1 User-Visible OpenAFS Changes
7 Introduce the rxgk Rx security class, initially limited to server-to-server
8 traffic and local keys (gerrit topic: rxgk-phase1).
10 Add options to the vos and pts commands for server-to-server rxgk support.
12 Add support to add and delete rxgk keys with asetkey. Add support to
13 generate random keys with asetkey, which can be useful to create certain
16 Raise implementation-defined anti-DoS length limits for prdb-related XDR
17 array types, which were being reached at some sites (13838).
19 Bring "-setpag" functionality in klog to parity with aklog (14146).
21 Fix potential Rx hang when an incoming call must wait (14158).
23 Do not leave empty directories behind in the file server vice partition
24 when running the "vos zap -force" command (12879, 12839).
26 Make non-verbose "vos remsite" output output more readable (14127).
28 Display the usage of simple commands (commands without subcommands) when run
29 only with the -help option (10983).
31 Replace SOURCE-MAP with a README.md (14003).
33 Remove unused definition of LINUX_PKGREL from configure.ac (14117).
35 Improve logging and diagnostic messages:
36 * Add a warning message to vos when performing an incremental volume
37 restore over an existing volume which is newer than the dump
39 * Log the binding address and port during startup in the cache manager and
40 all of the server processes (13272)
41 * Improve volume server logging to provide better information during
42 volume restore failures (13252)
43 * Improved cache manager syslog tracing (11858)
44 * Improved database server logging to log important messages at the
45 default logging level, log information during database synchronizations,
46 and log diagnostic messages during recovery aborts (12617, 13079, 12618)
47 * Set a thread name for rx listener threads (13600)
48 * Avoid truncating authentication information in vlserver log
50 * Log when ubik recovery aborts a running remote transaction (13862).
52 Fix warnings issued by static code analyzers:
53 * Fix possible undefined variable in disconnected mode (13207)
54 * Remove redundant conditionals (13158, 13157)
55 * Exit if out of memory while attempting to format command help
57 * Fix possible undefined variable when reading old vldb formats (13755)
59 Assorted memory-handling fixes (13461, 12293, 13355, 13395, 13396, 13161,
60 13659, 13714, 13715, 13760, 13716, 13761).
62 Fix many (but not all) of the new warnings issued by recent versions of
63 gcc and clang (12987..12989, 13010, 13287, 13462..13464, 13467..13468,
64 13470..13476, 13494, 13660..13664, 13684, 13726, 13754, 14049, 14092,
67 Added unit tests for functions mapping vide partition to id (13176).
69 Fix issues resulting in parallel "make install" to fail (13786, 14137).
71 Updated libauth test program (13394).
73 General code cleanup; remove unused code or obsolete code, old comments,
74 or refactor for clarity
75 (12988, 13204, 13209, 13210, 13213, 13226, 13227, 13260, 13271, 13277,
76 13309, 13310, 13324, 13325, 13339, 13345, 13346, 13351, 13361, 13362, 13363,
77 13390, 13397, 13408, 13414, 13458, 13490, 13500, 13509, 13514, 13557, 13640,
80 Build system clean up and fixes
81 (12956, 12961, 12962, 12963, 12992, 12993, 12994, 13237, 13275, 13338,
82 13357, 13360, 13387, 13419, 13594, 13652).
86 Improve database server logging by logging messages when and why a server
87 is marked as down (12616).
89 Log a warning message when starting server processes with no encryption
90 keys available (13911).
92 Remove redundant lseek system calls and use positional I/O in the database
93 servers to improve performance (12271, 12272).
95 Fix an edge case where writes were errneously allowed on readonly
98 Add an option to allow members of system:administrators to perform
99 write operations on otherwise readonly fileservers (13707).
101 Allow "vos rename" to be re-run to finish a previously interrupted
102 volume rename operation (13720).
104 Take volumes offline during convertROtoRW operations, since volume access
105 during the conversion can leave it in an inconsistent state (14066).
107 Do not overwrite the errno variable when logging certain database sendfile
110 Code migration to POSIX Threads (pthreads) from LWP.
111 * Convert upserver and upclient to pthreads (12754)
112 * Convert xstat libraries and related utilities to pthreads (12745, 12746,
113 12747, 12753, 13454, 13455)
117 Attempt to detect and report some common types of cache corruption
118 (13436, 13747, 13969, 14002).
120 Require opt-in to use the historical/deprecated single-DES krb5
121 encryption types, which are being removed from krb5 distributions (13689).
123 Fix incorrect informational messages when the AFSCELL environment
124 variable is set (13371).
126 Fix reading entries of historic vldb formats (13465).
130 Be more careful about overriding the current credentials for operations
131 on cache files, preventing spurious permission errors when systems like
132 AppArmor and SELinux are in use (13751, 14098).
134 Avoid panics from procfs when the kernel module is loaded but afsd is
137 Improve ppc64le support (13980, 14046).
141 Simplify background-move return-code processing (13280).
143 Support macOS Catalina (13935, 13936, 13668..13671, 13928, 14062).
147 Remove references to (unspported) SunOS 4 (13506).
149 Build system fixes for parallel make on Solaris.
151 Support function attributes when building with recent versions of Solaris
154 Fix many (but not all) of the compiler warnings when building on the Solaris
157 Add autoconf support for Studio 12.6 tools (13867).
161 General improvements to VFS compliance.
163 Create destination kernel module directory when installing on
164 FreeBSD (13653, 13690)).
166 Add param.h files and sysnames for FreeBSD 11.2 (13534) and 11.3 (13792).
168 Fix fcntl-style locks by adapting to quirky historical behavior (12579).
170 Support kernels that use VIMAGE support at runtime, not just at
173 Change LWP stack strategy to avoid SIGBUS errors (13691).
175 Skip SIGBUS test (for reasons unrelated to the previous) (14145).
182 * Fix OPENAFS-SA-2019-001: information leakage in failed RPC output
183 Generated RPC handler routines ran output variables through XDR encoding
184 even when the call had failed and would shortly be aborted (and for
185 which uninitialized output variables is common); any complete packets
186 assembled in the process would be sent to the peer, leaking the contents
187 of the uninitialized memory in question.
189 * Fix OPENAFS-SA-2019-002: information leakage from uninitialized scalars
190 Generated RPC handler routines did not initialize output variables of
191 scalar (fixed-length) type, since they did not require dedicated logic to
192 free. Such variables allocated on the stack could remain uninitialized
193 in some cases (including those affected by OPENAFS-SA-2019-001), and the
194 contents of uninitialized memory would be returned to the peer.
198 * Fix OPENAFS-SA-2019-003: fix crash in database servers
199 The ubik debugging RPCs prioritize being fast and non-disruptive to
200 database operations over strict correctness, and do not adhere to the
201 usual locking protocol for data access. A data race could cause a NULL
202 dereference if the second memory load was not optimized out by the
209 Build system updates to remove obsolete autoconf macros and remove missing
210 script warning during builds (13480, 13481, 13482, 13483, 13484, 13486,
213 Build system update to fix a conditional check in the pthread.m4 autoconf
216 Build system update to create the man3 subdirectory, fixing a
217 reported build failure (13535).
219 Remove the last reference to src/mcas in the documentation (13558).
223 Fix fileserver's parsing of the options -vlruthresh, -vlruinterval,
224 -vlrumax and -novbc (13680).
226 Fixes to make ptserver's behaviour when run in restricted mode consistent
227 with the documentation: Non-members of the system:administrators group
228 are no longer allowed to issue the adduser, setfields and delete pts
229 commands, and all members of system:administrators are now allowed to
230 issue pts commands in this mode, not just the admin principal (13686..88).
234 Fix missing Rx call clean-up after failing to read dcaches from a file
237 Fix an Rx call leak for calls aborted by a connection abort after the call
238 was initialized but before use (13517).
240 Remove the obsolete afs_xosi lock to remove unnecessary serialization of
241 VOP_GETATTR calls. This can lead to improved performance under heavy
244 Increase the size of the Directory Name Lookup Cache (DNLC) to improve
245 cache performance (13559).
247 Fix getting tokens for cells with a three character name (13679).
249 Avoid a misleading message about the cell being used when aklog is run
250 with the -cell parameter but the AFSCELL environment variable is set to
251 a different cell (13676).
253 Build system update to honor the CFLAGS environment variable when building
258 Support for mainline kernels up to 5.3 (13787, 13789).
260 More fixes for improper use of ENOENT fixes to avoid incorrect use of linux
261 negative dentry cache, which can lead to false ENOENT errors (13542, 13543,
262 13590, 13692) (RT #134904).
264 Return errors instead of returning incomplete directory listings when the
265 directory objects are incomplete in the cache (13591).
267 Add ppc64le_linux26 sysname for the ppc64le architecture (13636, 13637,
270 Fix configure check for a kernel time function in order to build on
273 RPM packaging update for RHEL8 adding a build requirement to ensure the
274 kernel module can be built from the SRPM (13563) (RT #134900).
276 On systemd based RHEL/Fedora systems, start the client after dkms startup
277 is finished if the latter is installed and enabled, to avoid attempting
278 starts without the kernel module being available yet (13674) (RT #134974).
282 Build system updates for MacOS (13584).
286 Add CTF debugging records to userspace objects to improve debugging
289 Convert the cache manager vnodes to be non-embebbed on Solaris 11 in order
290 to make the cache manager more resilient across Solaris 11 changes (13524,
291 13525, 13526, 13527, 13528).
298 * Improved diagnostics and error messages (13186 13411 13417)
300 * Avoid sending RX packets with random garbage in the userStatus field
303 * Fixed detection of the RX initialization status (13416)
305 * Assorted fixes to avoid segmentation faults and other potential problems
306 by detecting internal errors rather than letting them go unnoticed
311 * Fixed a build problem accidentally introduced in release 1.8.2 (13328)
313 * Assorted efficiency improvements in the ubik implementation (13153 13218
316 * Fixed locking around transaction list processing in volserver to avoid
317 segmentation faults and other potential problems (13336 13337)
319 * When the volserver attempts to remove a temporary volume after a
320 transaction, but the volume was already removed, e.g., by the salvager,
321 this is no longer treated as an error (13235)
325 * Update the CellServDB to the latest version from grand.central.org from
326 May 14th 2018 (13409)
328 * Avoid a panic during cache initialization when allocating the required
331 * Add back the packet counters and timestamps to "vos status" output
332 which had been missing since release 1.8.0 (13421)
334 * Correctly handle errors encountered while reading data from the server
335 and writing it to the cache, e.g., due to a full cache partition (13443)
337 * Avoid a panic due to a recoverable error while flushing cache items
342 * Support mainline kernels 4.20 and 5.0 and distribution kernels with
343 backports from those (13405 13406 13440 13441 13442)
345 * DKMS-related fixes in Red Hat packaging (13438 13479)
349 * Support building and packaging on macOS 10.14 "Mojave" (13412 13413)
356 * Fix OPENAFS-SA-2018-002: information leakage in RPC output variables
357 Various RPC routines did not always initialize all output fields,
358 exposing memory contents to network attackers. The relevant RPCs include
359 an AFSCB_ RPC, so cache managers are affected as well as servers.
363 * Fix OPENAFS-SA-2018-003: denial of service due to excess resource consumption
364 Various RPCs were defined as allowing unbounded arrays as input, allowing
365 an unauthenticated attacker to cause excess memory allocation and tie up
366 network bandwidth by sending (or claiming to send) large input arrays.
368 * Fix OPENAFS-SA-2018-001: unauthenticated volume operations via butc
369 On systems using the in-tree backup system, the butc process was running
370 with administrative credentials, but accepted incoming RPCs over
371 unauthenticated connections; these incoming RPCs in turn triggered
372 outgoing RPCs using the administrative credentials. Unauthenticated
373 attackers could construct volue dumps containing arbitrary contents
374 and cause these dumps to be restored and overwrite arbitrary volume
375 contents; afterward, the backup database could be restored to its
376 initial state, hiding evidence of the unauthorized changes.
378 Running butc with -localauth now requires authenticated incoming
379 connections, and the backup utility makes authenticated connections to
380 the butc. Audit capabilities have been added to the butc RPC handlers.
381 Command-line arguments are provided to retain the (insecure) historical
382 behavior until all systems have been upgraded.
388 * Support for mainline kernel 4.18 and distribution kernels with backports
395 * Improve the usability and consistency of the public API: install missing
396 headers, and add additional symbols to the export list for shared libraries.
398 * Improved Rx abort generation: use the proper serial number for an existing
399 connection if possible, and 0 otherwise (to improve debugging).
401 * Assorted minor fixes in response to static analysis of the codebase.
403 * Fix memory-safety error in XDR decoding of enumerated types.
407 * Fix reference counting error that could cause an assertion failure
410 * vldb_check -fix will no longer corrupt the vldb when multiple MH blocks are present.
412 * Assorted cleanups and efficiency improvements in the ubik implementation.
414 * Return a valid InlineBulkStatus response in error cases.
416 * The fileserver now rejects invalid partition names when attaching partitions.
420 * Fix volume callbacks (e.g., when running 'vos release').
422 * Treat failure to obtain a DSlot as a hard error for that cache partition,
423 avoiding a flood of "disk cache read error in CacheItems" log messages,
424 and reducing the chance of subsequent panic.
426 * Improve error messages for invalid values with -volume-ttl.
428 * Remove useless error message:
429 "find_preferred_connection: no connection and !create".
431 * Avoid passing NULL to a kernel memory deallocator, which is not guaranteed
432 to be safe on all systems.
436 * Add support for 64-bit ARM clients ("arm64").
438 * Fix panic when cache bypass is enabled.
440 * Improve cache manager behavior when unable to open cache files.
442 * Improvements to the RPM packaging.
444 * Detect out-of-memory when using kernel pages for writing.
448 * Fix various issues in the build process for recent Solaris versions.
452 * Fix clients on OS X 10.13.
454 FreeBSD / NetBSD / OpenBSD
456 * Fix panic triggered during periodic cleanup operations and shutdown.
462 * Substantial code quality improvements, largely spotted by Coverity and
463 clang's static analysis.
464 - Add new library for platform independent functions (opr).
465 - Remove arbitrary path name length limits.
466 - Convert to Heimdal's roken library for reliability.
467 - Avoid garbage in allocated buffers (calloc).
468 - Modernize signal handling in pthreaded server processes (softsig).
469 - Improve code comments and additional Doxygen style comments.
470 - Reduce compiler warnings, dead code, unused variables, and
472 - Fix bugs found by static code analyzer (clang-analyzer).
473 - Improved unit test coverage.
474 - Make VLDB flag definitions consistent.
475 - Improve use of run-time assertions and add static assertions.
476 - Add compiler attributes to assist static analyzers.
477 - Clean up include headers in the entire tree.
478 - Improve command-line handling library (libcmd).
479 - Replace hash functions with Jenkin's hash function for faster
480 and more evenly distributed lookups.
481 - Provide a red-black tree data structure to enable algorithmic speedups.
482 - Convert backup and salvage servers to the common logging API (libutil).
483 - Improve volume id data type consistency (VolumeId).
484 - Import APIs for kerberos-style profile configuration support.
485 - Add new APIs to support UserList identities.
486 - Add new APIs to support tabular output from command-line utilities.
487 - Convert vnode macros to inline-functions.
489 * Improved support for non-DES encryption types:
490 - Convert to Heimdal's hcrypto library to support RFC 3961.
491 - Add extended key file format replacing rxkad.keytab, and
492 new key management APIs.
493 - Add support for extended key types to asetkey.
494 - Add akeyconvert to assist in upgrading to OpenAFS 1.8.x by converting an
495 existing rxkad.keytab file to an extended key file.
496 - Do not install the kaserver and related utilities by default to
497 discourage the use of these DES-dependent components.
498 - Remove obsolete klogin and klogin.krb programs.
499 - Add new token APIs to support new rx security classes.
501 * Migrate from LWP to POSIX threads (pthreads):
502 - Convert the ptserver and vlserver from LWP to pthreads.
503 - Remove LWP version of the file server binary.
504 - Convert afsd, aklog, asetkey, klog.krb5, pts, udebug, and vos, from LWP
507 * Improvements to Rx:
508 - Restructure the Rx API to privatize the implementation.
509 - Convert rx events to a red-black tree data structure to improve
511 - Convert from mutexes to atomic operations for counters to reduce lock
513 - Provide per-opcode Rx statistics.
514 - Add an rx_opaque data type to support non-DES encryption types
515 and general code cleanup.
516 - Wake up the application thread after 'twind' is updated to avoid 100ms
517 transmit delays when the receive window transitions from closed to
519 - Fix for OPENAFS-SA-2017-001: sanity-check peer transport parmeters
520 received in ack trailers
522 * Libraries (both internal and installed) are built using libtool, including
523 libuafs. The resulting shared libraries for libafsrpc and libafsauthent
524 should be more usable than previously.
526 * Improvements to the build system:
527 - Convert to libtool to build shared libraries.
528 - Clean up and improve the build system.
529 - Support out of tree builds.
530 - Add a makefile target to generate Doxygen source code documentation.
531 - Link the Java API for OpenAFS with libuafs.a and remove the
533 - Always build the rxperf tool.
534 - Fix man-page generation by make after ./regen.sh -q
535 - Support the SOURCE_DATE_EPOCH environment variable to improve build
537 - Modernize language specific SWIG typemaps for libuafs Perl bindings.
538 - Refactor acinclude.m4 into a set of smaller m4 files (12876, 12877, 12878)
540 * Improvements to documentation:
541 - Document the new KeyFileExt file.
542 - Reorganized the README files.
543 - Improvements and fixes to documentation generation.
544 - Add experimental epub and mobi support
545 - Remove obsolete LWP information from the file server documentation.
546 - Update and reorganize the Quick Start Guide.
547 - Update the Admin Guide.
548 - Remove AIX, HP-UX, and IRIX information from the Quick Start Guide.
549 - Document the vldb and prdb (ubik) file formats.
550 - Add PtLog man page.
551 - Corrections and clarifications to man pages.
552 - Add ubik threading analysis doc.
553 - Normalize the location of text documents in the source tree.
555 * Improvements for troublshooting, debugging, and testing:
556 - Log more details on volume-server-to-fileserver communication errors
558 - Set thread names in pthreaded servers on platforms which support
560 - Add dynroot lock tracking to cmdebug
561 - Fix tracking of an fstrace call site in the cache manager background
563 - Add the afsload tool to simulate multiple cache managers for file server
565 - Add run-time checks for refcount imbalances in the cache manager.
566 - Fix missing newlines in afsd -debug output.
568 * Developer tool improvements:
569 - Improvements and fixes for rxgen (used to generate Rx RPC bindings).
570 - Add tool for man page verification of command options.
571 - Add tool to find Unix cache manager lock identification numbers.
572 - Add an option for pretty build output.
574 * RPM packaging updates:
575 - Update the spec file to keep up with accumulated changes.
576 - Move the klog.krb5 man page to the openafs-krb5 sub-package.
577 - Remove stray man pages. (12870, 12871)
578 - Prevent double-starting client on RHEL7
579 - Convert rpm spec file from deprecated 'make dest' to 'make install'.
580 - Fix rpmbuild command line option default handling.
581 - Support older versions of rpmbuild which do not support the
582 rpmbuild %exclude directive. (12873)
583 - Move the legacy kaserver and related programs to separate sub-packages,
584 which are only built when rpmbuild is given the '--with kauth' option
586 - Package the libuafs perl bindings (12921)
588 * Add a new protection error code (PRNAMETOOLONG) instead of silently
589 truncating names which exceed the maximum name length (PR_MAXNAMELEN).
591 * Add an implementation limit (50000) on the number of names/ids which can
592 be transmitted by unauthenticated clients to the ptserver, avoiding
593 excessive resource consumption from unauthenticated requests.
595 * Add the -config option to vos, pts, and aklog to specify the path to the
596 cell configuration files.
598 * Add more details in vos release -verbose output.
600 * Add the cacheout -encrypt option to encrypt communication between the
601 cacheout client and the fileserver.
603 * Add the command line options to the afsio program to enable encryption of
604 traffic between afsio and the fileserver (-clear, -crypt).
606 * Add the vos release -force-reclone option to force recloning the volume to
607 be released without forcing a full volume dump being transmitted to all
610 * Fix vos to avoid writing loopback addresses into the VLDB in
613 * Print bos and pts error messages to standard error instead of
616 * Improve formatting of the -help output of all commands.
618 * Change -n to -dryrun in all backup subcommands.
620 * Change the backup deletedump -port command line option to -portoffset.
622 * Add user and build host in the version string returned by
625 * Support recent versions of gcc (7.2.1) (12897)
629 * Ubik servers using pthreads are now available and are used by default
631 * As part of improving Ubik reliability in certain edge cases, an extra
632 election cycle (about 60 seconds) may be needed before writes are
633 permitted. This is a conservative change that may be removed in
636 * Avoid continually retransmitting the ubik database to remote sites when
637 a write transaction occurs as remote sites are attempting to rejoin the
638 ubik cluster. (12896)
640 * Ensure the ubik database version number is updated on remote sites at the
641 point the database is transferred to remote sites instead of waiting for
642 the next ubik beacon. This avoids write transaction failures during the
643 window between the database transfer and the next ubik beacon (12885).
645 * Remove periodic background fsync by the fileserver (ihandle fsync thread).
647 * Fix potential file handle leak in the file server ihandle caching layer.
649 * Disable the so-called "hot threads" feature in the file server. The hot
650 threads feature was intended as an optimization for dispatching incoming
651 calls to the current listener thread, but has been reported to incur a
652 performance penalty on modern multi-core systems.
654 * Do not permit creation of users with id of ANONYMOUSID.
656 * Do not save/restore host states in the fsstate.dat file for hosts which
657 are in the process of retrieving CPS information from the ptserver when
658 the fileserver is being shutdown. This fixes a bug in which the fileserver
659 will incorrectly block all threads following a restart.
661 * Add the ptserver -restrict_anonymous option to inhibit exposure of user
662 names from the ptserver.
664 * Do not truncate server log files by default when server processes
665 are started. The -transarc-logs option provides backward compatibility
666 with IBM AFS log handling on server startup. Log messages may be lost
667 in back-to- back restarts when a server is running in this mode.
669 * Reopen server logs on SIGUSR1. This may be used by third-party log
670 rotation tools, such as logrotate, to reopen the log file handles after
671 log files have been renamed.
673 * Fix various bugs when logging with -mrafslogs enabled.
675 * Dynamically reload the kerberos realm to AFS cell mapping (krb.conf) and
676 exclusions for mapping kerberos principals to AFS identities (krb.excl)
677 configuration when the CellServDB cell configuration file is touched.
678 Previously, a restart of the file server was required after updating the
679 kerberos mapping configuration files.
681 * Add a command line option (-restricted_query) to the vlserver and
682 volserver to restrict information queries about volumes to a specific
685 * Add a command line option to the server programs to specify an alternate
686 fully qualified log file name (-logfile).
688 * Add a command line option (-config) to the server programs to specify
689 an alternate path to the server configuration.
691 * Add a command line option to the ptserver and vlserver to specify an
692 alternate path to the database data files.
694 * Add a command line option to the volume server to enable encryption of
695 volume-server-to-volume-server-traffic (-s2scrypt).
697 * Increase the maximum number of LWP threads allowed for the ptserver and
698 vlserver from 16 to 64 (-lwp).
700 * Remove an unused file server command line option (-k).
702 * Fix an incorrect assertion in Demand Attach File Server which could cause
703 the file server process to abort in certain rare conditions.
705 * Deprecate the -bitmap-later configure option for non-Demand-Attach File
708 * Add -vhashsize support to non-Demand-Attach File Servers (DAFS).
710 * Add support for subnet ranges in the NetInfo and NetRestrict
713 * Add the GetXStats RPC to the audit log.
715 * Fix directory creation by bosserver when built for non-Transarc paths.
717 * Fix incomplete list of server addresses retreived by vos listaddr when the
718 vldb contains unreferenced multi-homed server entries.
720 * Remove obsolete bos blockscanner and unblockscanner commands that
721 were only needed for the removed MR-AFS functionality.
723 * Remove obsolete bos salvage options that were only used by the
724 removed MR-AFS functionality..
726 * Remove calls to the deprecated sbrk() function.
728 * Add an experimental feature to database servers to support ubik reads
729 while write transactions are in progress, enabled at build time with the
730 --enable-ubik-read-while-write configure option. This feature is not
731 considered ready for production usage at this time.
733 * Avoid filling the FileLog with "Volume x offline: not in service" when
734 a volume is administratively taken offline with vos offline.
736 * Print an error message when bosserver is started with an unknown
739 * Modify the volume updateDate when the volume is changed by a salvage.
741 * Volume usage statistics are now preserved during reclone and restore
742 operations by default, the behavior previously enabled by
743 the -preserve-vol-stats flag to the volserver. The historical behavior
744 can be retained via the -clear-vol-stats argument.
748 * Use rxkad_crypt by default for connections to fileservers. This matches
749 the existing behavior of the Windows client and has been applied by
750 the distribution packaging on many platforms already.
752 * Add support for relative ACL changes with fs setacl. If a single plus (+)
753 or minus (-) character is appended to the rights' letters argument, the
754 new rights are computed relatively to the existing ones.
756 * Remove afsd -settime and afsd -nosettime support.
758 * Add the afsd -inumcalc option to specify the method used to calculate
759 inode numbers presented by AFS.
761 * Add the afsd -volume-ttl option to specify set the maximum amount of time
762 information retrieved from the vlserver will be cached, regardless of
763 callback expiry times.
765 * Return EIO on internal errors instead of the misleading ENOENT.
767 * Log ICMP errors received, if any, for unreachable servers.
769 * Improve performance of clients with multiple PAGs for different cells.
771 * Fix race condition between changing and using user tokens among cache
774 * Fix fs sysname for users with UID 2748 and 2750 when not running
777 * Add Perl bindings for the user-space cache manager library (libuafs).
779 * Fixes to the bypasscache feature.
781 * Fix fs getcacheparms miscounts.
783 * Remove the obsolete Netscape plugin.
785 * Fix building gtx when ncurses is linked against libtinfo.
787 * Update to the GCO CellServDB update from 14 March 2017.
791 * Remove Linux 2.2 and 2.4 support.
793 * Changes to avoid EIO errors with multiple processes doing intensive mmap
794 writing. (Drop PageReclaim AOP_WRITEPAGE_ACTIVATE.)
796 * Prevent fakestat data inconsistencies in certain cases (131855).
798 * Fix dentry leak which can cause a crash on shutdown.
800 * Fix improper use of ENOENT and avoid incorrect use of linux negative
803 * Use a more correct (less aggressive) scheme to react to downward
804 pressure on cache usage, avoiding d_invalidate(), which can cause
805 getcwd() failures on RHEL 7.4.
807 * Apply a workaround to be compatible with RHEL 7.5's KABI preservation
808 strategy for reading directories.
810 * Improve error reporting when encountering corrupt directories.
812 * Improve rx error handling in the Linux cache manager.
814 * Rename kpasswd to kapasswd when packaging RPMs to avoid colliding with
817 * Do not use the obsolete --enable-largefile-fileservers configure option
820 * In Red Hat packaging, use a separate rpm for kmod debuginfo,
821 removing a needless tight version dependency on the userspace package.
824 * Use the RemainAfterExit systemd feature to avoid premature exit
825 when -afsdb is not given, for RPM packages.
827 * Remove Debian packaging files from the OpenAFS source tree. Debian
828 packaging files are currently maintained in the downstream Debian
831 * Add the sparc_linux26 sysname.
833 * Desupport 32-bit Linux kernels on s390/s390x.
835 * Fix Debian/Ubuntu build regression on kernel 3.16.39.
837 * Fix --enable-kernel-debug for linux 4.8+.
839 * Fix a hang encountered when accessing a previously removed
840 directory entry (12811).
842 * Support linux 4.10, 4.11, 4.12, 4.13, 4.14, 4.15
846 * Remove support for all Solaris and SunOS platforms prior to Solaris 8.
848 * Build 64-bit binaries for Solaris x86 by default.
850 * Use one-group PAGs on Solaris 11, which is required for PAG support
851 on Solaris 11 since supplemental groups must be sorted starting with
854 * Update search paths for solaris cc for recent versions Solaris Studio.
856 * Modernize declaration of module dependences by converting from the
857 deprecated _depends_on symbol to ELF dependencies.
859 * Avoid BAD TRAP panic due to invalid opcodes on x86 with Studio 12.5.
861 * Add ctf debug records to Solaris kernel modules when debug builds
862 are enabled and the ctf tools are present (ctfconvert/ctfmerge).
864 * Save kernel module function arguments on x86 for debugging purposes.
868 * Stop processing upcalls once rx shutdown starts.
870 * Enable atomics for the darwin kernel.
872 * Add a syscall to enable/disable bulkstat at run-time, which is
875 * Fix path to binaries in the prefpane.
877 * Fix builds on MacOS 10.12 by building only the active architecture
880 * Support versions up through 10.13 (High Sierra) and APFS
884 * Use the native kernel module build system instead of an ad hoc
885 replacement build system.
887 * Remove FreeBSD packaging files from the OpenAFS source tree. FreeBSD
888 packaging files are currently maintained in the downstream FreeBSD Ports
891 * Stay up to date with new FreeBSD releases (through 10.3).
893 * Do not claim AFS_VM_RDWR_ENV
895 * Add sysnames and files for i386 and amd64 10.4, 11.1, and 12.0
896 (12-CURRENT, at present). (12887, 12888)
898 * Remove trailing semicolons to fix the build on FreeBSD (12899)
902 * Stay up to date with new NetBSD releases (through 7.x)
904 * Update to use cprng(9) as the randomness source on NetBSD 6.99/7.x.
906 * Build system updates for NetBDS 6.99.x
908 * Do not claim AFS_VM_RDWR_ENV
912 * Stay up to date with new OpenBSD releases (through 4.7)
914 * Do not claim AFS_VM_RDWR_ENV
918 * Updates for AIX support.
920 * Fix build system for AIX exports.
922 * Add the uidpag and localuid runtime options to the aklog LAM plugin.
923 (These runtime options override the use of UID-based PAGs, which were
924 introduced to appease the CDE screensaver.)
931 * Avoid a possible 100ms transmit delay in the RX protocol when a peer's
932 receive window transitions from closed to open (12627)
934 * Documentation improvements (12476 12477 12559[RT #133339])
938 * When bosserver is started with an unknown option, print an error message
939 and exit with a non-zero value rather than failing silently (12631)
941 All DB server platforms
943 * Hold the DB lock while checking for an aborted write transaction (12516)
945 All file server platforms
947 * On demand attach fileservers, don't save or restore a client's host
948 state if CPS ("Current Protection Subdomain") recalculation for it is
949 in progress, to avoid fileserver thread exhaustion (12568)
951 * On demand attach fileservers, avoid flooding the log with error messages,
952 which could happen when the fileserver was restarted while a volume was
955 * Update a volume's "Last Update" time when its content is modified by
956 the salvager, to make the change visible in the output of "vos examine"
957 and to backup services (12633)
961 * Corrected the DCentries bucket counts for very large and zero length
962 files in the output of "fs getcacheparms -excessive" (12604 12605)
964 * Fixed a bug that prevented users with GID 2748 and 2750 from executing
965 the "fs sysname" command on clients running afsd with -rmtsys (12607)
967 * Provide a new -inumcalc switch for afsd to allow enabling the alternative
968 MD5 method of inode number calculation, which was previously only
969 possible on Linux and through the sysctl interface (12608 12632)
973 * Support for mainline kernel 4.12 and distribution kernels with backports
974 from it (12624 12626)
976 * Re-added the improved algorithm for freeing unused vcaches to reduce
977 memory consumption first introduced with the 1.6.18 release, together
978 with a fix for the issue leading to its removal in 1.6.18.2 (12448..12451)
982 * Fixed a crash while stopping the client on macOS 10.12 "Sierra" (12602)
989 * Build fixes required by recent compilers or platforms (12514 12521 12534
992 * Allow the bos server to start successfully in the presence of those, by
993 accepting a now checked return value indicating that the client ThisCell
994 and CellServDB already exist (12522)
998 * Support for mainline kernels 4.10 and - most likely - 4.11 and
999 distribution kernels with backports from them (12530 12588..12590 12598)
1001 * Support for distribution kernels with partial backports from 4.9 (12535)
1004 * In Red Hat packaging, moved the klog.krb5 manual page into the krb5
1007 * In Red Hat packaging, prevent systemd from double-starting the client
1010 * Allow aklog to function on current S390/S390x (12499)
1014 * Make process authentication groups work on Solaris 11, now using a single
1015 group ID (12524..12527)
1017 * Fix a BAD TRAP panic on Solaris 11 clients built with Studio 12.5 (12567)
1021 * Fixed the preference pane for OS X 10.11 and later (12512)
1027 * Build fixes required by recent compilers (12482..12484)
1031 * Support for mainline kernel 4.9 and distribution kernels with
1032 backports from it (12478..12480)
1034 * In Red Hat packaging, make systemd deal correctly with the client
1035 when no userland processes remain after starting it (12481)
1040 * Support for release 10.12 "Sierra" (12431 12432)
1042 * Avoid a crash in the Mounts tab of the OpenAFS preference pane (12447)
1044 OpenAFS 1.6.20 (Security Release)
1048 * Fix for OPENAFS-SA-2016-003: file and directory names leak due to
1049 reuse of directory objects without zeroing the contents
1050 (12461 12462 12463 12464 12465)
1056 * Documentation improvements (12304)
1057 * Fixes for test failures (12396 12415)
1059 All DB server platforms
1061 * Avoid potentially writing to an out of date volume location or protection
1062 database, or losing a database write, which could happen in rare cases
1063 under special conditions during database leader election
1068 * Allow the fsinfo::: DTrace provider to work with AFS files (12371)
1072 * Don't commit more data to a file than was actually copied during writes,
1073 which could happen on architectures with a page size > 4 KiB (12413)
1074 * Fixed build on PPC64 with GCC 6.1 (12388) (RT #133407)
1075 * Fixed build on x86_64 with recent GCC (12365 12366)
1082 * Support for mainline kernel 4.7 and distribution kernels with
1083 backports from it (12348)
1087 * Fixed memory mapped I/O on files >= 4 GiB (12349 12350)
1089 Note that there is a suspicion that this might break the client
1090 on very old Solaris releases (2.6). If it does, the breakage should
1091 occur at build time.
1095 * Added tooling to build a package for OS X 10.10 "Yosemite" and
1096 10.11 "El Capitan" (12335 12351)
1103 * Support for mainline kernel 4.6 and distribution kernels with
1104 backports from it (12332)
1106 * Switch back to the pre-1.6.18 algorithm for freeing unused vcaches.
1107 While the new algorithm is still believed to be correct, it turned
1108 out that at least on some kernels, including 4.5 and 4.6, the dentry
1109 for the current working directory may be erroneously invalidated.
1110 This could lead to errors like "Unable to read current working directory"
1111 when a directory wasn't accessed for a few minutes. (12323)
1113 * Use a secure URL to retrieve the CellServDB in the script to create
1114 the Red Hat source package (12330)
1118 * Added sysname IDs for 10.2 and 10.3 to fix the build on those platforms
1126 * Support for mainline kernel 4.5 and distribution kernels with
1127 backports from it (12300..12302)
1134 * Documentation improvements (12224 11675 11613 12197)
1136 * Improved diagnostics and error messages (12129 12207 12185 12211 12113
1139 * Check that CellServDB entries are valid IPv4 addresses, to avoid
1140 occasional hangs or potentially other erratic behaviour due to invalid
1141 entries (12210) (RT #131794)
1143 All client platforms
1145 * Gracefully handle cases where a client shutdown sequence is initiated
1146 while the client is already shutting down, rather than cause a panic
1149 * Fixed several bugs that could cause erratic behaviour when the write
1150 offset into a file was more than 2 GiB beyond the file's current end
1151 on the server (12213 12214)
1153 All server platforms
1155 * Avoid a possible volserver crash during volume dump or restore due
1156 to invalid ACL entries (12127)
1158 * Allow recovering from a DAFS fileserver operation which allocates a
1159 new vnode but fails to update the vnode index, rather than crashing the
1162 * Fixed a longstanding bug which could damage the volume location database
1163 when "vos changeaddr" was run with "-oldaddr" and "-newaddr" and the
1164 old address was present in a multi-homed entry (12089)
1168 * Added support for releases 10.2 and 10.3 (12232)
1172 * Support for mainline kernel 4.4 and distribution kernels with
1173 backports from it, alas at a performance penalty (12226 12227 12228)
1174 (RT #132677 #132819)
1176 * Avoid using excessive amounts of kernel memory for dynamically
1177 allocated vcaches, by improving the algorithm to free unused ones
1180 * In Red Hat packaging, make the init script use "ip" if available, with
1181 "ifconfig" as a fallback (12193)
1185 * Basic support for release 10.11 "El Capitan" (12212)
1189 * Fixed kernel module builds with optimization (12198) (RT #131261)
1192 OpenAFS 1.6.17 (Security Release)
1194 All server platforms
1196 * Fix for OPENAFS-SA-2016-001: foreign users can create groups as
1197 if they were an administrator (RT #132822) (CVE-2016-2860)
1199 All client platforms
1201 * Fix for OPENAFS-SA-2016-002: information leakage from sending
1202 uninitialized memory over the network. Multiple call sites
1203 were vulnerable, with potential for leaking both kernel and
1204 userland stack data (RT #132847)
1206 * Update to the GCO CellServDB update from 01 January 2016 (12188)
1210 * Fix a crash when the root volume is not found and dynroot is not
1211 in use, a regression introduced in 1.6.14.1 (12166)
1213 * Avoid introducing a dependency on the kernel-devel package corresponding
1214 to the currently running system while building the srpm (12195)
1216 * Create systemd unit files with mode 0644 instead of 0755
1217 (12196) (RT #132662)
1223 * Documentation improvements (11932 12096 12100 12112 12120)
1225 * Improved diagnostics and error messages (11586 11587)
1227 * Distribute the contributor code of conduct with the stable release (12056)
1229 All server platforms
1231 * Create PID files in the right location when bosserver is started with
1232 the "-pidfiles" argument and transarc paths are not being used (12086)
1234 * Several fixes regarding volume dump creation and restore (11433 11553
1237 * Avoid a reported bosserver crash, and potentially others, by replacing
1238 fixed size buffers with dynamically allocated ones in some user handling
1239 functions (11436) (RT #130719)
1241 * Obey the "-toname" parameter in "vos clone" operations (11434)
1243 * Avoid writing a loopback address into the server CellServDB - search
1244 for a non-loopback one, and fail if none is found (12083 12105)
1246 * Rebuild the vldb free list with "vldb_check -fix" (12084)
1248 * Fixed and improved the "check_sysid" utility (12090)
1250 * Fixed and improved the "prdb_check" utility (12101..04)
1252 All client platforms
1254 * Avoid a potential denial of service issue, by fixing a bug in pioctl
1255 logic that allowed a local user to overrun a kernel buffer with a single
1256 NUL byte (commit 2ef86372) (RT #132256) (CVE-2015-8312)
1258 * Refuse to change multi-homed server entries with "vos changeaddr",
1259 unless "-force" is given, to avoid corruption of those entries (12087)
1261 * Provide a new vos subcommand "remaddrs" for removing server entries, to
1262 replace the slightly confusing "vos changeaddr -remove" (12092 12094)
1264 * Make "fs flushall" actually invalidate all cached data (11894)
1266 * Prevent spurious call aborts due to erroneous idle timeouts (11594)
1268 * Provide a "--disable-gtx" configure switch to avoid building and
1269 installing libgtx and its header files as well as the depending
1270 "scout" and "afsmonitor" applications (12095)
1272 * Fixed building the gtx applications against newer ncurses (12125)
1274 * Allow pioctls to work in environments where the syscall emulation
1275 pseudo file is created in a read-only pseudo filesystem, like in
1276 containers under recent versions of docker (12124)
1280 * In Red Hat packaging, avoid following a symbolic link when writing
1281 the client CellServDB, which could overwrite the server CellServDB,
1282 by removing an existing symlink before writing the file (12081)
1284 * In Red Hat packaging, avoid a conflict of openafs-debuginfo with
1285 krb5-debuginfo by excluding our kpasswd executable from debuginfo
1286 processing (12128) (RT #131771)
1288 OpenAFS 1.6.15 (Security Release)
1290 All client and server platforms
1292 * Fix for OPENAFS-SA-2015-007 "Tattletale"
1294 When constructing an Rx acknowledgment (ACK) packet, Andrew-derived
1295 Rx implementations do not initialize three octets of data that are
1296 padding in the C language structure and were inadvertently included
1297 in the wire protocol (CVE-2015-7762). Additionally, OpenAFS Rx in
1298 versions 1.5.75 through 1.5.78, 1.6.0 through 1.6.14, and 1.7.0
1299 through 1.7.32 include a variable-length padding at the end of the
1300 ACK packet, in an attempt to detect the path MTU, but only four octets
1301 of the additional padding are initialized (CVE-2015-7763).
1307 * Support kernels up to 4.2
1309 Due to changes to internal data structures with this kernel release,
1310 the OpenAFS client can no longer reset the link count during path
1311 lookups. Since volume root directories must behave like symlinks
1312 instead of normal directories in order to satisfy Linux kernel
1313 invariants, looking up paths containing more than 40 mount points
1314 will fail with ELOOP on such kernels.
1318 All server platforms
1320 * Prior to the OpenAFS security release 1.6.13, the Volume Location
1321 Server (vlserver) RPC VL_ListAttributesN2() supported wildcard volume
1322 name lookups via regular expression (regex) pattern matching. This
1323 support was completely disabled in 1.6.13 because it was judged to be
1324 a security risk due to buffer overruns in the implementation, as well
1325 as the possibility of denial of service attacks where certain regular
1326 expressions could cause excessive CPU usage in some regex
1329 Unfortunately, after 1.6.13 was released, it was discovered that
1330 the native OpenAFS 'backup' system uses the VL_ListAttributesN2()
1331 regex support to evaluate configured volume sets. If you use the
1332 OpenAFS 'backup' system (or another backup system which relies on it,
1333 such as Tivoli Storage Manager (TSM, aka Tivoli ADSM)), and are using
1334 volume sets which require regular expressions for the volume name,
1335 then those volume sets cannot be resolved by OpenAFS 1.6.13. The next
1336 paragraph provides details on how to identify any affected volume sets.
1338 OpenAFS backup volume sets may be described by fileserver, partition
1339 name, and volume name. The fileserver and partition specifications
1340 never require regular expression support. The volume name specification
1341 always requires regular expression support except for when specifying
1342 _all_ volumes via two special cases: the universal wildcard ".*", or "".
1343 For example, volume name "proj" or "*.backup" or "homevol.*" all
1344 require regex support - even if the specification contains no wildcard
1345 characters and/or exactly matches an existing volume name.
1347 As a result of this issue, OpenAFS 1.6.14 replaces the 1.6.13 changes
1348 to VL_ListAttributesN2. 1.6.14 prevents the buffer overruns and
1349 reenables the regex support, but restricts it to OpenAFS super-users
1350 and -localauth only. This is sufficient to restore the OpenAFS 'backup'
1351 system's ability to work correctly with any previously supported volume
1352 set. The OpenAFS 'backup' commands are already documented to require
1353 super-user authorization, so this restriction is moot for the backup
1356 There are no other direct consumers of the VL_ListAttributesN2() regex
1357 support in the OpenAFS tree. However, the VL_ListAttributesN2 RPC is
1358 publicly accessible and might be used by third party tools directly or
1359 indirectly via OpenAFS's libadmin. Any such tools that issue
1360 VL_ListAttributesN2 RPCs must now be executed using super-user or
1363 None of the other security fixes in OpenAFS 1.6.13 are known to have
1364 any issues, and are still included unchanged in OpenAFS 1.6.14.
1366 If there are any questions concerning the possible impact of OpenAFS
1367 1.6.13 or 1.6.14 at your site, please contact your OpenAFS support
1368 provider or the openafs-info@openafs.org mailing list for further
1373 All server platforms
1375 * Fix for CVE-2015-3282: vos leaks stack data onto the wire in the
1376 clear when creating vldb entries
1378 * Workaround for CVE-2015-3283: bos commands can be spoofed, including
1379 some which alter server state
1381 * Disabled searching the VLDB by volume name regular expression to avoid
1382 possible buffer overruns in the volume location server
1384 All client platforms
1386 * Fix for CVE-2015-3284: pioctls leak kernel memory
1388 * Fix for CVE-2015-3285: kernel pioctl support for OSD command passing
1393 * Fix for CVE-2015-3286: Solaris grouplist modifications for PAGs can
1394 panic or overwrite memory
1398 All server platforms
1400 * Avoid database corruption if a database server is shut down and then
1401 brought up again quickly with an altered database (11773 11774)
1404 All client platforms
1406 * Fixed a potential buffer overflow in aklog (11808)
1408 * Avoid a bogus warning regarding the checkserver daemon, which could be
1409 logged during startup when the cache initialization was very fast (11680)
1411 * Added documentation of the inaccuracy of the 'partition' field in
1412 'fs listquota' output for partitions larger than 2 TiB (11626)
1416 * Support kernels up to 4.1 (11872 11873)
1418 * Avoid spurious EIO errors when writing large chunks of data to
1419 mmapped files (11877)
1423 * Build fixes required at least on OS X 10.10 Yosemite with the latest
1424 XCode (11859 11876 11842..11845 11863 11878 11879)
1430 * Support kernels up to 4.0 (11760 11761)
1434 * Fixed kernel module build on systems with an updated clang which no
1435 longer accepts the -mno-align-long-strings as a no-op (11809)
1441 * Allow aklog to succeed creating native K5 tokens even when mapping
1442 the K5 principal to a K4 one fails (11538)
1444 * Build fixes (11435 11636)
1446 All client platforms
1448 * Avoid a potential kernel panic due to connection reference overcounts
1449 (11645) (RT #131885)
1451 * Avoid potential corruption of files written using memory mapped I/O
1452 when the file is larger than the cache (11656) (RT #131976)
1456 * Support kernels at least up to 3.19 (11549 11550 11569 11570 11595
1457 11658..11662 11694 11752)
1459 Note: By default this excludes kernels 3.17 to 3.17.2, which will leak
1460 an inode reference when an error occurs in d_splice_alias(). The
1461 module will build and work, but leak kernel memory, leading to
1462 performance degradation and eventually system failure due to
1463 memory exhaustion. Since it's impossible to detect this condition
1464 automatically, the switch --enable-linux-d_splice_alias-extra-iput
1465 must be passed to configure when building the module for those
1466 kernels. The same would be necessary for any kernel with backports
1467 of commit 908790fa3b779d37365e6b28e3aa0f6e833020c3 or commit
1468 95ad5c291313b66a98a44dc92b57e0b37c1dd589 but not the fix in commit
1469 51486b900ee92856b977eacfc5bfbe6565028070 in the linux-stable repo
1470 (git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git) or
1471 the corresponding changes on other branches.
1473 * Fixed a regression introduced in OpenAFS release 1.6.10 which could
1474 make the spurious "getcwd: cannot access parent directories" problem
1475 return (11558 11568) (RT #131780)
1477 * Avoid leaking memory when scanning a corrupt directory (11707)
1481 * Support OS X 10.10 "Yosemite" (11571 11572 11611) (RT #131946)
1485 * Avoid reading random data rather than correct cache content when using
1486 ZFS as the cache file system on Solaris >= 11, and fix potential similar
1487 problems on other platforms (11713 11714)
1491 * Build fix for releases >= 11.0 (11610)
1495 * Support release 5.4 (11700)
1502 * Don't hide the "version" subcommand in help output (11214)
1504 * Documentation improvements (11126 11216 11222 11223 11225 11226)
1506 * Improved diagnostics and error messages (11154 11246 11247 11249 11181
1509 * Build system improvements (11158 11221 11224 11225 11227..11241 11282
1510 11342 11350 11353 11242 11367 11392)
1512 * Avoid potentially erratic behaviour under certain error conditions by
1513 either avoiding or at least not ignoring them, in various places (11008
1514 11010..11065 11112 11148 11196 11530)
1518 * Support releases 9.3 and 10.1 (11368 11369 11402 11403 11404)
1520 * Makes a disk cache more likely to work on FreeBSD, though such
1521 configurations remain not very tested (11448)
1523 All server platforms
1525 * Added volscan(8) (11252..11280 11387 11388)
1527 * Fixed a bug causing subgroups not to function correctly if their
1528 ptdb entry had more than one continuation entry (11352)
1530 * Logging improvements (10946 11153)
1532 * Allow log rotation via copy and truncate (11193)
1534 * Avoid a server crash during startup only observed on a single platform
1535 and when using a 3rd party library under certain circumstances, which is
1536 a collateral effect of the security improvements introduced in OpenAFS
1537 release 1.6.5 (11075) (RT #131852)
1539 All client platforms
1541 * Raised the free space reported for /afs to the maximum possible value of
1542 just under 2 TiB - the old value was 9 GiB on most platforms (10984)
1544 * Reduced the amount of stack space used (11162 11163 11203 11164..11167
1545 11338 11339 11364..11366 11381)
1547 * Sped up a periodic client task which could be problematically slow
1548 on systems with a large number of PAGs and files in use (11307)
1550 * Fixed failure of the up command with large ACLs (11111)
1552 * Avoid a potential crash of aklog (11218)
1554 * Avoid potential crashes of scout and xstat_fs_test (11155)
1558 * Support kernels up to 3.16 (11308 11309)
1560 * Fixed a regression introduced in OpenAFS release 1.6.6 that made
1561 checking for existing write locks incorrectly fail on readonly volumes
1564 * Fixed a regression introduced in OpenAFS release 1.6.8 that could
1565 cause VFS cache inconsistencies when a previously-accessed directory
1566 entry was removed and recreated with the same name but pointing to a
1567 different file on another client (11358)
1569 * Use the right path to depmod in Red Hat packaging to avoid dependency
1570 calculation incorrectly failing unless a link /sbin -> /usr/sbin is
1571 present on the system performing it (11171) (RT #131860)
1573 * Do not ignore kernel module build errors (11205)
1577 All server platforms
1579 * Fix for OPENAFS-SA-2014-002
1585 * Documentation improvements (10751 10875 10931 10897 10883 10954 10955)
1587 * Improved diagnostics and error messages (10756 10814 10949)
1589 * Fixed a bug in RX that could make errors during packet reception go
1592 * Fixed a bug that made "vos size -dump" display the wrong size for
1593 large volumes. (10933) (RT #131819)
1595 All server platforms
1597 * Change the default fileserver sync behavior from "delayed" to "onclose".
1598 This means that explicit syncing only happens when a volume is detached.
1601 * Added the -offline-timeout and -offline-shutdown-timeout options to the
1602 fileserver, to implement interrupting clients accessing volumes we are
1603 trying to take offline. (6266 10799)
1605 All client platforms
1607 * When a client is shut down, it will give up its callbacks. The Windows
1608 client has been doing this since 2007. Note that older fileservers
1609 (1.3.50 to 1.4.5 and 1.5.0 to 1.5.27) had a bug in the implementation of
1610 the relevant RPC that could cause crashes or other undefined behavior
1611 when this happens. (6272 8840 10855)
1613 * Restored the pre-1.6 behavior of "vos e" being an alias for "vos examine".
1616 * Avoid flooding logs with warnings about byte-range locks, by throttling
1617 them per file. Also, make the messages more useful by including the
1620 * Avoid a possible panic during shutdown while tracing. (10932)
1624 * Fixed a bug that could cause the "getcwd: cannot access parent
1625 directories" problem (10804 10984)
1627 * Avoid a delay when accessing uncached data in AFS in a confined
1628 context under SELinux. (10598)
1630 * Red Hat packaging improvements (10600 10767 10807)
1634 All server platforms
1636 * Fix for OPENAFS-SA-2014-001
1638 * Fix for a potential DOS attack against RX servers
1644 * As of this release, OpenAFS no longer ships uncompressed source tarballs.
1645 Tarballs are still shipped with both compression formats, gzip and bzip2.
1648 * Documentation improvements (10136 10314 10601)
1650 * Improved diagnostics and error messages (9412 10085 10274)
1652 * Avoid redefining "assert" in our public header files, which could
1653 cause failures when building some applications using them. (10096)
1655 * Fixes for parallel builds (10005 10309 10337)
1657 * Added a -s switch to afscp (not installed by default) to help simulate
1658 a slow client. (9416 9417)
1660 * Added a -probe switch to vlclient test program (not installed by default)
1661 to ping all vlservers in a cell in parallel. (9570)
1663 All server platforms
1664 * The fileserver now ignores any vice partitions with a NeverAttach flag
1665 file present in the root directory. (RT #130561) (9470 9471)
1667 * Restrict forcing CPS ("Current Protection Subdomain") recalculation in
1668 the fileserver to administrators. Also fixed a bug that could cause this
1669 operation to be incomplete. (9485 9487)
1671 * Allow non-DAFS fileservers to attach unusable volumes, restoring pre-1.6
1672 behaviour. (RT #131505) (9499)
1674 * Restored the pre-1.6 behaviour when running vos examine for a volume
1675 currently in a transaction, showing the volume as busy again rather than
1676 offline. (9685 9915 9916)
1678 * Reduced the minimum time a bos salvage takes from 5 seconds to 1. (9476)
1680 * Fixed buserver to not segfault when started with the -servers option.
1681 (RT #131706) (10166)
1683 * Salvager fixes, addressing a wide variety of possible problems from
1684 unnecessary salvaging to aborts (9282 9283 9457 9458 9459 9461 9462 9480
1687 * Fixed a bug that could cause saved state information to be discarded
1688 when restarting a large or busy fileserver, which negatively impacted
1691 * Fixed a bug that could have caused undefined behaviour in the vlserver
1692 in rare cases when a fileserver registered its addresses in the VLDB.
1695 * Added the -preserve-vol-stats switch to volserver, allowing it to keep
1696 the access statistics across volume restore and reclone operations
1697 instead of resetting them. (9477)
1699 * Inserted an exponential delay between retries when bosserver attempts to
1700 restart a server process. (9571 10199)
1702 * Improved vldb_check (not installed by default) to cope with broken
1703 vlentry names and volids, and provide more output to aid debugging.
1706 * Releasing a volume after adding a new RO site no longer touches any of
1707 the existing RO sites, if the RW data hasn't changed since the last
1710 * Make the copyDate field for RO clones have the same meaning as for
1711 remote RO volumes. Previously, the copyDate field for clones was updated
1712 every time we released. (9451)
1714 * Fixed potentially undefined behaviour in ptserver when too many pts
1715 ids are allocated. (10124)
1717 * Note that the server side NAT pings feature present in the prereleases
1718 was removed before the final release, since no positive feedback
1719 was provided during prerelease testing. (9420 10135)
1723 * Start bosserver with -nofork in the systemd unit file, to allow systemd
1724 to track its state (10093)
1726 All client platforms
1728 * No longer track file locks on read-only volumes. Write locks can't
1729 succeed, read locks always will. Avoids log messages about this kind
1732 * Added the "fs flushall" subcommand, which makes the client discard all
1733 cached data. This was previously available on Windows only. (9065 9388
1736 * Fixed a bug that could make the client incorrectly believe its cache
1737 is up to date. This change could negatively impact AFS <-> DFS
1738 translators, should those still be running anywhere. (8898)
1740 * Several changes to avoid panicing in certain error conditions.
1741 (9131 9287 10354 10355 10356 10357) (partially addressing RT #131747)
1743 * Added the -rxmaxfrags switch to afsd, allowing to limit the number
1744 of UDP fragments sent or received per RX packet. (9430)
1746 * Build fixes for aklog on several platforms (RT #131716) (9917 10107 10275)
1748 * Require that the AFS mountpoint specified in the cacheinfo file is
1749 an absolute path. Relative paths result in a client that basically
1750 works but is not fully functional. (10253)
1752 * Fixed a bug that could cause one of the afsd threads to enter an infinite
1753 loop (10431 .. 10436)
1757 * Support Linux kernels up to 3.13 (10241)
1759 * Fixed a bug that made readv/writev calls in AFS space fail with Linux
1760 kernels where generic_file_aio_read exists but those operations have
1761 not been switched to using aio_read/aio_write. This was a regression
1762 introduced with release 1.6.3 and affected at least RHEL 5.9 kernels.
1765 * Fixed a similar bug making core dumps fail in AFS space, affecting
1766 a much wider range of kernels including the most recent ones.
1767 (RT #131729) (10254)
1769 * Enhanced the keyring code to make PAGs work correctly on kernels with a
1770 distribution specific change to the Linux keyring code. This affected at
1771 least SLES 11 SP3 kernels. (10252)
1773 * Fixed a bug that could make failures during PAG instantiation go
1776 * Fixed a bug that made compilation fail for Linux kernels without
1777 keyring support. This affected at least the SLE 10 SDK and an
1778 OEM version of SLES 11 SP1. (10325)
1780 * Fixed build for kernels with user namespace support enabled. Likely
1781 to be required for Ubuntu 14.04 and eventually other distributions.
1782 (10456 10457 10458 10518 10472)
1784 * Support RHEL 6.5 kernels, and possibly others with changes backported
1785 from recent mainline kernels that touch getname/putname, by no longer
1786 using those functions. Previously, the client could cause a kernel
1787 panic when syscall auditing was enabled. (10578)
1789 * Make tmpfs usable as the cache filesystem again. This had been broken
1790 since kernel 3.1 (9950 10193)
1792 * When starting the client fails, clean up the backing device information
1793 created in sysfs, to avoid error messages during a subsequent start
1794 and possible system instability later on (10454)
1796 * Update Red Hat packaging to support Fedora >= 20, RHEL >= 7 and
1797 ELrepo kernels (10597 10619 10622 10703 10704)
1801 * Support OS X 10.9 "Mavericks" (10519 10541 10542 10543 10548 10549)
1805 * Fixed a bug that caused the 1.6 AIX client to never receive any RX
1806 packets in the kernel. (RT #131725)
1810 * Support Solaris 11 (9454 9455)
1812 * Allow other users to access filesystems mounted by root. (9452)
1816 * Build tvolser and dvolser on this platform (10122)
1817 * Several fixes to catch up with newer releases (10374 .. 10381)
1821 * Build tsalvaged, tvolser and dvolser on this platform (10121)
1822 * Fixed build on NetBSD 5 and newer. (10138)
1828 * Fixes for OpenAFS-SA-2013-0003 and OpenAFS-SA-2013-0004
1834 * Obey the jumbo/nojumbo settings for ubik servers (the DB servers)
1835 too. In previous releases, those servers may have used jumbograms
1836 even if they were not configured to do so. This change corrects
1837 the actual behaviour, and will improve performance and reliability
1838 for sites where jumbograms are problematic. It could cause a decrease
1839 in performance for sites where jumbograms work, but those can turn
1840 them back on manually.
1842 * Dozens of fixes for common coding problems like use after free,
1843 use of possibly uninitialised memory, reading or writing past the
1844 end of arrays and potential NULL pointer derefences. Spotted by
1845 code analysis tools or human inspection.
1847 * Documentation improvements.
1849 * Fixes and improvements to the diagnostic or log messages printed by
1850 vos, the fileserver and others.
1852 * Build fixes, making parallel builds more reliable with certain
1853 configuration options and helping various platforms including
1854 recent releases of IRIX, Solaris and several flavours of Linux.
1856 * Avoid sending a small amount of data over the wire unencrypted
1857 under certain conditions, and emit the correct error message in
1860 All server platforms
1862 * Avoid generating duplicate IDs for readonly and backup volumes,
1863 which could happen under certain conditions.
1865 * Allow the fileserver to return volume data like quota or free space,
1866 which is available publicly elsewhere, without the additional access
1867 check for read permissions on a volume's root directory the fileserver
1870 * The fileserver now emits a log message when it ran out of memory for
1873 * Avoid several potential fileserver problems, including memory
1874 corruption and segmentation faults, due to client bookkeeping.
1876 * Avoid known cases of silent data corruption due to background syncs
1877 on the fileserver, especially during Copy on Write.
1879 * Make the fileserver sync behaviour runtime configurable. Up to 1.4.5,
1880 we had synchronous syncs which were safe but really slow. Since 1.4.5,
1881 we've had asynchronous syncs which are much faster but believed to
1882 be the cause of rare data corruption issues, and while all known cases
1883 of these happening are believed to be fixed in the 1.6.3 release, doubts
1884 remain. This change allows choosing between those, and in addition allows
1885 to turn syncs by the fileserver off altogether, thus relying on the vice
1886 partition's backend filesystem and the operating system, or to just
1887 execute them when a volume is detached. The default behaviour is
1888 unchanged from releases since 1.4.5, but it's highly recommended to
1889 consider the additional options this change provides. Future OpenAFS
1890 releases will default to "-sync=none".
1892 * For dbservers, avoid a situation where misinterpreting transient
1893 network errors causes long-term issues with achieving ubik quorum.
1895 All UNIX client platforms
1897 * Improvements to the detection of an aklog-specific krb5 configuration
1898 file, for the purposes of turning on "weak crypto" for aklog.
1900 * Fixed a regression introduced in release 1.6.2 which caused the
1901 supposedly persistent disk cache to be discarded upon client start.
1906 * Support Linux kernels up to 3.10
1908 * Fixed two bugs making it impossible to unmount a disk cache filesystem
1909 after it has been used by the client. (RT #131613)
1911 * Fixed a bug that could cause an oops with kernels 3.6 and later
1915 * Improved support for OpenBSD 4.9 to 5.3
1919 This release number had to be skipped for technical reasons.
1925 * Support Linux kernels up to 3.8.
1927 * Make the init script cope with the output of ifconfig on recent Fedora.
1933 * Fix buffer overflows in fileserver and ptserver.
1935 * Abort an rx connection when given an unknown service (Gerrit 7593).
1937 * "idle dead" behavior improvements.
1939 * Documentation updates.
1941 All server platforms
1943 * Fix rare file corruption during background sync (Gerrit 8796).
1945 * Fix corrupting clients' metadata cache during certain errors (Gerrit
1948 * Avoid saying a volume doesn't exist when accessed as the volume is
1949 going offline (Gerrit 7488).
1951 * Fix fileservers to properly report >2 TiB partitions.
1953 * Fix stale volume info from vos examine on non-DAFS filservers.
1955 * Fix possible volume corruption with vos convertROtoRW.
1957 * Fix bosserver to preserve all command-line options over restart.
1959 * Fix bosserver to properly kill hung processes during shutdown.
1961 All UNIX client platforms
1963 * Fixes for memcache, especially on Solaris.
1965 * Increase the size of the DNS resolver answer buffer to allow sites
1966 with a long response list to use SRV and AFSDB records.
1968 * Fix a crash when a server appears to run out of addresses (Gerrit
1971 * Fix cache corruption when reading from a file another client is
1972 simultaneously writing to (Gerrit 7994).
1974 * Improve handling of disk cache disk errors.
1978 * fix DKMS configuration for DKMS 2.2.
1980 * Avoid generating inode number 0 with md5 inodes (Gerrit 7276).
1982 * Fix a crash when reading /proc/fs/openafs/unixusers (Gerrit 7914).
1984 * Make PAG-less access use the real UID of the calling process
1985 instead of the effective UID, when determining what credentials to
1988 * Fix possible abuse of fs mkmount.
1989 Prior to 1.6.2, users could crash a client by nesting volume mounts.
1991 * Fix fileserver memory corruption on RHEL 6
1992 Prior to 1.6.2, fileservers on RHEL 6 may crash under heavy load.
1994 * Fix client page cache corruption on Linux
1995 When multiple clients read and write to a file, the reading client
1996 may see first page (4096 bytes) of a file as nulls.
1998 * Support Linux kernels up to 3.7.
2000 * Support newer glibc versions.
2002 * Improve client systemd unit file.
2004 * Update Red Hat packaging.
2008 * Fix crashes on shutdown.
2010 * Prevent unloading the module before shutdown completes.
2012 * Security improvement for the OpenAFS preference pane.
2016 * Support newer versions of the Sun Studio compiler software.
2018 * Support compiling on newer versions of Solaris 11 and Solaris 10.
2021 OpenAFS 1.6.0 (2011-08-15)
2025 * Substantial Rx updates to correct erroneous behavior.
2027 * vos now properly deals with matching sites when servers are
2030 * Don't stop Rx keepalives after an ackall is received, avoiding
2031 spurious connection timeouts. (128848)
2033 * Don't retry Rx calls on channels returning busy errors and improve
2034 Rx busy call channel error handling. (128671)
2036 * Properly enable Rx connection hard timeouts.
2038 * Rx NAT pings are not enabled until peer has answered.
2040 * Initialize rx_multi lock before use.
2042 * Avoid spurious crashes when initializing in "backup" client.
2044 * Revert UUID support in vos.
2046 * pt_util fixed to properly create new databases.
2048 * MTU discovery now properly shut down on call reset.
2050 * Avoid leaking references to hosts during callback break multi-Rx
2051 operations. (129376)
2053 * xstat tools now cope with differing timeval structures between
2056 * Numerous fixes to command argument parsing.
2058 * Documentation updates.
2060 All server platforms
2062 * A file descriptor leak which could result in corrupted files in the
2063 fileserver was fixed. An IMMEDIATE upgrade from previous 1.5 release
2064 fileservers is recommended.
2066 * Fix ptserver supergroups support on 64 bit platforms.
2068 * Demand attach salvaging doesn't use freed volume pointers.
2070 * Properly hold host lock during host enumeration in fileserver.
2072 * Attempt to recovery more quickly from timed out volume release
2075 * Auditing now properly byte order swaps IP addresses when printing.
2077 * vos split now has improved error handling.
2079 * Many changes to again support Windows fileservers.
2081 * During volume removal, data removal speed improved.
2083 * Improve CPU utilization during volume attaching by DAFS.
2085 * In salvager check-only mode, avoid potentially fixing a vnode.
2087 * Fix support for large (greater than 2gb) volume special files.
2089 * Salvager will not crash if multiple or bad volume link tables are
2092 * Avoid erroneous full dump by remembering which sites were out of
2093 date at the start of the release.
2095 * A deleted volume can now be recreated properly.
2097 * Callbacks are again not broken during whole partition salvages.
2099 * Positional vectored IO fixed for largefile (>2GB) capable systems.
2101 * Fileserver per-client thread usage again properly enforced.
2103 * Anonymous dropbox support improved and drawbacks documented.
2105 * Demand attach: ensure vnodes are not reallocated while in use due to
2106 volume bitmap errors.
2108 * Properly support large volume numbers (larger than 2147483647).
2110 * Allow salvager to be run manually again when DAFS is being
2113 * Avoid leaking references to hosts during callback break multi-Rx
2114 operations. (129376)
2116 * Demand attach: unlink fileserver state file on standalone salvage.
2118 * Salvager tries harder to detect linktable issues.
2120 * Demand attach: don't attach volumes with special status set.
2122 * Avoid crashing on host table exhaustion. Instead, defer clients.
2126 * afs_config will not longer set the Tray Icon State in the registry
2127 if the checkbox is not present in the dialog. (128591)
2129 * AFS Explorer Shell Extension now works from folder backgrounds.
2130 Overlays for mount points and symlinks are present in the dll, but
2131 are not registered at present by the installers.
2133 * Do not use RankServerInterval registry value as the value for
2134 PerformanceTuningInterval.
2136 * When the data version of a mountpoint or symlink changes, the target
2137 string in the cm_scache_t object must be cleared.
2139 * "fs checkservers" now includes vldb servers in the output and only
2140 lists multi-homed servers once. A multi-homed server that has at
2141 least one up interface is no longer considered to be down.
2143 * When asynchronously storing dirty data buffers to the file server
2144 ensure that (a) the cm_scache_t object and the cm_buf_t object are
2145 for the same File ID so that locking and signalling work properly;
2146 and (b) if the FID no longer exists on the file server, do not
2147 panic, just discard the buffer.
2149 * When processing VNOVOL, VMOVED and VOFFLINE errors perform server
2150 comparisons by UUID or address and not simply by cm_server_t
2151 pointer. Otherwise, server failover may not succeed.
2153 * Do not preserve status information for cm_scache_t objects when the
2154 issuing server is multi-homed.
2156 * Giving up all callbacks when shutting down or suspending the machine
2157 is now significantly faster due to the use of an rx_multi
2158 implementation. (This functionality is still off by default and
2159 must be activated by a registry value.)
2161 * Race conditions were possible when updating the state of the
2162 cm_volume_t flags and when moving the volumes within the least
2165 * Ensure that the lanahelper library does not perform a NCBRESET of
2166 each lan adapter when enumerating the current network bindings.
2167 Correcting this permits OpenAFS to work on Windows 7 when the
2168 network adapter settings change.
2170 * Fix creation of mount points and symlinks as \\AFS\xxxx
2172 * Icon tray state now conditionally set. (128591)
2174 * Properly create new cell mount points in freelance mode.
2176 * Avoid recursive offline volume checks.
2178 * Fix caching of non-existent volumes. The test to trigger an
2179 immediate CM_ERROR_NOSUCHVOLUME in cm_UpdateVolumeLocation() was
2182 * Prevent the background daemon from checking the status of
2183 non-existent volumes. cm_CheckOfflineVolumes() should skip volume
2184 groups with the CM_VOLUMEFLAG_NOEXIST flag set.
2186 * The afskfw library should return an error immediately if the
2187 krb5_32.dll library cannot be loaded. Affects afslogon.dll and
2190 * No longer depend on leashw32.dll in afskfw library.
2192 * NPLogonNotify must provide the user password in all calls to
2193 KFW_AFS_get_cred(). It cannot count on a credential cache being
2194 preserved between calls. Permits tokens to be acquired for all
2195 cells listed in the TheseCells registry value for a domain.
2197 * Improve the trace logging from NPLogonNotify().
2199 * Avoid a race when writing the cm_scache_t mountPointString
2200 when acquiring mount point or symlink target data via
2201 cm_GetData(). The race could result in bogus target
2204 * Permit the use of des-cbc-md5 and des-cbc-md4 enctypes
2205 as DES keys in asetkey.exe.
2207 * aklog supports dotted Kerberos v5 principal names.
2209 * afskfw library always attempts afs/cell@USER-REALM
2211 * afskfw library must test return code from krb5_cc_start_seq_get() or
2212 will trigger a null pointer exception when using Heimdal.
2214 * Lock protected fields must be 32-bit in order to avoid memory
2217 * Add support for NTFS symlinks.
2219 * Handle file search requests for virtual syscall ioctl file.
2221 * Process SyncOps properly to enforce ordered operations.
2223 * Avoid recursing during NewServer operations.
2225 * Correct lock acquisition order during SMB locking.
2227 * Add shutdown message to event log.
2229 * Check offline volume status by policy rather than on each daemon
2232 * Return error on directory object not found instead of crashing.
2234 * Improve error message output.
2236 * afslogin.dll can start afsd_service if it's not starting or started.
2238 * Optimize away release lock RPCs for deleted files.
2240 * Background Daemon will not perform operations on deleted files.
2242 * Resort recently used directories to the top of the LRU if the
2243 directory is larger than the stat cache.
2245 * Resort deleted objects to the bottom of the LRU.
2247 * Use interlocked operations for state and queue fields to allow safe
2248 bit set and clear on multiprocessor systems.
2250 All UNIX client platforms
2252 * Servers now marked down when GetCapabilities returns error.
2254 * In-use vcache count is now properly tracked.
2256 * Check for /afs existance before starting, unless -nomount is
2259 * Avoid a potential panic when using /afs/.:mount syntax.
2261 * Avoid a panic in memcache mode due to missing CellItems file.
2263 * FUSE client support fixed for non-/afs mounts.
2265 * Avoid a potential deadlock (which times out) when we need to
2266 allocate more callback returns and must flush some already in use.
2268 * Deal with libcom_err conflicts with other packages using it
2269 (e.g. krb5) (128640)
2271 * Fall back to afs3-vlserver SRV record values when afs3-ptserver SRV
2272 record is not available.
2274 * Avoid holding unneeded locks when probing server capabilties.
2276 * Do not attempt page flushes for directories.
2278 * Rx connection reference counting is enabled.
2280 * An Rx connection reference count leak is fixed in bulkstat.
2282 * Handle unparsable directory objects.
2284 * Handle Kerberos cred cache errors in aklog.
2288 * Fix PAG usage to track by PAG identifier, not group list.
2292 * Fix socket termination on shutdown.
2294 * Support for 7.2, 7.3, 7.4 and 8.2 included.
2296 * References to vcaches are no longer leaked during root or reclaim.
2298 * Remove support for "Giant" lock as we no longer need to use it.
2300 * Don't sleep with AFS GLOCK.
2302 * Properly enable 64 bit long long support.
2304 * Restore support for FreeBSD 7 (128612)
2306 * Fix locking issues at shutdown and avoid panic at shutdown due to
2309 * Support for virtual network stacks.
2311 * New RC script, updated packaging.
2315 * Properly create new vnodes to avoid crashing in the client.
2319 * Support through kernel 2.6.39. Treat Linux 3.0 as Linux 2.6 for
2322 * Use rx_Readv in cache bypass to improve performance.
2324 * Properly handle 0-length replies during cache bypass operations.
2326 * Properly handle non-contiguous readpage cache bypass operations.
2328 * Do proper locking when transitioning to or from cache bypass.
2330 * Avoid extra runs of vcache freeing routine. (128756)
2332 * Perform vcache eviction via a fast path before visiting vcaches
2333 where sleep is needed.
2335 * setpag() errors are now properly reported.
2337 * Avoid attempting to free stat cache entries when we are below
2338 user-specified number of entries in use.
2340 * Properly track user-specified number of stat cache entries to use as
2341 a desired usage target.
2343 * Don't read pages beyond EOF in the cache. (128452)
2345 * Various corrections and improvements to Red Hat packaging, including
2346 modifying the init script to allow deferring for a new binary
2347 restart and properly supporting RHEL6.
2349 * Fix lockup in 2.6.38 due to erroneous kernel feature configure test.
2351 * Improve RPM building tools.
2353 * Attempt to properly handle SELinux in packaging.
2355 * Init script properly returns status as exit code.
2357 * RPM packaging fixes (executable libraries, no postinstall message)
2359 * Kill i386 from RPM packaging.
2363 * MacOS 10.7 support.
2365 * Properly handle setpag errors. PAGs are not supported.
2367 * Check for unloaded kernel extensions when decoding AFS panics.
2369 * Disable "get tokens at login" in prefs pane if AD authentication
2370 plugin is configured.
2372 * aklog AuthorizationPlugin now provided.
2374 * Preferences Pane behavior fixed for 1.6 series (version detection is
2375 used to select default behavior).
2377 * A potential kernel panic during bulkstat operations is
2380 * 64-bit MacOS kernel performance is greatly improved. (128934)
2382 * Properly shut down AFS, closing the Rx socket in the upcall handler
2383 to avoid attempting to process data after we can no longer do so.
2385 * Rework logic for bulk status operations to avoid a potential hang.
2387 * Avoid panic when doing FSEvent synthesis.
2389 * Fix bug when using non-dynroot.
2391 * Update Kerberos support in PreferencesPane.
2395 * Updates for platform support.
2399 * Bug fixes for issues introduced previously in 1.5 series.
2401 * Support through OpenBSD 4.8.
2405 * Switch to ioctl() syscall replacement for Solaris 11 since syscall
2408 * Fix support for Solaris pre-10.
2410 * Corrected Solaris 11 startup script.
2412 * vcache mappings freed on shutdown to avoid panic.
2414 * Properly report errors for AFS system call callers.
2416 * Don't leave dangling function references if kernel extension fails
2419 * Try harder to avoid deadlocks on file-larger-than-cache operations.
2421 * Avoid panic on shutdown when mount failed.
2426 All systems: Minor bugfixes.
2428 ADDITIONAL CHANGES IN 1.6.0PRE2
2432 - Documentation updates.
2434 - Don't stop Rx keepalives after an ackall is received, avoiding
2435 spurious connection timeouts. (128848)
2437 - Don't retry Rx calls on channels returning busy errors. (128671)
2439 - vos will not die with a double free error at command completion.
2441 - Properly enable Rx connection hard timeouts.
2443 - Initialize rx_multi lock before use.
2445 - Avoid spurious crashes when initializing in "backup" client.
2449 - Check for /afs existance before starting, unless -nomount is specified.
2451 - Avoid a potential panic when using /afs/.:mount syntax.
2453 - Avoid a panic in memcache mode due to missing CellItems file.
2455 All server platforms:
2457 - Attempt to recovery more quickly from timed out volume release
2460 - Auditing now properly byte order swaps IP addresses when printing.
2462 - vos split now has improved error handling.
2464 - Many changes to again support Windows fileservers.
2466 - During volume removal, data removal speed improved.
2468 - Improve CPU utilization during volume attaching by DAFS.
2470 - In salvager check-only mode, avoid potentially fixing a vnode.
2472 - Fix support for large (greater than 2gb) volume special files.
2474 - Salvager will not crash if multiple or bad volume link tables
2477 - Avoid erroneous full dump by remembering which sites were out of date
2478 at the start of the release.
2482 - Remove support for "Giant" lock as we no longer need to use it.
2484 - Don't sleep with AFS GLOCK.
2486 - Properly enable 64 bit long long support.
2488 - Restore support for FreeBSD 7 (128612)
2490 - Fix locking issues at shutdown.
2494 - support through kernel 2.6.38.
2496 - RedHat packaging now properly supports RHEL6.
2498 - Use rx_Readv in cache bypass to improve performance.
2500 - Properly handle 0-length replies during cache bypass operations.
2502 - Properly handle non-contiguous readpage cache bypass operations.
2504 - Do proper locking when transitioning to or from cache bypass.
2506 - Avoid extra runs of vcache freeing routine. (128756)
2510 - Check for unloaded kernel extensions when decoding AFS panics.
2512 - Properly handle setpag errors. PAGs are not supported.
2514 - Disable "get tokens at login" in prefs pane if AD authentication
2515 plugin is configured.
2519 - support through OpenBSD 4.8.
2523 - Fix support for Solaris pre-10.
2527 - afs_config will not longer set the Tray Icon State
2528 in the registry if the checkbox is not present in
2529 the dialog. (128591)
2531 - AFS Explorer Shell Extension now works from folder
2532 backgrounds. Overlays for mount points and symlinks
2533 are present in the dll, but are not registered at present
2536 - Do not use RankServerInterval registry value as the value for
2537 PerformanceTuningInterval.
2539 - When the data version of a mountpoint or symlink changes,
2540 the target string in the cm_scache_t object must be cleared.
2542 - "fs checkservers" now includes vldb servers in the output
2543 and only lists multi-homed servers once. A multi-homed
2544 server that has at least one up interface is no longer
2545 considered to be down.
2547 - When asynchronously storing dirty data buffers to the
2548 file server ensure that (a) the cm_scache_t object and
2549 the cm_buf_t object are for the same File ID so that
2550 locking and signalling work properly; and (b) if the
2551 FID no longer exists on the file server, do not panic,
2552 just discard the buffer.
2554 - When processing VNOVOL, VMOVED and VOFFLINE errors perform
2555 server comparisons by UUID or address and not simply by
2556 cm_server_t pointer. Otherwise, server failover may not
2559 - Do not preserve status information for cm_scache_t objects
2560 when the issuing server is multi-homed.
2562 - Giving up all callbacks when shutting down or suspending
2563 the machine is now significantly faster due to the use
2564 of an rx_multi implementation. (This functionality is
2565 still off by default and must be activated by a registry
2568 - Race conditions were possible when updating the state
2569 of the cm_volume_t flags and when moving the volumes
2570 within the least recently used list.
2572 - Ensure that the lanahelper library does not perform a
2573 NCBRESET of each lan adapter when enumerating the
2574 current network bindings. Correcting this permits OpenAFS
2575 to work on Windows 7 when the network adapter settings
2578 - Fix creation of mount points and symlinks as \\AFS\xxxx
2584 - vos now properly deals with matching sites when servers are multihomed.
2588 - Servers now marked down when GetCapabilities returns error.
2590 - In-use vcache count is now properly tracked.
2592 All server platforms:
2594 - Fix ptserver supergroups support on 64 bit platforms.
2596 - Demand attach salvaging doesn't use freed volume pointers.
2598 - Properly hold host lock during host enumeration in fileserver.
2602 - Fix socket termination on shutdown.
2604 - Support for 7.2, 7.3, 7.4 and 8.2 included.
2606 - References to vcaches are no longer leaked during root or reclaim.
2610 - Define llseek handler to avoid ESPIPE error in 2.6.37.
2612 - Mount interface replaces get_sb (new for 2.6.37, not yet required).
2614 - RedHat init script allows deferring for a new binary restart.
2616 - DEFINE_MUTEX replaces DECLARE_MUTEX for 2.6.37.
2620 - Correct return value from setpag syscall.
2624 - Bug fixes for issues introduced previously in 1.5 series.
2628 - Switch to ioctl() syscall replacement for Solaris 11 since syscall 65
2632 OpenAFS 1.5.78 (2010-11-04)
2636 * Revisions to Rx to fix performance issues.
2638 * Make fs getfid behave consistently across all platforms. (128372)
2640 * Properly check IDs handed to pts when creating users or groups so
2641 useful error messages can be provided. (128343)
2643 * Correct byte order handling of port in afsconf_LookupServer for SRV
2646 * Force a full dump when releasing to a site which was previously
2647 marked "don't use", in case the previous clone was out of date.
2649 All server platforms
2651 * Demand salvage of attached volumes now correctly track attachment
2654 * Avoid a potential crash due to failure to hold a lock when attaching
2659 * Track SMB connections by SID rather than username.
2661 * Error write attempts to known-readonly volumes earlier.
2663 * Validate directory buffers to avoid potential crashes.
2665 * Handle VIO errors from bulkstatus.
2667 * Make PMTU discovery configurable and register error handlers for it.
2669 All UNIX client platforms
2671 * Use larger I/O sizes in memcache to improve performance.
2673 * Avoid potential alignment issues doing I/O for pioctl calls.
2677 * Avoid panicing if the listener process is not findable.
2679 * Avoid deadlock issues while performing lookups.
2683 * Handle stale file handle errors for some cache partition types.
2685 * Avoid blocking with xvcache lock when attempting to free in-use
2688 * Build fixes for older kernels.
2690 * Properly configure LWP to use ucontext() on platforms where it
2693 * Eliminate spurious errors from AFS system call returns. (126230)
2697 * Attempt to honor configured Kerberos defaults in Preferences Pane.
2700 OpenAFS 1.5.77 (2010-09-08)
2704 * Rx path MTU detection will terminate detection in cases where the
2705 minimum required packet size cannot be transferred.
2707 * vos dryrun mode now shows effects for syncvldb single volume case.
2709 * vos dryrun mode now shows "status after" for syncvldb and syncserv.
2711 All server platforms
2713 * RXAFS_GetStatistics64 now returns statistics properly.
2717 * Attempt to properly identify the local system SMB connection for
2720 * Remap timeout and offline errors to proper NT RPC errors.
2722 * Properly fail over to other replicas on bulkstat IO errors.
2724 * Properly error delete-mode createfile if a file is set readonly.
2726 * Validate directory entry buffers to avoid crashing the service.
2728 * Log file modes properly.
2730 * Log cell name when logging server information.
2732 All UNIX client platforms
2734 * cacheout program for discarding callbacks is now built.
2736 * bulkstatus kernel locking is corrected to avoid a potential panic.
2740 * userspace support update
2744 * Updated vnode locking for children returned via lookup().
2746 * Avoid file open undercount with needed calls to
2747 FakeOpen/FakeClose().
2749 * Use vnode_pager_setsize to properly track file size during kernel
2752 * Update system call installation.
2754 * Fix shutdown of Rx kernel listener to avoid potential dereference
2757 * Avoid closing vnodes during vnode recycle.
2759 * Fix bogus call to FlushVS for vnode reclaims.
2763 * Packaging updated for current configure options and built files.
2765 * Cache bypass now holds reference on pages during readpage.
2767 * s390x setgroups32 patching update.
2771 * DNS resolver is reinitialized on IP address change. (126440)
2774 OpenAFS 1.5.76 (2010-08-16)
2778 * Updates to build-time configuration.
2780 * Fix XDR support in Rx to match header definition.
2782 * vos status now shows transaction creation, not action creation.
2784 * Rx avoids reporting loopback adapters when listing interfaces.
2786 All server platforms
2788 * Demand-Attach Fileserver always built and installed (dafileserver,
2789 davolserver, dasalvager).
2791 * Return VNOVOL from fileserver when a volume is deleted.
2793 * Ignore duplicate tags during volume restore operation.
2795 * Update inode array after salvage repairs volume.
2797 * Zero a corrupted header in memory during salvage to avoid further
2800 * Fix NAMEI backend to allow low-numbered volumes to work properly.
2802 * ptserver does not include cell name as part of length check for
2805 * Updated error messages for unblessed volumes.
2807 * vlserver avoids buffer overflow with regex pattern
2809 * Attach-time failures now note failures as the rest of the fileserver
2812 * Server argument logging will no longer overflow stack.
2814 * Provide fast-restart-like unsafe-nosalvage option for DAFS.
2816 * Deal with host hash collisions in the fileserver.
2820 * Avoid crashing when interpreting a drive letter as potentially
2821 matching a cell name.
2823 * Properly handle volume package errors.
2825 * Allow page recycling from known-readonly content without ensuring
2828 * 32 bit tools installer should not override client configuration.
2830 * Ensure root scache item has a valid callback when use is attempted.
2832 * Freelance directory changes now properly invalidate and replace the
2835 All UNIX client platforms
2837 * Support disconnected reconnecting with specified UID for PAGless
2840 * Proper disconnected vnode reference tracking.
2842 * Update server site blacklisting to not return success if nothing was
2845 * Avoid a panic during vcache contention due to CVInit vcache
2850 * Update for network stack in 8.1/9.0.
2860 * Disable PMTU error packet handling.
2864 * Debian packaging updated.
2866 * freezer interface updates.
2870 * Hold references to disconnected mode written vnodes properly.
2874 * Handle NFS translator module references for amd64.
2876 * INODE fileserver backend support now exists for amd64.
2879 OpenAFS 1.5.75 (2010-07-07)
2883 * Prevent rx_rpc_stats global lock from being a bottleneck.
2885 * Path MTU discovery is now provided to allow traffic to pass networks
2886 with sub-1500 byte MTUs and poor fragment handling.
2888 * Further reduce Rx NAT ping transmission when enabled.
2890 * Update Kerberos 5-based token handling in rxkad from upstream
2893 * Update version numbers emitted during build to reflect what is
2894 actually being built.
2896 * Add "-human" switch for human-readable units in fs diskfree and
2899 * vos provides reasons for locked volumes when known.
2901 * Do not count retransmission and ping acks as non-idle for Rx
2904 * Rx: provide service-specific data getter and setter routines.
2906 * Update build-time Kerberos detection.
2908 * Updated userspace AFS client.
2910 * Beginning of a modernized test suite.
2912 * Additional documentation.
2914 * Updated documentation, notably the Administrators Guide.
2916 * Substantial code cleanup.
2918 All server platforms
2920 * Update handling of vnode allocation failures.
2922 * DAFS: allow salvaging volumes not known to the fileserver, to allow
2923 cleanup of data not attached to a current volume.
2925 * Properly handle volumes slated for destruction.
2927 * Handle volumes with many files properly.
2929 * Force core file generation in bosserver by overriding default
2930 resource limits when possible.
2932 * Update vlclient and vldb_check.
2934 * Avoid potentially corrupting a volume on creation if files are left
2935 from previous failed cleanup.
2937 * Note volume changed during salvage as needed.
2939 * DAFS: do not assume invalid addresses are in fileserver address hash
2942 * Avoid tying up fileserver threads with volumes that are being taken
2945 * Do not set inUse on volumes for non-DAFS other than in fileserver.
2947 * Break origin's callback on target of rename operation.
2949 * Avoid unneeded parent directory link updates during some rename
2952 * Do not open /dev/console for writing in the fileserver.
2954 * DAFS: avoid spurious restarts when binary restarts are configured.
2956 * Avoid spurious and unneeded calls to sync(), which can slow down the
2961 * Revised SMB QuerySecurityInfo to address issues caused by MS10-020
2962 (http://support.microsoft.com/kb/980232)
2964 * Prevent use of the AFSCache file contents if mapped to a new
2967 * Make fs newcell include behavior compatible with the non-Windows
2970 * Provide a registry option (FreelanceImportCellServDB) to pre-create
2971 mount points in the AFS root for all cells in CellServDB.
2973 * Fix a memory leak in the cm_FreeServerList() routine.
2975 * Reduce privilege when reading registry CellServDB.
2977 * Add support for RPC Pipe Service NetWkstaGetInfo levels needed for
2980 * Prevent overflow when computing quota percentage in Explorer Shell.
2983 * Generate meaningful errors for ACL operations on freelance AFS root.
2985 * Fix error handling on InlineBulkStatus RPCs.
2987 * Show configuration pages for all types of MSI installations.
2989 * Improve freemount AFS root directory handling and operations.
2991 * Properly validate GetVolumeStatus pioctl responses.
2993 * Commit file length changes and dirty buffers when flushing a file.
2995 All UNIX client platforms
2997 * Update version of files for disk cache.
2999 * Do not call afs_FlushVCBs with xvcache lock held, to improve
3002 * Add mariner log messages for creating and removing files.
3004 * Don't hold xvcache lock while creating symlinks, to improve
3007 * Provide -dynroot-sparse mode to not show all cells in CellServDB in
3010 * Avoid a potential crash in aklog in linked cell handling.
3012 * Log MTU-caused packet retransmission.
3014 * Prevent crashes caused be fs checkservers while cache is being set
3017 * fs getserverprefs now has a buffer large enough for the default
3020 * Report server address when logging warnings.
3022 * Avoid panic in GetCapabilities when cell is not known.
3024 * Lock process name and id for advisory lock warnings when possible.
3026 * Handle need for allocating additional Rx packets.
3028 * Properly handle errors from InlineBulkStatus operations.
3030 * Fix errors returned from fcntl() on readonly files locked for write.
3032 * Flush pending changes to the server on LOCK_EX unlock.
3034 * Reflect length changes as a result of callbacks even when file is
3037 * Avoid hanging due to error exit when attempting to store a large
3038 file to a non-largefile fileserver.
3040 * Recover from afs_GetVolSlot errors.
3044 * Bugfixes for kernel VFS and network routines.
3048 * Provide makesname().
3052 * Avoid syscall probes when keyrings are present, by default. (125215)
3054 * Remove "Big Kernel Lock" from VFS operations.
3056 * Use filehandles for all Linux 2.6 versions to avoid need for matched
3059 * Updated RPM packaging.
3061 * Fix dkms configuration provided with RPMs.
3063 * Hold reference on pages during background I/O for cache bypass.
3065 * Fix cache bypass handling of non-largefile fileservers.
3067 * Protect truncate_inode_pages mappings with mutex or semaphore as
3070 * Fix pagevec use in cache bypass. (127505)
3072 * Updates for 2.6.35
3076 * Improve launchd configuration.
3078 * Avoid hanging on recursive cache file lock acquisition when user
3079 notification is enabled.
3081 * Fix and re-enable bulkstat mode.
3089 * Precluding unmount while AFS is busy.
3091 * Avoid deadlocking when releasing the VFS object.
3093 * Stop network interface poller in kernel on AFS shutdown.
3095 * Avoid issues with lookups on empty directory names. (127356)
3098 OpenAFS 1.5.74 (2010-04-22)
3102 * Add "vos setaddrs" command.
3104 * Rx library lock contention avoidance between rx_NewCall and
3107 * Rx library races due to inconsistent use of rx_connection
3108 conn_data_lock to protect the flags field.
3110 * Rx library inconsistent use of RX_CALL_TQ_WAIT which could result in
3113 * Rx library must signal transmit queue waiters when flushing.
3115 * afsmonitor shows busy counts now.
3117 * afsmonitor displays xstat callback statistics.
3119 * Provide expandgroups for pts mem on a supergroups server.
3121 * Provide supergroup option to liste nested groups during pts mem.
3123 All server platforms
3125 * Avoid volume lock contention during DAFS startup.
3129 * Avoid a race when updating cell vldb server lists that can result in
3132 * Avoid a deadlock when managing CM_SCACHESYNC_STOREDATA state
3133 operations for directory objects.
3135 * Add new Windows Application Event log messages for VBUSY,
3136 VRESTARTING, ALL_BUSY, ALL_OFFLINE, and ALL_DOWN.
3138 * Reduce lock contention by waiting for cm_buf_t I/O operations.
3140 * Split the cm_buf_t flags field to separate the flags that are
3141 protected by the cm_buf_t mutex from those protected by the
3144 * In cm_UpdateVolumeLocation, avoid searching for a ".readonly" volume
3145 on a numeric volume name.
3147 * File buffer allocations whose offsets are beyond server EOF should
3148 be locally allocated and zero filled. The file server should not be
3149 issued a FetchData rpc which is guaranteed to fail.
3151 * Enable integrated logon to work with Windows 7/2008 when user logons
3152 are performed with a non-Domain Kerberos principal.
3154 * Add Protection Error messages to aklog output.
3156 All UNIX client platforms
3158 * Provide a FUSE-interfacing userspace afs client.
3160 * Updates to libuafs userspace cache manager.
3162 * Probe servers using GetCapabilities instead of GetTime, thus
3163 requiring fewer RPCs.
3165 * Fix DNS SRV record handling for cell lookup.
3169 * Fix sleep/wakeup routines.
3171 * Update for 8.0 release.
3175 * Handle high memory addresses correctly.
3179 * Make 32 bit AFS syscalls work again.
3181 * Work around finder "Duplicate" failure (caused by setting modes on
3184 * Disable bulkstat again (will be re-enabled at or before .75).
3186 * Provide symlink type hints during readdir.
3189 OpenAFS 1.5.73 (2010-03-24)
3191 All systems: Minor bugfixes. New features.
3193 * New functionality:
3197 - NAT keepalive support at Rx level.
3203 - Corrected server IP address output in vos syncvldb verbose mode.
3205 - Corrected server IP address output for last "yes" host in udebug.
3207 - Corrected SRV record support for canonicalizing cell names.
3211 - Fixed a potential race in Disconnected AFS "remove" support.
3213 - Fix a potential blocking condition in fakestat mode.
3215 - Avoid some errors and stack overflow reports when vos is interrupted.
3218 - Clean up several minor memory leaks.
3220 - If a large file is stored to a non-largefile fileserver, avoid
3221 a potential deadlock.
3223 - Increase maximum number of sysnames to 32.
3225 - Readd fs mariner "storing" message, missing since AFS 3.3.
3227 - Attempt timeouts on AFSDB lookups in userspace.
3229 - Avoid interrupting writes due to an idle deadtime timeout.
3231 All server platforms:
3233 - Properly notify only affected hosts for volume callbacks. (126497)
3235 - Allow volumes with trashed root directory to be recovered. (94658)
3237 - Hold lock in file and volservers when traversing partition list.
3239 - Use finer-grained locking in DAFS: volume, instead of partition locks.
3241 - Schedule all DAFS salvages via FSSYNC.
3243 - Avoid stale ptserver credential caching issue on keyfile update.
3245 - Improve callback table overflow handling. (126451)
3247 - Preclude deadlocks on when attempting to save DAFS state.
3249 - Avoid races deleting hosts. (126454)
3251 - Improve salvage speed for DAFS (124488)
3253 - The bosserver now handles SIGTERM.
3257 - Prevent the Explorer Shell extension from crashing if symlink
3258 creation failed. (126406)
3260 - A Rx level NAT ping has been implemented. A registry value enables.
3262 - Adds krb5 error message translation to aklog, afscreds,
3263 afslogon.dll, the network identity manager afs provider and
3266 - Default mode bit settings for file and directory creation are now
3267 provided, and can be configured.
3269 - An SMB request trace facility is provided and can be enabled for
3274 - Clean up properly on mount failure.
3276 - Add entry to /etc/vfs to allow umount to work.
3280 - Additional work to support FreeBSD 8-current.
3284 - Fix build issues with library order.
3288 - Fix s390 support conflict with executable stack patches.
3290 - Don't count root's AFS session keyrings against quota.
3292 - Correct dkms support in RPM config file.
3294 - Keyring destructor now properly cleans up all tokens.
3296 - Build again on old 2.6 kernels.
3298 - Avoid GLOCK leak when updating CellServDB in-core.
3300 - Fix byte-range lock handling.
3302 - Attempt to deal with bdi issues. (126514)
3306 - Some FSEvents hinting for authentication events now done. (23781)
3308 - Update uninstaller. (125634)
3310 - Rewrite afssettings and fstab code to avoid licensing issue with APSL.
3312 - Growl client for user monitoring of AFS events included.
3314 - Properly support insert-only dropboxes.
3316 - Add bulkstat support.
3318 - Include support for moving in Finder across mount points.
3320 - Preferences Pane includes support for Kerberos 5 ticket renewal.
3324 - Some support for OpenBSD 4.7.
3327 OpenAFS 1.5.72 (2010-02-15)
3331 * Provide internationalization support in com_err.
3333 * Fix array length checking to avoid crashes when checking for a
3334 volume type based on name in vos.
3336 All server platforms
3338 * Provide backward compatible "-f" flag to salvager for force mode.
3342 * Restore use of DNS AFSDB and SRV records by kaserver clients.
3344 All UNIX client platforms
3346 * Fix client cache file truncation to not lose chunks when truncating
3349 * Ensure a cache writeback hook is installed in the client (bug from
3352 * Avoid spurious free memory warnings during clean shutdown.
3354 * Fakestat mode avoids AFSDB lookups.
3356 * "fs storebehind" now correctly reports errors on readonly volumes.
3358 * Additional documentation for "fs getcacheparms"
3360 * Forced new uuid generation with "fs uuid -generate" now works
3361 enforced permission correctly.
3365 * Add optimized Rx event handler in kernel.
3367 * Installer now allows installing an older version.
3369 * Panic decoder can now deal with MacOS 10.5 again.
3371 * MacOS ._ files are now correctly not looked up as cellnames.
3375 * To deal with SELinux file labeling, try cache accesses with current
3376 credentials in event of failure.
3378 * Rx XDR encoding bug on i386 Linux is fixed (bug introduced in
3383 * Code compilation fixes.
3387 * Update for OpenBSD 4.6.
3390 OpenAFS 1.5.69 (2010-01-19)
3394 * Configuration of BOSserver no longer defaults to weekly restarts
3397 * Provide BOS restricted mode by default.
3399 * Add support for "vos endtrans" command.
3401 * Default to providing full output from vos listvol.
3403 * Correct additional-address tracking in the fileserver.
3405 * Improve Rx performance by not unnecessarily dropping and reacquiring
3406 call locks in read and write processes.
3408 * Avoid crashes when monitoring volserver transactions across
3409 potential transaction garbage collection.
3411 * Numerous warning fixes.
3413 All server platforms
3415 * Avoid saving fileserver state in demand attach fileserver when
3418 * Demand attach fileserver allows other callers to schedule salvages.
3420 * Demand attach "bos salvage" now works correctly with restricted
3425 * Numerous changes to the client-internal btree directory handling to
3428 * fs examine reports owner and group ids as signed values (PTS groups
3431 * Preclude corruption due to races writing to smb buffers.
3433 * Allow MTU settings in registry to be used.
3435 * Apply MTU to both send and receive sizes.
3437 All UNIX client platforms
3439 * Avoid double-freeing Rx call structure if reading a response from
3440 the file server results in a short read.
3442 * Handle negative lengths in FetchStatus results correctly.
3444 * Properly clean up allocated memory at shutdown.
3446 * Default to AFSDB compiled into the cache manager.
3448 * Avoid inadvertant disclosure of stat() information to clients not so
3451 * Correct a bug with AFSDB lookups introduced with SRV record support.
3455 * Install kernel panic processing tool in /Library/OpenAFS/Tools.
3457 * Include debugging symbols for kernel extension in additional package.
3459 * Support "Application Firewall" users.
3461 * Avoid ._cellname AFSDB lookups.
3463 * Compile preferences pane as a universal binary.
3467 * Use splice to speed up storing files.
3469 * When using memcache, avoid duplicating work in readpages.
3471 * Use dget_parent to safely find an inode's parent.
3473 * Disable access time updates in our superblock.
3475 * Avoid crashing doing writeback if no credentials were stashed at
3478 * Simplify keyring support.
3480 * Properly clean up vcache in event of failed mount.
3484 * Update for current FreeBSD 8.
3488 * Abstractly manipulate groups as now required.
3490 * Abstractly access time instead of using lbolt directly.
3493 OpenAFS 1.5.68 (2009-12-08)
3497 * aklog now attempts to convert non-AFS errors to human-readable
3500 * Make stack not executable when compiling assembler source with GCC.
3502 * Numerous source warning cleanups and code reorganization.
3504 All server platforms
3506 * Compute midnight for volume statistics calculation from local time.
3508 * Salvager now orphans duplicate special inodes when running to allow
3509 recovery in event of a problem, instead of simply ignoring the
3512 * Support to ensure a server panic attempt leaves a core and thus
3513 restarts in a timely manner, rather than potentially hanging. Use
3514 panic to attempt cleanup before leaving a core when possible.
3516 * Volume sync data reported during bulkstatus is now set correctly.
3518 * Provide better tuning for fileserver file descriptor caching.
3520 * Allow more than 128 threads in fileserver by modifying host
3521 structure in-use tracking.
3523 * Avoid crashes getting volume server status during transaction
3526 * Improved logging of offline volume conditions.
3528 * Correct volume statistics when cloning a volume.
3530 * Avoid referencing host structures in the fileserver which are marked
3533 * Demand attach fileserver corrections to avoid coring during an
3536 * host array bounds checking corrections to avoid buffer overflow.
3538 * Handle special inodes correctly when promoting an inode fileserver
3539 readonly volume to read-write.
3543 * Set the DOS Readonly attribute on a file/directory whenever the unix
3544 mode combined with the mask 0200 is true. Previously there was a
3545 discrepency between the mask used for testing for readonly behavior
3546 and that used for setting the attribute.
3548 * Disable AFSVolSync based .readonly "whole-volume callback" support
3549 because the all file servers prior to 1.5.67 (and perhaps 1.4.12) do
3550 not properly assign a value to the AFSVolSync structure in bulk
3551 status RPC responses.
3553 * Improve the error output from aklog to output the value from krb5
3554 error_message() if the afs_com_err output indicates an unknown
3557 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and do not permit
3558 them to be exposed to the smb redirector.
3560 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT to avoid
3561 confusion within the smb redirector.
3563 * Fix the byte order assigned to port numbers associated with AFSDB
3564 record lookups. They must be network byte order not host byte
3567 * Add dynamic server ranking based on RPC round trip time
3570 All UNIX client platforms
3572 * Additional shutdown-time memory leaks removed.
3574 * Improved logging of resource contention.
3576 * Provide dumping for Rx debug packet tracking support in source.
3578 * Update afscp test client to build, and provide an unlock client.
3580 * Client buffers for directory parsing can now be allocated beyond the
3581 fixed set formerly provided.
3583 * Work around race condition when manipulating read-only volume
3586 * Bugfixes to get PAG value pioctl.
3588 * Bugfixes to SRV record support.
3592 * Path MTU tracking code cleanup.
3594 * Avoid an oops due to racing with vcache recycling thread.
3596 * Changes to keyring PAG handling: for sufficiently new kernels, use
3597 only keyring-based PAGs, and disable group PAGs entirely.
3599 * Updates to the kernel page cache interface: writing pages will now
3600 not spuriously leak page locks, and will avoid requiring duplicate
3603 * Credential references are now tracked using native atomic counters.
3605 * Kernel mutex/semaphore lock ordering fix to avoid deadlocks.
3607 * Manipulate disk cache with credentials used to initialize it, to
3608 avoid security issues.
3612 * Fix fstrace message catalog location.
3614 * Fix kernel fstrace logging.
3617 OpenAFS 1.5.66 (2009-10-25)
3621 * Avoid calling exit() in library code.
3623 * Add rx window size and peer timeout tuning APIs.
3625 * Correct rx peer timeout handling to disallow 0ms timeouts.
3627 * Correct calculation of rx RTT by disregarding retransmitted packets.
3629 * vos manpages updated to reflect changes in recent versions.
3631 * GNU-style long options (e.g. --cell) are now supported in all
3634 * fs listacl can now print a command to recreate the current ACL.
3636 All server platforms
3638 * Fix a race on transaction objects in the volserver which can cause a
3641 * Avoid destroying and setting to NULL the callback connection when it
3642 could still be being used.
3644 * Correct unlink handling in salvager.
3646 * Improve error messages due to I/O errors in the volserver.
3648 * Correct an issue which caused converted RO to RW volumes on namei
3649 fileservers to not come online immediately.
3653 * Official support for Windows 7 and Server 2008 R2.
3655 * Prevent a file server bug (FetchData returning an invalid length
3656 instead of zero) from causing an "unexpected network error" when
3659 * Promote DNS SRV records as superior to DNS AFSDB records. Support
3660 arbitrary port numbers for vldb servers.
3662 * Add AFSVolSync based .readonly "whole-volume callback" support.
3663 With this functionality, multiple objects from a .readonly volume
3664 can have their status validated by issuing a single
3665 RXAFS_FetchStatus RPC.
3667 * Remove drive mapping functionality and service start/stop from
3670 * Remove drive mapping functionality from afs_config.exe.
3672 * Use {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab" to
3673 restore access to drive mapping functionality in afscreds.exe and
3676 * Adjust SMB error return codes to avoid returning STATUS_TIMEOUT
3677 which results in the SMB redirector disconnecting.
3679 * Network Identity Manager OpenAFS Provider now provides its own "AFS
3680 lock" notification icon to report the status of "have tokens, have
3681 no tokens, service not started, service started but inaccessible".
3682 Hovering over the icon lists the cells for which tokens exist (if
3683 any) and the OpenAFS version number. Double-clicking executes the
3684 Network Identity Manager default action.
3686 * Prevent pioctl calls from retrying indefinitely when a sharing
3687 violation error occurs.
3689 All UNIX client platforms
3691 * Correct a condition which could discard the error from initializing
3694 * Avoid using invalid references to afs_Conn connection structures,
3695 and thus potentially producing invalid data when a retry is needed.
3697 * SRV records are now supported for discovering AFS servers.
3701 * Correct writepage behavior.
3703 * Fix error code handling in the writepage code.
3705 * Avoid leaking page locks, which could potentially hang a machine.
3709 * Preferences Pane improvements.
3713 * Avoiding attempting to handle critical signals in servers, so that
3714 core file handling works correctly.
3717 OpenAFS 1.5.65 (2009-10-06)
3721 * Code compilation warning fixes, to enable better finding and
3724 * Provide configure-time switch to enable code warning compilation.
3726 All server platforms
3728 * Demand-attach fileserver now makes volume LRU list operations
3729 exclusive operations to avoid races during adding to the list.
3731 * Fileservers now avoid potential "negative length" fetches.
3733 * A leak in host tracking objects in the fileserver has been fixed.
3735 * Salvager now unlinks all files by full path, to deal with the change
3736 to not chdir for core file tracking.
3738 * Salvager avoids asserting if the volume header is unreadable.
3740 * Demand-attach fileserver puts back volume references from fssync
3745 * Improved service response to suspend and shutdown event
3748 * Avoid a bug in the file server that can result in an invalid length
3749 being returned as part of a fetch data response if the client
3750 attempts to read beyond the length of the file.
3752 * Do not publish a default stream object for directories and mount
3753 point objects. This was impacting the ability of some Windows XP
3754 systems to save roaming profiles.
3756 All UNIX client platforms
3758 * A bug which could cause erroneous handling of lengths on data reads
3761 * A bug where erroneous length returns from the fileserver could
3762 result in a false error has been fixed.
3766 * Background page copies are now supported for enhanced disk cache
3769 * Blocking readahead is supported in readpages() to reduce overhead.
3771 * Use readpage() instead of read() to access cache data to avail disk
3772 cache users of the kernel backing cache for improved performance.
3774 * Minimize credential handling for improved performance.
3778 * Preferences Pane cleanup.
3782 * Provide a fs_pathconf method with sensible defaults.
3784 * Provide a _PC_FILESIZEBITS method to fix some NFS translator
3788 OpenAFS 1.5.64 (2009-09-22)
3790 All server platforms
3792 * The demand attach fileserver now puts back volume references gotten
3793 via the fssync interface.
3795 * The demand attach fileserver had a structure reference error, which
3796 has been correected.
3800 * Restores Windows 2000 compatibility.
3802 * Fixes a data consistency error between the output of NetWkstaGetInfo
3803 and NetServerGetInfo RPCs, specify the Lan workstation group name
3804 "AFS", and report server name as "AFS" instead of "\\AFS" when the
3805 caller asks for "\\AFS".
3807 * Enables executables to be run from \\AFS on Windows 7. Returns
3808 "Name not found" instead of "File not found" when a directory or
3809 file name cannot be found. This avoids loader errors when system
3810 dlls cannot be located in the executable directory.
3812 * Prevents cache manager from marking the file server "down" when the
3813 data returned in response to either RXAFS_FetchData64 or
3814 RXAFS_StoreData64 is invalid.
3816 * Adds pioctl data validation to the AFS Explorer Shell extension.
3818 All UNIX client platforms
3820 * A bug which could cause a kernel panic in 1.5.63 has been corrected.
3821 This would manifest as a GetDCache panic or oops.
3825 * aklog -setpag works again with recent kernels when keyring is in
3830 * When Fast User Switch is in use, AFS login is now handled correctly
3831 by the integration tool included with the preferences pane.
3833 * Several packaging bugs have been corrected.
3836 OpenAFS 1.5.63 (2009-09-11)
3840 * The restorevol command is now documented and installed as a user
3843 * The uss command now properly translates vldb entries to its expected
3844 format when handling them in all cases.
3846 * Documentation now refers to Kerberos instead of kaserver.
3848 All server platforms
3850 * bosserver now handles BosConfig.new when restarting, allowing
3851 configuration to be replaced at restart time rather than with bos
3852 delete and bos create. Documentation is updated to reflect this.
3854 * The demand attach fileservice not longer potentially hangs trying to
3855 terminate demand-salvages which have already exited.
3857 * The demand attach fileservice has been modified to avoid spurious
3858 'SYNC_putRes: write failed' warnings when some protocol messages
3859 cannot be acknowledged due to the sender terminating the connection.
3861 * In the event of failure to contact the vlserver or ptserver, the
3862 fileserver will not exit and trigger a forced salvage. It will
3863 continue to try in the background to contact the needed services.
3865 * The salvager can now repair certain cases of a damamged vnode index.
3867 * The accessDate metadata for a volume is now updated correctly.
3871 * CRITICAL: Some applications for example those based on Cygwin were
3872 unable to access data stored in the AFS name space. Explorer Shell
3873 also experienced inconsistent behavior. This is fixed.
3875 * CRITICAL: Multiple AFS pioctl requests issued nearly simultaneously
3876 by applications could result in pioctl responses being received by
3877 the wrong requester. This in turn could result in application
3878 crashes. symlink.exe, fs.exe, afslogon.dll, afscreds.exe, and the
3879 netidmgr afscred.dll plugin were all affected.
3881 * Some XP machines running 1.5.62 had trouble saving roaming profile
3882 data. This is fixed.
3884 * Integrated Logon (afslogon.dll) did not function with domain
3885 specific configurations.
3887 * Ensure that access denied and over quota errors experienced while
3888 storing data to the file server do not result in on-going retry
3891 All UNIX client platforms
3893 * Except on Solaris and AIX, the compiler may now be overriden at
3894 configure time by setting the CC environment variable.
3896 * afsd now properly deals with large cache partitions.
3900 * Build shared libafsauthent and libafsrpc.
3904 * Kernel module DKMS support now installs an unstripped module to
3905 allow debugging information to be collected.
3909 * Preferences pane properly updates token information.
3913 * klog will now properly handle passwords of 8 or fewer characters
3914 with an AFS string to key on hosts able to run 64 bit binaries.
3916 * A panic at AFS shutdown due to "NO PCB" on a udp_lock has been
3919 * The panic decoder script included in the source now properly handles
3920 32 and 64 bit panics.
3924 * Avoid defining AFS_KERBEROS_ENV globally as it creates a circular
3927 * Build shared libafsauthent and libafsrpc.
3931 * Build shared libafsauthent and libafsrpc.
3934 OpenAFS 1.5.62 (2009-08-28)
3938 * Numerous invisible changes to improve code maintainability,
3939 portability and enhanceability.
3943 * CRITICAL: Fixes two errors that can result in data loss when storing
3944 data to the file server.
3945 1. Failure to Store Portions of Unaligned Writes
3946 2. Failure to Store Data to File Servers Lacking Large File Support
3947 Read the announcement for more details:
3948 http://www.openafs.org/pipermail/openafs-announce/2009/000305.html
3950 * CRITICAL: The cache manager daemon thread could terminate when the
3951 machine enters suspend mode. This daemon thread performs the
3952 background check of down servers, offline volumes, callback
3955 * CRITICAL: Integrated Logon (afslogon.dll) was terminating
3956 unexpectedly. Error checking has been improved and NULL pointer
3957 dereferences after Lsa API calls fail have been eliminated.
3959 * For the first time, the OpenAFS SMB Server supports the DCE RPC
3960 services SRVSVC and WKSSVC. Browsing \\AFS with the Explorer Shell
3961 or NET VIEW will now be faster and provide additional functionality.
3962 No longer will cell names longer than 12 characters be truncated.
3964 * Improvements to DFS Referral request processing have been
3967 * Unnecessary DNS lookups of share names are avoided improving
3970 All UNIX client platforms
3972 * Non-Kerberos PAM modules work correctly again.
3976 * MacOS 10.6 (Snowleopard) is now supported, both 32 and 64 bit mode.
3978 * Updates to the AFSCommander preferences pane.
3980 * Installer now permits cell names with dashes.
3983 OpenAFS 1.5.61 (2009-08-06)
3987 * Correct another race condition in the Rx library that could result
3988 in an unexpected panic while freeing the Rx call iovq.
3990 * rx packet resend and data packets sent counts were incorrect.
3992 * fs setquota, fs setcachesize, vos setfields, and vos create now
3993 accept human readable orders of magnitude. (K, M, G)
3995 * fs listquota fixed to permit large quota sizes to be displayed.
3997 * Correct documentation of bosserver permissions requirements.
3999 * Modify vlserver to avoid potentially corrupting the database through
4002 * Generalized support for fast Rx timeout due to network
4005 All server platforms
4007 * Allow audit logs to be sent via sys5 IPC message queues instead of
4012 * If a file server becomes inaccessible while the cache manager has
4013 dirty buffers to write, the afsd_service buf_IncrSync thread can
4014 attempt to use 100% of the cpu.
4016 * Fix "fs newcell" which was broken in 1.5.60.
4018 * Do not attempt to synchronize dirty buffers if the associated volume
4019 is known to be unavailable.
4021 * Modify behavior of a Freelance mountpoint target that does not
4022 specify a cell. Instead of assuming the target volume is in the
4023 Freelance.Local cell, use the workstation "Cell" specified in the
4024 registry. A mountpoint target of "#root.cell." will now mean the
4025 root.cell volume in the workstation cell for the current session.
4026 If the workstation cell changes from "athena.mit.edu" to
4027 "andrew.cmu.edu", the referenced volume will also change without
4028 requiring that the mount point targets be altered.
4030 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
4031 to permit the server Uuid to be used to lookup the server object and
4032 from that determine the cell. This permits callbacks that are
4033 received from alternate addresses to be processed with a known
4034 server object. Previously a request from an unknown server would
4035 clear all callbacks from all cells.
4037 * Fix a bug that prevented optimal performance when using a non-zero
4038 value for 'daemonCheckVolCBInterval'. As a reminder, when
4039 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
4040 volume callbacks are automatically renewed 90 minutes before their
4043 * Fix automatic ranking of vldb servers whose values are obtained from
4044 the CellServDB file.
4046 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
4047 or the call is to a vldb server.
4049 * Add registry based cell search functionality to NetIdMgr,
4050 afs_config.exe, and klog.exe.
4052 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
4053 lookups on the host names in the CellServDB instead of using the
4054 specified IP addresses. This provides aklog, pts, vos, etc. the
4055 same CellServDB behavior that the Windows Cache Manager uses.
4057 * When updating the stat cache entry callback of a .readonly object
4058 from the volume group object, update the file server reference to
4059 ensure it matches the most update to date callback.
4061 * Add proper support for processing callbacks from multi-homed file
4062 servers. Instead of comparing servers by cm_server_t pointer,
4063 compare them by UUID when the UUID is known.
4065 * During a shutdown short circuit the offline volume check daemon
4068 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
4069 preference to an error code reported by rx_EndCall.
4071 * Add "PerFileAccessCheck" registry value to permit testing against
4072 experimental file servers that include per-file acl support. This
4073 value is intentionally undocumented. It is not to be used by
4074 production environment deployments.
4076 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
4077 provider from obtaining tokens when referrals are in play.
4079 * Add "fs chown" and "fs chgrp" commands to permit the owner and group
4080 of objects stored in AFS to be set from Windows.
4082 * Avoid performing background daemon operations when the machine is
4083 going into suspend mode.
4085 * Perform offline volume checks in most recently used order.
4087 * Prevent crash when a data version for a cache object goes backwards.
4089 * Multi-thread safe library versions are now being generated and used.
4090 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
4092 * Microsoft SMB Redirector (mrxsmb.sys) support for
4093 ExtendedSessTimeout values are now available on XP through Windows
4094 7. Add functionality to autodetect if such support is present on
4095 the machine. If so, configure it if necessary and dynamically
4096 adjust the AFS Rx timeout values accordingly.
4098 All UNIX client platforms
4100 * Fix out-of-tree source builds.
4104 * GUI installer now asks for local cell information.
4106 * AFS Commander preferences pane is now installed by default.
4110 * Avoid kernel panics due to null pointer dereferences in the network
4111 interface poller kernel thread.
4114 OpenAFS 1.5.60 (2009-05-31)
4118 * Retry volserver transaction creation on failure.
4120 * Allow building HTML and PDF documentation from included XML copies
4121 of User Guide, Admin Guide and Quick Start Guide for Unix.
4123 * Documentation updates and additional documentation.
4125 * Add -encrypt support to pts client.
4127 * Convert MR-AFS fs commands to OSD commands.
4129 All server platforms
4131 * Updated background sync process in fileserver to avoid a race which
4132 could result in a volume being taken offline.
4136 * On April 9th Microsoft released a Hot Fix for Windows Server 2003
4137 SP2 that corrects a deadlock in the smb redirector and also adds new
4138 functionality that permits the AFS SMB server to be given a longer
4139 timeout than is normally the case. New functionality has been added
4140 to configure these additional LanmanWorkstation\Parameter values.
4141 (This functionality has been backported to XP SP3 and is scheduled
4142 to be released on June 5th.)
4144 * Fix RT#124787, a race condition between "fs flush <dir>", "fs
4145 flushvolume", or "fs flushall" and on-going directory operations
4146 that can result in afsd_service.exe crashing.
4148 * Release Notes, User and Administrator guides are now shipped as
4149 indexed Windows HtmlHelp Files. (.chm). Shortcuts are provided from
4152 * A method of specifying Client CellServDB information within the
4153 registry has been added that can be used to either override the
4154 CellServDB file or force the use of DNS lookups for a given cell.
4155 See the release notes for details.
4157 * The pioctl interface now properly handles drive letter substitution
4158 to UNC paths. (SUBST <d:> <\\afs\cell\path>)
4160 * The BackConnectionHostNames registry value configuration was broken
4161 when dynamic re-establishment of Netbios Name registrations was
4162 added. This release restores the functionality.
4164 * All hidden vos.exe commands are now revealed.
4166 * Attempts to store the same dirty file chunk from multiple threads
4169 * The IsPathInAfs test in Explorer Shell Extension and fs.exe now
4170 permits broken symlinks to be treated as being in AFS.
4172 * vos.exe commands that output 64-bit integer values once again do so.
4173 This was broken in 1.5.59.
4175 * Cygwin Import Libraries are provided in the SDK for all OpenAFS
4176 DLLs. This permits building cygwin applications against OpenAFS
4179 * NSIS installer does a much better job of cleaning up files left over
4180 from previous installs.
4182 * libafsconf.dll moved from Client\Program to Common directory as is
4183 is now used by all modules for CellServDB processing.
4185 All UNIX client platforms
4187 * Write back changes on last store for memcache to avoid discarding
4190 * Abstract disk cache support to allow for path, fh, inode based
4191 caches with no need for messy ifdef structures each time a new type
4196 * Support as a userland port.
4200 * Corrected structure definition for userspace cache manager to allow
4205 * Corrected client locking support.
4207 * Updated patch to stop deadlocking in the kernel during mmap.
4209 * Avoid oops when setting up groups for PAGs to match keyrings.
4211 * Use Linux fh-based cache in cases where possible by default.
4215 * Corrected structure definition for userspace cache manager to allow
4220 * Support for OpenBSD 4.5.
4224 * Corrected support for server-side vos split interface.
4227 OpenAFS 1.5.59 (2009-04-06)
4231 * Increased service priority class to "High" to match the priority of
4232 system components that are dependent upon the a timely response.
4234 * SMB error responses avoid returning errors that could confuse the
4235 Microsoft SMB redirector into disconnecting the connection to \\AFS.
4237 All UNIX client platforms (except MacOS X 10.4 and 10.5)
4239 * OpenAFS 1.5.59 contains fixes for the client issues addressed by the
4240 security advisories OPENAFS-SA-2009-001 and OPENAFS-SA-2009-002.
4244 * Support for prerelease Linux 2.6.30 kernels.
4247 OpenAFS 1.5.58 (2009-03-30)
4251 * Code cleanup and prototyping.
4253 * Avoid unnecessary blocking in Rx periodic cleanup code.
4255 All server platforms
4257 * Fileserver CopyOnWrite routine optimized for performance.
4259 * Make fileserver callback dumps 64 bit safe.
4261 * Fix byte order issues with fileserver host hashing.
4263 * Fix buffer size issues with butc.
4265 * Fix several Ubik recovery issues.
4267 * Avoid leaking file references in the fileserver.
4269 * Fix a race in DAFS while closing vnodes, and another offlining
4272 * volserver interfaces for volume splitting client.
4276 * [RT 124293] A race condition exists which can result in a crash.
4278 * [RT 124276] If the vldb is out of sync with the contents of the file
4279 servers, afsd_service will retry too many times when a file server
4280 reports a volume as not being present. Now if the list reported by
4281 the vldb is the same as the previously seen list, then the retry is
4284 * [RT 124276] Read-only volume failover was broken in 1.5.53 whenever
4285 accessing a volume results in VNOVOL or VMOVED.
4287 * [RT 124276] Prior to 1.3.70 the volume server reference list was not
4288 reference counted and would be prematurely freed while in use. When
4289 reference counting was added in 1.3.70 a bug was introduced that
4290 could result in service reference list corruption.
4292 * Add Windows Application Event Log warning messages for "Client SMB
4293 MPX value too large" and "Client SMB Buffer Size too small".
4295 * Renaming of files across directory boundaries would result in an
4296 invalid handle error when attempting to access the files after the
4299 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail when a
4300 smb file descriptor cannot be found. The whole point of using a
4301 Path Info function is because an smb file descriptor wasn't
4304 * More edge cases in which dynamic addition of Freelance root.afs
4305 entries would get the wrong FID or where the root.afs directory
4306 would not be refreshed.
4308 * Buffer overflow could occur if the workstation cell name was longer
4309 than 64 characters. Crashes could occur in afscreds.exe,
4310 afslogon.dll, and afsd_service.exe.
4312 * VNOSERVICE and VOFFLINE errors were leaking and were exposed to the
4315 * Log file server uuid values as part of the cm_server object when
4316 available. Dump the cm_server object list in response to "fs
4319 * Optimize the performance of resetting access control lists when
4320 tokens are set or removed.
4322 * Remove symlink recursion tests and increase max symlink count to 64
4325 * Windows specific Rx performance improvements.
4327 * Support for Network Identity Manager 2.0
4329 All UNIX client platforms
4331 * Avoid issues with freeing resources at shutdown.
4333 * Numerous fixes to disconnected AFS.
4335 * Disconnected AFS fixes for replaying changes without double-freeing.
4337 * Attempt to use krb524 principal conversion in aklog if available.
4341 * Kerberos configuration at build time corrected.
4345 * Default to dynamic allocation of AFS kernel cache entries to allow
4346 growth for inotify()-pinned entries. (beagle, famd, etc)
4348 * Change client truncation routines to avoid locking issues.
4350 * IA64 port clients fixed on Linux 2.6.
4352 * RPMs now install fstrace message catalog.
4354 * Support through kernel 2.6.29 tested.
4356 * Fix locking issues on CellServDB file in /proc.
4360 * Support OpenBSD 4.4
4363 OpenAFS 1.5.57 (2009-01-23)
4367 * Conditional compilation of rxdebug support is now possible.
4369 * Documentation updates.
4371 * Further race connditions in Rx have been corrected.
4373 All server platforms
4375 * Salvager no longer attempts to recreate headers in the wrong
4378 * Volumes are properly marked in use on creation and subsequently on
4379 examination with vos.
4383 * Undo the "UAC manifest fix" applied to afs_config.exe.
4385 * Ensure that Freelance allocation of vnodes follow the AFS convention
4386 of odd vnodes are directories and everything else is an even vnode.
4388 * Add Freelance logic to mount point and symlink evaluation functions.
4390 * Enhance smb_ParseASCIIBlock() so that it can handle all of the
4391 STRING formats defined by the CIFS Technical Report 1.0.
4393 * Validate the output of smb_ParseASCIIBlock() in all callers. Return
4394 CM_ERROR_BADSMB if the STRING field cannot be parsed.
4395 CM_ERROR_BADSMB will cause the contents of the packet to be logged.
4397 * If multiple SMB Raw Write operations were taking place at the same
4398 time, there could be data corruption because unique event objects
4399 were not generated for each Netbios receive operation.
4401 All UNIX client platforms
4403 * Userspace AFS library can now deal with large files when supported
4406 * Numerous updates to disconnected AFS support, including changes to
4407 allow reconnection to work in more circumstances.
4411 * FreeBSD unstrategy code has been updated.
4415 * A race during file truncation has been corrected.
4417 * System call probing routines have been updated.
4419 * 2.6.29 is now supported.
4423 * 10.3 support has been corrected.
4427 * Initial OpenBSD 4.4 support.
4431 * Updates to allow compiling on newer OpenSolaris are now included.
4434 OpenAFS 1.5.56 (2008-12-30)
4438 * libuafs (userspace cache manager) updated to correct several errors.
4440 * Additional rx debugging support is available as a conditional
4443 * A race condition in Rx leading to a panic has been corrected.
4445 * Rx idle time tracking has been corrected.
4447 * ubik clone server support has been corrected.
4449 All server platforms
4451 * Salvager no longer leaves cores in vice partitions.
4453 * The vol-dump tool now supports dumps larger than 2gb where possible.
4455 * Operations on multiple files now report all FIDs in the audit log.
4457 * butc XBSA support now works correctly on amd64.
4461 * The NetIDMgr AFS Provider automated configuration logic was broken
4462 by the introduction of Kerberos referrals. If the realm of the
4463 identity cannot be determined, the workstation cell is now assumed
4464 to belong to the newly created identity.
4466 * Avoid a reference count under flow during rename operations.
4468 * Avoid a crash caused by treating an arbitrary length directory
4469 search mask as an 8.3 mask.
4471 * Prevent rename operations if a case insensitive match for the target
4472 name already exists and does not refer to the object being renamed.
4474 * Increase the maximum number of background daemons to 64.
4476 * Fix the UAC manifest applied to afs_config.exe
4478 All UNIX client platforms
4480 * Updates to disconnected AFS support.
4484 * FreeBSD 7.1 is now supported.
4486 * amd64 FreeBSD is now supported.
4490 * Generic fh (exportfs API) cache type is now available.
4492 * Avoid some oopses due to backing_dev_info inode fields not being
4495 * 2.6.28 is now supported.
4499 * 10.3 support has been corrected.
4503 * Large partition support has been corrected.
4505 * Filesystem-agnostic cache is now available on Solaris 10 and 11.
4508 OpenAFS 1.5.55 (2009-11-10)
4512 * Salvager avoids leaving core files in vice partitions.
4514 * NFS translator fixes.
4516 * Unresponsive server handling fixes.
4518 * A volserver race which could result in duplicate transactions is
4523 * Fixes a panic caused by corruption of the SMB virtual circuit list.
4526 * Fixes a panic caused by receipt of a UTF-16 string that cannot be
4529 * Implements a more aggressive recovery algorithm for Netbios errors
4530 that result in loss of communication to the AFS SMB server.
4532 * Improve pioctl response time when testing whether or not a PATH is
4535 * Adds support for linked cells.
4537 * Increases the length of the cell and realm names that can be input
4538 into the Network Identity Manage AFS provider configuration dialog.
4540 All UNIX client platforms
4542 * Disconnected AFS avoids infinite recursion during rmdir.
4546 * Support for 2.6.28 (not complete for NFS translator modules).
4548 * Support for using exportfs API for filesystem-agnostic cache.
4550 * Disable backing store readahead.
4552 * Avoid deadlocks when writing back mmapped files larger than the
4555 * Avoid Oops when doing PAG garbage collection.
4558 OpenAFS 1.5.54 (2008-10-08)
4562 * Updates for new Tivoli X/Open API finding.
4564 * A double-free is corrected in Rx.
4566 All server platforms
4568 * Ubik cleans up file descriptor cache correctly when doing recovery.
4570 * Enhanced vldb error checker included.
4574 * Prevent a crash that could occur when multiple file / directory
4575 change notifications are processed simultaneously.
4579 * AFS claims more free space so Finder will attempt file copies.
4583 * Avoid spurious ENOENT when calling gwtcwd() on a volume root.
4585 * Avoid spurious ENOTDIR during fakestat.
4588 OpenAFS 1.5.53 (2008-09-26)
4592 * rx avoids many packet leaks.
4594 * rx jumbogram disabling now works (and is the default).
4596 All server platforms
4598 * Demand Attach fileserver tries to avoid issues tracking offline
4603 * Many potential deadlock conditions due to out of order lock
4604 acquisitions have been corrected.
4606 * A race condition resulting in an undercount on the cm_scache_t
4607 reference counts has been corrected.
4609 * An empty string when sent as an ioctl path is now properly
4610 interpreted as meaning the current directory. This affects "fs
4611 lsm", "symlink list", etc.
4613 * Fix smb string parsing differences where the smb protocol
4614 documentation does not match the actual Windows implementation.
4616 * Random access denied errors fixed.
4618 * A file server lock synchronization issue was corrected in SMB
4619 NTCreateX and NTTranCreate operations. This bug prevented properly
4620 operation when loading roaming profiles.
4622 * Fix a heap overwrite error during server probe operations if a new
4623 server is added while a probe operation is in progress.
4625 * Fix an LSA memory leak that was the result of an LSA error.
4627 * Do not leak cm_cell_t objects if the VLDB server lookup fails.
4629 * Only initialize rx mutex/lock objects once.
4631 * Do not nul terminate the AFS volume name when reported to Windows.
4633 * Improve VNOVOL error handling. Prevent updated vl information from
4634 being destroyed immediately after it was acquired. This bug
4635 prevented proper fail over when volumes are moved or removed from a
4638 * Remove volume id from the server volume list in response to VMOVED
4641 * "fs flushXXX" commands now destroy locally built B+ directory trees.
4643 * Prevent mixture of locally modified directory pages and file server
4646 * Fail over to alternate vl servers if a ubik error is returned.
4648 All UNIX client platforms
4650 * Disconnected AFS now supports read-write mode.
4652 * volserver now builds correctly.
4656 * AIX 6 is now supported.
4660 * FreeBSD 7 is now supported.
4664 * cache bypass is now supported.
4666 * 2.6.x kmod compilation now uses kernel compile options always.
4668 * Support through 2.6.27.
4672 * Show more space free so Finder doesn't get confused.
4676 * Default to namei rather than inode.
4679 OpenAFS 1.5.52 (2008-08-18)
4683 * Initialize volume updateDate at volume creation.
4685 * Avoid potential corruption of directories during salvage.
4687 * Check for out of memory condition during allocation of additional Rx
4692 * Restore support for Windows 2000 (broken in 1.5.50).
4694 * Perform additional validation on volume names in mount points during
4695 creation and evaluation.
4697 * Fix several deadlocks, races, and reference count issues.
4699 * Further optimize SMB Directory Search processing and minimize the
4700 number of InlineBulkStatus RPCs sent to the file server.
4702 * Enable "bos restricted" operations.
4704 * Fix the create of submounts used by the AFSCreds and afs_config
4707 * Fix a short name truncation bug. (1.5.50)
4709 * Fix the error code reported when attempting to delete a file on a
4710 readonly volume or one that is marked with the readonly DOS
4713 * Fix a heap corruption error when reading the CellServDB file
4716 * Add the "RxUdpBufSize" registry value. The new default is 256KB.
4718 * Do not include trailing NULs in the directory search output.
4721 * Pre-allocate 64 Rx Packet buffers per thread in order to improve
4724 * For debugging: add smb lock requests and stat cache lock allocations
4725 to the output from "fs memdump".
4729 * Workaround broken sigwait() to allow fileserver to shut down
4730 correctly pre NetBSD 5.0.
4734 * Default to namei fileserver; Allow inode fileserver at configure
4739 * Try harder to avoid kernel panics for malformed requests.
4742 OpenAFS 1.5.51 (2008-07-29)
4746 * salvager tries harder to arrange for clients to get VBUSY while
4747 salvaging single volumes.
4749 * salvager avoids certain corruption when salvaging directories.
4751 * Rx connection clones disabled.
4755 * The 32-bit EXE 1.5.50 installer failed to properly install the C
4756 Runtime library. When used as an upgrade OpenAFS would continue to
4757 work but when used as a new installation, OpenAFS binaries would
4760 * Fixes the "fs" and "symlink" commands to properly parse Unicode path
4761 prefixes during the pioctl remote procedure call. This bug would
4762 result in file not found errors for files and directories that
4763 clearly exist. (Bug introduced in 1.5.50)
4765 * Large File support is disabled. (Bug introduced in 1.5.50)
4767 * Removes the possibility of a deadlock during volume location update
4768 operation if all of the reported file servers are unreachable at the
4771 * Ensures that reference counts are properly incremented/decremented
4772 on Rx connection objects used for volume location RPCs.
4774 * Over Quota errors during cm_FSync() calls would lead to an infinite
4775 loop as the error was never propagated to the caller.
4777 All UNIX client platforms
4779 * Bugfixes to disconnected AFS support in the cache manager.
4782 OpenAFS 1.5.50 (2008-07-16)
4786 * volserver puts recloned volumes back online before returning the
4787 volume to the fileserver, avoiding spurious VNOVOL errors.
4789 * Updated TSM X/Open API support available.
4791 * Demand Attach fileserver will not crash due to accesses during
4794 * Substantial documentation updates.
4796 * Demand Attach fileserver state tracking and analyzer tool
4799 * UAFS userspace cachemanager updates.
4801 * Corrected support for anti-meltdown protection in the client.
4805 * UNICODE Character Set Support.
4807 * Pioctl interfaces to the cache manager have been refactored to
4808 provide layering between the SMB specific code and the general
4809 purpose ioctl operation.
4811 * Garbage collect dead SMB virtual circuits as soon as they are no
4812 longer being referenced. This avoids problems with outstanding
4813 locks not being dropped when the virtual circuit becomes invalid.
4815 * Remove the IBM Administration Reference documentation and replace it
4816 with the OpenAFS Command Reference Manual.
4818 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime functions each
4819 time a connection is about to be used. Do not hold a lock on the rx
4820 connection object while it is being selected. This avoids a race
4821 between threads attempting to set the timeout values and removes a
4822 bottleneck that was hampering performance.
4824 * Ensure that the smb directory attribute is set for all directory
4827 * Replace the VC Runtime EXE installer with the MSI installer in the
4828 NSIS installer scripts.
4832 * Support for updates to OpenSolaris.
4836 * Correct dentry revalidation for cross-directory renames.
4838 * Updated rpm packaging materials for 1.5 release series and 2.4
4841 * Corrected syscall table probing.
4843 * NFS translator updates for current kernels.
4846 OpenAFS 1.5.39 (2008-06-24)
4850 * Updates for Demand Attach fileserver.
4854 * Fix two memory leaks.
4856 * Fix one missing lock.
4858 * Handle access denied errors when writing dirty buffers.
4860 * Fix two errors that would cause the *experimental* AFS Servers
4864 OpenAFS 1.5.38 (2008-05-24)
4868 * Add read-only disconnected support.
4871 OpenAFS 1.5.37 (2008-05-21)
4875 * Includes a number of optimizations for testing.
4878 OpenAFS 1.5.36 (2008-05-09)
4882 * Rx optimizations now attempt to deal with high latency WANs.
4884 * Client will not wait infinitely for a server which is not providing
4885 data. Additional servers will be polled without marking the server
4886 which is not providing data down.
4888 * vos move will not erroneously unlock locked vldb entries on failure.
4890 All server platforms
4892 * Fileserver avoids a potential infinite loop when a client
4893 relinquishes an address.
4895 * Fileserver large setting now configures more threads.
4897 * Fileserver properly registers uuids of new clients.
4899 * Ubik servers do not improperly hide updates from clients.
4901 * Fileservers reserve enough file descriptors such that each thread
4902 can cache one to avoid spurious errors.
4906 * Fix a cm_buf_t reference count leak when attempts to write dirty
4907 buffers to the file server from within cm_IncrSyncer() fail.
4909 * Prevent udebug from crashing.
4911 * Another VNOVNODE issue fixed. When writing a dirty buffer to the
4912 file server, if VNOVNODE is received, mark all buffers as invalid
4913 without further attempts to contact the file server.
4918 All client systems: Major bugfixes.
4919 File servers: Major bugfixes.
4921 * New functionality:
4925 - Provide portable (pioctl) method for discovering what PAG a user is
4926 in. Required to support userspace PAG information collection on AIX
4927 5.1, and knowing whether Linux uses one group, two group, or only
4928 keyring based PAGs. (124709)
4934 - Fixes to avoid issues cleaning up deleted hosts in the fileserver (126454)
4936 - Fixes to avoid dropping writes due to server idle timeouts.
4938 - Don't miss cache chunks of large files while truncating.
4940 - Avoid null pointer dereference for unexpected volume names in volume
4943 - Don't mark connections waiting for additional packet window availability
4946 - Kerberos 5 utilities (klog.krb5, aklog) enable weak encryption support.
4948 - Avoid a double-free of an Rx call structure during a client fetch error.
4950 - Avoid losing hosts during address changes. (125215)
4952 - Clients shouldn't trust Fetchdata replies for the size of returned data.
4954 - fileserver will not hang when attempting to cleanup and dump core.
4956 - salvager will not leave core files in random directories.
4958 - avoid letting retransmit timer get to 0 seconds.
4960 - in event of dbserver contact failure, shut fileserver down cleanly.
4962 - handle large partitions during check for needed disk space at
4965 - time out Rx connections if network unreachable error received.
4967 - avoid dereferencing NULL pointer freeing Rx packets in receive. (125110)
4969 - mark stack not executable in LWP. (125491)
4971 - return a correct VolumeSync structure from Bulkstat RPCs in fileserver.
4973 - client attempts to better free memory at shutdown.
4975 - clear rx call queue safely. (125110)
4977 - retry VLserver registration on failure in fileserver.
4979 - update accessdate for volumes on access in fileserver.
4981 - additional safety checks on vlserver operations to avoid
4982 database corruption.
4984 - make ktc_curpag available on all builds. (125155)
4992 - Handle kernel changes through 2.6.33.
4994 - Fix oops in clear_inode due to missed locking. (125589)
4996 - Better handle /afs mount failures.
4998 - Clean up after failures creating our kernel kmem cache.
5000 - Work around memory management issues with some kernels when configuring
5001 the buffer cache/bdi (126514)
5003 - Rename compile_et to afs_compile_et to avoid RPM conflicts.
5005 - Handle whole-file locks properly. (126561)
5007 - Deal with kernel autoconf header renaming.
5009 - Handle SELinux cache backing file labels better to avoid potential oops.
5014 - klog now works correctly on 64 bit machines.
5016 - launchd now used to launch AFS at boot.
5018 - Preferences pane included for 10.4 and later.
5020 - Older versions can now be installed from packages.
5022 - Finder does not trigger bogus AFSDB lookups in /afs in dynroot mode.
5024 - Include package with debug kernel module symbols.
5030 - Support for x86_64.
5034 - Handle ZFS caches usefully. (125365)
5036 - Implement additional pathconf support.
5040 Linux client systems: Major bugfixes.
5041 All client systems: Minor bugfixes.
5042 File servers: Major bugfixes.
5044 * New functionality:
5048 - Provide portable (pioctl) method for discovering what PAG a user is
5049 in. Required to support userspace PAG information collection on AIX
5050 5.1, and knowing whether Linux uses one group, two group, or only
5051 keyring based PAGs. (124709)
5057 - Fix bosserver to invoke salvager with "-force" instead of ambiguous
5060 - Cleanup for ptserver argument parsing to allow debug mode to work. (124893)
5062 - Sanity checking for ptserver log levels. (124894)
5064 - Fix for uninitialized memory dereference in klog.krb5.
5066 - Fix an overflow in the cellconfig code used by client and server. (124891)
5068 - Fix an erroneous vos verbose mode format string.
5070 - Avoid losing writes on mmap()ed files when cache is memcache. (124671)
5072 - Provide an afsd switch to allow override of the maximum MTU. (124880)
5074 - Provide support for encrypt mode in pts.
5076 - Fix race in background sync code which could cause volumes to go offline.
5079 - Fix fileserver to avoid a null pointer dereference in client identity
5080 lookup routines. (125020)
5082 - Improve handling of moves of volumes from 1.5 series fileservers. (18349)
5086 - UKERNEL build fix. (124681)
5090 - Allow syscall probing to be disabled by switch to configure at build time.
5092 - Fix bug in anti-recursion protection for mmap clients. (124627)
5094 - Avoid a panic caused by changing credentials during VFS operations. (124737)
5096 - Avoid need for rcu subsystem when unavailable. (124986)
5098 - Improve keyring PAG setup code. (125001)
5100 - Avoid possible ext3 cache truncation issues. (124942)
5104 - MacOS 10.3 UKERNEL build fix. (124681)
5108 - Update support for 4.5 (124719)
5112 - Updates for newer OpenSolaris kernels. (124116, 124924)
5116 All client systems: Security fixes.
5117 File servers: Major bugfixes.
5118 All systems: Minor bugfixes.
5124 - Avoid a potential kernel memory overrun if more items than requested are
5125 returned from an InlineBulk or BulkStatus message. (124579)
5127 Linux client systems:
5129 - Avoid converting negative errors into invalid kernel memory pointers. (124580)
5135 - Don't build aklog NAS module when krb5 is not available. (124522)
5139 - Additional fixes and support. (124107, 123917)
5143 - Support 2.6.28. (123580)
5145 - Support 2.6.29. (124115)
5147 - Attempt to support 2.6.30 (124560)
5149 - Avoid race during truncation. (124094)
5151 - Dynamic vcache allocation support, to deal with inotify vcache pinning. (124334)
5153 - Correct use of truncate_inode_pages to vmtruncate for locking issues. (124128)
5155 - Update RPM configuration. (123650, 102673, 124272)
5157 - Update kernel feature detection. (124507, 123604)
5159 - Do appropriate locking for CellServDB in /proc. (124407)
5163 - Fix MacOS 10.3 support.
5165 - Add candidate Darwin 10 support.
5169 - Corrected NetBSD version tests. (123647)
5173 - Update support for 4.4 (124541)
5177 - Support cache filesystems which do not allow open by inode number, enabled by
5178 default on Solaris 9 and later. (123677)
5180 - Improve error code return quality. (124426)
5182 - Allow large partitions on Solaris servers.
5186 - Avoid improper error messages about key version when krb5 is in use. (124220)
5188 - Avoid attempting to free kernel memory which was already freed. (124531)
5190 - Properly count offline volumes in vos client. (124333)
5194 - Avoid 64 bit time issues in callback dump files. (124451)
5196 - Support more than one local Kerberos realm; Usernames are assumed to be the
5199 - Ubik recovery is corrected to avoid spurious errors. (123723)
5201 - Do proper host address hashing for little endian machines in fileserver. (124447)
5203 - Update backup utility to properly compute header needs in the backup buffer.
5206 - Avoid blocking during Rx unused connection reaping.
5208 - Avoid leaking file handles in the fileserver when closing a volume. (124359)
5210 - Fix bosserver corefile naming to be y2k-safe. (124340)
5212 - Avoid potential infinite loop in deleted host handling in the fileserver.
5214 - Support large volume dumps in vol-dump. (123984)
5216 - Build butc XBSA support on 64 bit systems.
5220 - Properly track Rx connection idleness for timeouts.
5222 - Additional documentation. (124472)
5224 - Avoid a race which may result in an in-use Rx packet being freed. (123799)
5228 All client systems: Security fixes.
5234 - Avoid a potential kernel memory overrun if more items than requested are
5235 returned from an InlineBulk or BulkStatus message. (124579)
5237 Linux client systems:
5239 - Avoid converting negative errors into invalid kernel memory pointers. (124580)
5243 File servers: Major bugfixes.
5244 All systems: Minor bugfixes.
5250 - AIX 6.1 is now supported.
5252 - Unpin kernel memory references after free. (99456)
5256 - FreeBSD 7 is now supported.
5260 - Avoid deadlock when writing back pages in an mmap()ed file larger than
5263 - Update process tree walking for PAG garbage collection to avoid oopses.
5266 - fakestat mode now correctly avoids spurious ENOTDIR errors.
5268 - Use kernel build system for all platforms.
5270 - Remove openafs directory from proc in correct order. (112910)
5272 - Handle renames across directories correctly in the linux dcache. (74672)
5274 - Probe syscall table when possible. (105457)
5276 - Mount point parsing is now updated to handle only well-formed mount
5277 points rather than similarly-formed symlinks. (113558, 100836)
5279 - ucontext-style LWP is now anbled for glibc versions newer than 2.3.
5283 - Update available space shown. (112910)
5287 - Work around broken SIGWAIT. (111404)
5291 - Solaris 10 now defaults to namei fileservers.
5293 - NFS translator issues fixed.
5295 - Changes to address Solaris updates. (105495)
5299 - udebug correctly displays the last "yes" host.
5301 - Allow more vldb lookups to be cached in the client.
5303 - Fix aklog to not be excessively verbose when not requested.
5305 - Add support for timing out accesses which are not completing in a timely
5308 - Properly flag backup volumes being added to the vldb by vos syncvldb.
5312 - fileserver "large" setting now implies 128 threads instead of 12.
5314 - fileserver check for duplicate uuids is now applied correctly.
5316 - Newer xbsa APIs are supported for TSM integration in butc.
5318 - salvager avoids corrupting length of directory objects. (111585, 107767)
5320 - volserver avoids a race during volume release so a volume will not
5321 appear to be offline. (107258)
5323 - Add support for returning errors when accesses are not completing
5326 - Avoid potential race in the volserver when creating transactions. (121263)
5328 - Return sensible error when a release or restore exhausts server
5331 - volserver now returns EXDEV if a new replica would duplicate one which
5332 already existed elsewhere on the server.
5334 - Disable jumbograms by default.
5336 - volserver updates a volume's updateDate on volume creation. (110943)
5338 - Partitions over 2tb are now supported. (88811)
5342 - Re-enable Rx client keepalives. (20727)
5344 - Support autoconf 2.62. (118058)
5346 - Update Rx to avoid leaking packets.
5348 - vos supports the -noresolve options to avoid issues with 127.0.0.1 being
5349 named in /etc/hosts.
5351 - Additional documentation. (104110)
5355 File servers: Major bugfixes.
5356 All systems: Minor bugfixes.
5362 - Kerberos as included in AIX has missing symbols. AFS krb5 tools now
5365 - AIX LAM aklog plugin can now be used CDE screenlocker.
5367 - Add support for getting the current PAG in pagsh and PAM.
5369 - Avoid sending a terminal hangup to STREAMS in aklog.
5371 - Fix afsdb support in the client.
5375 - Kernels through 2.6.25 are now supported. (77370,88000,83716,83890,80463)
5377 - Client now only hashes dirty inodes. (78544)
5379 - Fix to avoid returning invalid mount point data when -fakestat-all is in use.
5382 - RPM build system updates. (93616)
5384 - Restored write-on-close-or-fsync semantics when possible. (17509)
5386 - Enabled support for flock() on files in AFS. (53457)
5388 - ARM Linux now supported.
5390 - Kernel keyring support updated.
5392 - Fix client-displayed timestamp ordering by zeroing nanosecond field.
5396 - Boot time init script now uses afs.conf to store config options. (81825)
5398 - Avoid kernel panic due to excessive lock tracking when removing files.
5400 - Avoid leaking kernel memory when trying to read() a directory.
5404 - Avoid potential kernel panic if the root vnode of AFS changes.
5406 - Avoid potential kernel panic when shutting down if contracts are in use.
5408 - Avoid potential delays when creating new PAGs if the system clock
5413 - At client shutdown, try harder to clean up in-use resources. (74479)
5415 - When fakestat is in use, enable optimization for Gnome Nautilus lookups.
5417 - Properly hold lock when updating disk cache metadata to avoid
5418 kernel panic. (59136)
5420 - Avoid wrapping to the start of a file when attempting to write a large file
5421 to a pre-largefile fileserver. (73720)
5425 - Fixed to avoid truncating ubik databases during recovery. (77183)
5427 - fileserver issue with internal file cache filling has been fixed. (87977)
5429 - fileserver thread quota enforcement now done in all cases. (87416)
5431 - fileserver avoids potential network-related deadlock when breaking
5434 - fileserver avoids crash due to race of resource creation and user requests
5437 - fileserver avoids crash when reinitializing Ubik connections.
5439 - volserver fixed to avoid leaving orphaned files during restore. (46937)
5441 - volserver now supports convertROtoRW for inode fileservers.
5443 - Support disabling kerberos 4 style username protection in servers. (75101)
5447 - Fix to butc to avoid crash due to threaded library variant.
5449 - Fix to avoid network retransmission issues if the system clock goes
5452 - vos syncvldb and syncserver now support a dryrun (do nothing) mode.
5454 - vos addsite now supports adding a site where a replica is already
5457 - vos clone now supports creating properly-named readonly and backup clones.
5459 - vos restore now allows an older copy of a volume at an alternate site to
5462 - cmdebug now supports dumping a client's CellServDB.
5464 - cmdebug now supports showing human-readable expiration times.
5466 - aklog now handles Kerberos referrals.
5468 - Additional documentation now included. (89288,89289,86677)
5472 All systems: Major bugfixes.
5478 - fileserver host tracking code had a missing lock on a host structure;
5479 that lock has been added.
5481 - fileserver handling for clients which are giving up callbacks did not
5482 hold a lock, making it unsafe and allowing clients to potentially
5483 crash the server by racing.
5485 - fileserver will now leave a corefile when it is doing a shutdown due
5486 to error conditions.
5488 - fileserver again allows ufs as a valid filesystem type (regression in 1.4.5).
5490 - cbd handler for fileserver status data has an interpretation error which
5491 could cause crashes corrected.
5493 - fileserver accurately tracks number of callbacks on a given file.
5497 All systems: Minor bugfixes.
5498 New systems: MacOS 10.5.
5502 - fileserver address tracking is improved to avoid potentially merging
5505 - Documentation updates.
5507 - namei fileserver now does fsync()s in background batches for performance
5510 - Kerberos ticket support corrected in bundled Kerberos 4 utilities on 64
5513 - fileserver includes limited per-host thread quota support to avoid
5514 resource starvation.
5516 - fileserver deals with more damaged volumes without asserting.
5518 - vos validates dumpfiles before attempting restores.
5520 - vos clone will no longer potentially delete the parent volume.
5522 - Client no longer permits empty UUID to be created.
5524 - fs uuid command for checking, regenerating UUID added.
5526 - Updates for gcc 4.2.
5528 - fileserver treats w (write) permission as granting read lock permission
5529 in addition to write.
5533 - Bundled NAS Kerberos is now supported. (5.x)
5535 - LAM aklog module is provided. (5.x)
5537 - Associate cache files with correct filesystem to avoid snapshots when
5538 performing maintenance on local filesystems.
5546 - Kernels throigh 2.6.23 are known to work.
5548 - Updates to syscall table probing.
5550 - Bug fix in keyring PAG support to avoid oops.
5552 - updated sample RPM configuration.
5554 - sparc32 lwp support updated.
5556 - Avoid potential oops in symlink support in certain older kernels.
5558 - Avoid potential deadlock during vmalloc.
5560 - Corrected locking while interacting with kernel task list.
5564 - Several panics fixed, including remove_fsref. (10.4 and later)
5566 - IP address changes now tracked.
5568 - Corrected support for dropboxes (li access without r) with cp and Finder.
5570 - fstrace and ancillary files now included.
5572 - man pages are now installed in the default MANPATH.
5574 - Servers will be timed out quickly if there is no route available.
5576 - Temporary files from remove-while-busy now cleaned up correctly.
5580 - Updates to accomodate kernel interface changes. (10u4 and later)
5582 - knfs and NFS translator support updated.
5584 - Changes to accomodate version 5.4 xbsa library.
5586 - Updates for new SunStudio defaults.
5590 All systems: Major bugfixes.
5596 * A bug in the namei volserver which could erroneously make a replicated or
5597 moved volume go offline has been fixed.
5599 * Volume package users (fileserver, volserver, salvager) avoid using lockf to
5600 avoid leaking byte range locks on volume internal files.
5604 * A bug where the client kernel module could free stack memory (which caused
5605 issues with 64 bit Intel most commonly) has been fixed.
5609 * A missing kernel feature test has been fixed.
5611 * group based PAG support is still enabled when possible.
5613 * ia32 syscall table support for amd64 has build fixes for modern kernels.
5617 * fopen() is not safe for use with more than 255 file descriptors open;
5618 Emulate it in the afsconf package so afsconf can be used in the fileserver.
5622 * DNS registration is disabled for the loopback adapter, and we make sure
5623 Netbios is turned on.
5628 * Remove use of ubik_Call in the source code so prototypes are used.
5630 * Avoid synchrony in call from the fileserver to the ptserver.
5632 * Fix a bug in the backup suite when restoring.
5634 * fileserver and volserver now log for error conditions which may cause
5637 * rx avoids a stack overrun when more packets are needed.
5639 * volserver avoids holding a lock too long when purging volumes.
5641 * volserver lock initialization fixes
5643 * volserver volume nuke fixes to avoid leaving files behind
5645 * fileserver avoids error when authenticating ptserver requests
5647 * fileserver no longer crashes when GetCPS fails
5649 * salvager enhancements to deal better with corrupt volumes for namei
5653 * aklog deals with KDCs which give "generic" replies to principals not
5656 * Fix bug in cache parameter autotuning
5664 * amd64 pthread library family updates.
5666 * autoconf fixes for kernel feature testing
5668 * keyring PAG support now only enabled if needed features are present
5671 * inline a version of BUG() so we get better oopses
5675 * tsm is updated to work with the new AIX 5 interface.
5679 * Cross compile fixes
5681 * Packaging improvements
5683 * Large file support fixed (Thanks to Chaskiel Grundman)
5685 * Fixes for Leopard seed.
5687 * Installer image updates
5691 * Removes race conditions and a deadlock introduced in 1.4.1
5693 * Fixes ANSI filename option.
5695 * Establishes new connections to file servers when
5696 IP address configuration changes are detected.
5698 * Improved CIFS compatibility
5700 * Cache Manager optimizations
5702 * Fixes vlserver failover when mounting 'root.afs'
5703 (Freelance mode disabled)
5705 * Installs help files in the correct location for use by afscreds.exe
5706 and afs control panel.
5708 * Improve reporting of "over quota" and "disk full" errors.
5710 * Prevent crash when evaluating mount points to volumes that do
5713 * Removes auto-registration of AFS ID in foreign ptservers from
5714 Integrated Logon DLL. This prevents crashes if the DLL is loaded
5715 and unloaded prior to termination of the process.
5717 * SDK moved to \Program Files\OpenAFS\SDK
5719 * NSIS and WiX Installer Frameworks update to the latest versions
5721 * Improvements to the Kerberos Logon Integration
5723 * Prevents exception in Integrated Logon DLL during SysPrep
5725 * Prevents displays of MessageBox dialogs in response to Network Adapter
5728 * Hard Dead and Connection Timeout values restricted to the CIFS Session
5731 * Correct writing of BackConnectionsHostNames registry value.
5733 * Properly recycles Volume entries
5735 * The AFS Explorer Shell Extension always finds its resource library.
5737 * The export list for AFSAUTHENT.DLL has been corrected. (The AFS
5738 plugin for NetIDMgr will no longer use 100% of CPU.)
5740 * Renaming files on Microsoft Vista Build 5536 works.
5742 * Better handling of "." directory in fs commands
5744 * Add OpenAFS License text to installers
5746 * fs setquota and fs mkmount commands behave the same as the UNIX
5751 All systems: Major bugfixes.
5757 - Volume dump parsing code in the volserver has better error checking.
5759 - salvager has improved damaged volume handling on namei fileservers.
5761 - fileserver has size validity checks for when large file support is disabled.
5763 - fileserver avoids potentially multiply adding a host to its hash table.
5765 - rxkad client private data storage is allocated dynamically on ticket size.
5767 - Handle universal error code translation for file locking.
5769 - fileserver needs to swap callback connections on a client IP change.
5771 - fileserver host package revised to reduce lock contention.
5773 - Rx has been fixed to count hard acks, thus opening the congestion window.
5775 - All servers support bound Rx sockets (on one interface).
5777 - namei fileserver no longer use lockf() to avoid range locking issues.
5779 - most binaries now support the -version switch.
5781 - backup suite fixes for 64 bit platforms.
5783 - volserver avoids holding holds during volume purges.
5785 - volserver avoids losing files on namei during vos zap.
5789 - fileserver now properly supports large files.
5791 - TSM updates for AIX 5
5793 - Kernel module avoids leaking Rx packets.
5795 - Avoid use of global ubik client structure in fileserver.
5797 - Update ubik call client interface to allow for prototyping.
5799 - audit logging fixes when stdarg does not provide integral va_list type.
5803 - A bug where the client kernel module could free stack memory (which
5804 caused issues with 64 bit Intel most commonly) has been fixed.
5806 - Packaging fixes and updates.
5808 - Uninstaller added.
5810 - Fix large file support.
5814 - autoconf kernel feature testing has been restructured.
5816 - PAG garbage collection is enabled by default.
5818 - Kerberos updates for RHEL3.
5820 - Fix POSIX lock enrollment for older Linux kernels.
5822 - Updates for new 2.6 kernels.
5824 - Avoid deadlocks in put_inode handler.
5826 - Keyring-based PAG support.
5828 - Fixes to avoid getting better oops info in the kernel.
5832 - Remove some kernel symbol bindings for symbols we don't use.
5834 - Cleanup for loopback mount of AFS on Solaris 10.
5836 - Avoid issues with stdio not supporting file descriptors above 255 on
5837 Solaris 8 and lower in the fileserver.
5841 All systems: Major bugfixes.
5842 New systems: MacOS 10.4 (PowerPC and Intel)
5848 - Several race conditions in the host tracking and handling in the fileserver
5849 which could cause inconsistent behavior and crashes have been fixed.
5851 - A fileserver bug where a reference to a volume could be leaked and later
5852 cause a deadlock as a result of a bulk status call
5854 - Reference counting of fileserver objects in unsigned 32 bit integers
5855 instead of signed 16 bit integers.
5857 - Avoid type mismatches when handling time values (betweemn 32 bit and 64 bit
5860 - Fix a memory leak during multilevel packet queue handling.
5862 - Audit log output had been updated to include FIDs for newly created files.
5866 - 64 bit (large file) inodes are supported.
5868 - Salvager will now handle large (>4gb) partitions.
5874 - asetkey is now included to ease Kerberos 5 integration for server
5877 - A new fileserver statistics collection including callback statistics was
5880 - man pages are now generated.
5884 - Fixes error message problems experienced by fs.exe and the AFS Explorer
5885 Shell Extensions related to the use of Universal Error Codes by the
5888 - Adds full SMB/CIFS support for byte range locking. In this implementation
5889 all locks are allocated locally and the AFS lock privilege is ignored.
5890 While this will not prevent two processes on different machines from
5891 simultaneously writing to the same file, it will prevent two processes
5892 on the same machine from doing so.
5894 - The UP server check period has been reduced to once every ten minutes to
5895 match the period used by the UNIX clients. The shorter period will
5896 assist clients maintain RX connections through NATs.
5898 - Fixes the DOWN server check logic to ensure that any server that responds
5899 to a check is marked UP unless it is in the process of restarting.
5901 - Add logic to better handle objects that no longer exist on the file server.
5904 - Prevent the removal of existing drive mappings by "afscreds.exe -M"
5906 - Fixes the procmgmt library so that it doesn't cause applications that
5909 - Improves the warnings written to the afsd_init.log file when the
5910 Windows RPC Protocol drivers are improperly configured.
5912 - Fixes "fs setserverprefs -vlserver". Multiple calls with the same
5913 server parameter could result in a crash of afsd_service.exe.
5915 - The SMB/CIFS layer was audited for reference miscounts and memory leaks.
5916 All SMB objects are now properly counted, locked, and released when
5919 - Prevent file truncation of the user does not have the appropriate access.
5921 - Token management was re-written to allow user tokens to be preserved
5922 during integrated login and freed after logoff is complete.
5924 - Added a mechanism by which abandoned SMB virtual circuits can be
5925 detected and the associated resources cleaned up.
5927 - Prevent the allocation of SMB file handles with a value of 0 or 0xFFFF
5928 which would be considered invalid by Windows applications.
5930 - Fixed the processing of cell names to ensure that they are always
5931 treated as case insensitive strings.
5933 - Fixed the network provider code to avoid querying the profile location
5934 if integrated login is disabled.
5936 - If a mount point string is empty, return Path Not Found to the application.
5938 - Windows returns WSAECONNRESET when an ICMP packet is received in response
5939 to a transmitted UDP packet that cannot be delivered. Do not mark the
5940 connection as bad but instead retry the request.
5942 - Fix the data written to the registry as part of the BackConnectionHostnames
5945 - Fixed the rx-lwp implementation to always generate unique rx call
5948 - The default "fs minidump" type now includes data segments.
5953 * -nosettime is now the default for afsd. Use "-settime" to get the
5956 * OpenBSD is now supported.
5958 * Mountpoint directory information is now only faked for cross-cell
5959 mountpoints when using the -fakestat flag (e.g. for the directories
5960 under /afs, but not for most other volumes mounted inside the cell).
5961 The -fakestat-all switch can be used to fake information for all
5964 * When fakestat is enabled on MacOSX, the Finder can be used to browse
5965 a fully-populated /afs directory. However, this precludes reliable
5966 use of entire volumes as MacOS bundles (i.e. containing a Contents
5967 directory in the root of the volume).
5969 * Mountpoint directory information can be faked by the cache manager,
5970 making operations such as stat'ing all cells under /afs much faster.
5971 This is enabled by passing -fakestat to afsd, but might not be stable
5977 * The kaserver now defaults to not allowing interrealm authentication,
5978 due to security vulnerabilities in the krb4 protocol. The new
5979 "-crossrealm" flag to the kaserver is provided to reenable interrealm
5980 authentication if desired.
5982 * RedHat Linux 9.0 is now supported.
5984 * Solaris 9 12/02 is now supported. Solaris 7 and 8 x86 should now
5987 * On Linux machines using 2.2 series kernels, 2.2.19 or higher is now
5990 * An OpenAFS 1.2.9 afsd will not work with kernel modules built from
5991 an earlier OpenAFS release. In general, using a mismatched afsd and
5992 kernel modules set is unsupported; it is not recommended that you use
5993 such a configuration on a regular basis.
5998 * Mountpoint directory information is now only faked for cross-cell
5999 mountpoits when using the -fakestat flag (e.g. for the directories
6000 under /afs, but not for most other volumes mounted inside the cell).
6001 The -fakestat-all switch can be used to fake information for all
6004 * HPUX 11.0 is now supported.
6006 * It is now possible for AFS to use Kerberos 5 directly, via rxkad 2b.
6007 See the OpenAFS 1.2.8 Release Notes for more information on using this
6010 * An NFS translator kernel module is now included and compiled by
6011 default for Solaris only.
6016 * MacOS X 10.2 is now supported. FreeBSD 4.3 and later support is
6017 included in this release, but is still under active development and
6018 should only be used by those doing active development on the OpenAFS
6021 * When fakestat is enabled on MacOSX, the Finder can be used to browse a
6022 fully-populated /afs directory. However, this precludes reliable use
6023 of entire volumes as MacOS bundles (i.e. containing a Contents
6024 directory in the root of the volume).
6026 * The fileserver will now use Rx pings to determine if clients are
6027 reachable prior to allocating resources to them, to prevent asymmetric
6028 clients from consuming all fileserver resources.
6033 * Mountpoint directory information can be faked by the cache manager,
6034 making operations such as stat'ing all cells under /afs much faster.
6035 This is enabled by passing -fakestat to afsd.
6037 * Solaris 9 FCS and Solaris 7 and 8 x86 are now supported.
6042 * A remote denial of service attack in the AIX and IRIX clients has been
6043 fixed. Users of those platforms are strongly encouraged to upgrade.
6045 * Fixed race conditions in fileserver that could result in crash.
6050 * Server logfiles now more consistant about format in which hosts are
6053 * vfsck on Solaris will now allow force runs (using -y flag) even if old
6059 * Cell aliases for dynroot can be specified in the CellAlias file in
6060 /usr/vice/etc or /usr/local/etc/openafs, in format "realname alias",
6061 one per line. They can also be managed at runtime with "fs newalias"
6062 and "fs listaliases".
6067 * Solaris 9 and Linux PA-RISC are now supported.
6069 * fileserver will not erroneously delay legitimate errors for 3 seconds
6070 after 10 errors are returned (e.g. stat() on a directory you can't
6073 * Rx MTU calculation now works for Irix, Solaris and Linux
6075 * If afsd is started with the -dynroot flag, /afs will be locally
6076 generated from the CellServDB. AFSDB cells will be mounted
6077 automatically upon access.
6079 * The namei fileserver allows vice "partitions" to be directories
6080 instead of partitions and will attach and display accordingly.
6081 Creating the file "AlwaysAttach" in the /vicepX directory is used as
6082 the trigger to attach it.
6084 * TSM support for butc no longer requires editing a Makefile, simply
6085 specify the --enable-tivoli-tsm configure option.
6087 * Linux builds no longer require source changes every time the kernel
6088 inode structure changes; the OpenAFS sources will now configure itself
6089 to the actual inode structure as defined in the kernel sources.
6094 * vfsck on Digital UNIX and Solaris will now refuse to fsck mounted
6100 * AFS now supports --prefix and the other directory options of
6101 configure. By default AFS builds assuming it will be installed in
6102 /usr/local. In order to get traditional AFS directory paths (/usr/afs
6103 and /usr/vice/etc) use the --enable-transarc-paths option to
6104 configure. More details on the new directory layout are found in
6110 * Windows 95/98/ME/NT/2000 - Consistent versioning: Installation, AFS
6111 Control Center, Client dialog boxes and properties pages for
6112 executables display a consistent OpenAFS version number. Installation
6113 detects previous installation and prompts the user for upgrade
6116 * Windows 95/98/ME/NT/2000 - Installation features: During installation
6117 the user can select the source of the CellservDB file, AFS home cell,
6118 and drive mappings. During installation a drive path mapping can
6119 include a variable that will be substituted with the current UserName
6122 * Windows 2000/NT - Integrated logon: The Integrated Logon feature
6125 * Windows 95/98/ME - Logon script features: The Windows 95/98/ME client
6126 now offers a command-line option for starting up the AFS client
6127 without authenication. It is now possilbe to start the AFS client
6128 first and obtain tokens, and map drives all through Windows scripts.
6129 This helps using Windows 95/98/ME client in Kerberos 5 environment.
6131 * Windows 2000/NT - LANA numbers: AFS client now scans the LANA numbers
6132 to establish the correct NETBIOS connection. NetBEUI is no longer
6133 needed. The user no longer needs to find the correct LANA number.
6135 * Windows 2000/NT - OpenAFS naming consistancy: Further progress has
6136 been made to remove references to "Transarc AFS" and replace with
6142 * AFS now builds with configure. The README for building has been
6143 updated and includes full details.
6145 * A client system can now have multiple sysname values for @sys. They
6146 will be searched in order when looking up files in AFS. The
6147 -newsysname argument to fs sysname can be repeated to set multiple
6150 * A new system group is created for new cells (system:ptsviewers with id
6151 -203). If this group exists, members of this group can examine and
6152 read the entire protection database. They can examine all users and
6153 groups and can get the membership of any group.
6155 * A new program, pt_util has been added to the distribution. This
6156 program allows users to print the contents of the protection database
6157 or to edit the protection database without running a ptserver. It can
6158 be used to set up a new cell without ever running in noauth mode. Run
6159 pt_util -h for help.
6161 * The fs setcrypt and fs getcrypt commands have been added. These
6162 commands allow the system administrator to require that the client
6163 encrypt all authenticated traffic between the client workstation and
6164 AFS. The encryption used is weak, but is likely better than sending
6165 unencrypted traffic in most environments. Some functions, such as
6166 looking for a volume may not be encrypted, but data transfer certainly
6167 is. By default data is not encrypted. At this time no significant
6168 experimentation with server performance has been conducted.
6170 * By default AFS is compiled with AFS_AFSDB_ENV, enabling the -afsdb
6171 option to be given to afsd on startup. If this option is used, then
6172 new cells will be looked up using AFSDB records stored in DNS if they
6173 are not found in CellServDB. This means that users can create
6174 cross-cell mountpoints in directories they control to access cells not
6175 in root.afs, and that cells in root.afs need not be in the client's
6178 * AFS database servers can be marked as read-only clones. Surround the
6179 hostname in square brackets on the bos addhost command and the
6180 database server will never be elected sync site. This is useful for
6181 cells distributed over a wide region.
6183 * The AFS servers now support the -syslog flag. This flag causes them
6184 to log to syslog rather than to files. This flag is not supported on
6185 NT. For all servers besides the salvager, the flag can also be
6186 specified as -syslog=facility, where facility is an integer facility
6187 code from syslog.h. A -syslogfacility option is provided for the
6188 salvager to accomplish the same goal.
6190 * If the --enable-fast-restart flag is given when configuring AFS, then
6191 the salvager supports the -dontsalvage flag which causes it to exit
6192 without salvaging any volumes. If this is configured into the third
6193 command of a fs process, then the fileserver will start without
6194 salvaging. It will fail to attach volumes that need salvaging and
6195 they can be salvaged manually. This provides significantly better
6196 server startup performance at the cost of administrative complexity.
6198 * If the --enable-bitmap-later flag is given when configuring AFS, then
6199 the fileserver creates bitmaps for free vnodes on demand, allowing
6202 * If bosserver finds a BosConfig.new file at startup, it reads this file
6203 and renames it to BosConfig. This allows bosserver to be reconfigured
6206 * The bosserver can be placed in a restricted mode in which AFS
6207 superusers are only granted limited access to the server host. The
6208 following functionality is disabled when restricted mode is in use:
6211 bos getlog (except for files with no '/'s in their name)*
6217 specific exceptions are made for functionality that "bos salvage"
6220 A cron bnode who's name is "salvage-tmp", time is now, and command
6221 begins with "/usr/afs/bin/salvager" may be created. This bnode deletes
6222 itself when complete, so no special "delete" support is needed. This
6223 functionality may be removed in the future if a "Salvage" RPC is
6226 The file with the exact path /usr/afs/logs/SalvageLog may be fetched,
6227 since that is how bos salvage [...] -showlog is implimented.
6229 Restricted mode is enabled using a new bos command (bos setrestricted)
6230 or bossever command line switch (bosserver -restricted). Restricted
6231 mode can be disabled by a) sending the bosserver process a SIGFPE
6232 (which will then allow restricted operations until the next restart or
6233 setrestricted command) or b) editing /usr/afs/local/BosConfig (or
6234 BosConfig.new), and restarting the bosserver.
6236 * The bos UserList of trusted administrators can now contain cross-realm
6237 Kerberos principals.
6239 * udebug now takes --server not --servers.
6241 * Several error messages have been improved to include volume numbers.
6243 * Several new ports have been included for UNIX platforms: Darwin
6244 (ppc_darwin_12 and ppc_darwin_13), Linux 2.4 (i386_linux24), Linux on
6245 the Powerpc (ppc_linux22 and ppc_linux24), Linux on the Sparc
6246 (sparc_linux22, sparc64_linux22 and sparc64_linux24).
6248 * Incomplete FreeBSD and Alpha Linux ports are included. The FreeBSD
6249 port has a working server and the Alpha Linux port has a partially
6252 * A native client for Windows 95/98/ME has been added to the
6253 distribution. With this program, a gateway machine is no longer
6254 required for Windows 9x to access AFS files. One drive letter will be
6255 created on your machine by default - Z:. The Z: drive will be the
6256 root of the AFS tree, allowing you to browse all sites that have AFS
6257 servers available. Additional drive letters can be defined for other
6258 AFS directories. A Windows Explorer shell extension is included that
6259 allows you to right click on items within an AFS tree to bring up an
6260 "AFS" menu item and perform various operations on a file or directory.
6261 The most useful item is "Access Control Lists", which allows you to
6262 view and edit the permissions of a particular directory. Command line
6263 tools are also available in the install directory. These commands
6264 include klog, unlog, tokens, kpasswd, symlink, fs and pts. The
6265 installable includes a readme file that contains more information on
6266 how to use the client program and known issues.
6268 * Support for large caches in afsd. Cachefiles are stored in
6269 subdirectories. The default is 2048 files per subdirectory, which
6270 should work fine in most situations. You can use the new afsd option
6271 -files_per_subdir to change this number. Note that the first time you
6272 run afsd with this patch, your cachefiles will get moved into
6273 subdirectories. If you subsequently run an older version of afsd, you
6274 will lose all your cached files.