1 User-Visible OpenAFS Changes
3 OpenAFS 1.6.0 (in progress)
7 * vos now properly deals with matching sites when servers are
10 * Don't stop Rx keepalives after an ackall is received, avoiding
11 spurious connection timeouts. (128848)
13 * Don't retry Rx calls on channels returning busy errors and improve
14 Rx busy call channel error handling. (128671)
16 * Properly enable Rx connection hard timeouts.
18 * Initialize rx_multi lock before use.
20 * Avoid spurious crashes when initializing in "backup" client.
22 * Revert UUID support in vos.
24 * pt_util fixed to properly create new databases.
26 * MTU discovery now properly shut down on call reset.
30 * Fix ptserver supergroups support on 64 bit platforms.
32 * Demand attach salvaging doesn't use freed volume pointers.
34 * Properly hold host lock during host enumeration in fileserver.
36 * Attempt to recovery more quickly from timed out volume release
39 * Auditing now properly byte order swaps IP addresses when printing.
41 * vos split now has improved error handling.
43 * Many changes to again support Windows fileservers.
45 * During volume removal, data removal speed improved.
47 * Improve CPU utilization during volume attaching by DAFS.
49 * In salvager check-only mode, avoid potentially fixing a vnode.
51 * Fix support for large (greater than 2gb) volume special files.
53 * Salvager will not crash if multiple or bad volume link tables are
56 * Avoid erroneous full dump by remembering which sites were out of
57 date at the start of the release.
59 * A deleted volume can now be recreated properly.
61 * Callbacks are again not broken during whole partition salvages.
63 * Positional vectored IO fixed for largefile (>2GB) capable systems.
65 * Fileserver per-client thread usage again properly enforced.
67 * Anonymous dropbox support improved and drawbacks documented.
69 * Demand attach: ensure vnodes are not reallocated while in use due to
74 * afs_config will not longer set the Tray Icon State in the registry
75 if the checkbox is not present in the dialog. (128591)
77 * AFS Explorer Shell Extension now works from folder backgrounds.
78 Overlays for mount points and symlinks are present in the dll, but
79 are not registered at present by the installers.
81 * Do not use RankServerInterval registry value as the value for
82 PerformanceTuningInterval.
84 * When the data version of a mountpoint or symlink changes, the target
85 string in the cm_scache_t object must be cleared.
87 * "fs checkservers" now includes vldb servers in the output and only
88 lists multi-homed servers once. A multi-homed server that has at
89 least one up interface is no longer considered to be down.
91 * When asynchronously storing dirty data buffers to the file server
92 ensure that (a) the cm_scache_t object and the cm_buf_t object are
93 for the same File ID so that locking and signalling work properly;
94 and (b) if the FID no longer exists on the file server, do not
95 panic, just discard the buffer.
97 * When processing VNOVOL, VMOVED and VOFFLINE errors perform server
98 comparisons by UUID or address and not simply by cm_server_t
99 pointer. Otherwise, server failover may not succeed.
101 * Do not preserve status information for cm_scache_t objects when the
102 issuing server is multi-homed.
104 * Giving up all callbacks when shutting down or suspending the machine
105 is now significantly faster due to the use of an rx_multi
106 implementation. (This functionality is still off by default and
107 must be activated by a registry value.)
109 * Race conditions were possible when updating the state of the
110 cm_volume_t flags and when moving the volumes within the least
113 * Ensure that the lanahelper library does not perform a NCBRESET of
114 each lan adapter when enumerating the current network bindings.
115 Correcting this permits OpenAFS to work on Windows 7 when the
116 network adapter settings change.
118 * Fix creation of mount points and symlinks as \\AFS\xxxx
120 * Icon tray state now conditionally set. (128591)
122 All UNIX client platforms
124 * Servers now marked down when GetCapabilities returns error.
126 * In-use vcache count is now properly tracked.
128 * Check for /afs existance before starting, unless -nomount is
131 * Avoid a potential panic when using /afs/.:mount syntax.
133 * Avoid a panic in memcache mode due to missing CellItems file.
135 * FUSE client support fixed for non-/afs mounts.
139 * Fix socket termination on shutdown.
141 * Support for 7.2, 7.3, 7.4 and 8.2 included.
143 * References to vcaches are no longer leaked during root or reclaim.
145 * Remove support for "Giant" lock as we no longer need to use it.
147 * Don't sleep with AFS GLOCK.
149 * Properly enable 64 bit long long support.
151 * Restore support for FreeBSD 7 (128612)
153 * Fix locking issues at shutdown.
157 * Support through kernel 2.6.38.
159 * Red Hat init script allows deferring for a new binary restart.
161 * Red Hat packaging now properly supports RHEL6.
163 * Use rx_Readv in cache bypass to improve performance.
165 * Properly handle 0-length replies during cache bypass operations.
167 * Properly handle non-contiguous readpage cache bypass operations.
169 * Do proper locking when transitioning to or from cache bypass.
171 * Avoid extra runs of vcache freeing routine. (128756)
173 * Perform vcache eviction via a fast path before visiting vcaches
174 where sleep is needed.
178 * Properly handle setpag errors. PAGs are not supported.
180 * Check for unloaded kernel extensions when decoding AFS panics.
182 * Disable "get tokens at login" in prefs pane if AD authentication
183 plugin is configured.
185 * aklog AuthorizationPlugin now provided.
189 * Bug fixes for issues introduced previously in 1.5 series.
191 * Support through OpenBSD 4.8.
195 * Switch to ioctl() syscall replacement for Solaris 11 since syscall
198 * Fix support for Solaris pre-10.
200 * Corrected Solaris 11 startup script.
202 * vcache mappings freed on shutdown to avoid panic.
205 OpenAFS 1.5.78 (2010-11-04)
209 * Revisions to Rx to fix performance issues.
211 * Make fs getfid behave consistently across all platforms. (128372)
213 * Properly check IDs handed to pts when creating users or groups so
214 useful error messages can be provided. (128343)
216 * Correct byte order handling of port in afsconf_LookupServer for SRV
219 * Force a full dump when releasing to a site which was previously
220 marked "don't use", in case the previous clone was out of date.
224 * Demand salvage of attached volumes now correctly track attachment
227 * Avoid a potential crash due to failure to hold a lock when attaching
232 * Track SMB connections by SID rather than username.
234 * Error write attempts to known-readonly volumes earlier.
236 * Validate directory buffers to avoid potential crashes.
238 * Handle VIO errors from bulkstatus.
240 * Make PMTU discovery configurable and register error handlers for it.
242 All UNIX client platforms
244 * Use larger I/O sizes in memcache to improve performance.
246 * Avoid potential alignment issues doing I/O for pioctl calls.
250 * Avoid panicing if the listener process is not findable.
252 * Avoid deadlock issues while performing lookups.
256 * Handle stale file handle errors for some cache partition types.
258 * Avoid blocking with xvcache lock when attempting to free in-use
261 * Build fixes for older kernels.
263 * Properly configure LWP to use ucontext() on platforms where it
266 * Eliminate spurious errors from AFS system call returns. (126230)
270 * Attempt to honor configured Kerberos defaults in Preferences Pane.
273 OpenAFS 1.5.77 (2010-09-08)
277 * Rx path MTU detection will terminate detection in cases where the
278 minimum required packet size cannot be transferred.
280 * vos dryrun mode now shows effects for syncvldb single volume case.
282 * vos dryrun mode now shows "status after" for syncvldb and syncserv.
286 * RXAFS_GetStatistics64 now returns statistics properly.
290 * Attempt to properly identify the local system SMB connection for
293 * Remap timeout and offline errors to proper NT RPC errors.
295 * Properly fail over to other replicas on bulkstat IO errors.
297 * Properly error delete-mode createfile if a file is set readonly.
299 * Validate directory entry buffers to avoid crashing the service.
301 * Log file modes properly.
303 * Log cell name when logging server information.
305 All UNIX client platforms
307 * cacheout program for discarding callbacks is now built.
309 * bulkstatus kernel locking is corrected to avoid a potential panic.
313 * userspace support update
317 * Updated vnode locking for children returned via lookup().
319 * Avoid file open undercount with needed calls to
320 FakeOpen/FakeClose().
322 * Use vnode_pager_setsize to properly track file size during kernel
325 * Update system call installation.
327 * Fix shutdown of Rx kernel listener to avoid potential dereference
330 * Avoid closing vnodes during vnode recycle.
332 * Fix bogus call to FlushVS for vnode reclaims.
336 * Packaging updated for current configure options and built files.
338 * Cache bypass now holds reference on pages during readpage.
340 * s390x setgroups32 patching update.
344 * DNS resolver is reinitialized on IP address change. (126440)
347 OpenAFS 1.5.76 (2010-08-16)
351 * Updates to build-time configuration.
353 * Fix XDR support in Rx to match header definition.
355 * vos status now shows transaction creation, not action creation.
357 * Rx avoids reporting loopback adapters when listing interfaces.
361 * Demand-Attach Fileserver always built and installed (dafileserver,
362 davolserver, dasalvager).
364 * Return VNOVOL from fileserver when a volume is deleted.
366 * Ignore duplicate tags during volume restore operation.
368 * Update inode array after salvage repairs volume.
370 * Zero a corrupted header in memory during salvage to avoid further
373 * Fix NAMEI backend to allow low-numbered volumes to work properly.
375 * ptserver does not include cell name as part of length check for
378 * Updated error messages for unblessed volumes.
380 * vlserver avoids buffer overflow with regex pattern
382 * Attach-time failures now note failures as the rest of the fileserver
385 * Server argument logging will no longer overflow stack.
387 * Provide fast-restart-like unsafe-nosalvage option for DAFS.
389 * Deal with host hash collisions in the fileserver.
393 * Avoid crashing when interpreting a drive letter as potentially
394 matching a cell name.
396 * Properly handle volume package errors.
398 * Allow page recycling from known-readonly content without ensuring
401 * 32 bit tools installer should not override client configuration.
403 * Ensure root scache item has a valid callback when use is attempted.
405 * Freelance directory changes now properly invalidate and replace the
408 All UNIX client platforms
410 * Support disconnected reconnecting with specified UID for PAGless
413 * Proper disconnected vnode reference tracking.
415 * Update server site blacklisting to not return success if nothing was
418 * Avoid a panic during vcache contention due to CVInit vcache
423 * Update for network stack in 8.1/9.0.
433 * Disable PMTU error packet handling.
437 * Debian packaging updated.
439 * freezer interface updates.
443 * Hold references to disconnected mode written vnodes properly.
447 * Handle NFS translator module references for amd64.
449 * INODE fileserver backend support now exists for amd64.
452 OpenAFS 1.5.75 (2010-07-07)
456 * Prevent rx_rpc_stats global lock from being a bottleneck.
458 * Path MTU discovery is now provided to allow traffic to pass networks
459 with sub-1500 byte MTUs and poor fragment handling.
461 * Further reduce Rx NAT ping transmission when enabled.
463 * Update Kerberos 5-based token handling in rxkad from upstream
466 * Update version numbers emitted during build to reflect what is
467 actually being built.
469 * Add "-human" switch for human-readable units in fs diskfree and
472 * vos provides reasons for locked volumes when known.
474 * Do not count retransmission and ping acks as non-idle for Rx
477 * Rx: provide service-specific data getter and setter routines.
479 * Update build-time Kerberos detection.
481 * Updated userspace AFS client.
483 * Beginning of a modernized test suite.
485 * Additional documentation.
487 * Updated documentation, notably the Administrators Guide.
489 * Substantial code cleanup.
493 * Update handling of vnode allocation failures.
495 * DAFS: allow salvaging volumes not known to the fileserver, to allow
496 cleanup of data not attached to a current volume.
498 * Properly handle volumes slated for destruction.
500 * Handle volumes with many files properly.
502 * Force core file generation in bosserver by overriding default
503 resource limits when possible.
505 * Update vlclient and vldb_check.
507 * Avoid potentially corrupting a volume on creation if files are left
508 from previous failed cleanup.
510 * Note volume changed during salvage as needed.
512 * DAFS: do not assume invalid addresses are in fileserver address hash
515 * Avoid tying up fileserver threads with volumes that are being taken
518 * Do not set inUse on volumes for non-DAFS other than in fileserver.
520 * Break origin's callback on target of rename operation.
522 * Avoid unneeded parent directory link updates during some rename
525 * Do not open /dev/console for writing in the fileserver.
527 * DAFS: avoid spurious restarts when binary restarts are configured.
529 * Avoid spurious and unneeded calls to sync(), which can slow down the
534 * Revised SMB QuerySecurityInfo to address issues caused by MS10-020
535 (http://support.microsoft.com/kb/980232)
537 * Prevent use of the AFSCache file contents if mapped to a new
540 * Make fs newcell include behavior compatible with the non-Windows
543 * Provide a registry option (FreelanceImportCellServDB) to pre-create
544 mount points in the AFS root for all cells in CellServDB.
546 * Fix a memory leak in the cm_FreeServerList() routine.
548 * Reduce privilege when reading registry CellServDB.
550 * Add support for RPC Pipe Service NetWkstaGetInfo levels needed for
553 * Prevent overflow when computing quota percentage in Explorer Shell.
556 * Generate meaningful errors for ACL operations on freelance AFS root.
558 * Fix error handling on InlineBulkStatus RPCs.
560 * Show configuration pages for all types of MSI installations.
562 * Improve freemount AFS root directory handling and operations.
564 * Properly validate GetVolumeStatus pioctl responses.
566 * Commit file length changes and dirty buffers when flushing a file.
568 All UNIX client platforms
570 * Update version of files for disk cache.
572 * Do not call afs_FlushVCBs with xvcache lock held, to improve
575 * Add mariner log messages for creating and removing files.
577 * Don't hold xvcache lock while creating symlinks, to improve
580 * Provide -dynroot-sparse mode to not show all cells in CellServDB in
583 * Avoid a potential crash in aklog in linked cell handling.
585 * Log MTU-caused packet retransmission.
587 * Prevent crashes caused be fs checkservers while cache is being set
590 * fs getserverprefs now has a buffer large enough for the default
593 * Report server address when logging warnings.
595 * Avoid panic in GetCapabilities when cell is not known.
597 * Lock process name and id for advisory lock warnings when possible.
599 * Handle need for allocating additional Rx packets.
601 * Properly handle errors from InlineBulkStatus operations.
603 * Fix errors returned from fcntl() on readonly files locked for write.
605 * Flush pending changes to the server on LOCK_EX unlock.
607 * Reflect length changes as a result of callbacks even when file is
610 * Avoid hanging due to error exit when attempting to store a large
611 file to a non-largefile fileserver.
613 * Recover from afs_GetVolSlot errors.
617 * Bugfixes for kernel VFS and network routines.
621 * Provide makesname().
625 * Avoid syscall probes when keyrings are present, by default. (125215)
627 * Remove "Big Kernel Lock" from VFS operations.
629 * Use filehandles for all Linux 2.6 versions to avoid need for matched
632 * Updated RPM packaging.
634 * Fix dkms configuration provided with RPMs.
636 * Hold reference on pages during background I/O for cache bypass.
638 * Fix cache bypass handling of non-largefile fileservers.
640 * Protect truncate_inode_pages mappings with mutex or semaphore as
643 * Fix pagevec use in cache bypass. (127505)
649 * Improve launchd configuration.
651 * Avoid hanging on recursive cache file lock acquisition when user
652 notification is enabled.
654 * Fix and re-enable bulkstat mode.
662 * Precluding unmount while AFS is busy.
664 * Avoid deadlocking when releasing the VFS object.
666 * Stop network interface poller in kernel on AFS shutdown.
668 * Avoid issues with lookups on empty directory names. (127356)
671 OpenAFS 1.5.74 (2010-04-22)
675 * Add "vos setaddrs" command.
677 * Rx library lock contention avoidance between rx_NewCall and
680 * Rx library races due to inconsistent use of rx_connection
681 conn_data_lock to protect the flags field.
683 * Rx library inconsistent use of RX_CALL_TQ_WAIT which could result in
686 * Rx library must signal transmit queue waiters when flushing.
688 * afsmonitor shows busy counts now.
690 * afsmonitor displays xstat callback statistics.
692 * Provide expandgroups for pts mem on a supergroups server.
694 * Provide supergroup option to liste nested groups during pts mem.
698 * Avoid volume lock contention during DAFS startup.
702 * Avoid a race when updating cell vldb server lists that can result in
705 * Avoid a deadlock when managing CM_SCACHESYNC_STOREDATA state
706 operations for directory objects.
708 * Add new Windows Application Event log messages for VBUSY,
709 VRESTARTING, ALL_BUSY, ALL_OFFLINE, and ALL_DOWN.
711 * Reduce lock contention by waiting for cm_buf_t I/O operations.
713 * Split the cm_buf_t flags field to separate the flags that are
714 protected by the cm_buf_t mutex from those protected by the
717 * In cm_UpdateVolumeLocation, avoid searching for a ".readonly" volume
718 on a numeric volume name.
720 * File buffer allocations whose offsets are beyond server EOF should
721 be locally allocated and zero filled. The file server should not be
722 issued a FetchData rpc which is guaranteed to fail.
724 * Enable integrated logon to work with Windows 7/2008 when user logons
725 are performed with a non-Domain Kerberos principal.
727 * Add Protection Error messages to aklog output.
729 All UNIX client platforms
731 * Provide a FUSE-interfacing userspace afs client.
733 * Updates to libuafs userspace cache manager.
735 * Probe servers using GetCapabilities instead of GetTime, thus
736 requiring fewer RPCs.
738 * Fix DNS SRV record handling for cell lookup.
742 * Fix sleep/wakeup routines.
744 * Update for 8.0 release.
748 * Handle high memory addresses correctly.
752 * Make 32 bit AFS syscalls work again.
754 * Work around finder "Duplicate" failure (caused by setting modes on
757 * Disable bulkstat again (will be re-enabled at or before .75).
759 * Provide symlink type hints during readdir.
762 OpenAFS 1.5.73 (2010-03-24)
766 * NAT keepalive support at Rx level.
768 * Corrected SRV record support for cell name canonicalization.
772 * Fix volume callback notification to not notify unaffected clients.
775 * Allow root directory recreation by salvager. (94658)
777 * Numerous DAFS fixes.
779 * Improvements to callback table overflow handling. (126451)
781 * bosserver now shuts down cleanly on SIGTERM.
785 * Prevent the Explorer Shell extension from crashing if symlink
786 creation failed. (126406)
788 * A Rx level NAT ping has been implemented. A registry value enables.
790 * Adds krb5 error message translation to aklog, afscreds,
791 afslogon.dll, the network identity manager afs provider and
794 * Default mode bit settings for file and directory creation are now
795 provided, and can be configured.
797 * An SMB request trace facility is provided and can be enabled for
800 All UNIX client platforms
802 * Rx idle deadtime does not stop file writes.
804 * Disconnected AFS no longer has a race condition during remove ops.
806 * Fakestat avoids a condition which could cause it to block on network
809 * Several fixes to handle interruptions in vos operations. (33360,
812 * Allow more sysnames in a sysname list.
814 * Attempt to enforce timeouts on AFSDB lookups.
818 * Clean up properly on mount failure.
820 * Add entry to /etc/vfs to allow umount to work.
824 * Several issues to deal with older kernels.
826 * Avoid leaking the global lock in the /proc cellservdb code.
828 * Keyring destruction now cleans up all tokens.
830 * Keyring quotas are not enforced against root.
834 * Some FSEvents hinting for authentication events now done. (23781)
836 * Update uninstaller. (125634)
838 * Rewrite afssettings and fstab code to avoid licensing issue with
841 * Growl client for user monitoring of AFS events included.
843 * Properly support insert-only dropboxes.
845 * Add bulkstat support.
847 * Include support for moving in Finder across mount points.
849 * Preferences Pane includes support for Kerberos 5 ticket renewal.
852 OpenAFS 1.5.72 (2010-02-15)
856 * Provide internationalization support in com_err.
858 * Fix array length checking to avoid crashes when checking for a
859 volume type based on name in vos.
863 * Provide backward compatible "-f" flag to salvager for force mode.
867 * Restore use of DNS AFSDB and SRV records by kaserver clients.
869 All UNIX client platforms
871 * Fix client cache file truncation to not lose chunks when truncating
874 * Ensure a cache writeback hook is installed in the client (bug from
877 * Avoid spurious free memory warnings during clean shutdown.
879 * Fakestat mode avoids AFSDB lookups.
881 * "fs storebehind" now correctly reports errors on readonly volumes.
883 * Additional documentation for "fs getcacheparms"
885 * Forced new uuid generation with "fs uuid -generate" now works
886 enforced permission correctly.
890 * Add optimized Rx event handler in kernel.
892 * Installer now allows installing an older version.
894 * Panic decoder can now deal with MacOS 10.5 again.
896 * MacOS ._ files are now correctly not looked up as cellnames.
900 * To deal with SELinux file labeling, try cache accesses with current
901 credentials in event of failure.
903 * Rx XDR encoding bug on i386 Linux is fixed (bug introduced in
908 * Code compilation fixes.
912 * Update for OpenBSD 4.6.
915 OpenAFS 1.5.69 (2010-01-19)
919 * Configuration of BOSserver no longer defaults to weekly restarts
922 * Provide BOS restricted mode by default.
924 * Add support for "vos endtrans" command.
926 * Default to providing full output from vos listvol.
928 * Correct additional-address tracking in the fileserver.
930 * Improve Rx performance by not unnecessarily dropping and reacquiring
931 call locks in read and write processes.
933 * Avoid crashes when monitoring volserver transactions across
934 potential transaction garbage collection.
936 * Numerous warning fixes.
940 * Avoid saving fileserver state in demand attach fileserver when
943 * Demand attach fileserver allows other callers to schedule salvages.
945 * Demand attach "bos salvage" now works correctly with restricted
950 * Numerous changes to the client-internal btree directory handling to
953 * fs examine reports owner and group ids as signed values (PTS groups
956 * Preclude corruption due to races writing to smb buffers.
958 * Allow MTU settings in registry to be used.
960 * Apply MTU to both send and receive sizes.
962 All UNIX client platforms
964 * Avoid double-freeing Rx call structure if reading a response from
965 the file server results in a short read.
967 * Handle negative lengths in FetchStatus results correctly.
969 * Properly clean up allocated memory at shutdown.
971 * Default to AFSDB compiled into the cache manager.
973 * Avoid inadvertant disclosure of stat() information to clients not so
976 * Correct a bug with AFSDB lookups introduced with SRV record support.
980 * Install kernel panic processing tool in /Library/OpenAFS/Tools.
982 * Include debugging symbols for kernel extension in additional package.
984 * Support "Application Firewall" users.
986 * Avoid ._cellname AFSDB lookups.
988 * Compile preferences pane as a universal binary.
992 * Use splice to speed up storing files.
994 * When using memcache, avoid duplicating work in readpages.
996 * Use dget_parent to safely find an inode's parent.
998 * Disable access time updates in our superblock.
1000 * Avoid crashing doing writeback if no credentials were stashed at
1003 * Simplify keyring support.
1005 * Properly clean up vcache in event of failed mount.
1009 * Update for current FreeBSD 8.
1013 * Abstractly manipulate groups as now required.
1015 * Abstractly access time instead of using lbolt directly.
1018 OpenAFS 1.5.68 (2009-12-08)
1022 * aklog now attempts to convert non-AFS errors to human-readable
1025 * Make stack not executable when compiling assembler source with GCC.
1027 * Numerous source warning cleanups and code reorganization.
1029 All server platforms
1031 * Compute midnight for volume statistics calculation from local time.
1033 * Salvager now orphans duplicate special inodes when running to allow
1034 recovery in event of a problem, instead of simply ignoring the
1037 * Support to ensure a server panic attempt leaves a core and thus
1038 restarts in a timely manner, rather than potentially hanging. Use
1039 panic to attempt cleanup before leaving a core when possible.
1041 * Volume sync data reported during bulkstatus is now set correctly.
1043 * Provide better tuning for fileserver file descriptor caching.
1045 * Allow more than 128 threads in fileserver by modifying host
1046 structure in-use tracking.
1048 * Avoid crashes getting volume server status during transaction
1051 * Improved logging of offline volume conditions.
1053 * Correct volume statistics when cloning a volume.
1055 * Avoid referencing host structures in the fileserver which are marked
1058 * Demand attach fileserver corrections to avoid coring during an
1061 * host array bounds checking corrections to avoid buffer overflow.
1063 * Handle special inodes correctly when promoting an inode fileserver
1064 readonly volume to read-write.
1068 * Set the DOS Readonly attribute on a file/directory whenever the unix
1069 mode combined with the mask 0200 is true. Previously there was a
1070 discrepency between the mask used for testing for readonly behavior
1071 and that used for setting the attribute.
1073 * Disable AFSVolSync based .readonly "whole-volume callback" support
1074 because the all file servers prior to 1.5.67 (and perhaps 1.4.12) do
1075 not properly assign a value to the AFSVolSync structure in bulk
1076 status RPC responses.
1078 * Improve the error output from aklog to output the value from krb5
1079 error_message() if the afs_com_err output indicates an unknown
1082 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and do not permit
1083 them to be exposed to the smb redirector.
1085 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT to avoid
1086 confusion within the smb redirector.
1088 * Fix the byte order assigned to port numbers associated with AFSDB
1089 record lookups. They must be network byte order not host byte
1092 * Add dynamic server ranking based on RPC round trip time
1095 All UNIX client platforms
1097 * Additional shutdown-time memory leaks removed.
1099 * Improved logging of resource contention.
1101 * Provide dumping for Rx debug packet tracking support in source.
1103 * Update afscp test client to build, and provide an unlock client.
1105 * Client buffers for directory parsing can now be allocated beyond the
1106 fixed set formerly provided.
1108 * Work around race condition when manipulating read-only volume
1111 * Bugfixes to get PAG value pioctl.
1113 * Bugfixes to SRV record support.
1117 * Path MTU tracking code cleanup.
1119 * Avoid an oops due to racing with vcache recycling thread.
1121 * Changes to keyring PAG handling: for sufficiently new kernels, use
1122 only keyring-based PAGs, and disable group PAGs entirely.
1124 * Updates to the kernel page cache interface: writing pages will now
1125 not spuriously leak page locks, and will avoid requiring duplicate
1128 * Credential references are now tracked using native atomic counters.
1130 * Kernel mutex/semaphore lock ordering fix to avoid deadlocks.
1132 * Manipulate disk cache with credentials used to initialize it, to
1133 avoid security issues.
1137 * Fix fstrace message catalog location.
1139 * Fix kernel fstrace logging.
1142 OpenAFS 1.5.66 (2009-10-25)
1146 * Avoid calling exit() in library code.
1148 * Add rx window size and peer timeout tuning APIs.
1150 * Correct rx peer timeout handling to disallow 0ms timeouts.
1152 * Correct calculation of rx RTT by disregarding retransmitted packets.
1154 * vos manpages updated to reflect changes in recent versions.
1156 * GNU-style long options (e.g. --cell) are now supported in all
1159 * fs listacl can now print a command to recreate the current ACL.
1161 All server platforms
1163 * Fix a race on transaction objects in the volserver which can cause a
1166 * Avoid destroying and setting to NULL the callback connection when it
1167 could still be being used.
1169 * Correct unlink handling in salvager.
1171 * Improve error messages due to I/O errors in the volserver.
1173 * Correct an issue which caused converted RO to RW volumes on namei
1174 fileservers to not come online immediately.
1178 * Official support for Windows 7 and Server 2008 R2.
1180 * Prevent a file server bug (FetchData returning an invalid length
1181 instead of zero) from causing an "unexpected network error" when
1184 * Promote DNS SRV records as superior to DNS AFSDB records. Support
1185 arbitrary port numbers for vldb servers.
1187 * Add AFSVolSync based .readonly "whole-volume callback" support.
1188 With this functionality, multiple objects from a .readonly volume
1189 can have their status validated by issuing a single
1190 RXAFS_FetchStatus RPC.
1192 * Remove drive mapping functionality and service start/stop from
1195 * Remove drive mapping functionality from afs_config.exe.
1197 * Use {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab" to
1198 restore access to drive mapping functionality in afscreds.exe and
1201 * Adjust SMB error return codes to avoid returning STATUS_TIMEOUT
1202 which results in the SMB redirector disconnecting.
1204 * Network Identity Manager OpenAFS Provider now provides its own "AFS
1205 lock" notification icon to report the status of "have tokens, have
1206 no tokens, service not started, service started but inaccessible".
1207 Hovering over the icon lists the cells for which tokens exist (if
1208 any) and the OpenAFS version number. Double-clicking executes the
1209 Network Identity Manager default action.
1211 * Prevent pioctl calls from retrying indefinitely when a sharing
1212 violation error occurs.
1214 All UNIX client platforms
1216 * Correct a condition which could discard the error from initializing
1219 * Avoid using invalid references to afs_Conn connection structures,
1220 and thus potentially producing invalid data when a retry is needed.
1222 * SRV records are now supported for discovering AFS servers.
1226 * Correct writepage behavior.
1228 * Fix error code handling in the writepage code.
1230 * Avoid leaking page locks, which could potentially hang a machine.
1234 * Preferences Pane improvements.
1238 * Avoiding attempting to handle critical signals in servers, so that
1239 core file handling works correctly.
1242 OpenAFS 1.5.65 (2009-10-06)
1246 * Code compilation warning fixes, to enable better finding and
1249 * Provide configure-time switch to enable code warning compilation.
1251 All server platforms
1253 * Demand-attach fileserver now makes volume LRU list operations
1254 exclusive operations to avoid races during adding to the list.
1256 * Fileservers now avoid potential "negative length" fetches.
1258 * A leak in host tracking objects in the fileserver has been fixed.
1260 * Salvager now unlinks all files by full path, to deal with the change
1261 to not chdir for core file tracking.
1263 * Salvager avoids asserting if the volume header is unreadable.
1265 * Demand-attach fileserver puts back volume references from fssync
1270 * Improved service response to suspend and shutdown event
1273 * Avoid a bug in the file server that can result in an invalid length
1274 being returned as part of a fetch data response if the client
1275 attempts to read beyond the length of the file.
1277 * Do not publish a default stream object for directories and mount
1278 point objects. This was impacting the ability of some Windows XP
1279 systems to save roaming profiles.
1281 All UNIX client platforms
1283 * A bug which could cause erroneous handling of lengths on data reads
1286 * A bug where erroneous length returns from the fileserver could
1287 result in a false error has been fixed.
1291 * Background page copies are now supported for enhanced disk cache
1294 * Blocking readahead is supported in readpages() to reduce overhead.
1296 * Use readpage() instead of read() to access cache data to avail disk
1297 cache users of the kernel backing cache for improved performance.
1299 * Minimize credential handling for improved performance.
1303 * Preferences Pane cleanup.
1307 * Provide a fs_pathconf method with sensible defaults.
1309 * Provide a _PC_FILESIZEBITS method to fix some NFS translator
1313 OpenAFS 1.5.64 (2009-09-22)
1315 All server platforms
1317 * The demand attach fileserver now puts back volume references gotten
1318 via the fssync interface.
1320 * The demand attach fileserver had a structure reference error, which
1321 has been correected.
1325 * Restores Windows 2000 compatibility.
1327 * Fixes a data consistency error between the output of NetWkstaGetInfo
1328 and NetServerGetInfo RPCs, specify the Lan workstation group name
1329 "AFS", and report server name as "AFS" instead of "\\AFS" when the
1330 caller asks for "\\AFS".
1332 * Enables executables to be run from \\AFS on Windows 7. Returns
1333 "Name not found" instead of "File not found" when a directory or
1334 file name cannot be found. This avoids loader errors when system
1335 dlls cannot be located in the executable directory.
1337 * Prevents cache manager from marking the file server "down" when the
1338 data returned in response to either RXAFS_FetchData64 or
1339 RXAFS_StoreData64 is invalid.
1341 * Adds pioctl data validation to the AFS Explorer Shell extension.
1343 All UNIX client platforms
1345 * A bug which could cause a kernel panic in 1.5.63 has been corrected.
1346 This would manifest as a GetDCache panic or oops.
1350 * aklog -setpag works again with recent kernels when keyring is in
1355 * When Fast User Switch is in use, AFS login is now handled correctly
1356 by the integration tool included with the preferences pane.
1358 * Several packaging bugs have been corrected.
1361 OpenAFS 1.5.63 (2009-09-11)
1365 * The restorevol command is now documented and installed as a user
1368 * The uss command now properly translates vldb entries to its expected
1369 format when handling them in all cases.
1371 * Documentation now refers to Kerberos instead of kaserver.
1373 All server platforms
1375 * bosserver now handles BosConfig.new when restarting, allowing
1376 configuration to be replaced at restart time rather than with bos
1377 delete and bos create. Documentation is updated to reflect this.
1379 * The demand attach fileservice not longer potentially hangs trying to
1380 terminate demand-salvages which have already exited.
1382 * The demand attach fileservice has been modified to avoid spurious
1383 'SYNC_putRes: write failed' warnings when some protocol messages
1384 cannot be acknowledged due to the sender terminating the connection.
1386 * In the event of failure to contact the vlserver or ptserver, the
1387 fileserver will not exit and trigger a forced salvage. It will
1388 continue to try in the background to contact the needed services.
1390 * The salvager can now repair certain cases of a damamged vnode index.
1392 * The accessDate metadata for a volume is now updated correctly.
1396 * CRITICAL: Some applications for example those based on Cygwin were
1397 unable to access data stored in the AFS name space. Explorer Shell
1398 also experienced inconsistent behavior. This is fixed.
1400 * CRITICAL: Multiple AFS pioctl requests issued nearly simultaneously
1401 by applications could result in pioctl responses being received by
1402 the wrong requester. This in turn could result in application
1403 crashes. symlink.exe, fs.exe, afslogon.dll, afscreds.exe, and the
1404 netidmgr afscred.dll plugin were all affected.
1406 * Some XP machines running 1.5.62 had trouble saving roaming profile
1407 data. This is fixed.
1409 * Integrated Logon (afslogon.dll) did not function with domain
1410 specific configurations.
1412 * Ensure that access denied and over quota errors experienced while
1413 storing data to the file server do not result in on-going retry
1416 All UNIX client platforms
1418 * Except on Solaris and AIX, the compiler may now be overriden at
1419 configure time by setting the CC environment variable.
1421 * afsd now properly deals with large cache partitions.
1425 * Build shared libafsauthent and libafsrpc.
1429 * Kernel module DKMS support now installs an unstripped module to
1430 allow debugging information to be collected.
1434 * Preferences pane properly updates token information.
1438 * klog will now properly handle passwords of 8 or fewer characters
1439 with an AFS string to key on hosts able to run 64 bit binaries.
1441 * A panic at AFS shutdown due to "NO PCB" on a udp_lock has been
1444 * The panic decoder script included in the source now properly handles
1445 32 and 64 bit panics.
1449 * Avoid defining AFS_KERBEROS_ENV globally as it creates a circular
1452 * Build shared libafsauthent and libafsrpc.
1456 * Build shared libafsauthent and libafsrpc.
1459 OpenAFS 1.5.62 (2009-08-28)
1463 * Numerous invisible changes to improve code maintainability,
1464 portability and enhanceability.
1468 * CRITICAL: Fixes two errors that can result in data loss when storing
1469 data to the file server.
1470 1. Failure to Store Portions of Unaligned Writes
1471 2. Failure to Store Data to File Servers Lacking Large File Support
1472 Read the announcement for more details:
1473 http://www.openafs.org/pipermail/openafs-announce/2009/000305.html
1475 * CRITICAL: The cache manager daemon thread could terminate when the
1476 machine enters suspend mode. This daemon thread performs the
1477 background check of down servers, offline volumes, callback
1480 * CRITICAL: Integrated Logon (afslogon.dll) was terminating
1481 unexpectedly. Error checking has been improved and NULL pointer
1482 dereferences after Lsa API calls fail have been eliminated.
1484 * For the first time, the OpenAFS SMB Server supports the DCE RPC
1485 services SRVSVC and WKSSVC. Browsing \\AFS with the Explorer Shell
1486 or NET VIEW will now be faster and provide additional functionality.
1487 No longer will cell names longer than 12 characters be truncated.
1489 * Improvements to DFS Referral request processing have been
1492 * Unnecessary DNS lookups of share names are avoided improving
1495 All UNIX client platforms
1497 * Non-Kerberos PAM modules work correctly again.
1501 * MacOS 10.6 (Snowleopard) is now supported, both 32 and 64 bit mode.
1503 * Updates to the AFSCommander preferences pane.
1505 * Installer now permits cell names with dashes.
1508 OpenAFS 1.5.61 (2009-08-06)
1512 * Correct another race condition in the Rx library that could result
1513 in an unexpected panic while freeing the Rx call iovq.
1515 * rx packet resend and data packets sent counts were incorrect.
1517 * fs setquota, fs setcachesize, vos setfields, and vos create now
1518 accept human readable orders of magnitude. (K, M, G)
1520 * fs listquota fixed to permit large quota sizes to be displayed.
1522 * Correct documentation of bosserver permissions requirements.
1524 * Modify vlserver to avoid potentially corrupting the database through
1527 * Generalized support for fast Rx timeout due to network
1530 All server platforms
1532 * Allow audit logs to be sent via sys5 IPC message queues instead of
1537 * If a file server becomes inaccessible while the cache manager has
1538 dirty buffers to write, the afsd_service buf_IncrSync thread can
1539 attempt to use 100% of the cpu.
1541 * Fix "fs newcell" which was broken in 1.5.60.
1543 * Do not attempt to synchronize dirty buffers if the associated volume
1544 is known to be unavailable.
1546 * Modify behavior of a Freelance mountpoint target that does not
1547 specify a cell. Instead of assuming the target volume is in the
1548 Freelance.Local cell, use the workstation "Cell" specified in the
1549 registry. A mountpoint target of "#root.cell." will now mean the
1550 root.cell volume in the workstation cell for the current session.
1551 If the workstation cell changes from "athena.mit.edu" to
1552 "andrew.cmu.edu", the referenced volume will also change without
1553 requiring that the mount point targets be altered.
1555 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
1556 to permit the server Uuid to be used to lookup the server object and
1557 from that determine the cell. This permits callbacks that are
1558 received from alternate addresses to be processed with a known
1559 server object. Previously a request from an unknown server would
1560 clear all callbacks from all cells.
1562 * Fix a bug that prevented optimal performance when using a non-zero
1563 value for 'daemonCheckVolCBInterval'. As a reminder, when
1564 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
1565 volume callbacks are automatically renewed 90 minutes before their
1568 * Fix automatic ranking of vldb servers whose values are obtained from
1569 the CellServDB file.
1571 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
1572 or the call is to a vldb server.
1574 * Add registry based cell search functionality to NetIdMgr,
1575 afs_config.exe, and klog.exe.
1577 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
1578 lookups on the host names in the CellServDB instead of using the
1579 specified IP addresses. This provides aklog, pts, vos, etc. the
1580 same CellServDB behavior that the Windows Cache Manager uses.
1582 * When updating the stat cache entry callback of a .readonly object
1583 from the volume group object, update the file server reference to
1584 ensure it matches the most update to date callback.
1586 * Add proper support for processing callbacks from multi-homed file
1587 servers. Instead of comparing servers by cm_server_t pointer,
1588 compare them by UUID when the UUID is known.
1590 * During a shutdown short circuit the offline volume check daemon
1593 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
1594 preference to an error code reported by rx_EndCall.
1596 * Add "PerFileAccessCheck" registry value to permit testing against
1597 experimental file servers that include per-file acl support. This
1598 value is intentionally undocumented. It is not to be used by
1599 production environment deployments.
1601 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
1602 provider from obtaining tokens when referrals are in play.
1604 * Add "fs chown" and "fs chgrp" commands to permit the owner and group
1605 of objects stored in AFS to be set from Windows.
1607 * Avoid performing background daemon operations when the machine is
1608 going into suspend mode.
1610 * Perform offline volume checks in most recently used order.
1612 * Prevent crash when a data version for a cache object goes backwards.
1614 * Multi-thread safe library versions are now being generated and used.
1615 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
1617 * Microsoft SMB Redirector (mrxsmb.sys) support for
1618 ExtendedSessTimeout values are now available on XP through Windows
1619 7. Add functionality to autodetect if such support is present on
1620 the machine. If so, configure it if necessary and dynamically
1621 adjust the AFS Rx timeout values accordingly.
1623 All UNIX client platforms
1625 * Fix out-of-tree source builds.
1629 * GUI installer now asks for local cell information.
1631 * AFS Commander preferences pane is now installed by default.
1635 * Avoid kernel panics due to null pointer dereferences in the network
1636 interface poller kernel thread.
1639 OpenAFS 1.5.60 (2009-05-31)
1643 * Retry volserver transaction creation on failure.
1645 * Allow building HTML and PDF documentation from included XML copies
1646 of User Guide, Admin Guide and Quick Start Guide for Unix.
1648 * Documentation updates and additional documentation.
1650 * Add -encrypt support to pts client.
1652 * Convert MR-AFS fs commands to OSD commands.
1654 All server platforms
1656 * Updated background sync process in fileserver to avoid a race which
1657 could result in a volume being taken offline.
1661 * On April 9th Microsoft released a Hot Fix for Windows Server 2003
1662 SP2 that corrects a deadlock in the smb redirector and also adds new
1663 functionality that permits the AFS SMB server to be given a longer
1664 timeout than is normally the case. New functionality has been added
1665 to configure these additional LanmanWorkstation\Parameter values.
1666 (This functionality has been backported to XP SP3 and is scheduled
1667 to be released on June 5th.)
1669 * Fix RT#124787, a race condition between "fs flush <dir>", "fs
1670 flushvolume", or "fs flushall" and on-going directory operations
1671 that can result in afsd_service.exe crashing.
1673 * Release Notes, User and Administrator guides are now shipped as
1674 indexed Windows HtmlHelp Files. (.chm). Shortcuts are provided from
1677 * A method of specifying Client CellServDB information within the
1678 registry has been added that can be used to either override the
1679 CellServDB file or force the use of DNS lookups for a given cell.
1680 See the release notes for details.
1682 * The pioctl interface now properly handles drive letter substitution
1683 to UNC paths. (SUBST <d:> <\\afs\cell\path>)
1685 * The BackConnectionHostNames registry value configuration was broken
1686 when dynamic re-establishment of Netbios Name registrations was
1687 added. This release restores the functionality.
1689 * All hidden vos.exe commands are now revealed.
1691 * Attempts to store the same dirty file chunk from multiple threads
1694 * The IsPathInAfs test in Explorer Shell Extension and fs.exe now
1695 permits broken symlinks to be treated as being in AFS.
1697 * vos.exe commands that output 64-bit integer values once again do so.
1698 This was broken in 1.5.59.
1700 * Cygwin Import Libraries are provided in the SDK for all OpenAFS
1701 DLLs. This permits building cygwin applications against OpenAFS
1704 * NSIS installer does a much better job of cleaning up files left over
1705 from previous installs.
1707 * libafsconf.dll moved from Client\Program to Common directory as is
1708 is now used by all modules for CellServDB processing.
1710 All UNIX client platforms
1712 * Write back changes on last store for memcache to avoid discarding
1715 * Abstract disk cache support to allow for path, fh, inode based
1716 caches with no need for messy ifdef structures each time a new type
1721 * Support as a userland port.
1725 * Corrected structure definition for userspace cache manager to allow
1730 * Corrected client locking support.
1732 * Updated patch to stop deadlocking in the kernel during mmap.
1734 * Avoid oops when setting up groups for PAGs to match keyrings.
1736 * Use Linux fh-based cache in cases where possible by default.
1740 * Corrected structure definition for userspace cache manager to allow
1745 * Support for OpenBSD 4.5.
1749 * Corrected support for server-side vos split interface.
1752 OpenAFS 1.5.59 (2009-04-06)
1756 * Increased service priority class to "High" to match the priority of
1757 system components that are dependent upon the a timely response.
1759 * SMB error responses avoid returning errors that could confuse the
1760 Microsoft SMB redirector into disconnecting the connection to \\AFS.
1762 All UNIX client platforms (except MacOS X 10.4 and 10.5)
1764 * OpenAFS 1.5.59 contains fixes for the client issues addressed by the
1765 security advisories OPENAFS-SA-2009-001 and OPENAFS-SA-2009-002.
1769 * Support for prerelease Linux 2.6.30 kernels.
1772 OpenAFS 1.5.58 (2009-03-30)
1776 * Code cleanup and prototyping.
1778 * Avoid unnecessary blocking in Rx periodic cleanup code.
1780 All server platforms
1782 * Fileserver CopyOnWrite routine optimized for performance.
1784 * Make fileserver callback dumps 64 bit safe.
1786 * Fix byte order issues with fileserver host hashing.
1788 * Fix buffer size issues with butc.
1790 * Fix several Ubik recovery issues.
1792 * Avoid leaking file references in the fileserver.
1794 * Fix a race in DAFS while closing vnodes, and another offlining
1797 * volserver interfaces for volume splitting client.
1801 * [RT 124293] A race condition exists which can result in a crash.
1803 * [RT 124276] If the vldb is out of sync with the contents of the file
1804 servers, afsd_service will retry too many times when a file server
1805 reports a volume as not being present. Now if the list reported by
1806 the vldb is the same as the previously seen list, then the retry is
1809 * [RT 124276] Read-only volume failover was broken in 1.5.53 whenever
1810 accessing a volume results in VNOVOL or VMOVED.
1812 * [RT 124276] Prior to 1.3.70 the volume server reference list was not
1813 reference counted and would be prematurely freed while in use. When
1814 reference counting was added in 1.3.70 a bug was introduced that
1815 could result in service reference list corruption.
1817 * Add Windows Application Event Log warning messages for "Client SMB
1818 MPX value too large" and "Client SMB Buffer Size too small".
1820 * Renaming of files across directory boundaries would result in an
1821 invalid handle error when attempting to access the files after the
1824 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail when a
1825 smb file descriptor cannot be found. The whole point of using a
1826 Path Info function is because an smb file descriptor wasn't
1829 * More edge cases in which dynamic addition of Freelance root.afs
1830 entries would get the wrong FID or where the root.afs directory
1831 would not be refreshed.
1833 * Buffer overflow could occur if the workstation cell name was longer
1834 than 64 characters. Crashes could occur in afscreds.exe,
1835 afslogon.dll, and afsd_service.exe.
1837 * VNOSERVICE and VOFFLINE errors were leaking and were exposed to the
1840 * Log file server uuid values as part of the cm_server object when
1841 available. Dump the cm_server object list in response to "fs
1844 * Optimize the performance of resetting access control lists when
1845 tokens are set or removed.
1847 * Remove symlink recursion tests and increase max symlink count to 64
1850 * Windows specific Rx performance improvements.
1852 * Support for Network Identity Manager 2.0
1854 All UNIX client platforms
1856 * Avoid issues with freeing resources at shutdown.
1858 * Numerous fixes to disconnected AFS.
1860 * Disconnected AFS fixes for replaying changes without double-freeing.
1862 * Attempt to use krb524 principal conversion in aklog if available.
1866 * Kerberos configuration at build time corrected.
1870 * Default to dynamic allocation of AFS kernel cache entries to allow
1871 growth for inotify()-pinned entries. (beagle, famd, etc)
1873 * Change client truncation routines to avoid locking issues.
1875 * IA64 port clients fixed on Linux 2.6.
1877 * RPMs now install fstrace message catalog.
1879 * Support through kernel 2.6.29 tested.
1881 * Fix locking issues on CellServDB file in /proc.
1885 * Support OpenBSD 4.4
1888 OpenAFS 1.5.57 (2009-01-23)
1892 * Conditional compilation of rxdebug support is now possible.
1894 * Documentation updates.
1896 * Further race connditions in Rx have been corrected.
1898 All server platforms
1900 * Salvager no longer attempts to recreate headers in the wrong
1903 * Volumes are properly marked in use on creation and subsequently on
1904 examination with vos.
1908 * Undo the "UAC manifest fix" applied to afs_config.exe.
1910 * Ensure that Freelance allocation of vnodes follow the AFS convention
1911 of odd vnodes are directories and everything else is an even vnode.
1913 * Add Freelance logic to mount point and symlink evaluation functions.
1915 * Enhance smb_ParseASCIIBlock() so that it can handle all of the
1916 STRING formats defined by the CIFS Technical Report 1.0.
1918 * Validate the output of smb_ParseASCIIBlock() in all callers. Return
1919 CM_ERROR_BADSMB if the STRING field cannot be parsed.
1920 CM_ERROR_BADSMB will cause the contents of the packet to be logged.
1922 * If multiple SMB Raw Write operations were taking place at the same
1923 time, there could be data corruption because unique event objects
1924 were not generated for each Netbios receive operation.
1926 All UNIX client platforms
1928 * Userspace AFS library can now deal with large files when supported
1931 * Numerous updates to disconnected AFS support, including changes to
1932 allow reconnection to work in more circumstances.
1936 * FreeBSD unstrategy code has been updated.
1940 * A race during file truncation has been corrected.
1942 * System call probing routines have been updated.
1944 * 2.6.29 is now supported.
1948 * 10.3 support has been corrected.
1952 * Initial OpenBSD 4.4 support.
1956 * Updates to allow compiling on newer OpenSolaris are now included.
1959 OpenAFS 1.5.56 (2008-12-30)
1963 * libuafs (userspace cache manager) updated to correct several errors.
1965 * Additional rx debugging support is available as a conditional
1968 * A race condition in Rx leading to a panic has been corrected.
1970 * Rx idle time tracking has been corrected.
1972 * ubik clone server support has been corrected.
1974 All server platforms
1976 * Salvager no longer leaves cores in vice partitions.
1978 * The vol-dump tool now supports dumps larger than 2gb where possible.
1980 * Operations on multiple files now report all FIDs in the audit log.
1982 * butc XBSA support now works correctly on amd64.
1986 * The NetIDMgr AFS Provider automated configuration logic was broken
1987 by the introduction of Kerberos referrals. If the realm of the
1988 identity cannot be determined, the workstation cell is now assumed
1989 to belong to the newly created identity.
1991 * Avoid a reference count under flow during rename operations.
1993 * Avoid a crash caused by treating an arbitrary length directory
1994 search mask as an 8.3 mask.
1996 * Prevent rename operations if a case insensitive match for the target
1997 name already exists and does not refer to the object being renamed.
1999 * Increase the maximum number of background daemons to 64.
2001 * Fix the UAC manifest applied to afs_config.exe
2003 All UNIX client platforms
2005 * Updates to disconnected AFS support.
2009 * FreeBSD 7.1 is now supported.
2011 * amd64 FreeBSD is now supported.
2015 * Generic fh (exportfs API) cache type is now available.
2017 * Avoid some oopses due to backing_dev_info inode fields not being
2020 * 2.6.28 is now supported.
2024 * 10.3 support has been corrected.
2028 * Large partition support has been corrected.
2030 * Filesystem-agnostic cache is now available on Solaris 10 and 11.
2033 OpenAFS 1.5.55 (2009-11-10)
2037 * Salvager avoids leaving core files in vice partitions.
2039 * NFS translator fixes.
2041 * Unresponsive server handling fixes.
2043 * A volserver race which could result in duplicate transactions is
2048 * Fixes a panic caused by corruption of the SMB virtual circuit list.
2051 * Fixes a panic caused by receipt of a UTF-16 string that cannot be
2054 * Implements a more aggressive recovery algorithm for Netbios errors
2055 that result in loss of communication to the AFS SMB server.
2057 * Improve pioctl response time when testing whether or not a PATH is
2060 * Adds support for linked cells.
2062 * Increases the length of the cell and realm names that can be input
2063 into the Network Identity Manage AFS provider configuration dialog.
2065 All UNIX client platforms
2067 * Disconnected AFS avoids infinite recursion during rmdir.
2071 * Support for 2.6.28 (not complete for NFS translator modules).
2073 * Support for using exportfs API for filesystem-agnostic cache.
2075 * Disable backing store readahead.
2077 * Avoid deadlocks when writing back mmapped files larger than the
2080 * Avoid Oops when doing PAG garbage collection.
2083 OpenAFS 1.5.54 (2008-10-08)
2087 * Updates for new Tivoli X/Open API finding.
2089 * A double-free is corrected in Rx.
2091 All server platforms
2093 * Ubik cleans up file descriptor cache correctly when doing recovery.
2095 * Enhanced vldb error checker included.
2099 * Prevent a crash that could occur when multiple file / directory
2100 change notifications are processed simultaneously.
2104 * AFS claims more free space so Finder will attempt file copies.
2108 * Avoid spurious ENOENT when calling gwtcwd() on a volume root.
2110 * Avoid spurious ENOTDIR during fakestat.
2113 OpenAFS 1.5.53 (2008-09-26)
2117 * rx avoids many packet leaks.
2119 * rx jumbogram disabling now works (and is the default).
2121 All server platforms
2123 * Demand Attach fileserver tries to avoid issues tracking offline
2128 * Many potential deadlock conditions due to out of order lock
2129 acquisitions have been corrected.
2131 * A race condition resulting in an undercount on the cm_scache_t
2132 reference counts has been corrected.
2134 * An empty string when sent as an ioctl path is now properly
2135 interpreted as meaning the current directory. This affects "fs
2136 lsm", "symlink list", etc.
2138 * Fix smb string parsing differences where the smb protocol
2139 documentation does not match the actual Windows implementation.
2141 * Random access denied errors fixed.
2143 * A file server lock synchronization issue was corrected in SMB
2144 NTCreateX and NTTranCreate operations. This bug prevented properly
2145 operation when loading roaming profiles.
2147 * Fix a heap overwrite error during server probe operations if a new
2148 server is added while a probe operation is in progress.
2150 * Fix an LSA memory leak that was the result of an LSA error.
2152 * Do not leak cm_cell_t objects if the VLDB server lookup fails.
2154 * Only initialize rx mutex/lock objects once.
2156 * Do not nul terminate the AFS volume name when reported to Windows.
2158 * Improve VNOVOL error handling. Prevent updated vl information from
2159 being destroyed immediately after it was acquired. This bug
2160 prevented proper fail over when volumes are moved or removed from a
2163 * Remove volume id from the server volume list in response to VMOVED
2166 * "fs flushXXX" commands now destroy locally built B+ directory trees.
2168 * Prevent mixture of locally modified directory pages and file server
2171 * Fail over to alternate vl servers if a ubik error is returned.
2173 All UNIX client platforms
2175 * Disconnected AFS now supports read-write mode.
2177 * volserver now builds correctly.
2181 * AIX 6 is now supported.
2185 * FreeBSD 7 is now supported.
2189 * cache bypass is now supported.
2191 * 2.6.x kmod compilation now uses kernel compile options always.
2193 * Support through 2.6.27.
2197 * Show more space free so Finder doesn't get confused.
2201 * Default to namei rather than inode.
2204 OpenAFS 1.5.52 (2008-08-18)
2208 * Initialize volume updateDate at volume creation.
2210 * Avoid potential corruption of directories during salvage.
2212 * Check for out of memory condition during allocation of additional Rx
2217 * Restore support for Windows 2000 (broken in 1.5.50).
2219 * Perform additional validation on volume names in mount points during
2220 creation and evaluation.
2222 * Fix several deadlocks, races, and reference count issues.
2224 * Further optimize SMB Directory Search processing and minimize the
2225 number of InlineBulkStatus RPCs sent to the file server.
2227 * Enable "bos restricted" operations.
2229 * Fix the create of submounts used by the AFSCreds and afs_config
2232 * Fix a short name truncation bug. (1.5.50)
2234 * Fix the error code reported when attempting to delete a file on a
2235 readonly volume or one that is marked with the readonly DOS
2238 * Fix a heap corruption error when reading the CellServDB file
2241 * Add the "RxUdpBufSize" registry value. The new default is 256KB.
2243 * Do not include trailing NULs in the directory search output.
2246 * Pre-allocate 64 Rx Packet buffers per thread in order to improve
2249 * For debugging: add smb lock requests and stat cache lock allocations
2250 to the output from "fs memdump".
2254 * Workaround broken sigwait() to allow fileserver to shut down
2255 correctly pre NetBSD 5.0.
2259 * Default to namei fileserver; Allow inode fileserver at configure
2264 * Try harder to avoid kernel panics for malformed requests.
2267 OpenAFS 1.5.51 (2008-07-29)
2271 * salvager tries harder to arrange for clients to get VBUSY while
2272 salvaging single volumes.
2274 * salvager avoids certain corruption when salvaging directories.
2276 * Rx connection clones disabled.
2280 * The 32-bit EXE 1.5.50 installer failed to properly install the C
2281 Runtime library. When used as an upgrade OpenAFS would continue to
2282 work but when used as a new installation, OpenAFS binaries would
2285 * Fixes the "fs" and "symlink" commands to properly parse Unicode path
2286 prefixes during the pioctl remote procedure call. This bug would
2287 result in file not found errors for files and directories that
2288 clearly exist. (Bug introduced in 1.5.50)
2290 * Large File support is disabled. (Bug introduced in 1.5.50)
2292 * Removes the possibility of a deadlock during volume location update
2293 operation if all of the reported file servers are unreachable at the
2296 * Ensures that reference counts are properly incremented/decremented
2297 on Rx connection objects used for volume location RPCs.
2299 * Over Quota errors during cm_FSync() calls would lead to an infinite
2300 loop as the error was never propagated to the caller.
2302 All UNIX client platforms
2304 * Bugfixes to disconnected AFS support in the cache manager.
2307 OpenAFS 1.5.50 (2008-07-16)
2311 * volserver puts recloned volumes back online before returning the
2312 volume to the fileserver, avoiding spurious VNOVOL errors.
2314 * Updated TSM X/Open API support available.
2316 * Demand Attach fileserver will not crash due to accesses during
2319 * Substantial documentation updates.
2321 * Demand Attach fileserver state tracking and analyzer tool
2324 * UAFS userspace cachemanager updates.
2326 * Corrected support for anti-meltdown protection in the client.
2330 * UNICODE Character Set Support.
2332 * Pioctl interfaces to the cache manager have been refactored to
2333 provide layering between the SMB specific code and the general
2334 purpose ioctl operation.
2336 * Garbage collect dead SMB virtual circuits as soon as they are no
2337 longer being referenced. This avoids problems with outstanding
2338 locks not being dropped when the virtual circuit becomes invalid.
2340 * Remove the IBM Administration Reference documentation and replace it
2341 with the OpenAFS Command Reference Manual.
2343 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime functions each
2344 time a connection is about to be used. Do not hold a lock on the rx
2345 connection object while it is being selected. This avoids a race
2346 between threads attempting to set the timeout values and removes a
2347 bottleneck that was hampering performance.
2349 * Ensure that the smb directory attribute is set for all directory
2352 * Replace the VC Runtime EXE installer with the MSI installer in the
2353 NSIS installer scripts.
2357 * Support for updates to OpenSolaris.
2361 * Correct dentry revalidation for cross-directory renames.
2363 * Updated rpm packaging materials for 1.5 release series and 2.4
2366 * Corrected syscall table probing.
2368 * NFS translator updates for current kernels.
2371 OpenAFS 1.5.39 (2008-06-24)
2375 * Updates for Demand Attach fileserver.
2379 * Fix two memory leaks.
2381 * Fix one missing lock.
2383 * Handle access denied errors when writing dirty buffers.
2385 * Fix two errors that would cause the *experimental* AFS Servers
2389 OpenAFS 1.5.38 (2008-05-24)
2393 * Add read-only disconnected support.
2396 OpenAFS 1.5.37 (2008-05-21)
2400 * Includes a number of optimizations for testing.
2403 OpenAFS 1.5.36 (2008-05-09)
2407 * Rx optimizations now attempt to deal with high latency WANs.
2409 * Client will not wait infinitely for a server which is not providing
2410 data. Additional servers will be polled without marking the server
2411 which is not providing data down.
2413 * vos move will not erroneously unlock locked vldb entries on failure.
2415 All server platforms
2417 * Fileserver avoids a potential infinite loop when a client
2418 relinquishes an address.
2420 * Fileserver large setting now configures more threads.
2422 * Fileserver properly registers uuids of new clients.
2424 * Ubik servers do not improperly hide updates from clients.
2426 * Fileservers reserve enough file descriptors such that each thread
2427 can cache one to avoid spurious errors.
2431 * Fix a cm_buf_t reference count leak when attempts to write dirty
2432 buffers to the file server from within cm_IncrSyncer() fail.
2434 * Prevent udebug from crashing.
2436 * Another VNOVNODE issue fixed. When writing a dirty buffer to the
2437 file server, if VNOVNODE is received, mark all buffers as invalid
2438 without further attempts to contact the file server.
2441 =======> Changes back to 1.3 have not yet been incorporated here <=======
2446 * -nosettime is now the default for afsd. Use "-settime" to get the
2449 * OpenBSD is now supported.
2451 * Mountpoint directory information is now only faked for cross-cell
2452 mountpoints when using the -fakestat flag (e.g. for the directories
2453 under /afs, but not for most other volumes mounted inside the cell).
2454 The -fakestat-all switch can be used to fake information for all
2457 * When fakestat is enabled on MacOSX, the Finder can be used to browse
2458 a fully-populated /afs directory. However, this precludes reliable
2459 use of entire volumes as MacOS bundles (i.e. containing a Contents
2460 directory in the root of the volume).
2462 * Mountpoint directory information can be faked by the cache manager,
2463 making operations such as stat'ing all cells under /afs much faster.
2464 This is enabled by passing -fakestat to afsd, but might not be stable
2470 * The kaserver now defaults to not allowing interrealm authentication,
2471 due to security vulnerabilities in the krb4 protocol. The new
2472 "-crossrealm" flag to the kaserver is provided to reenable interrealm
2473 authentication if desired.
2475 * RedHat Linux 9.0 is now supported.
2477 * Solaris 9 12/02 is now supported. Solaris 7 and 8 x86 should now
2480 * On Linux machines using 2.2 series kernels, 2.2.19 or higher is now
2483 * An OpenAFS 1.2.9 afsd will not work with kernel modules built from
2484 an earlier OpenAFS release. In general, using a mismatched afsd and
2485 kernel modules set is unsupported; it is not recommended that you use
2486 such a configuration on a regular basis.
2491 * Mountpoint directory information is now only faked for cross-cell
2492 mountpoits when using the -fakestat flag (e.g. for the directories
2493 under /afs, but not for most other volumes mounted inside the cell).
2494 The -fakestat-all switch can be used to fake information for all
2497 * HPUX 11.0 is now supported.
2499 * It is now possible for AFS to use Kerberos 5 directly, via rxkad 2b.
2500 See the OpenAFS 1.2.8 Release Notes for more information on using this
2503 * An NFS translator kernel module is now included and compiled by
2504 default for Solaris only.
2509 * MacOS X 10.2 is now supported. FreeBSD 4.3 and later support is
2510 included in this release, but is still under active development and
2511 should only be used by those doing active development on the OpenAFS
2514 * When fakestat is enabled on MacOSX, the Finder can be used to browse a
2515 fully-populated /afs directory. However, this precludes reliable use
2516 of entire volumes as MacOS bundles (i.e. containing a Contents
2517 directory in the root of the volume).
2519 * The fileserver will now use Rx pings to determine if clients are
2520 reachable prior to allocating resources to them, to prevent asymmetric
2521 clients from consuming all fileserver resources.
2526 * Mountpoint directory information can be faked by the cache manager,
2527 making operations such as stat'ing all cells under /afs much faster.
2528 This is enabled by passing -fakestat to afsd.
2530 * Solaris 9 FCS and Solaris 7 and 8 x86 are now supported.
2535 * A remote denial of service attack in the AIX and IRIX clients has been
2536 fixed. Users of those platforms are strongly encouraged to upgrade.
2538 * Fixed race conditions in fileserver that could result in crash.
2543 * Server logfiles now more consistant about format in which hosts are
2546 * vfsck on Solaris will now allow force runs (using -y flag) even if old
2552 * Cell aliases for dynroot can be specified in the CellAlias file in
2553 /usr/vice/etc or /usr/local/etc/openafs, in format "realname alias",
2554 one per line. They can also be managed at runtime with "fs newalias"
2555 and "fs listaliases".
2560 * Solaris 9 and Linux PA-RISC are now supported.
2562 * fileserver will not erroneously delay legitimate errors for 3 seconds
2563 after 10 errors are returned (e.g. stat() on a directory you can't
2566 * Rx MTU calculation now works for Irix, Solaris and Linux
2568 * If afsd is started with the -dynroot flag, /afs will be locally
2569 generated from the CellServDB. AFSDB cells will be mounted
2570 automatically upon access.
2572 * The namei fileserver allows vice "partitions" to be directories
2573 instead of partitions and will attach and display accordingly.
2574 Creating the file "AlwaysAttach" in the /vicepX directory is used as
2575 the trigger to attach it.
2577 * TSM support for butc no longer requires editing a Makefile, simply
2578 specify the --enable-tivoli-tsm configure option.
2580 * Linux builds no longer require source changes every time the kernel
2581 inode structure changes; the OpenAFS sources will now configure itself
2582 to the actual inode structure as defined in the kernel sources.
2587 * vfsck on Digital UNIX and Solaris will now refuse to fsck mounted
2593 * AFS now supports --prefix and the other directory options of
2594 configure. By default AFS builds assuming it will be installed in
2595 /usr/local. In order to get traditional AFS directory paths (/usr/afs
2596 and /usr/vice/etc) use the --enable-transarc-paths option to
2597 configure. More details on the new directory layout are found in
2603 * Windows 95/98/ME/NT/2000 - Consistent versioning: Installation, AFS
2604 Control Center, Client dialog boxes and properties pages for
2605 executables display a consistent OpenAFS version number. Installation
2606 detects previous installation and prompts the user for upgrade
2609 * Windows 95/98/ME/NT/2000 - Installation features: During installation
2610 the user can select the source of the CellservDB file, AFS home cell,
2611 and drive mappings. During installation a drive path mapping can
2612 include a variable that will be substituted with the current UserName
2615 * Windows 2000/NT - Integrated logon: The Integrated Logon feature
2618 * Windows 95/98/ME - Logon script features: The Windows 95/98/ME client
2619 now offers a command-line option for starting up the AFS client
2620 without authenication. It is now possilbe to start the AFS client
2621 first and obtain tokens, and map drives all through Windows scripts.
2622 This helps using Windows 95/98/ME client in Kerberos 5 environment.
2624 * Windows 2000/NT - LANA numbers: AFS client now scans the LANA numbers
2625 to establish the correct NETBIOS connection. NetBEUI is no longer
2626 needed. The user no longer needs to find the correct LANA number.
2628 * Windows 2000/NT - OpenAFS naming consistancy: Further progress has
2629 been made to remove references to "Transarc AFS" and replace with
2635 * AFS now builds with configure. The README for building has been
2636 updated and includes full details.
2638 * A client system can now have multiple sysname values for @sys. They
2639 will be searched in order when looking up files in AFS. The
2640 -newsysname argument to fs sysname can be repeated to set multiple
2643 * A new system group is created for new cells (system:ptsviewers with id
2644 -203). If this group exists, members of this group can examine and
2645 read the entire protection database. They can examine all users and
2646 groups and can get the membership of any group.
2648 * A new program, pt_util has been added to the distribution. This
2649 program allows users to print the contents of the protection database
2650 or to edit the protection database without running a ptserver. It can
2651 be used to set up a new cell without ever running in noauth mode. Run
2652 pt_util -h for help.
2654 * The fs setcrypt and fs getcrypt commands have been added. These
2655 commands allow the system administrator to require that the client
2656 encrypt all authenticated traffic between the client workstation and
2657 AFS. The encryption used is weak, but is likely better than sending
2658 unencrypted traffic in most environments. Some functions, such as
2659 looking for a volume may not be encrypted, but data transfer certainly
2660 is. By default data is not encrypted. At this time no significant
2661 experimentation with server performance has been conducted.
2663 * By default AFS is compiled with AFS_AFSDB_ENV, enabling the -afsdb
2664 option to be given to afsd on startup. If this option is used, then
2665 new cells will be looked up using AFSDB records stored in DNS if they
2666 are not found in CellServDB. This means that users can create
2667 cross-cell mountpoints in directories they control to access cells not
2668 in root.afs, and that cells in root.afs need not be in the client's
2671 * AFS database servers can be marked as read-only clones. Surround the
2672 hostname in square brackets on the bos addhost command and the
2673 database server will never be elected sync site. This is useful for
2674 cells distributed over a wide region.
2676 * The AFS servers now support the -syslog flag. This flag causes them
2677 to log to syslog rather than to files. This flag is not supported on
2678 NT. For all servers besides the salvager, the flag can also be
2679 specified as -syslog=facility, where facility is an integer facility
2680 code from syslog.h. A -syslogfacility option is provided for the
2681 salvager to accomplish the same goal.
2683 * If the --enable-fast-restart flag is given when configuring AFS, then
2684 the salvager supports the -dontsalvage flag which causes it to exit
2685 without salvaging any volumes. If this is configured into the third
2686 command of a fs process, then the fileserver will start without
2687 salvaging. It will fail to attach volumes that need salvaging and
2688 they can be salvaged manually. This provides significantly better
2689 server startup performance at the cost of administrative complexity.
2691 * If the --enable-bitmap-later flag is given when configuring AFS, then
2692 the fileserver creates bitmaps for free vnodes on demand, allowing
2695 * If bosserver finds a BosConfig.new file at startup, it reads this file
2696 and renames it to BosConfig. This allows bosserver to be reconfigured
2699 * The bosserver can be placed in a restricted mode in which AFS
2700 superusers are only granted limited access to the server host. The
2701 following functionality is disabled when restricted mode is in use:
2704 bos getlog (except for files with no '/'s in their name)*
2710 specific exceptions are made for functionality that "bos salvage"
2713 A cron bnode who's name is "salvage-tmp", time is now, and command
2714 begins with "/usr/afs/bin/salvager" may be created. This bnode deletes
2715 itself when complete, so no special "delete" support is needed. This
2716 functionality may be removed in the future if a "Salvage" RPC is
2719 The file with the exact path /usr/afs/logs/SalvageLog may be fetched,
2720 since that is how bos salvage [...] -showlog is implimented.
2722 Restricted mode is enabled using a new bos command (bos setrestricted)
2723 or bossever command line switch (bosserver -restricted). Restricted
2724 mode can be disabled by a) sending the bosserver process a SIGFPE
2725 (which will then allow restricted operations until the next restart or
2726 setrestricted command) or b) editing /usr/afs/local/BosConfig (or
2727 BosConfig.new), and restarting the bosserver.
2729 * The bos UserList of trusted administrators can now contain cross-realm
2730 Kerberos principals.
2732 * udebug now takes --server not --servers.
2734 * Several error messages have been improved to include volume numbers.
2736 * Several new ports have been included for UNIX platforms: Darwin
2737 (ppc_darwin_12 and ppc_darwin_13), Linux 2.4 (i386_linux24), Linux on
2738 the Powerpc (ppc_linux22 and ppc_linux24), Linux on the Sparc
2739 (sparc_linux22, sparc64_linux22 and sparc64_linux24).
2741 * Incomplete FreeBSD and Alpha Linux ports are included. The FreeBSD
2742 port has a working server and the Alpha Linux port has a partially
2745 * A native client for Windows 95/98/ME has been added to the
2746 distribution. With this program, a gateway machine is no longer
2747 required for Windows 9x to access AFS files. One drive letter will be
2748 created on your machine by default - Z:. The Z: drive will be the
2749 root of the AFS tree, allowing you to browse all sites that have AFS
2750 servers available. Additional drive letters can be defined for other
2751 AFS directories. A Windows Explorer shell extension is included that
2752 allows you to right click on items within an AFS tree to bring up an
2753 "AFS" menu item and perform various operations on a file or directory.
2754 The most useful item is "Access Control Lists", which allows you to
2755 view and edit the permissions of a particular directory. Command line
2756 tools are also available in the install directory. These commands
2757 include klog, unlog, tokens, kpasswd, symlink, fs and pts. The
2758 installable includes a readme file that contains more information on
2759 how to use the client program and known issues.
2761 * Support for large caches in afsd. Cachefiles are stored in
2762 subdirectories. The default is 2048 files per subdirectory, which
2763 should work fine in most situations. You can use the new afsd option
2764 -files_per_subdir to change this number. Note that the first time you
2765 run afsd with this patch, your cachefiles will get moved into
2766 subdirectories. If you subsequently run an older version of afsd, you
2767 will lose all your cached files.