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
20 Severity : enhancement
22 Description: acceptor.c cleanup
23 Details : Code duplication in acceptor.c for the cases of kernel and
24 user-space removed. User-space libcfs tcpip primitives
25 uniformed to have prototypes similar to kernel ones. Minor
26 cosmetic changes in usocklnd to use cfs_socket_t as
27 representation of socket.
31 Description: IB path MTU mistakenly set to 1st path MTU when ib_mtu is off
32 Details : See comment 46 in bug 11245 for details - it's indeed a bug
33 introduced by the original 11245 fix.
37 Description: uptllnd credit overflow fix
38 Details : kptl_msg_t::ptlm_credits could be overflown by uptllnd since
43 Description: socklnd protocol version 3
44 Details : With current protocol V2, connections on router can be
45 blocked and can't receive any incoming messages when there is no
46 more router buffer, so ZC-ACK can't be handled (LNet message
47 can't be finalized) and will cause deadlock on router.
48 Protocol V3 has a dedicated connection for emergency messages
49 like ZC-ACK to router, messages on this dedicated connection
50 don't need any credit so will never be blocked. Also, V3 can send
51 keepalive ping in specified period for router healthy checking.
53 -------------------------------------------------------------------------------
55 12-31-2008 Sun Microsystems, Inc.
57 * Support for networks:
58 socklnd - any kernel supported by Lustre,
59 qswlnd - Qsnet kernel modules 5.20 and later,
60 openiblnd - IbGold 1.8.2,
61 o2iblnd - OFED 1.1, 1.2.0, 1.2.5, and 1.3
62 viblnd - Voltaire ibhost 3.4.5 and later,
63 ciblnd - Topspin 3.2.0,
64 iiblnd - Infiniserv 3.3 + PathBits patch,
65 gmlnd - GM 2.1.22 and later,
66 mxlnd - MX 1.2.1 or later,
67 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
71 Description: workaround for OOM from o2iblnd
72 Details : OFED needs allocate big chunk of memory for QP while creating
73 connection for o2iblnd, OOM can happen if no such a contiguous
75 QP size is decided by concurrent_sends and max_fragments of
76 o2iblnd, now we permit user to specify smaller value for
77 concurrent_sends of o2iblnd(i.e: concurrent_sends=7), which
78 will decrease memory block size required by creating QP.
82 Description: Support Zerocopy receive of Chelsio device
83 Details : Chelsio driver can support zerocopy for iov[1] if it's
84 contiguous and large enough.
88 Description: fix credit flow deadlock in uptllnd
92 Description: finalize network operation in reasonable time
93 Details : conf-sanity test_32a couldn't stop ost and mds because it
94 tried to access non-existent peer and tcp connect took
95 quite long before timing out.
99 Description: Continuous recovery on 33 of 413 nodes after lustre oss failure
100 Details : Lost reference on conn prevents peer from being destroyed, which
101 could prevent new peer creation if peer count has reached upper
106 Description: LNET Selftest results in Soft lockup on OSS CPU
107 Details : only hits when 8 or more o2ib clients involved and a session is
108 torn down with 'lst end_session' without preceeding 'lst stop'.
112 Description: concurrent_sends in IB LNDs should not be changeable at run time
113 Details : concurrent_sends in IB LNDs should not be changeable at run time
117 Description: ptl_send_rpc hits LASSERT when ptl_send_buf fails
118 Details : only hits under out-of-memory situations
121 -------------------------------------------------------------------------------
123 2009-02-07 Sun Microsystems, Inc.
125 * Support for networks:
126 socklnd - any kernel supported by Lustre,
127 qswlnd - Qsnet kernel modules 5.20 and later,
128 openiblnd - IbGold 1.8.2,
129 o2iblnd - OFED 1.1, 1.2.0, 1.2.5, and 1.3
130 viblnd - Voltaire ibhost 3.4.5 and later,
131 ciblnd - Topspin 3.2.0,
132 iiblnd - Infiniserv 3.3 + PathBits patch,
133 gmlnd - GM 2.1.22 and later,
134 mxlnd - MX 1.2.1 or later,
135 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
138 Description: workaround for OOM from o2iblnd
139 Details : OFED needs allocate big chunk of memory for QP while creating
140 connection for o2iblnd, OOM can happen if no such a contiguous
142 QP size is decided by concurrent_sends and max_fragments of
143 o2iblnd, now we permit user to specify smaller value for
144 concurrent_sends of o2iblnd(i.e: concurrent_sends=7), which
145 will decrease memory block size required by creating QP.
149 Description: Support Zerocopy receive of Chelsio device
150 Details : Chelsio driver can support zerocopy for iov[1] if it's
151 contiguous and large enough.
154 Description: fix credit flow deadlock in uptllnd
158 Description: finalize network operation in reasonable time
159 Details : conf-sanity test_32a couldn't stop ost and mds because it
160 tried to access non-existent peer and tcp connect took
161 quite long before timing out.
165 Description: Continuous recovery on 33 of 413 nodes after lustre oss failure
166 Details : Lost reference on conn prevents peer from being destroyed, which
167 could prevent new peer creation if peer count has reached upper
172 Description: LNET Selftest results in Soft lockup on OSS CPU
173 Details : only hits when 8 or more o2ib clients involved and a session is
174 torn down with 'lst end_session' without preceeding 'lst stop'.
178 Description: concurrent_sends in IB LNDs should not be changeable at run time
179 Details : concurrent_sends in IB LNDs should not be changeable at run time
181 -------------------------------------------------------------------------------
183 11-03-2008 Sun Microsystems, Inc.
185 * Support for networks:
186 socklnd - any kernel supported by Lustre,
187 qswlnd - Qsnet kernel modules 5.20 and later,
188 openiblnd - IbGold 1.8.2,
189 o2iblnd - OFED 1.1, 1.2.0, 1.2.5, and 1.3
190 viblnd - Voltaire ibhost 3.4.5 and later,
191 ciblnd - Topspin 3.2.0,
192 iiblnd - Infiniserv 3.3 + PathBits patch,
193 gmlnd - GM 2.1.22 and later,
194 mxlnd - MX 1.2.1 or later,
195 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
199 Description: ptl_send_rpc hits LASSERT when ptl_send_buf fails
200 Details : only hits under out-of-memory situations
203 -------------------------------------------------------------------------------
206 04-26-2008 Sun Microsystems, Inc.
208 * Support for networks:
209 socklnd - any kernel supported by Lustre,
210 qswlnd - Qsnet kernel modules 5.20 and later,
211 openiblnd - IbGold 1.8.2,
212 o2iblnd - OFED 1.1 and 1.2.0, 1.2.5
213 viblnd - Voltaire ibhost 3.4.5 and later,
214 ciblnd - Topspin 3.2.0,
215 iiblnd - Infiniserv 3.3 + PathBits patch,
216 gmlnd - GM 2.1.22 and later,
217 mxlnd - MX 1.2.1 or later,
218 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
222 Description: excessive debug information removed
223 Details : excessive debug information removed
227 Description: ksocknal_create_conn() hit ASSERTION during connection race
228 Details : ksocknal_create_conn() hit ASSERTION during connection race
232 Description: ksocknal_send_hello() hit ASSERTION while connecting race
233 Details : ksocknal_send_hello() hit ASSERTION while connecting race
237 Description: o2iblnd/ptllnd credit deadlock in a routed config.
238 Details : o2iblnd/ptllnd credit deadlock in a routed config.
242 Description: High load after starting lnet
243 Details : gmlnd should sleep in rx thread in interruptible way. Otherwise,
244 uptime utility reports high load that looks confusingly.
248 Description: ksocklnd fails to establish connection if accept_port is high
249 Details : PID remapping must not be done for active (outgoing) connections
251 --------------------------------------------------------------------------------
253 2008-01-11 Sun Microsystems, Inc.
255 * Support for networks:
256 socklnd - any kernel supported by Lustre,
257 qswlnd - Qsnet kernel modules 5.20 and later,
258 openiblnd - IbGold 1.8.2,
259 o2iblnd - OFED 1.1 and 1.2.0, 1.2.5
260 viblnd - Voltaire ibhost 3.4.5 and later,
261 ciblnd - Topspin 3.2.0,
262 iiblnd - Infiniserv 3.3 + PathBits patch,
263 gmlnd - GM 2.1.22 and later,
264 mxlnd - MX 1.2.1 or later,
265 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
268 Description: liblustre network error
269 Details : liblustre clients should understand LNET_ACCEPT_PORT environment
270 variable even if they don't start lnet acceptor.
274 Description: Strange message from lnet (Ignoring prediction from the future)
275 Details : Incorrect calculation of peer's last_alive value in ksocklnd
277 --------------------------------------------------------------------------------
279 2007-12-07 Cluster File Systems, Inc. <info@clusterfs.com>
281 * Support for networks:
282 socklnd - any kernel supported by Lustre,
283 qswlnd - Qsnet kernel modules 5.20 and later,
284 openiblnd - IbGold 1.8.2,
285 o2iblnd - OFED 1.1 and 1.2.0, 1.2.5.
286 viblnd - Voltaire ibhost 3.4.5 and later,
287 ciblnd - Topspin 3.2.0,
288 iiblnd - Infiniserv 3.3 + PathBits patch,
289 gmlnd - GM 2.1.22 and later,
290 mxlnd - MX 1.2.1 or later,
291 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
295 Description: ASSERTION(me == md->md_me) failed in lnet_match_md()
299 Description: increase send queue size for ciblnd/openiblnd
303 Description: new userspace socklnd
304 Details : Old userspace tcpnal that resided in lnet/ulnds/socklnd replaced
305 with new one - usocklnd.
307 Severity : enhancement
309 Description: Console message flood
310 Details : Make cdls ratelimiting more tunable by adding several tunable in
311 procfs /proc/sys/lnet/console_{min,max}_delay_centisecs and
312 /proc/sys/lnet/console_backoff.
314 --------------------------------------------------------------------------------
316 2007-09-27 Cluster File Systems, Inc. <info@clusterfs.com>
318 * Support for networks:
319 socklnd - any kernel supported by Lustre,
320 qswlnd - Qsnet kernel modules 5.20 and later,
321 openiblnd - IbGold 1.8.2,
322 o2iblnd - OFED 1.1 and 1.2,
323 viblnd - Voltaire ibhost 3.4.5 and later,
324 ciblnd - Topspin 3.2.0,
325 iiblnd - Infiniserv 3.3 + PathBits patch,
326 gmlnd - GM 2.1.22 and later,
327 mxlnd - MX 1.2.1 or later,
328 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
332 Description: /proc/sys/lnet has non-sysctl entries
333 Details : Updating dump_kernel/daemon_file/debug_mb to use sysctl variables
337 Description: TOE Kernel panic by ksocklnd
338 Details : offloaded sockets provide their own implementation of sendpage,
339 can't call tcp_sendpage() directly
343 Description: kibnal_shutdown() doesn't finish; lconf --cleanup hangs
344 Details : races between lnd_shutdown and peer creation prevent
345 lnd_shutdown from finishing.
349 Description: open files rlimit 1024 reached while liblustre testing
350 Details : ulnds/socklnd must close open socket after unsuccessful
355 Description: build error
356 Details : fix typos in gmlnd, ptllnd and viblnd
358 ------------------------------------------------------------------------------
360 2007-07-30 Cluster File Systems, Inc. <info@clusterfs.com>
362 * Support for networks:
363 socklnd - kernels up to 2.6.16,
364 qswlnd - Qsnet kernel modules 5.20 and later,
365 openiblnd - IbGold 1.8.2,
366 o2iblnd - OFED 1.1 and 1.2
367 viblnd - Voltaire ibhost 3.4.5 and later,
368 ciblnd - Topspin 3.2.0,
369 iiblnd - Infiniserv 3.3 + PathBits patch,
370 gmlnd - GM 2.1.22 and later,
371 mxlnd - MX 1.2.1 or later,
372 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
374 2007-06-21 Cluster File Systems, Inc. <info@clusterfs.com>
376 * Support for networks:
377 socklnd - kernels up to 2.6.16,
378 qswlnd - Qsnet kernel modules 5.20 and later,
379 openiblnd - IbGold 1.8.2,
381 viblnd - Voltaire ibhost 3.4.5 and later,
382 ciblnd - Topspin 3.2.0,
383 iiblnd - Infiniserv 3.3 + PathBits patch,
384 gmlnd - GM 2.1.22 and later,
385 mxlnd - MX 1.2.1 or later,
386 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
390 Description: Initialize cpumask before use
394 Description: ASSERTION failures when upgrading to the patchless zero-copy
396 Details : This bug affects "rolling upgrades", causing an inconsistent
397 protocol version negotiation and subsequent assertion failure
398 during rolling upgrades after the first wave of upgrades.
402 Details : Change "dropped message" CERRORs to D_NETERROR so they are
403 logged instead of creating "console chatter" when a lustre
404 timeout races with normal RPC completion.
407 Details : lnet_clear_peer_table can wait forever if user forgets to
411 Details : libcfs_id2str should check pid against LNET_PID_ANY.
415 Description: added LNET self test
416 Details : landing b_self_test
421 Description: cfs_duration_{u,n}sec() wrongly calculate nanosecond part of
423 Details : do_div() macro is used incorrectly.
425 2007-04-23 Cluster File Systems, Inc. <info@clusterfs.com>
429 Description: make panic on lbug configurable
433 Description: Add OFED1.2 support to o2iblnd
434 Details : o2iblnd depends on OFED's modules, if out-tree OFED's modules
435 are installed (other than kernel's in-tree infiniband), there
436 could be some problem while insmod o2iblnd (mismatch CRC of
438 If extra Module.symvers is supported in kernel (i.e, 2.6.17),
439 this link provides solution:
440 https://bugs.openfabrics.org/show_bug.cgi?id=355
441 if extra Module.symvers is not supported in kernel, we will
442 have to run the script in bug 12316 to update
443 $LINUX/module.symvers before building o2iblnd.
444 More details about this are in bug 12316.
446 ------------------------------------------------------------------------------
448 2007-04-01 Cluster File Systems, Inc. <info@clusterfs.com>
449 * version 1.4.10 / 1.6.0
450 * Support for networks:
451 socklnd - kernels up to 2.6.16,
452 qswlnd - Qsnet kernel modules 5.20 and later,
453 openiblnd - IbGold 1.8.2,
455 viblnd - Voltaire ibhost 3.4.5 and later,
456 ciblnd - Topspin 3.2.0,
457 iiblnd - Infiniserv 3.3 + PathBits patch,
458 gmlnd - GM 2.1.22 and later,
459 mxlnd - MX 1.2.1 or later,
460 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
464 Description: Ptllnd didn't init kptllnd_data.kptl_idle_txs before it could be
465 possibly accessed in kptllnd_shutdown. Ptllnd should init
466 kptllnd_data.kptl_ptlid2str_lock before calling kptllnd_ptlid2str.
470 Description: gmlnd ignored some transmit errors when finalizing lnet messages.
474 Description: ptllnd logs a piece of incorrect debug info in kptllnd_peer_handle_hello.
478 Description: the_lnet.ln_finalizing was not set when the current thread is
479 about to complete messages. It only affects multi-threaded
485 Description: Changed the default kqswlnd ntxmsg=512
490 Description: Assertion failure in kernel ptllnd caused by posting passive
491 bulk buffers before connection establishment complete.
496 Description: A race in kernel ptllnd between deleting a peer and posting
497 new communications for it could hang communications -
498 manifesting as "Unexpectedly long timeout" messages.
503 Description: Kernel ptllnd lock ordering issue could hang a node.
508 Description: node crash on socket teardown race
511 Frequency : 'lctl peer_list' issued on a mx net
513 Description: Enable lctl's peer_list for MXLND
516 Frequency : after Ptllnd timeouts and portals congestion
518 Description: Credit overflows
519 Details : This was a bug in ptllnd connection establishment. The fix
520 implements better peer stamps to disambiguate connection
521 establishment and ensure both peers enter the credit flow
522 state machine consistently.
527 Description: kptllnd didn't propagate some network errors up to LNET
528 Details : This bug was spotted while investigating 11394. The fix
529 ensures network errors on sends and bulk transfers are
530 propagated to LNET/lustre correctly.
532 Severity : enhancement
534 Description: Fixed console chatter in case of -ETIMEDOUT.
536 Severity : enhancement
538 Description: Added D_NETTRACE for recording network packet history
539 (initially only for ptllnd). Also a separate userspace
540 ptllnd facility to gather history which should really be
541 covered by D_NETTRACE too, if only CDEBUG recorded history in
547 Description: o2iblnd handle early RDMA_CM_EVENT_DISCONNECTED.
548 Details : If the fabric is lossy, an RDMA_CM_EVENT_DISCONNECTED
549 callback can occur before a connection has actually been
550 established. This caused an assertion failure previously.
552 Severity : enhancement
554 Description: Multiple instances for o2iblnd
555 Details : Allow multiple instances of o2iblnd to enable networking over
556 multiple HCAs and routing between them.
560 Description: lnet deadlock in router_checker
561 Details : turned ksnd_connd_lock, ksnd_reaper_lock, and ksock_net_t:ksnd_lock
562 into BH locks to eliminate potential deadlock caused by
563 ksocknal_data_ready() preempting code holding these locks.
567 Description: Millions of failed socklnd connection attempts cause a very slow FS
568 Details : added a new route flag ksnr_scheduled to distinguish from
569 ksnr_connecting, so that a peer connection request is only turned
570 down for race concerns when an active connection to the same peer
571 is under progress (instead of just being scheduled).
573 ------------------------------------------------------------------------------
575 2007-02-09 Cluster File Systems, Inc. <info@clusterfs.com>
577 * Support for networks:
578 socklnd - kernels up to 2.6.16
579 qswlnd - Qsnet kernel modules 5.20 and later
580 openiblnd - IbGold 1.8.2
582 viblnd - Voltaire ibhost 3.4.5 and later
583 ciblnd - Topspin 3.2.0
584 iiblnd - Infiniserv 3.3 + PathBits patch
585 gmlnd - GM 2.1.22 and later
586 mxlnd - MX 1.2.1 or later
587 ptllnd - Portals 3.3 / UNICOS/lc 1.5.x, 2.0.x
590 Severity : major on XT3
592 Description: libcfs overwrites /proc/sys/portals
593 Details : libcfs created a symlink from /proc/sys/portals to
594 /proc/sys/lnet for backwards compatibility. This is no
595 longer required and makes the Cray portals /proc variables
600 Description: OFED FMR API change
601 Details : This changes parameter usage to reflect a change in
602 ib_fmr_pool_map_phys() between OFED 1.0 and OFED 1.1. Note
603 that FMR support is only used in experimental versions of the
604 o2iblnd - this change does not affect standard usage at all.
606 Severity : enhancement
608 Description: new ko2iblnd module parameter: ib_mtu
609 Details : the default IB MTU of 2048 performs badly on 23108 Tavor
610 HCAs. You can avoid this problem by setting the MTU to 1024
611 using this module parameter.
613 Severity : enhancement
614 Bugzilla : 11118/11620
615 Description: ptllnd small request message buffer alignment fix
616 Details : Set the PTL_MD_LOCAL_ALIGN8 option on small message receives.
617 Round up small message size on sends in case this option
618 is not supported. 11620 was a defect in the initial
619 implementation which effectively asserted all peers had to be
620 running the correct protocol version which was fixed by always
621 NAK-ing such requests and handling any misalignments they
626 Description: When kib(nal|lnd)_del_peer() is called upon a peer whose
627 ibp_tx_queue is not empty, kib(nal|lnd)_destroy_peer()'s
628 'LASSERT(list_empty(&peer->ibp_tx_queue))' will fail.
630 Severity : enhancement
632 Description: Patchless ZC(zero copy) socklnd
633 Details : New protocol for socklnd, socklnd can support zero copy without
634 kernel patch, it's compatible with old socklnd. Checksum is
635 moved from tunables to modparams.
639 Description: When ksocknal_del_peer() is called upon a peer whose
640 ksnp_tx_queue is not empty, ksocknal_destroy_peer()'s
641 'LASSERT(list_empty(&peer->ksnp_tx_queue))' will fail.
644 Frequency : when ptlrpc is under heavy use and runs out of request buffer
646 Description: In lnet_match_blocked_msg(), md can be used without holding a
650 Frequency : very rarely
652 Description: If ksocknal_lib_setup_sock() fails, a ref on peer is lost.
653 If connd connects a route which has been closed by
654 ksocknal_shutdown(), ksocknal_create_routes() may create new
655 routes which hold references on the peer, causing shutdown
656 process to wait for peer to disappear forever.
658 Severity : enhancement
660 Description: Dump XT3 portals traces on kptllnd timeout
661 Details : Set the kptllnd module parameter "ptltrace_on_timeout=1" to
662 dump Cray portals debug traces to a file. The kptllnd module
663 parameter "ptltrace_basename", default "/tmp/lnet-ptltrace",
664 is the basename of the dump file.
667 Frequency : infrequent
669 Description: kernel ptllnd fix bug in connection re-establishment
670 Details : Kernel ptllnd could produce protocol errors e.g. illegal
671 matchbits and/or violate the credit flow protocol when trying
672 to re-establish a connection with a peer after an error or
675 Severity : enhancement
677 Description: Allow /proc/sys/lnet/debug to be set symbolically
678 Details : Allow debug and subsystem debug values to be read/set by name
679 in addition to numerically, for ease of use.
682 Frequency : only in configurations with LNET routers
684 Description: routes automatically marked down and recovered
685 Details : In configurations with LNET routers if a router fails routers
686 now actively try to recover routes that are down, unless they
687 are marked down by an administrator.
689 ------------------------------------------------------------------------------
691 2006-12-09 Cluster File Systems, Inc. <info@clusterfs.com>
694 Frequency : very rarely, in configurations with LNET routers and TCP
696 Description: incorrect data written to files on OSTs
697 Details : In certain high-load conditions incorrect data may be written
698 to files on the OST when using TCP networks.
700 ------------------------------------------------------------------------------
702 2006-07-31 Cluster File Systems, Inc. <info@clusterfs.com>
704 - rework CDEBUG messages rate-limiting mechanism b=10375
705 - add per-socket tunables for socklnd if the kernel is patched b=10327
707 ------------------------------------------------------------------------------
709 2006-02-15 Cluster File Systems, Inc. <info@clusterfs.com>
711 - fix use of portals/lnet pid to avoid dropping RPCs b=10074
712 - iiblnd wasn't mapping all memory, resulting in comms errors b=9776
713 - quiet LNET startup LNI message for liblustre b=10128
714 - Better console error messages if 'ip2nets' can't match an IP address
715 - Fixed overflow/use-before-set bugs in linux-time.h
716 - Fixed ptllnd bug that wasn't initialising rx descriptors completely
717 - LNET teardown failed an assertion about the route table being empty
718 - Fixed a crash in LNetEQPoll(<invalid handle>)
719 - Future protocol compatibility work (b_rls146_lnetprotovrsn)
720 - improve debug message for liblustre/Catamount nodes (b=10116)
722 2005-10-10 Cluster File Systems, Inc. <info@clusterfs.com>
723 * Configuration change for the XT3
724 The PTLLND is now used to run Lustre over Portals on the XT3.
725 The configure option(s) --with-cray-portals are no longer
726 used. Rather --with-portals=<path-to-portals-includes> is
727 used to enable building on the XT3. In addition to enable
728 XT3 specific features the option --enable-cray-xt3 must be
731 2005-10-10 Cluster File Systems, Inc. <info@clusterfs.com>
732 * Portals has been removed, replaced by LNET.
733 LNET is new networking infrastructure for Lustre, it includes a
734 reorganized network configuration mode (see the user
735 documentation for full details) as well as support for routing
736 between different network fabrics. Lustre Networking Devices
737 (LNDS) for the supported network fabrics have also been created
738 for this new infrastructure.
740 2005-08-08 Cluster File Systems, Inc. <info@clusterfs.com>
745 Frequency : rare (large Voltaire clusters only)
747 Description: the default number of reserved transmit descriptors was too low
748 for some large clusters
749 Details : As a workaround, the number was increased. A proper fix includes
752 2005-06-02 Cluster File Systems, Inc. <info@clusterfs.com>
757 Frequency : occasional (large-scale events, cluster reboot, network failure)
759 Description: too many error messages on console obscure actual problem and
760 can slow down/panic server, or cause recovery to fail repeatedly
761 Details : enable rate-limiting of console error messages, and some messages
762 that were console errors now only go to the kernel log
764 Severity : enhancement
766 Description: add /proc/sys/portals/catastrophe entry which will report if
767 that node has previously LBUGged
769 2005-04-06 Cluster File Systems, Inc. <info@clusterfs.com>
771 - update gmnal to use PTL_MTU, fix module refcounting (b=5786)
773 2005-04-04 Cluster File Systems, Inc. <info@clusterfs.com>
775 - handle error return code in kranal_check_fma_rx() (5915,6054)
777 2005-02-04 Cluster File Systems, Inc. <info@clusterfs.com>
779 - update vibnal (Voltaire IB NAL)
780 - update gmnal (Myrinet NAL), gmnalid
782 2005-02-04 Eric Barton <eeb@bartonsoftware.com>
784 * Landed portals:b_port_step as follows...
786 - removed CFS_DECL_SPIN*
787 just use 'spinlock_t' and initialise with spin_lock_init()
789 - removed CFS_DECL_MUTEX*
790 just use 'struct semaphore' and initialise with init_mutex()
792 - removed CFS_DECL_RWSEM*
793 just use 'struct rw_semaphore' and initialise with init_rwsem()
795 - renamed cfs_sleep_chan -> cfs_waitq
796 cfs_sleep_link -> cfs_waitlink
798 - fixed race in linux version of arch-independent socknal
799 (the ENOMEM/EAGAIN decision).
801 - Didn't fix problems in Darwin version of arch-independent socknal
802 (resetting socket callbacks, eager ack hack, ENOMEM/EAGAIN decision)
804 - removed libcfs types from non-socknal header files (only some types
805 in the header files had been changed; the .c files hadn't been