1 tbd Sun Microsystems, Inc.
3 * Support for networks:
4 socklnd - any kernel supported by Lustre,
5 qswlnd - Qsnet kernel modules 5.20 and later,
6 openiblnd - IbGold 1.8.2,
7 o2iblnd - OFED 1.1, 1.2.0, 1.2.5, and 1.3
8 viblnd - Voltaire ibhost 3.4.5 and later,
9 ciblnd - Topspin 3.2.0,
10 iiblnd - Infiniserv 3.3 + PathBits patch,
11 gmlnd - GM 2.1.22 and later,
12 mxlnd - MX 1.2.1 or later,
13 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
22 Description: Change ptllnd timeout and watchdog timers
23 Details : Add ptltrace_on_nal_failed and bump ptllnd timeout to match
28 Description: One down Lustre FS hangs ALL mounted Lustre filesystems
29 Details : Shared routing enhancements - peer health detection.
31 Severity : enhancement
33 Description: acceptor.c cleanup
34 Details : Code duplication in acceptor.c for the cases of kernel and
35 user-space removed. User-space libcfs tcpip primitives
36 uniformed to have prototypes similar to kernel ones. Minor
37 cosmetic changes in usocklnd to use cfs_socket_t as
38 representation of socket.
42 Description: IB path MTU mistakenly set to 1st path MTU when ib_mtu is off
43 Details : See comment 46 in bug 11245 for details - it's indeed a bug
44 introduced by the original 11245 fix.
48 Description: uptllnd credit overflow fix
49 Details : kptl_msg_t::ptlm_credits could be overflown by uptllnd since
54 Description: socklnd protocol version 3
55 Details : With current protocol V2, connections on router can be
56 blocked and can't receive any incoming messages when there is no
57 more router buffer, so ZC-ACK can't be handled (LNet message
58 can't be finalized) and will cause deadlock on router.
59 Protocol V3 has a dedicated connection for emergency messages
60 like ZC-ACK to router, messages on this dedicated connection
61 don't need any credit so will never be blocked. Also, V3 can send
62 keepalive ping in specified period for router healthy checking.
64 -------------------------------------------------------------------------------
66 12-31-2008 Sun Microsystems, Inc.
68 * Support for networks:
69 socklnd - any kernel supported by Lustre,
70 qswlnd - Qsnet kernel modules 5.20 and later,
71 openiblnd - IbGold 1.8.2,
72 o2iblnd - OFED 1.1, 1.2.0, 1.2.5, and 1.3
73 viblnd - Voltaire ibhost 3.4.5 and later,
74 ciblnd - Topspin 3.2.0,
75 iiblnd - Infiniserv 3.3 + PathBits patch,
76 gmlnd - GM 2.1.22 and later,
77 mxlnd - MX 1.2.1 or later,
78 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
82 Description: workaround for OOM from o2iblnd
83 Details : OFED needs allocate big chunk of memory for QP while creating
84 connection for o2iblnd, OOM can happen if no such a contiguous
86 QP size is decided by concurrent_sends and max_fragments of
87 o2iblnd, now we permit user to specify smaller value for
88 concurrent_sends of o2iblnd(i.e: concurrent_sends=7), which
89 will decrease memory block size required by creating QP.
93 Description: Support Zerocopy receive of Chelsio device
94 Details : Chelsio driver can support zerocopy for iov[1] if it's
95 contiguous and large enough.
99 Description: fix credit flow deadlock in uptllnd
103 Description: finalize network operation in reasonable time
104 Details : conf-sanity test_32a couldn't stop ost and mds because it
105 tried to access non-existent peer and tcp connect took
106 quite long before timing out.
110 Description: Continuous recovery on 33 of 413 nodes after lustre oss failure
111 Details : Lost reference on conn prevents peer from being destroyed, which
112 could prevent new peer creation if peer count has reached upper
117 Description: LNET Selftest results in Soft lockup on OSS CPU
118 Details : only hits when 8 or more o2ib clients involved and a session is
119 torn down with 'lst end_session' without preceeding 'lst stop'.
123 Description: concurrent_sends in IB LNDs should not be changeable at run time
124 Details : concurrent_sends in IB LNDs should not be changeable at run time
128 Description: ptl_send_rpc hits LASSERT when ptl_send_buf fails
129 Details : only hits under out-of-memory situations
132 -------------------------------------------------------------------------------
134 2009-02-07 Sun Microsystems, Inc.
136 * Support for networks:
137 socklnd - any kernel supported by Lustre,
138 qswlnd - Qsnet kernel modules 5.20 and later,
139 openiblnd - IbGold 1.8.2,
140 o2iblnd - OFED 1.1, 1.2.0, 1.2.5, and 1.3
141 viblnd - Voltaire ibhost 3.4.5 and later,
142 ciblnd - Topspin 3.2.0,
143 iiblnd - Infiniserv 3.3 + PathBits patch,
144 gmlnd - GM 2.1.22 and later,
145 mxlnd - MX 1.2.1 or later,
146 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
149 Description: workaround for OOM from o2iblnd
150 Details : OFED needs allocate big chunk of memory for QP while creating
151 connection for o2iblnd, OOM can happen if no such a contiguous
153 QP size is decided by concurrent_sends and max_fragments of
154 o2iblnd, now we permit user to specify smaller value for
155 concurrent_sends of o2iblnd(i.e: concurrent_sends=7), which
156 will decrease memory block size required by creating QP.
160 Description: Support Zerocopy receive of Chelsio device
161 Details : Chelsio driver can support zerocopy for iov[1] if it's
162 contiguous and large enough.
165 Description: fix credit flow deadlock in uptllnd
169 Description: finalize network operation in reasonable time
170 Details : conf-sanity test_32a couldn't stop ost and mds because it
171 tried to access non-existent peer and tcp connect took
172 quite long before timing out.
176 Description: Continuous recovery on 33 of 413 nodes after lustre oss failure
177 Details : Lost reference on conn prevents peer from being destroyed, which
178 could prevent new peer creation if peer count has reached upper
183 Description: LNET Selftest results in Soft lockup on OSS CPU
184 Details : only hits when 8 or more o2ib clients involved and a session is
185 torn down with 'lst end_session' without preceeding 'lst stop'.
189 Description: concurrent_sends in IB LNDs should not be changeable at run time
190 Details : concurrent_sends in IB LNDs should not be changeable at run time
192 -------------------------------------------------------------------------------
194 11-03-2008 Sun Microsystems, Inc.
196 * Support for networks:
197 socklnd - any kernel supported by Lustre,
198 qswlnd - Qsnet kernel modules 5.20 and later,
199 openiblnd - IbGold 1.8.2,
200 o2iblnd - OFED 1.1, 1.2.0, 1.2.5, and 1.3
201 viblnd - Voltaire ibhost 3.4.5 and later,
202 ciblnd - Topspin 3.2.0,
203 iiblnd - Infiniserv 3.3 + PathBits patch,
204 gmlnd - GM 2.1.22 and later,
205 mxlnd - MX 1.2.1 or later,
206 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
210 Description: ptl_send_rpc hits LASSERT when ptl_send_buf fails
211 Details : only hits under out-of-memory situations
213 -------------------------------------------------------------------------------
215 04-26-2008 Sun Microsystems, Inc.
217 * Support for networks:
218 socklnd - any kernel supported by Lustre,
219 qswlnd - Qsnet kernel modules 5.20 and later,
220 openiblnd - IbGold 1.8.2,
221 o2iblnd - OFED 1.1 and 1.2.0, 1.2.5
222 viblnd - Voltaire ibhost 3.4.5 and later,
223 ciblnd - Topspin 3.2.0,
224 iiblnd - Infiniserv 3.3 + PathBits patch,
225 gmlnd - GM 2.1.22 and later,
226 mxlnd - MX 1.2.1 or later,
227 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
231 Description: excessive debug information removed
232 Details : excessive debug information removed
236 Description: ksocknal_create_conn() hit ASSERTION during connection race
237 Details : ksocknal_create_conn() hit ASSERTION during connection race
241 Description: ksocknal_send_hello() hit ASSERTION while connecting race
242 Details : ksocknal_send_hello() hit ASSERTION while connecting race
246 Description: o2iblnd/ptllnd credit deadlock in a routed config.
247 Details : o2iblnd/ptllnd credit deadlock in a routed config.
251 Description: High load after starting lnet
252 Details : gmlnd should sleep in rx thread in interruptible way. Otherwise,
253 uptime utility reports high load that looks confusingly.
257 Description: ksocklnd fails to establish connection if accept_port is high
258 Details : PID remapping must not be done for active (outgoing) connections
261 --------------------------------------------------------------------------------
263 2008-01-11 Sun Microsystems, Inc.
265 * Support for networks:
266 socklnd - any kernel supported by Lustre,
267 qswlnd - Qsnet kernel modules 5.20 and later,
268 openiblnd - IbGold 1.8.2,
269 o2iblnd - OFED 1.1 and 1.2.0, 1.2.5
270 viblnd - Voltaire ibhost 3.4.5 and later,
271 ciblnd - Topspin 3.2.0,
272 iiblnd - Infiniserv 3.3 + PathBits patch,
273 gmlnd - GM 2.1.22 and later,
274 mxlnd - MX 1.2.1 or later,
275 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
279 Description: liblustre network error
280 Details : liblustre clients should understand LNET_ACCEPT_PORT environment
281 variable even if they don't start lnet acceptor.
285 Description: Strange message from lnet (Ignoring prediction from the future)
286 Details : Incorrect calculation of peer's last_alive value in ksocklnd
288 --------------------------------------------------------------------------------
290 2007-12-07 Cluster File Systems, Inc. <info@clusterfs.com>
292 * Support for networks:
293 socklnd - any kernel supported by Lustre,
294 qswlnd - Qsnet kernel modules 5.20 and later,
295 openiblnd - IbGold 1.8.2,
296 o2iblnd - OFED 1.1 and 1.2.0, 1.2.5.
297 viblnd - Voltaire ibhost 3.4.5 and later,
298 ciblnd - Topspin 3.2.0,
299 iiblnd - Infiniserv 3.3 + PathBits patch,
300 gmlnd - GM 2.1.22 and later,
301 mxlnd - MX 1.2.1 or later,
302 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
306 Description: ASSERTION(me == md->md_me) failed in lnet_match_md()
310 Description: increase send queue size for ciblnd/openiblnd
314 Description: new userspace socklnd
315 Details : Old userspace tcpnal that resided in lnet/ulnds/socklnd replaced
316 with new one - usocklnd.
318 Severity : enhancement
320 Description: Console message flood
321 Details : Make cdls ratelimiting more tunable by adding several tunable in
322 procfs /proc/sys/lnet/console_{min,max}_delay_centisecs and
323 /proc/sys/lnet/console_backoff.
325 --------------------------------------------------------------------------------
327 2007-09-27 Cluster File Systems, Inc. <info@clusterfs.com>
329 * Support for networks:
330 socklnd - any kernel supported by Lustre,
331 qswlnd - Qsnet kernel modules 5.20 and later,
332 openiblnd - IbGold 1.8.2,
333 o2iblnd - OFED 1.1 and 1.2,
334 viblnd - Voltaire ibhost 3.4.5 and later,
335 ciblnd - Topspin 3.2.0,
336 iiblnd - Infiniserv 3.3 + PathBits patch,
337 gmlnd - GM 2.1.22 and later,
338 mxlnd - MX 1.2.1 or later,
339 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
343 Description: /proc/sys/lnet has non-sysctl entries
344 Details : Updating dump_kernel/daemon_file/debug_mb to use sysctl variables
348 Description: TOE Kernel panic by ksocklnd
349 Details : offloaded sockets provide their own implementation of sendpage,
350 can't call tcp_sendpage() directly
354 Description: kibnal_shutdown() doesn't finish; lconf --cleanup hangs
355 Details : races between lnd_shutdown and peer creation prevent
356 lnd_shutdown from finishing.
360 Description: open files rlimit 1024 reached while liblustre testing
361 Details : ulnds/socklnd must close open socket after unsuccessful
366 Description: build error
367 Details : fix typos in gmlnd, ptllnd and viblnd
369 --------------------------------------------------------------------------------
371 2007-07-30 Cluster File Systems, Inc. <info@clusterfs.com>
373 * Support for networks:
374 socklnd - kernels up to 2.6.16,
375 qswlnd - Qsnet kernel modules 5.20 and later,
376 openiblnd - IbGold 1.8.2,
377 o2iblnd - OFED 1.1 and 1.2
378 viblnd - Voltaire ibhost 3.4.5 and later,
379 ciblnd - Topspin 3.2.0,
380 iiblnd - Infiniserv 3.3 + PathBits patch,
381 gmlnd - GM 2.1.22 and later,
382 mxlnd - MX 1.2.1 or later,
383 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
385 --------------------------------------------------------------------------------
387 2007-06-21 Cluster File Systems, Inc. <info@clusterfs.com>
389 * Support for networks:
390 socklnd - kernels up to 2.6.16,
391 qswlnd - Qsnet kernel modules 5.20 and later,
392 openiblnd - IbGold 1.8.2,
394 viblnd - Voltaire ibhost 3.4.5 and later,
395 ciblnd - Topspin 3.2.0,
396 iiblnd - Infiniserv 3.3 + PathBits patch,
397 gmlnd - GM 2.1.22 and later,
398 mxlnd - MX 1.2.1 or later,
399 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
403 Description: Initialize cpumask before use
407 Description: ASSERTION failures when upgrading to the patchless zero-copy
409 Details : This bug affects "rolling upgrades", causing an inconsistent
410 protocol version negotiation and subsequent assertion failure
411 during rolling upgrades after the first wave of upgrades.
415 Details : Change "dropped message" CERRORs to D_NETERROR so they are
416 logged instead of creating "console chatter" when a lustre
417 timeout races with normal RPC completion.
420 Details : lnet_clear_peer_table can wait forever if user forgets to
424 Details : libcfs_id2str should check pid against LNET_PID_ANY.
428 Description: added LNET self test
429 Details : landing b_self_test
434 Description: cfs_duration_{u,n}sec() wrongly calculate nanosecond part of
436 Details : do_div() macro is used incorrectly.
438 2007-04-23 Cluster File Systems, Inc. <info@clusterfs.com>
442 Description: make panic on lbug configurable
446 Description: Add OFED1.2 support to o2iblnd
447 Details : o2iblnd depends on OFED's modules, if out-tree OFED's modules
448 are installed (other than kernel's in-tree infiniband), there
449 could be some problem while insmod o2iblnd (mismatch CRC of
451 If extra Module.symvers is supported in kernel (i.e, 2.6.17),
452 this link provides solution:
453 https://bugs.openfabrics.org/show_bug.cgi?id=355
454 if extra Module.symvers is not supported in kernel, we will
455 have to run the script in bug 12316 to update
456 $LINUX/module.symvers before building o2iblnd.
457 More details about this are in bug 12316.
459 ------------------------------------------------------------------------------
461 2007-04-01 Cluster File Systems, Inc. <info@clusterfs.com>
462 * version 1.4.10 / 1.6.0
463 * Support for networks:
464 socklnd - kernels up to 2.6.16,
465 qswlnd - Qsnet kernel modules 5.20 and later,
466 openiblnd - IbGold 1.8.2,
468 viblnd - Voltaire ibhost 3.4.5 and later,
469 ciblnd - Topspin 3.2.0,
470 iiblnd - Infiniserv 3.3 + PathBits patch,
471 gmlnd - GM 2.1.22 and later,
472 mxlnd - MX 1.2.1 or later,
473 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
477 Description: Ptllnd didn't init kptllnd_data.kptl_idle_txs before it could be
478 possibly accessed in kptllnd_shutdown. Ptllnd should init
479 kptllnd_data.kptl_ptlid2str_lock before calling kptllnd_ptlid2str.
483 Description: gmlnd ignored some transmit errors when finalizing lnet messages.
487 Description: ptllnd logs a piece of incorrect debug info in kptllnd_peer_handle_hello.
491 Description: the_lnet.ln_finalizing was not set when the current thread is
492 about to complete messages. It only affects multi-threaded
498 Description: Changed the default kqswlnd ntxmsg=512
503 Description: Assertion failure in kernel ptllnd caused by posting passive
504 bulk buffers before connection establishment complete.
509 Description: A race in kernel ptllnd between deleting a peer and posting
510 new communications for it could hang communications -
511 manifesting as "Unexpectedly long timeout" messages.
516 Description: Kernel ptllnd lock ordering issue could hang a node.
521 Description: node crash on socket teardown race
524 Frequency : 'lctl peer_list' issued on a mx net
526 Description: Enable lctl's peer_list for MXLND
529 Frequency : after Ptllnd timeouts and portals congestion
531 Description: Credit overflows
532 Details : This was a bug in ptllnd connection establishment. The fix
533 implements better peer stamps to disambiguate connection
534 establishment and ensure both peers enter the credit flow
535 state machine consistently.
540 Description: kptllnd didn't propagate some network errors up to LNET
541 Details : This bug was spotted while investigating 11394. The fix
542 ensures network errors on sends and bulk transfers are
543 propagated to LNET/lustre correctly.
545 Severity : enhancement
547 Description: Fixed console chatter in case of -ETIMEDOUT.
549 Severity : enhancement
551 Description: Added D_NETTRACE for recording network packet history
552 (initially only for ptllnd). Also a separate userspace
553 ptllnd facility to gather history which should really be
554 covered by D_NETTRACE too, if only CDEBUG recorded history in
560 Description: o2iblnd handle early RDMA_CM_EVENT_DISCONNECTED.
561 Details : If the fabric is lossy, an RDMA_CM_EVENT_DISCONNECTED
562 callback can occur before a connection has actually been
563 established. This caused an assertion failure previously.
565 Severity : enhancement
567 Description: Multiple instances for o2iblnd
568 Details : Allow multiple instances of o2iblnd to enable networking over
569 multiple HCAs and routing between them.
573 Description: lnet deadlock in router_checker
574 Details : turned ksnd_connd_lock, ksnd_reaper_lock, and ksock_net_t:ksnd_lock
575 into BH locks to eliminate potential deadlock caused by
576 ksocknal_data_ready() preempting code holding these locks.
580 Description: Millions of failed socklnd connection attempts cause a very slow FS
581 Details : added a new route flag ksnr_scheduled to distinguish from
582 ksnr_connecting, so that a peer connection request is only turned
583 down for race concerns when an active connection to the same peer
584 is under progress (instead of just being scheduled).
586 ------------------------------------------------------------------------------
588 2007-02-09 Cluster File Systems, Inc. <info@clusterfs.com>
590 * Support for networks:
591 socklnd - kernels up to 2.6.16
592 qswlnd - Qsnet kernel modules 5.20 and later
593 openiblnd - IbGold 1.8.2
595 viblnd - Voltaire ibhost 3.4.5 and later
596 ciblnd - Topspin 3.2.0
597 iiblnd - Infiniserv 3.3 + PathBits patch
598 gmlnd - GM 2.1.22 and later
599 mxlnd - MX 1.2.1 or later
600 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
603 Severity : major on XT3
605 Description: libcfs overwrites /proc/sys/portals
606 Details : libcfs created a symlink from /proc/sys/portals to
607 /proc/sys/lnet for backwards compatibility. This is no
608 longer required and makes the Cray portals /proc variables
613 Description: OFED FMR API change
614 Details : This changes parameter usage to reflect a change in
615 ib_fmr_pool_map_phys() between OFED 1.0 and OFED 1.1. Note
616 that FMR support is only used in experimental versions of the
617 o2iblnd - this change does not affect standard usage at all.
619 Severity : enhancement
621 Description: new ko2iblnd module parameter: ib_mtu
622 Details : the default IB MTU of 2048 performs badly on 23108 Tavor
623 HCAs. You can avoid this problem by setting the MTU to 1024
624 using this module parameter.
626 Severity : enhancement
627 Bugzilla : 11118/11620
628 Description: ptllnd small request message buffer alignment fix
629 Details : Set the PTL_MD_LOCAL_ALIGN8 option on small message receives.
630 Round up small message size on sends in case this option
631 is not supported. 11620 was a defect in the initial
632 implementation which effectively asserted all peers had to be
633 running the correct protocol version which was fixed by always
634 NAK-ing such requests and handling any misalignments they
639 Description: When kib(nal|lnd)_del_peer() is called upon a peer whose
640 ibp_tx_queue is not empty, kib(nal|lnd)_destroy_peer()'s
641 'LASSERT(list_empty(&peer->ibp_tx_queue))' will fail.
643 Severity : enhancement
645 Description: Patchless ZC(zero copy) socklnd
646 Details : New protocol for socklnd, socklnd can support zero copy without
647 kernel patch, it's compatible with old socklnd. Checksum is
648 moved from tunables to modparams.
652 Description: When ksocknal_del_peer() is called upon a peer whose
653 ksnp_tx_queue is not empty, ksocknal_destroy_peer()'s
654 'LASSERT(list_empty(&peer->ksnp_tx_queue))' will fail.
657 Frequency : when ptlrpc is under heavy use and runs out of request buffer
659 Description: In lnet_match_blocked_msg(), md can be used without holding a
663 Frequency : very rarely
665 Description: If ksocknal_lib_setup_sock() fails, a ref on peer is lost.
666 If connd connects a route which has been closed by
667 ksocknal_shutdown(), ksocknal_create_routes() may create new
668 routes which hold references on the peer, causing shutdown
669 process to wait for peer to disappear forever.
671 Severity : enhancement
673 Description: Dump XT3 portals traces on kptllnd timeout
674 Details : Set the kptllnd module parameter "ptltrace_on_timeout=1" to
675 dump Cray portals debug traces to a file. The kptllnd module
676 parameter "ptltrace_basename", default "/tmp/lnet-ptltrace",
677 is the basename of the dump file.
680 Frequency : infrequent
682 Description: kernel ptllnd fix bug in connection re-establishment
683 Details : Kernel ptllnd could produce protocol errors e.g. illegal
684 matchbits and/or violate the credit flow protocol when trying
685 to re-establish a connection with a peer after an error or
688 Severity : enhancement
690 Description: Allow /proc/sys/lnet/debug to be set symbolically
691 Details : Allow debug and subsystem debug values to be read/set by name
692 in addition to numerically, for ease of use.
695 Frequency : only in configurations with LNET routers
697 Description: routes automatically marked down and recovered
698 Details : In configurations with LNET routers if a router fails routers
699 now actively try to recover routes that are down, unless they
700 are marked down by an administrator.
702 ------------------------------------------------------------------------------
704 2006-12-09 Cluster File Systems, Inc. <info@clusterfs.com>
707 Frequency : very rarely, in configurations with LNET routers and TCP
709 Description: incorrect data written to files on OSTs
710 Details : In certain high-load conditions incorrect data may be written
711 to files on the OST when using TCP networks.
713 ------------------------------------------------------------------------------
715 2006-07-31 Cluster File Systems, Inc. <info@clusterfs.com>
717 - rework CDEBUG messages rate-limiting mechanism b=10375
718 - add per-socket tunables for socklnd if the kernel is patched b=10327
720 ------------------------------------------------------------------------------
722 2006-02-15 Cluster File Systems, Inc. <info@clusterfs.com>
724 - fix use of portals/lnet pid to avoid dropping RPCs b=10074
725 - iiblnd wasn't mapping all memory, resulting in comms errors b=9776
726 - quiet LNET startup LNI message for liblustre b=10128
727 - Better console error messages if 'ip2nets' can't match an IP address
728 - Fixed overflow/use-before-set bugs in linux-time.h
729 - Fixed ptllnd bug that wasn't initialising rx descriptors completely
730 - LNET teardown failed an assertion about the route table being empty
731 - Fixed a crash in LNetEQPoll(<invalid handle>)
732 - Future protocol compatibility work (b_rls146_lnetprotovrsn)
733 - improve debug message for liblustre/Catamount nodes (b=10116)
735 2005-10-10 Cluster File Systems, Inc. <info@clusterfs.com>
736 * Configuration change for the XT3
737 The PTLLND is now used to run Lustre over Portals on the XT3.
738 The configure option(s) --with-cray-portals are no longer
739 used. Rather --with-portals=<path-to-portals-includes> is
740 used to enable building on the XT3. In addition to enable
741 XT3 specific features the option --enable-cray-xt3 must be
744 2005-10-10 Cluster File Systems, Inc. <info@clusterfs.com>
745 * Portals has been removed, replaced by LNET.
746 LNET is new networking infrastructure for Lustre, it includes a
747 reorganized network configuration mode (see the user
748 documentation for full details) as well as support for routing
749 between different network fabrics. Lustre Networking Devices
750 (LNDS) for the supported network fabrics have also been created
751 for this new infrastructure.
753 2005-08-08 Cluster File Systems, Inc. <info@clusterfs.com>
758 Frequency : rare (large Voltaire clusters only)
760 Description: the default number of reserved transmit descriptors was too low
761 for some large clusters
762 Details : As a workaround, the number was increased. A proper fix includes
765 2005-06-02 Cluster File Systems, Inc. <info@clusterfs.com>
770 Frequency : occasional (large-scale events, cluster reboot, network failure)
772 Description: too many error messages on console obscure actual problem and
773 can slow down/panic server, or cause recovery to fail repeatedly
774 Details : enable rate-limiting of console error messages, and some messages
775 that were console errors now only go to the kernel log
777 Severity : enhancement
779 Description: add /proc/sys/portals/catastrophe entry which will report if
780 that node has previously LBUGged
782 2005-04-06 Cluster File Systems, Inc. <info@clusterfs.com>
784 - update gmnal to use PTL_MTU, fix module refcounting (b=5786)
786 2005-04-04 Cluster File Systems, Inc. <info@clusterfs.com>
788 - handle error return code in kranal_check_fma_rx() (5915,6054)
790 2005-02-04 Cluster File Systems, Inc. <info@clusterfs.com>
792 - update vibnal (Voltaire IB NAL)
793 - update gmnal (Myrinet NAL), gmnalid
795 2005-02-04 Eric Barton <eeb@bartonsoftware.com>
797 * Landed portals:b_port_step as follows...
799 - removed CFS_DECL_SPIN*
800 just use 'spinlock_t' and initialise with spin_lock_init()
802 - removed CFS_DECL_MUTEX*
803 just use 'struct semaphore' and initialise with init_mutex()
805 - removed CFS_DECL_RWSEM*
806 just use 'struct rw_semaphore' and initialise with init_rwsem()
808 - renamed cfs_sleep_chan -> cfs_waitq
809 cfs_sleep_link -> cfs_waitlink
811 - fixed race in linux version of arch-independent socknal
812 (the ENOMEM/EAGAIN decision).
814 - Didn't fix problems in Darwin version of arch-independent socknal
815 (resetting socket callbacks, eager ack hack, ENOMEM/EAGAIN decision)
817 - removed libcfs types from non-socknal header files (only some types
818 in the header files had been changed; the .c files hadn't been