_strtoul() for 2.6.24
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
commit 5a846b3bebd5a5eaf66876a95c243714c30d1399
Author: Amir Vadai <[EMAIL PROTECTED]>
Date: Sun Oct 5 15:05:29 2008 +0200
SDP: removed calling proc_net_create twice
Signed-off-by: Amir Vadai <[EMAIL PROTECTED]>
commit 5476bbd2ce16f4bd1f2f0302c94a47dc01800006
Author: Amir Vadai <[EMAIL PROTECTED]>
Date: Tue Jul 15 21:17:00 2008 +0300
SDP: fix initial recv buffer size
Set initial recv buffer according to incoming hha header.
Fixed bugzilla 1086: SDP Linux and SDP windows don't work togeather
Signed-off-by: Amir Vadai <[EMAIL PROTECTED]>
commit 82d85889abcf7e1d79a1867f2feac05954aa5354
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Sun Oct 5 18:57:19 2008 -0700
RDMA/nes: Fix slab corruption
Referencing cm_node after it is freed via rem_ref_cm_node() caused a slab
corruption. There is no need to set cm_node->cm_id to NULL in
mini_cm_close().
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
commit d730e929cb7072fd23680965124cc6cf715b35c5
Author: Bob Sharp <[EMAIL PROTECTED]>
Date: Fri Oct 3 12:21:19 2008 -0700
RDMA/nes: Fix routed RDMA connections
Fix routed RDMA connections to destinations where the next hop is not
the final destination. Use neigh_*() to properly locate neighbor.
Signed-off-by: Bob Sharp <[EMAIL PROTECTED]>
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
commit 57d68bf4bb0ce07665dc42abcaba755296cdbd6c
Author: Vadim Makhervaks <[EMAIL PROTECTED]>
Date: Fri Oct 3 12:21:18 2008 -0700
RDMA/nes: Enhanced PFT management scheme
Change management of perfect filter table to allow enhanced
performance applications.
Signed-off-by: Vadim Makhervaks <[EMAIL PROTECTED]>
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 8216fefaa78d53c1becebe29dc4d02d6a07ed99d
Author: Faisal Latif <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Handle AE bounds violation
Handle async error NES_AEQE_AEID_AMP_BOUNDS_VIOLATION.
Signed-off-by: Faisal Latif <[EMAIL PROTECTED]>
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 822b57df90257c64739deb198ae0801ceb4e78cc
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Limit critical error interrupts
Mask off a critical error after 100 critical error interrupts to
keep the system "sane".
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 262c0a2f0c8e077ecd1ea3e2b1aa2f6f8c9fd82b
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Stop spurious MAC interrupts
Mask off MAC interrupts on netdev_stop to prevent spurious MAC interrupts
on unload/reload of iw_nes.
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 653c23a84991b9447645477a4f819709000c42ef
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Correct tso_wqe_length
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 26ef70fdd7e327273d574c70209243c6844f4215
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Fill in firmware version for ethtool
Fill in firmware version for ethtool_drvinfo.
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 8b56970ac077a0e426f69a303c548522676c7eba
Author: John Lacombe <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Use ethtool timer value
Use timer value set via ethtool intead of #defines.
Signed-off-by: John Lacombe <[EMAIL PROTECTED]>
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit e00a10ef0cae8a8273c8635cb0e4e6a68925f137
Author: Bob Sharp <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Correct MAX TSO frags value
Use correct define for max TSO fragments.
Signed-off-by: Bob Sharp <[EMAIL PROTECTED]>
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 318eb1b80b45fa3ccd01b02e429cbd5640ddcffd
Author: Bob Sharp <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Enable MC/UC after changing MTU
Re-enable multicast and unicast after changing MTU.
Signed-off-by: Bob Sharp <[EMAIL PROTECTED]>
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit d168429b5b53ed610997f813787bbe0b9102d682
Author: Bob Sharp <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Free NIC TX buffers when destroying NIC QP
Signed-off-by: Bob Sharp <[EMAIL PROTECTED]>
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 54191b594b766a019042e994fa9ca6c89d777f8b
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Fix MDC setting
Clear MDC bits before setting them to a new value. Adjust MDC value
for 10G.
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 930a0f8da620eb17326e00afc81d77c29e816981
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Add wqm_quanta module option
Add a module parameter wqm_quanta. It controls the number of segments
transmitted at a time.
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 0d0831a904b735d3671b91a124aa6c3da4993a5a
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Fri Sep 26 15:08:10 2008 -0500
RDMA/nes: Module parameter permissions
Change permission to 0644 so root can set mpa_version, disable_mpa_crc,
send_first, and nes_drv_opt at runtime.
Signed-off-by: Sweta Bhatt <[EMAIL PROTECTED]>
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 7021120a7691302fd166e6a81a68ff2675b040f4
Author: Jon Mason <[EMAIL PROTECTED]>
Date: Tue Sep 30 14:51:19 2008 -0700
RDMA/cxgb3: Set active_mtu in ib_port_attr
When running ibv_devinfo, the active_mtu returned is garbage. This is
due to the field not being populated in the query_port function in the
driver. The patch below populates the active_mtu field with a MTU of
2k. It also zeros the struct, so that any new additions to it will
return 0.
Signed-off-by: Jon Mason <[EMAIL PROTECTED]>
Acked-by: Steve Wise <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit f3835ec9637e7d6bed1a6d4dd849bb742d6489bf
Author: Chien Tung <[EMAIL PROTECTED]>
Date: Tue Sep 30 14:49:44 2008 -0700
RDMA/nes: Add support for 4-port 1G HP blade card
Add support for NetEffect 4 port 1G HP blade card. The mapping
between physical port and MAC is different from the standup card.
Signed-off-by: Chien Tung <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit b28ce16d74e187e3858e5621ab65b315b4806309
Author: Faisal Latif <[EMAIL PROTECTED]>
Date: Tue Sep 30 14:47:27 2008 -0700
RDMA/nes: Make mini_cm_connect() static
Signed-off-by: Faisal Latif <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 31f71fd61491070a0d74d9ed84f75027c437b99d
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Sun Oct 5 16:11:05 2008 +0200
IPoIB backport: set partition key for the IPv6 multicast GID
for kernels <= 2.6.24.
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
commit aca1fe22ec37d01ff25ad60fb98b5ea18e79cd06
Merge: ae867e9... 94aca1d...
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Sun Oct 5 15:36:11 2008 +0200
Merge commit 'v2.6.27-rc8' into ofed_kernel
commit ae867e9d664a42d94825a545e319f5aafe9d8d27
Merge: 9d27888... 33f1701...
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Sun Oct 5 15:28:02 2008 +0200
Merge branch 'code-drop/20081003' of
git://git.openfabrics.org/~agrover/ofed_1_4/linux-2.6 into ofed_kernel
commit 33f17013a7f34a7f7f793428d5e62dc15f7a9ef3
Author: Andy Grover <[EMAIL PROTECTED]>
Date: Fri Oct 3 14:06:13 2008 -0700
RDS: rewrite connection and ipaddr list handling for IB devices
From a patch from Vlad, this patch includes the addition of tracking
connections associated with an IB device. This allows them to be properly
cleaned up on exit and thus fixes
https://bugs.openfabrics.org/show_bug.cgi?id=826 .
In addition, Pradeep found another bug in ipaddr connection tracking.
They now are properly removed from an old device when migrated to a new
one.
Finally, since some of the new calls allocate memory, we need to check
their return values at callsites.
Signed-off-by: Andy Grover <[EMAIL PROTECTED]>
commit 24859898688678c9fc936836b94c07b171394cf6
Author: Pradeep <[EMAIL PROTECTED]>
Date: Tue Sep 30 14:17:40 2008 -0700
RDS: Fix race between incoming packet and socket close
Race between incoming packet and socket close: rds_recv_incoming()
calls rds_find_bound() which doesn't increment the refcnt in case
the socket is marked DEAD. But the 'rs' pointer is returned. Since
the pointer is not NULL, rds_recv_incoming() decrements the refcnt.
This could cause corruption as the socket structure would have been
free'd by now.
Signed-off-by: Andy Grover <[EMAIL PROTECTED]>
commit 924da143a37ec3ccb9d9025821231a502310aaf8
Author: Andy Grover <[EMAIL PROTECTED]>
Date: Tue Sep 30 13:08:36 2008 -0700
RDS: Fix send ring leak
If mapping fails then we need to unalloc the send ring spots we
allocated.
Signed-off-by: Andy Grover <[EMAIL PROTECTED]>
commit 9ee3ee6456078ac21b45c1eb04391a8c2e4d375e
Author: Andy Grover <[EMAIL PROTECTED]>
Date: Wed Sep 24 12:55:48 2008 -0700
RDS: Fix spin between poll() and sendmsg() each returning -EAGAIN
If there's a small amount of room in the send buffer, we won't send
it (-EAGAIN) but poll() will say to try again right away. Of course
we should be getting acks that free up send space, but if we don't
then we burn a CPU polling. This patch lets the socket go slightly
over its buffsize, but poll() will now DTRT.
Signed-off-by: Andy Grover <[EMAIL PROTECTED]>
commit 9d27888f55ad8e8c8c6c88110b75e93b31da73cc
Merge: 207c51f... 5b43eab...
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Thu Oct 2 13:13:54 2008 -0700
Merge branch 'ofed_kernel' of git://git.openfabrics.org/ofed_1_4/linux-2.6
into ofed_kernel
commit 207c51f36240366ad0e0363e6d3f698bf5aea5d2
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Thu Oct 2 11:09:41 2008 -0700
NFSRDMA RHEL51 backport: builds completely
Signed-off by: Jeff Becker <[EMAIL PROTECTED]>
commit eacea840cf75ead76b3170d3577d5117a2e07c2d
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Thu Oct 2 10:28:24 2008 -0700
NFSRDMA RHEL51 backport: net/sunrpc/auth_gss builds
Signed-off by Jeff Becker <[EMAIL PROTECTED]>
commit 5b43eab9f58bc71d66815f1a78acc62cc5174da6
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Oct 2 14:55:57 2008 +0300
mlx4_en: Fixed LRO backport patch format
Had duplicated changes in the patch
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit aefa7ce23651ce5224d74a7db2edf50b0d04858a
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Oct 2 11:41:29 2008 +0300
mlx4_en: Get stats count Ethtool interface backport
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 4ad2854286ec51e0dfeb90c01f844429faece610
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Oct 2 11:16:12 2008 +0300
mlx4_en: Fixed compilation errors for non existing Macros
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 31b7b0ec27e6badb57d00570f4573f77a1d9ca66
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Oct 2 10:35:15 2008 +0300
mlx4_en: Removed debug messages from backport patches
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 5797162791efe0fb09fbc6566e9be0338e484cb8
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Sep 25 13:40:07 2008 +0300
mlx4_en: Don't init device if bad MAC address burned
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 075b88a397e790fec99b0ea7ec04a36a6e71de63
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Tue Sep 16 17:36:10 2008 +0300
mlx4_en: Addaptive moderation policy change on RHAS4.
When in sender mode, stop moderation to avoid long
timeouts for acks.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit b74481137e9f09b77b8f5a393e3be6486e9bc83c
Author: Eli Cohen <[EMAIL PROTECTED]>
Date: Thu Sep 18 14:34:19 2008 +0300
mlx4_en: fix inline hanlding
The previous code assumed that for inline WQEs for which skb->len is
greater then the space allowed in the first inline segment,
skb_headlen() is greated then that space. This appeared to be wrong in
CentOS 4.6.
Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
commit 10988e43e07b233327d2c60415a4bbf9ea1b5796
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Sep 25 13:27:46 2008 +0300
mlx4_en: Use 1K rings size for RX rings
Don't separate between 64 and 32 bits systems, 1K for all
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 8574360101b789cfd6cf055c342860b3ed364edf
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Mon Sep 15 17:09:52 2008 +0300
mlx4_en: Default TX ring size to 1K
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 6720cf305cfdea94426ff61077544abf8155960e
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Tue Sep 23 11:11:32 2008 +0300
mlx4_en: Synchronize all TX polling
Added a completion lock to sync between interrupts and polling tx cqs.
Queue is not being opened if cq is armed.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 973db6e701712f02c5b5f6603648718b057be73a
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Sep 14 12:52:53 2008 +0300
mlx4_en: Go back to using polling timer to avoid uncompleted skbs
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 8136605db78c8bf9fc6d75919a5c911faad9154c
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Sep 11 15:36:08 2008 +0300
mlx4_en: Don't try to delete timer from xmit function
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit ab95fb29eca4aaea044e8af328e3ac77e44f8cfc
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Sep 11 15:25:21 2008 +0300
mlx4_core: Check number of ports before validating it
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 147a759b8ad48db12018aebc50517a080bda4d12
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Tue Sep 23 11:01:21 2008 +0300
mlx4_en: Poll tx cq only from xmit function
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit c84682541e30a2f009b3cc7f6e39979162e4d75b
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Wed Sep 10 16:03:20 2008 +0300
mlx4_fc: Added list.h backport to sles10_sp1
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit c97e1cf6c293f1054c6ca1b6e834f835e7d622ad
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Tue Sep 23 10:47:07 2008 +0300
mlx4_en: limit lro to frags mode
Current code uses both frag arrays and linking SKBs to generate LRO
SKBs. Since this is not supported by skb_seq_read(), we do not allow
this in the our lro implementation too.
Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
commit f254421bb75918fcae85cc366535fb3a958595fd
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Wed Sep 10 09:59:23 2008 +0300
mlx4_en: Check that TX ring size is not under lower limit
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 98dc9d5bfa4fbb91edebf33ba5acd6352597ffe2
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Wed Sep 10 09:49:41 2008 +0300
mlx4_en: Return to polling after one interrupt
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 16fb28db8a6b3482055e5f4a459623966fcf2128
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Wed Sep 10 10:59:46 2008 +0300
mlx4_en: Added tx interrupts when queue stopped
Once queue is stopped we are using interrupts to open it as
soon as possible. Polling after 1 jiffie is to long period on
some OS's and causes TCP throughpu degradation.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit e8a9b25ab7bdce95243166bd72436a3801080257
Author: Eli Cohen <[EMAIL PROTECTED]>
Date: Tue Sep 9 16:54:20 2008 +0300
mlx4_en: Use standard API in receive flow
Use skb_copy_to_linear_data to copy data to an SKB's linear data
instead of using memcpy.
Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
commit 86f6af764f8c18fd62e194601bb92c772f531a4b
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Wed Sep 10 10:50:15 2008 +0300
mlx4_en: Add read memory barrier when polling receive CQ
Add rmb() after deciding we have a valid CQE to ensure we reference
valid CQE fields.
Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
commit a1f572e8cd6c84521d62464293d16a9243c0e5d0
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Sep 4 10:23:37 2008 +0300
mlx4_en: Removed 'PREFETCH' macro
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit fe5fe90e586d302069a50d8a3f81f147b1e99604
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Thu Sep 4 10:20:01 2008 +0300
mlx4_en: Added get/set_msglevel support
Using Ethtool interaface to set debug printing level
Debug messages are printed according to the message
level, using netif_msg* macros
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit f89cf4fb1e354ecbc9b3c10750c85c07fe3b3db3
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Mon Sep 1 17:31:58 2008 +0300
mlx4_en: Implemented 'get_wol' ethtool option
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 90b041cecac7980037d62809387e7333f3e97292
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Mon Sep 1 17:24:25 2008 +0300
mlx4_en: replace 'get_strings_count' with 'get_sset_count'
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit d924e9c17ff679c6c36c3b64191f47ff5dbff0b4
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Wed Sep 3 18:21:57 2008 +0300
mlx4_en: Show ring statistics
Ethtool -S will now present number of bytes and packets on
both TX and RX rings. Removed the compile-time option.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 8922bb35f900f0feb2f0165c5fef7bc9654e21c4
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Mon Sep 1 10:09:27 2008 +0300
mlx4_en:Added tx checksum offload counter
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit e141b4c8e2b391e49f17117d3fcc851ebc88e2ab
Author: Yevgeny Petrilin <[EMAIL PROTECTED]>
Date: Sun Aug 31 18:22:36 2008 +0300
mlx4_en: Changed 'ethtool -S' output
The output for this commands no longer shows all the statistics
retrieved from HW, only the relevanr one. Added LRO statistics
as well.
Signed-off-by: Yevgeny Petrilin <[EMAIL PROTECTED]>
commit 1ea844465a768399c5a709f421792e4163f0f2e5
Author: Yossi Etigin <[EMAIL PROTECTED]>
Date: Wed Oct 1 22:00:18 2008 +0300
backports: fix csum_tcpudp_nofold backport for ia64
On ia64, kernels <= 2.6.22, csum_tcpudp_nofold is an unexported symbol
(on other platforms it's inline and in ia64 later kernels it's exported).
So need to put our own inline declaration.
Signed-off-by: Yossi Etigin <[EMAIL PROTECTED]>
commit 2e815bc1f314599f1b12ce47ee730aa932e65232
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Wed Oct 1 10:22:01 2008 -0700
NFSRDMA RHEL51 backport: net/sunrpc builds
Signed-off by: Jeff Becker <[EMAIL PROTECTED]>
commit c51f83eae7bb994cd89a2f7bfe6a2afc09c37cd9
Merge: 65d0778... 7891514...
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Mon Sep 29 11:57:29 2008 -0700
Merge branch 'ofed_kernel' of git://git.openfabrics.org/ofed_1_4/linux-2.6
into ofed_kernel
commit 65d07786a3ed79ba432872d1a76e125a9f3855f0
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Mon Sep 29 11:56:02 2008 -0700
NFSRDMA RHEL51 backport: fs/nfsd builds
Signed-off by: Jeff Becker <[EMAIL PROTECTED]>
commit 1a3290dbcbef9023838c5e78e3be03bce447b646
Author: Yannick Cote <[EMAIL PROTECTED]>
Date: Sun Sep 28 09:06:43 2008 -0700
IB/ipath: fix backport for 2.6.21
Signed-off-by: Yannick Cote <[EMAIL PROTECTED]>
commit 7891514239b97fe3396e92d2fcf24f7ab18e60e9
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Sun Sep 28 17:20:05 2008 +0300
Update ipoib_0100_to_2.6.21.patch backport patch after
'47595f49915ac51116916c640f3c0d98df521789' commit.
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
commit 47595f49915ac51116916c640f3c0d98df521789
Author: Roland Dreier <[EMAIL PROTECTED]>
Date: Sun Sep 28 14:58:30 2008 +0300
IPoIB: Continue of "defer skb_orphan() until irqs enabled"
Instead of trying to rely on some complicated and
fragile reasoning about when some race might occur, let's just do what
we want to do anyway and get rid of LLTX. We change from priv->tx_lock
(taken with IRQ disabling) to netif_tx_lock (taken on with
BH-disabling). And then we can keep the skb_orphan in the place it is,
since our xmit routine runs with IRQs enabled.
Most of this patch is just compensating for the fact that the tx_lock
regions are now IRQ-enabled, and so we have to convert places that take
priv->lock to disable IRQs too.
If we could change ipoib_cm_rx_event_handler to not need priv->lock,
then we could change priv->lock to a BH-disabling lock too and simplify
things a bit further.
I've tested this patch some in both datagram and connected mode with a
kernel with lockdep and other debugging enabled, so it is at least
somewhat sane. However more stress testing would definitely be helpful
if we want to put this in 2.6.28. Also it would be interesting to see
if there are any performance effects.
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 48cd0b5312c6fbc7807ddeabaca3e0acd743488e
Author: Hoang-Nam Nguyen <[EMAIL PROTECTED]>
Date: Fri Sep 26 08:34:29 2008 +0200
IB/ehca: Add flush error cqe generation
Signed-off-by: Alexander Schmidt <[EMAIL PROTECTED]>
commit 6ef190cc92e33565accff6a320f0e7d90480bfe7
Author: Roland Dreier <[EMAIL PROTECTED]>
Date: Thu Sep 25 15:28:08 2008 -0700
IPoIB: Fix crash when path record fails after path flush
Commit ee1e2c82 ("IPoIB: Refresh paths instead of flushing them on SM
change events") changed how paths are flushed on an SM event. This
change introduces a problem if the path record query triggered by
fails, causing path->ah to become NULL. A later successful path query
will then trigger WARN_ON() in path_rec_completion(), and crash
because path->ah has already been freed, so the ipoib_put_ah() inside
the lock in path_rec_completion() may actually drop the last reference
(contrary to the comment that claims this is safe).
Fix this by updating path->ah and freeing old_ah only when the path
record query is successful. This prevents the neighbour AH and that
path AH from getting out of sync.
This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=1194>
Reported-by: Rabah Salem <[EMAIL PROTECTED]>
Debugged-by: Eli Cohen <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
Signed-off-by: Linus Torvalds <[EMAIL PROTECTED]>
commit 73b7b54786d5a494d53f983272b5c63d9851af1f
Merge: 0cf19ff... a30e7fa...
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Thu Sep 25 15:39:53 2008 -0700
Merge branch 'ofed_kernel' of git://git.openfabrics.org/ofed_1_4/linux-2.6
into ofed_kernel
commit 0cf19ff6fb1372686d8a1fd14e274eb011aa820b
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Thu Sep 25 15:38:42 2008 -0700
NFSRDMA RHEL51 Backport: fs/nfs builds
Signed-off by: Jeff Becker <[EMAIL PROTECTED]>
commit a30e7fa3e7b6fb5d9f2daeb870206f4768ae3f58
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Thu Sep 25 15:40:58 2008 +0300
RDS: Avoid ib_dereg_mr for NULL memory region causing kernel panic.
If the driver and FW support IB_DEVICE_LOCAL_DMA_LKEY then there will be no
DMA MR allocations.
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
commit 606307deb43e7c54b116c7ed13123b6c136d57c0
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Wed Sep 24 15:48:13 2008 +0300
RDS: destroy cm_id if rdma_bind_addr failed.
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
commit a70311f7c8e0391f88f4679ffbd252b9bb81e6c8
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Wed Sep 24 15:01:44 2008 +0300
Fix sysctl backport header.
Commit '69f7c56b84ba52cdc41b6f885be8131a4ccf5c89' broke sysctl backport
for SLES10SP2 causing kernel oops when removing rds module.
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
commit 7dcc216ebcb275a13b829855f3b2ba26517ba89d
Author: Yannick Cote <[EMAIL PROTECTED]>
Date: Mon Sep 22 13:05:12 2008 -0700
IB/ipath: ib_ipath module hangs on unload
If a send request is posted after the link is brought down, ib_ipath waits
on a CQE that never comes. This fix handles the case where posting a send
is requested when the link is down.
Signed-off-by: Yannick Cote <[EMAIL PROTECTED]>
commit b85142c1efd9c2a22dc10b9dbb4e6b3f060cd560
Author: Aleksey Senin <[EMAIL PROTECTED]>
Date: Tue Sep 23 12:18:40 2008 +0300
Fix RH4.7 backport headers for IA64.
Signed-off-by: Aleksey Senin <[EMAIL PROTECTED]>
commit d0edb7ad5d103122a11593c98758a70116f01e98
Merge: 3861b52... 72d3105...
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Tue Sep 23 10:50:52 2008 +0300
Merge commit 'v2.6.27-rc7' into ofed_kernel
commit 3861b525d6a9227ae76ac22983e5534878dfafa5
Author: Eli Cohen <[EMAIL PROTECTED]>
Date: Tue Sep 23 09:47:29 2008 +0300
IB/ipoib: avoid WARN_ON on NULL path->ah
Commit ee1e2c82c245a5fb2864e9dbcdaab3390fde3fcc introduced an
optimization on path flushing. This caused a new possible scenario in
which unicast_arp_send triggers path query which could fail, causing
path->ah to become NULL. A successive successfull path query will then
trigger WARN_ON() in path_rec_completion(). This fix requires old_ah
to differ from NULL as a prerequsite to trigger the WARN_ON().
Moreover, that commit also allowed path resolution to be triggered for
an invalid path; if that path resolution failed, old_ah would be freed
outside priv->lock violating the assumption that dropping references
inside the lock are guaranteed not to reach zero reference.
Fixes bug #1194
Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
commit d7a0f805e83eaac39b503466d115f50cf3f4fbd2
Author: Eli Cohen <[EMAIL PROTECTED]>
Date: Sun Sep 21 18:14:40 2008 +0300
IB/ipoib: Fix neigh destructor oops
For kernels 2.6.20 and older, it may happen that the pointer to
ipoib_neigh_cleanup() is called after IPoIB has been unloades,
causing a kernel oops. This problem has been fixed for 2.6.21 with
the following commit: ecbb416939da77c0d107409976499724baddce7b
The idea with this patch is to have a helper module which remains
always loaded, and this modules provides the destructor for
neighbours which calls IPoIB's destructor through a function pointer.
When IPoIB is unloaded, the function pointer is cleared so subsequent
calls to a neighbour destructor will be made to valid addresses but
IPoIB's destructor won't get called.
This fixes bug #1188
Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
commit 405ee8d84f8573e6d91de55a7de077933facfe12
Author: Eli Cohen <[EMAIL PROTECTED]>
Date: Sun Sep 21 18:06:13 2008 +0300
mlx4_core: Print a message if a command fails
Print a message with the value returned from the FW in case a command
interface invocation fails. This should ease with debugging problems
that might occur a customer's site.
Signed-off-by: Eli Cohen <[EMAIL PROTECTED]>
commit 4481aefaf9a439d0dd98ddb61b2525254c4143f5
Merge: fd9b2b8... 2997bde...
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Sun Sep 21 00:32:18 2008 +0300
Merge branch 'ofed_kernel' of
git://git.openfabrics.org/~jbecker/nfs_rdma_ofed_1_4 into ofed_kernel
commit 5a0cd4eb661fea095ff9962060c21c161a9ed43f
Merge: b4df9d8... 7e2c232...
Author: Linus Torvalds <[EMAIL PROTECTED]>
Date: Fri Sep 19 16:18:21 2008 -0700
Merge branch 'for-linus' of
git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband
* 'for-linus' of
git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
IPoIB: Fix deadlock on RTNL between bcast join comp and ipoib_stop()
RDMA/nes: Fix client side QP destroy
IB/mlx4: Fix up fast register page list format
mlx4_core: Set RAE and init mtt_sz field in FRMR MPT entries
commit 2997bdedad1a90728dff00d1f2edbc3daedf4f8b
Author: Jeff Becker <[EMAIL PROTECTED]>
Date: Fri Sep 19 15:47:48 2008 -0700
NFSRDMA backports: fix SLES10SP2 bug, plus interim RHEL51 update
Signed-off by: Jeff Becker <[EMAIL PROTECTED]>
commit fd9b2b818006bfde19491e183aedcf3ef10b1c32
Author: Ramachandra Kuchimanchi (QLogic) <[EMAIL PROTECTED]>
Date: Fri Sep 19 07:00:13 2008 -0700
QLogic VNIC: Fix for small memory leak in driver code.
This patch fixes the memory leak detected in function
vnic_ib_mc_mod_qp_to_rts in vnic_ib.c, where qp_attr structure
being allocated is not freed. Also, device files create_primary
and create_secondary are not being removed while cleanup. This
fix thus asked for change in the backport patches too.
Signed-off-by: Ramachandra K <[EMAIL PROTECTED]>
Signed-off-by: Poornima Kamath <[EMAIL PROTECTED]>
Signed-off-by: Amar Mudrankit <[EMAIL PROTECTED]>
commit 7e2c23285454ecc8c8159b86d6343efe222ca3ae
Merge: e8224e4... 29bdc88... d7ffd50...
Author: Roland Dreier <[EMAIL PROTECTED]>
Date: Tue Sep 16 11:57:52 2008 -0700
Merge branches 'ipoib', 'mlx4' and 'nes' into for-linus
commit e8224e4b804b4fd26723191c1891101a5959bb8a
Author: Yossi Etigin <[EMAIL PROTECTED]>
Date: Tue Sep 16 11:57:45 2008 -0700
IPoIB: Fix deadlock on RTNL between bcast join comp and ipoib_stop()
Taking rtnl_lock in ipoib_mcast_join_complete() causes a deadlock with
ipoib_stop(). We avoid it by scheduling the piece of code that takes
the lock on ipoib_workqueue instead of executing it directly. This
works because we only flush the ipoib_workqueue with the RTNL not held.
The deadlock happens because ipoib_stop() calls ipoib_ib_dev_down()
which calls ipoib_mcast_dev_flush(), which calls ipoib_mcast_free(),
which calls ipoib_mcast_leave(). The latter calls
ib_sa_free_multicast(), and this waits until the multicast completion
handler finishes. This handler is ipoib_mcast_join_complete(), which
waits for the rtnl_lock(), which was already taken by ipoib_stop().
This bug was introduced in commit a77a57a1 ("IPoIB: Fix deadlock on
RTNL in ipoib_stop()").
Signed-off-by: Yossi Etigin <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit d7ffd5076d4407d54b25bc4b25f3002f74fbafde
Author: Faisal Latif <[EMAIL PROTECTED]>
Date: Tue Sep 16 11:56:26 2008 -0700
RDMA/nes: Fix client side QP destroy
Fix QP not being destroyed properly on the client, which leads to
userspace programs hanging on exit. This is a missing chunk from the
connection management rewrite in commit 6492cdf3 ("RDMA/nes: CM
connection setup/teardown rework").
Signed-off-by: Faisal Latif <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit 29bdc88384c2b24e37e5760df0dc898546083d6b
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Mon Sep 15 14:25:23 2008 -0700
IB/mlx4: Fix up fast register page list format
Byte swap the addresses in the page list for fast register work requests
to big endian to match what the HCA expectx. Also, the addresses must
have the "present" bit set so that the HCA knows it can access them.
Otherwise the HCA will fault the first time it accesses the memory
region.
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
commit c9257433f2eaf8803a1f3d3be5d984232db41ffe
Author: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Date: Tue Sep 2 13:38:29 2008 -0700
mlx4_core: Set RAE and init mtt_sz field in FRMR MPT entries
Set the RAE (remote access enable) bit and correctly initialize the
MTT size in MPT entries being set up for fast register memory
regions. Otherwise the callers can't enable remote access and in fact
can't fast register at all (since the HCA will think no MTT entries
are allocated).
Signed-off-by: Vladimir Sokolovsky <[EMAIL PROTECTED]>
Signed-off-by: Roland Dreier <[EMAIL PROTECTED]>
_______________________________________________
ewg mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
