1 User-Visible OpenAFS Changes
3 OpenAFS 1.6.0 (in progress)
7 * vos now properly deals with matching sites when servers are
12 * Fix ptserver supergroups support on 64 bit platforms.
14 * Demand attach salvaging doesn't use freed volume pointers.
16 * Properly hold host lock during host enumeration in fileserver.
18 All UNIX client platforms
20 * Servers now marked down when GetCapabilities returns error.
22 * In-use vcache count is now properly tracked.
26 * Fix socket termination on shutdown.
28 * Support for 7.2, 7.3, 7.4 and 8.2 included.
30 * References to vcaches are no longer leaked during root or reclaim.
34 * Define llseek handler to avoid ESPIPE error in 2.6.37.
36 * Mount interface replaces get_sb (new for 2.6.37, not yet required).
38 * RedHat init script allows deferring for a new binary restart.
40 * DEFINE_MUTEX replaces DECLARE_MUTEX for 2.6.37.
44 * Correct return value from setpag syscall.
48 * Bug fixes for issues introduced previously in 1.5 series.
52 * Switch to ioctl() syscall replacement for Solaris 11 since syscall
56 OpenAFS 1.5.78 (2010-11-04)
60 * Revisions to Rx to fix performance issues.
62 * Make fs getfid behave consistently across all platforms. (128372)
64 * Properly check IDs handed to pts when creating users or groups so
65 useful error messages can be provided. (128343)
67 * Correct byte order handling of port in afsconf_LookupServer for SRV
70 * Force a full dump when releasing to a site which was previously
71 marked "don't use", in case the previous clone was out of date.
75 * Demand salvage of attached volumes now correctly track attachment
78 * Avoid a potential crash due to failure to hold a lock when attaching
83 * Track SMB connections by SID rather than username.
85 * Error write attempts to known-readonly volumes earlier.
87 * Validate directory buffers to avoid potential crashes.
89 * Handle VIO errors from bulkstatus.
91 * Make PMTU discovery configurable and register error handlers for it.
93 All UNIX client platforms
95 * Use larger I/O sizes in memcache to improve performance.
97 * Avoid potential alignment issues doing I/O for pioctl calls.
101 * Avoid panicing if the listener process is not findable.
103 * Avoid deadlock issues while performing lookups.
107 * Handle stale file handle errors for some cache partition types.
109 * Avoid blocking with xvcache lock when attempting to free in-use
112 * Build fixes for older kernels.
114 * Properly configure LWP to use ucontext() on platforms where it
117 * Eliminate spurious errors from AFS system call returns. (126230)
121 * Attempt to honor configured Kerberos defaults in Preferences Pane.
124 OpenAFS 1.5.77 (2010-09-08)
128 * Rx path MTU detection will terminate detection in cases where the
129 minimum required packet size cannot be transferred.
131 * vos dryrun mode now shows effects for syncvldb single volume case.
133 * vos dryrun mode now shows "status after" for syncvldb and syncserv.
137 * RXAFS_GetStatistics64 now returns statistics properly.
141 * Attempt to properly identify the local system SMB connection for
144 * Remap timeout and offline errors to proper NT RPC errors.
146 * Properly fail over to other replicas on bulkstat IO errors.
148 * Properly error delete-mode createfile if a file is set readonly.
150 * Validate directory entry buffers to avoid crashing the service.
152 * Log file modes properly.
154 * Log cell name when logging server information.
156 All UNIX client platforms
158 * cacheout program for discarding callbacks is now built.
160 * bulkstatus kernel locking is corrected to avoid a potential panic.
164 * userspace support update
168 * Updated vnode locking for children returned via lookup().
170 * Avoid file open undercount with needed calls to
171 FakeOpen/FakeClose().
173 * Use vnode_pager_setsize to properly track file size during kernel
176 * Update system call installation.
178 * Fix shutdown of Rx kernel listener to avoid potential dereference
181 * Avoid closing vnodes during vnode recycle.
183 * Fix bogus call to FlushVS for vnode reclaims.
187 * Packaging updated for current configure options and built files.
189 * Cache bypass now holds reference on pages during readpage.
191 * s390x setgroups32 patching update.
195 * DNS resolver is reinitialized on IP address change. (126440)
198 OpenAFS 1.5.76 (2010-08-16)
202 * Updates to build-time configuration.
204 * Fix XDR support in Rx to match header definition.
206 * vos status now shows transaction creation, not action creation.
208 * Rx avoids reporting loopback adapters when listing interfaces.
212 * Demand-Attach Fileserver always built and installed (dafileserver,
213 davolserver, dasalvager).
215 * Return VNOVOL from fileserver when a volume is deleted.
217 * Ignore duplicate tags during volume restore operation.
219 * Update inode array after salvage repairs volume.
221 * Zero a corrupted header in memory during salvage to avoid further
224 * Fix NAMEI backend to allow low-numbered volumes to work properly.
226 * ptserver does not include cell name as part of length check for
229 * Updated error messages for unblessed volumes.
231 * vlserver avoids buffer overflow with regex pattern
233 * Attach-time failures now note failures as the rest of the fileserver
236 * Server argument logging will no longer overflow stack.
238 * Provide fast-restart-like unsafe-nosalvage option for DAFS.
240 * Deal with host hash collisions in the fileserver.
244 * Avoid crashing when interpreting a drive letter as potentially
245 matching a cell name.
247 * Properly handle volume package errors.
249 * Allow page recycling from known-readonly content without ensuring
252 * 32 bit tools installer should not override client configuration.
254 * Ensure root scache item has a valid callback when use is attempted.
256 * Freelance directory changes now properly invalidate and replace the
259 All UNIX client platforms
261 * Support disconnected reconnecting with specified UID for PAGless
264 * Proper disconnected vnode reference tracking.
266 * Update server site blacklisting to not return success if nothing was
269 * Avoid a panic during vcache contention due to CVInit vcache
274 * Update for network stack in 8.1/9.0.
284 * Disable PMTU error packet handling.
288 * Debian packaging updated.
290 * freezer interface updates.
294 * Hold references to disconnected mode written vnodes properly.
298 * Handle NFS translator module references for amd64.
300 * INODE fileserver backend support now exists for amd64.
303 OpenAFS 1.5.75 (2010-07-07)
307 * Prevent rx_rpc_stats global lock from being a bottleneck.
309 * Path MTU discovery is now provided to allow traffic to pass networks
310 with sub-1500 byte MTUs and poor fragment handling.
312 * Further reduce Rx NAT ping transmission when enabled.
314 * Update Kerberos 5-based token handling in rxkad from upstream
317 * Update version numbers emitted during build to reflect what is
318 actually being built.
320 * Add "-human" switch for human-readable units in fs diskfree and
323 * vos provides reasons for locked volumes when known.
325 * Do not count retransmission and ping acks as non-idle for Rx
328 * Rx: provide service-specific data getter and setter routines.
330 * Update build-time Kerberos detection.
332 * Updated userspace AFS client.
334 * Beginning of a modernized test suite.
336 * Additional documentation.
338 * Updated documentation, notably the Administrators Guide.
340 * Substantial code cleanup.
344 * Update handling of vnode allocation failures.
346 * DAFS: allow salvaging volumes not known to the fileserver, to allow
347 cleanup of data not attached to a current volume.
349 * Properly handle volumes slated for destruction.
351 * Handle volumes with many files properly.
353 * Force core file generation in bosserver by overriding default
354 resource limits when possible.
356 * Update vlclient and vldb_check.
358 * Avoid potentially corrupting a volume on creation if files are left
359 from previous failed cleanup.
361 * Note volume changed during salvage as needed.
363 * DAFS: do not assume invalid addresses are in fileserver address hash
366 * Avoid tying up fileserver threads with volumes that are being taken
369 * Do not set inUse on volumes for non-DAFS other than in fileserver.
371 * Break origin's callback on target of rename operation.
373 * Avoid unneeded parent directory link updates during some rename
376 * Do not open /dev/console for writing in the fileserver.
378 * DAFS: avoid spurious restarts when binary restarts are configured.
380 * Avoid spurious and unneeded calls to sync(), which can slow down the
385 * Revised SMB QuerySecurityInfo to address issues caused by MS10-020
386 (http://support.microsoft.com/kb/980232)
388 * Prevent use of the AFSCache file contents if mapped to a new
391 * Make fs newcell include behavior compatible with the non-Windows
394 * Provide a registry option (FreelanceImportCellServDB) to pre-create
395 mount points in the AFS root for all cells in CellServDB.
397 * Fix a memory leak in the cm_FreeServerList() routine.
399 * Reduce privilege when reading registry CellServDB.
401 * Add support for RPC Pipe Service NetWkstaGetInfo levels needed for
404 * Prevent overflow when computing quota percentage in Explorer Shell.
407 * Generate meaningful errors for ACL operations on freelance AFS root.
409 * Fix error handling on InlineBulkStatus RPCs.
411 * Show configuration pages for all types of MSI installations.
413 * Improve freemount AFS root directory handling and operations.
415 * Properly validate GetVolumeStatus pioctl responses.
417 * Commit file length changes and dirty buffers when flushing a file.
419 All UNIX client platforms
421 * Update version of files for disk cache.
423 * Do not call afs_FlushVCBs with xvcache lock held, to improve
426 * Add mariner log messages for creating and removing files.
428 * Don't hold xvcache lock while creating symlinks, to improve
431 * Provide -dynroot-sparse mode to not show all cells in CellServDB in
434 * Avoid a potential crash in aklog in linked cell handling.
436 * Log MTU-caused packet retransmission.
438 * Prevent crashes caused be fs checkservers while cache is being set
441 * fs getserverprefs now has a buffer large enough for the default
444 * Report server address when logging warnings.
446 * Avoid panic in GetCapabilities when cell is not known.
448 * Lock process name and id for advisory lock warnings when possible.
450 * Handle need for allocating additional Rx packets.
452 * Properly handle errors from InlineBulkStatus operations.
454 * Fix errors returned from fcntl() on readonly files locked for write.
456 * Flush pending changes to the server on LOCK_EX unlock.
458 * Reflect length changes as a result of callbacks even when file is
461 * Avoid hanging due to error exit when attempting to store a large
462 file to a non-largefile fileserver.
464 * Recover from afs_GetVolSlot errors.
468 * Bugfixes for kernel VFS and network routines.
472 * Provide makesname().
476 * Avoid syscall probes when keyrings are present, by default. (125215)
478 * Remove "Big Kernel Lock" from VFS operations.
480 * Use filehandles for all Linux 2.6 versions to avoid need for matched
483 * Updated RPM packaging.
485 * Fix dkms configuration provided with RPMs.
487 * Hold reference on pages during background I/O for cache bypass.
489 * Fix cache bypass handling of non-largefile fileservers.
491 * Protect truncate_inode_pages mappings with mutex or semaphore as
494 * Fix pagevec use in cache bypass. (127505)
500 * Improve launchd configuration.
502 * Avoid hanging on recursive cache file lock acquisition when user
503 notification is enabled.
505 * Fix and re-enable bulkstat mode.
513 * Precluding unmount while AFS is busy.
515 * Avoid deadlocking when releasing the VFS object.
517 * Stop network interface poller in kernel on AFS shutdown.
519 * Avoid issues with lookups on empty directory names. (127356)
522 OpenAFS 1.5.74 (2010-04-22)
526 * Add "vos setaddrs" command.
528 * Rx library lock contention avoidance between rx_NewCall and
531 * Rx library races due to inconsistent use of rx_connection
532 conn_data_lock to protect the flags field.
534 * Rx library inconsistent use of RX_CALL_TQ_WAIT which could result in
537 * Rx library must signal transmit queue waiters when flushing.
539 * afsmonitor shows busy counts now.
541 * afsmonitor displays xstat callback statistics.
543 * Provide expandgroups for pts mem on a supergroups server.
545 * Provide supergroup option to liste nested groups during pts mem.
549 * Avoid volume lock contention during DAFS startup.
553 * Avoid a race when updating cell vldb server lists that can result in
556 * Avoid a deadlock when managing CM_SCACHESYNC_STOREDATA state
557 operations for directory objects.
559 * Add new Windows Application Event log messages for VBUSY,
560 VRESTARTING, ALL_BUSY, ALL_OFFLINE, and ALL_DOWN.
562 * Reduce lock contention by waiting for cm_buf_t I/O operations.
564 * Split the cm_buf_t flags field to separate the flags that are
565 protected by the cm_buf_t mutex from those protected by the
568 * In cm_UpdateVolumeLocation, avoid searching for a ".readonly" volume
569 on a numeric volume name.
571 * File buffer allocations whose offsets are beyond server EOF should
572 be locally allocated and zero filled. The file server should not be
573 issued a FetchData rpc which is guaranteed to fail.
575 * Enable integrated logon to work with Windows 7/2008 when user logons
576 are performed with a non-Domain Kerberos principal.
578 * Add Protection Error messages to aklog output.
580 All UNIX client platforms
582 * Provide a FUSE-interfacing userspace afs client.
584 * Updates to libuafs userspace cache manager.
586 * Probe servers using GetCapabilities instead of GetTime, thus
587 requiring fewer RPCs.
589 * Fix DNS SRV record handling for cell lookup.
593 * Fix sleep/wakeup routines.
595 * Update for 8.0 release.
599 * Handle high memory addresses correctly.
603 * Make 32 bit AFS syscalls work again.
605 * Work around finder "Duplicate" failure (caused by setting modes on
608 * Disable bulkstat again (will be re-enabled at or before .75).
610 * Provide symlink type hints during readdir.
613 OpenAFS 1.5.73 (2010-03-24)
617 * NAT keepalive support at Rx level.
619 * Corrected SRV record support for cell name canonicalization.
623 * Fix volume callback notification to not notify unaffected clients.
626 * Allow root directory recreation by salvager. (94658)
628 * Numerous DAFS fixes.
630 * Improvements to callback table overflow handling. (126451)
632 * bosserver now shuts down cleanly on SIGTERM.
636 * Prevent the Explorer Shell extension from crashing if symlink
637 creation failed. (126406)
639 * A Rx level NAT ping has been implemented. A registry value enables.
641 * Adds krb5 error message translation to aklog, afscreds,
642 afslogon.dll, the network identity manager afs provider and
645 * Default mode bit settings for file and directory creation are now
646 provided, and can be configured.
648 * An SMB request trace facility is provided and can be enabled for
651 All UNIX client platforms
653 * Rx idle deadtime does not stop file writes.
655 * Disconnected AFS no longer has a race condition during remove ops.
657 * Fakestat avoids a condition which could cause it to block on network
660 * Several fixes to handle interruptions in vos operations. (33360,
663 * Allow more sysnames in a sysname list.
665 * Attempt to enforce timeouts on AFSDB lookups.
669 * Clean up properly on mount failure.
671 * Add entry to /etc/vfs to allow umount to work.
675 * Several issues to deal with older kernels.
677 * Avoid leaking the global lock in the /proc cellservdb code.
679 * Keyring destruction now cleans up all tokens.
681 * Keyring quotas are not enforced against root.
685 * Some FSEvents hinting for authentication events now done. (23781)
687 * Update uninstaller. (125634)
689 * Rewrite afssettings and fstab code to avoid licensing issue with
692 * Growl client for user monitoring of AFS events included.
694 * Properly support insert-only dropboxes.
696 * Add bulkstat support.
698 * Include support for moving in Finder across mount points.
700 * Preferences Pane includes support for Kerberos 5 ticket renewal.
703 OpenAFS 1.5.72 (2010-02-15)
707 * Provide internationalization support in com_err.
709 * Fix array length checking to avoid crashes when checking for a
710 volume type based on name in vos.
714 * Provide backward compatible "-f" flag to salvager for force mode.
718 * Restore use of DNS AFSDB and SRV records by kaserver clients.
720 All UNIX client platforms
722 * Fix client cache file truncation to not lose chunks when truncating
725 * Ensure a cache writeback hook is installed in the client (bug from
728 * Avoid spurious free memory warnings during clean shutdown.
730 * Fakestat mode avoids AFSDB lookups.
732 * "fs storebehind" now correctly reports errors on readonly volumes.
734 * Additional documentation for "fs getcacheparms"
736 * Forced new uuid generation with "fs uuid -generate" now works
737 enforced permission correctly.
741 * Add optimized Rx event handler in kernel.
743 * Installer now allows installing an older version.
745 * Panic decoder can now deal with MacOS 10.5 again.
747 * MacOS ._ files are now correctly not looked up as cellnames.
751 * To deal with SELinux file labeling, try cache accesses with current
752 credentials in event of failure.
754 * Rx XDR encoding bug on i386 Linux is fixed (bug introduced in
759 * Code compilation fixes.
763 * Update for OpenBSD 4.6.
766 OpenAFS 1.5.69 (2010-01-19)
770 * Configuration of BOSserver no longer defaults to weekly restarts
773 * Provide BOS restricted mode by default.
775 * Add support for "vos endtrans" command.
777 * Default to providing full output from vos listvol.
779 * Correct additional-address tracking in the fileserver.
781 * Improve Rx performance by not unnecessarily dropping and reacquiring
782 call locks in read and write processes.
784 * Avoid crashes when monitoring volserver transactions across
785 potential transaction garbage collection.
787 * Numerous warning fixes.
791 * Avoid saving fileserver state in demand attach fileserver when
794 * Demand attach fileserver allows other callers to schedule salvages.
796 * Demand attach "bos salvage" now works correctly with restricted
801 * Numerous changes to the client-internal btree directory handling to
804 * fs examine reports owner and group ids as signed values (PTS groups
807 * Preclude corruption due to races writing to smb buffers.
809 * Allow MTU settings in registry to be used.
811 * Apply MTU to both send and receive sizes.
813 All UNIX client platforms
815 * Avoid double-freeing Rx call structure if reading a response from
816 the file server results in a short read.
818 * Handle negative lengths in FetchStatus results correctly.
820 * Properly clean up allocated memory at shutdown.
822 * Default to AFSDB compiled into the cache manager.
824 * Avoid inadvertant disclosure of stat() information to clients not so
827 * Correct a bug with AFSDB lookups introduced with SRV record support.
831 * Install kernel panic processing tool in /Library/OpenAFS/Tools.
833 * Include debugging symbols for kernel extension in additional package.
835 * Support "Application Firewall" users.
837 * Avoid ._cellname AFSDB lookups.
839 * Compile preferences pane as a universal binary.
843 * Use splice to speed up storing files.
845 * When using memcache, avoid duplicating work in readpages.
847 * Use dget_parent to safely find an inode's parent.
849 * Disable access time updates in our superblock.
851 * Avoid crashing doing writeback if no credentials were stashed at
854 * Simplify keyring support.
856 * Properly clean up vcache in event of failed mount.
860 * Update for current FreeBSD 8.
864 * Abstractly manipulate groups as now required.
866 * Abstractly access time instead of using lbolt directly.
869 OpenAFS 1.5.68 (2009-12-08)
873 * aklog now attempts to convert non-AFS errors to human-readable
876 * Make stack not executable when compiling assembler source with GCC.
878 * Numerous source warning cleanups and code reorganization.
882 * Compute midnight for volume statistics calculation from local time.
884 * Salvager now orphans duplicate special inodes when running to allow
885 recovery in event of a problem, instead of simply ignoring the
888 * Support to ensure a server panic attempt leaves a core and thus
889 restarts in a timely manner, rather than potentially hanging. Use
890 panic to attempt cleanup before leaving a core when possible.
892 * Volume sync data reported during bulkstatus is now set correctly.
894 * Provide better tuning for fileserver file descriptor caching.
896 * Allow more than 128 threads in fileserver by modifying host
897 structure in-use tracking.
899 * Avoid crashes getting volume server status during transaction
902 * Improved logging of offline volume conditions.
904 * Correct volume statistics when cloning a volume.
906 * Avoid referencing host structures in the fileserver which are marked
909 * Demand attach fileserver corrections to avoid coring during an
912 * host array bounds checking corrections to avoid buffer overflow.
914 * Handle special inodes correctly when promoting an inode fileserver
915 readonly volume to read-write.
919 * Set the DOS Readonly attribute on a file/directory whenever the unix
920 mode combined with the mask 0200 is true. Previously there was a
921 discrepency between the mask used for testing for readonly behavior
922 and that used for setting the attribute.
924 * Disable AFSVolSync based .readonly "whole-volume callback" support
925 because the all file servers prior to 1.5.67 (and perhaps 1.4.12) do
926 not properly assign a value to the AFSVolSync structure in bulk
927 status RPC responses.
929 * Improve the error output from aklog to output the value from krb5
930 error_message() if the afs_com_err output indicates an unknown
933 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and do not permit
934 them to be exposed to the smb redirector.
936 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT to avoid
937 confusion within the smb redirector.
939 * Fix the byte order assigned to port numbers associated with AFSDB
940 record lookups. They must be network byte order not host byte
943 * Add dynamic server ranking based on RPC round trip time
946 All UNIX client platforms
948 * Additional shutdown-time memory leaks removed.
950 * Improved logging of resource contention.
952 * Provide dumping for Rx debug packet tracking support in source.
954 * Update afscp test client to build, and provide an unlock client.
956 * Client buffers for directory parsing can now be allocated beyond the
957 fixed set formerly provided.
959 * Work around race condition when manipulating read-only volume
962 * Bugfixes to get PAG value pioctl.
964 * Bugfixes to SRV record support.
968 * Path MTU tracking code cleanup.
970 * Avoid an oops due to racing with vcache recycling thread.
972 * Changes to keyring PAG handling: for sufficiently new kernels, use
973 only keyring-based PAGs, and disable group PAGs entirely.
975 * Updates to the kernel page cache interface: writing pages will now
976 not spuriously leak page locks, and will avoid requiring duplicate
979 * Credential references are now tracked using native atomic counters.
981 * Kernel mutex/semaphore lock ordering fix to avoid deadlocks.
983 * Manipulate disk cache with credentials used to initialize it, to
984 avoid security issues.
988 * Fix fstrace message catalog location.
990 * Fix kernel fstrace logging.
993 OpenAFS 1.5.66 (2009-10-25)
997 * Avoid calling exit() in library code.
999 * Add rx window size and peer timeout tuning APIs.
1001 * Correct rx peer timeout handling to disallow 0ms timeouts.
1003 * Correct calculation of rx RTT by disregarding retransmitted packets.
1005 * vos manpages updated to reflect changes in recent versions.
1007 * GNU-style long options (e.g. --cell) are now supported in all
1010 * fs listacl can now print a command to recreate the current ACL.
1012 All server platforms
1014 * Fix a race on transaction objects in the volserver which can cause a
1017 * Avoid destroying and setting to NULL the callback connection when it
1018 could still be being used.
1020 * Correct unlink handling in salvager.
1022 * Improve error messages due to I/O errors in the volserver.
1024 * Correct an issue which caused converted RO to RW volumes on namei
1025 fileservers to not come online immediately.
1029 * Official support for Windows 7 and Server 2008 R2.
1031 * Prevent a file server bug (FetchData returning an invalid length
1032 instead of zero) from causing an "unexpected network error" when
1035 * Promote DNS SRV records as superior to DNS AFSDB records. Support
1036 arbitrary port numbers for vldb servers.
1038 * Add AFSVolSync based .readonly "whole-volume callback" support.
1039 With this functionality, multiple objects from a .readonly volume
1040 can have their status validated by issuing a single
1041 RXAFS_FetchStatus RPC.
1043 * Remove drive mapping functionality and service start/stop from
1046 * Remove drive mapping functionality from afs_config.exe.
1048 * Use {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab" to
1049 restore access to drive mapping functionality in afscreds.exe and
1052 * Adjust SMB error return codes to avoid returning STATUS_TIMEOUT
1053 which results in the SMB redirector disconnecting.
1055 * Network Identity Manager OpenAFS Provider now provides its own "AFS
1056 lock" notification icon to report the status of "have tokens, have
1057 no tokens, service not started, service started but inaccessible".
1058 Hovering over the icon lists the cells for which tokens exist (if
1059 any) and the OpenAFS version number. Double-clicking executes the
1060 Network Identity Manager default action.
1062 * Prevent pioctl calls from retrying indefinitely when a sharing
1063 violation error occurs.
1065 All UNIX client platforms
1067 * Correct a condition which could discard the error from initializing
1070 * Avoid using invalid references to afs_Conn connection structures,
1071 and thus potentially producing invalid data when a retry is needed.
1073 * SRV records are now supported for discovering AFS servers.
1077 * Correct writepage behavior.
1079 * Fix error code handling in the writepage code.
1081 * Avoid leaking page locks, which could potentially hang a machine.
1085 * Preferences Pane improvements.
1089 * Avoiding attempting to handle critical signals in servers, so that
1090 core file handling works correctly.
1093 OpenAFS 1.5.65 (2009-10-06)
1097 * Code compilation warning fixes, to enable better finding and
1100 * Provide configure-time switch to enable code warning compilation.
1102 All server platforms
1104 * Demand-attach fileserver now makes volume LRU list operations
1105 exclusive operations to avoid races during adding to the list.
1107 * Fileservers now avoid potential "negative length" fetches.
1109 * A leak in host tracking objects in the fileserver has been fixed.
1111 * Salvager now unlinks all files by full path, to deal with the change
1112 to not chdir for core file tracking.
1114 * Salvager avoids asserting if the volume header is unreadable.
1116 * Demand-attach fileserver puts back volume references from fssync
1121 * Improved service response to suspend and shutdown event
1124 * Avoid a bug in the file server that can result in an invalid length
1125 being returned as part of a fetch data response if the client
1126 attempts to read beyond the length of the file.
1128 * Do not publish a default stream object for directories and mount
1129 point objects. This was impacting the ability of some Windows XP
1130 systems to save roaming profiles.
1132 All UNIX client platforms
1134 * A bug which could cause erroneous handling of lengths on data reads
1137 * A bug where erroneous length returns from the fileserver could
1138 result in a false error has been fixed.
1142 * Background page copies are now supported for enhanced disk cache
1145 * Blocking readahead is supported in readpages() to reduce overhead.
1147 * Use readpage() instead of read() to access cache data to avail disk
1148 cache users of the kernel backing cache for improved performance.
1150 * Minimize credential handling for improved performance.
1154 * Preferences Pane cleanup.
1158 * Provide a fs_pathconf method with sensible defaults.
1160 * Provide a _PC_FILESIZEBITS method to fix some NFS translator
1164 OpenAFS 1.5.64 (2009-09-22)
1166 All server platforms
1168 * The demand attach fileserver now puts back volume references gotten
1169 via the fssync interface.
1171 * The demand attach fileserver had a structure reference error, which
1172 has been correected.
1176 * Restores Windows 2000 compatibility.
1178 * Fixes a data consistency error between the output of NetWkstaGetInfo
1179 and NetServerGetInfo RPCs, specify the Lan workstation group name
1180 "AFS", and report server name as "AFS" instead of "\\AFS" when the
1181 caller asks for "\\AFS".
1183 * Enables executables to be run from \\AFS on Windows 7. Returns
1184 "Name not found" instead of "File not found" when a directory or
1185 file name cannot be found. This avoids loader errors when system
1186 dlls cannot be located in the executable directory.
1188 * Prevents cache manager from marking the file server "down" when the
1189 data returned in response to either RXAFS_FetchData64 or
1190 RXAFS_StoreData64 is invalid.
1192 * Adds pioctl data validation to the AFS Explorer Shell extension.
1194 All UNIX client platforms
1196 * A bug which could cause a kernel panic in 1.5.63 has been corrected.
1197 This would manifest as a GetDCache panic or oops.
1201 * aklog -setpag works again with recent kernels when keyring is in
1206 * When Fast User Switch is in use, AFS login is now handled correctly
1207 by the integration tool included with the preferences pane.
1209 * Several packaging bugs have been corrected.
1212 OpenAFS 1.5.63 (2009-09-11)
1216 * The restorevol command is now documented and installed as a user
1219 * The uss command now properly translates vldb entries to its expected
1220 format when handling them in all cases.
1222 * Documentation now refers to Kerberos instead of kaserver.
1224 All server platforms
1226 * bosserver now handles BosConfig.new when restarting, allowing
1227 configuration to be replaced at restart time rather than with bos
1228 delete and bos create. Documentation is updated to reflect this.
1230 * The demand attach fileservice not longer potentially hangs trying to
1231 terminate demand-salvages which have already exited.
1233 * The demand attach fileservice has been modified to avoid spurious
1234 'SYNC_putRes: write failed' warnings when some protocol messages
1235 cannot be acknowledged due to the sender terminating the connection.
1237 * In the event of failure to contact the vlserver or ptserver, the
1238 fileserver will not exit and trigger a forced salvage. It will
1239 continue to try in the background to contact the needed services.
1241 * The salvager can now repair certain cases of a damamged vnode index.
1243 * The accessDate metadata for a volume is now updated correctly.
1247 * CRITICAL: Some applications for example those based on Cygwin were
1248 unable to access data stored in the AFS name space. Explorer Shell
1249 also experienced inconsistent behavior. This is fixed.
1251 * CRITICAL: Multiple AFS pioctl requests issued nearly simultaneously
1252 by applications could result in pioctl responses being received by
1253 the wrong requester. This in turn could result in application
1254 crashes. symlink.exe, fs.exe, afslogon.dll, afscreds.exe, and the
1255 netidmgr afscred.dll plugin were all affected.
1257 * Some XP machines running 1.5.62 had trouble saving roaming profile
1258 data. This is fixed.
1260 * Integrated Logon (afslogon.dll) did not function with domain
1261 specific configurations.
1263 * Ensure that access denied and over quota errors experienced while
1264 storing data to the file server do not result in on-going retry
1267 All UNIX client platforms
1269 * Except on Solaris and AIX, the compiler may now be overriden at
1270 configure time by setting the CC environment variable.
1272 * afsd now properly deals with large cache partitions.
1276 * Build shared libafsauthent and libafsrpc.
1280 * Kernel module DKMS support now installs an unstripped module to
1281 allow debugging information to be collected.
1285 * Preferences pane properly updates token information.
1289 * klog will now properly handle passwords of 8 or fewer characters
1290 with an AFS string to key on hosts able to run 64 bit binaries.
1292 * A panic at AFS shutdown due to "NO PCB" on a udp_lock has been
1295 * The panic decoder script included in the source now properly handles
1296 32 and 64 bit panics.
1300 * Avoid defining AFS_KERBEROS_ENV globally as it creates a circular
1303 * Build shared libafsauthent and libafsrpc.
1307 * Build shared libafsauthent and libafsrpc.
1310 OpenAFS 1.5.62 (2009-08-28)
1314 * Numerous invisible changes to improve code maintainability,
1315 portability and enhanceability.
1319 * CRITICAL: Fixes two errors that can result in data loss when storing
1320 data to the file server.
1321 1. Failure to Store Portions of Unaligned Writes
1322 2. Failure to Store Data to File Servers Lacking Large File Support
1323 Read the announcement for more details:
1324 http://www.openafs.org/pipermail/openafs-announce/2009/000305.html
1326 * CRITICAL: The cache manager daemon thread could terminate when the
1327 machine enters suspend mode. This daemon thread performs the
1328 background check of down servers, offline volumes, callback
1331 * CRITICAL: Integrated Logon (afslogon.dll) was terminating
1332 unexpectedly. Error checking has been improved and NULL pointer
1333 dereferences after Lsa API calls fail have been eliminated.
1335 * For the first time, the OpenAFS SMB Server supports the DCE RPC
1336 services SRVSVC and WKSSVC. Browsing \\AFS with the Explorer Shell
1337 or NET VIEW will now be faster and provide additional functionality.
1338 No longer will cell names longer than 12 characters be truncated.
1340 * Improvements to DFS Referral request processing have been
1343 * Unnecessary DNS lookups of share names are avoided improving
1346 All UNIX client platforms
1348 * Non-Kerberos PAM modules work correctly again.
1352 * MacOS 10.6 (Snowleopard) is now supported, both 32 and 64 bit mode.
1354 * Updates to the AFSCommander preferences pane.
1356 * Installer now permits cell names with dashes.
1359 OpenAFS 1.5.61 (2009-08-06)
1363 * Correct another race condition in the Rx library that could result
1364 in an unexpected panic while freeing the Rx call iovq.
1366 * rx packet resend and data packets sent counts were incorrect.
1368 * fs setquota, fs setcachesize, vos setfields, and vos create now
1369 accept human readable orders of magnitude. (K, M, G)
1371 * fs listquota fixed to permit large quota sizes to be displayed.
1373 * Correct documentation of bosserver permissions requirements.
1375 * Modify vlserver to avoid potentially corrupting the database through
1378 * Generalized support for fast Rx timeout due to network
1381 All server platforms
1383 * Allow audit logs to be sent via sys5 IPC message queues instead of
1388 * If a file server becomes inaccessible while the cache manager has
1389 dirty buffers to write, the afsd_service buf_IncrSync thread can
1390 attempt to use 100% of the cpu.
1392 * Fix "fs newcell" which was broken in 1.5.60.
1394 * Do not attempt to synchronize dirty buffers if the associated volume
1395 is known to be unavailable.
1397 * Modify behavior of a Freelance mountpoint target that does not
1398 specify a cell. Instead of assuming the target volume is in the
1399 Freelance.Local cell, use the workstation "Cell" specified in the
1400 registry. A mountpoint target of "#root.cell." will now mean the
1401 root.cell volume in the workstation cell for the current session.
1402 If the workstation cell changes from "athena.mit.edu" to
1403 "andrew.cmu.edu", the referenced volume will also change without
1404 requiring that the mount point targets be altered.
1406 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
1407 to permit the server Uuid to be used to lookup the server object and
1408 from that determine the cell. This permits callbacks that are
1409 received from alternate addresses to be processed with a known
1410 server object. Previously a request from an unknown server would
1411 clear all callbacks from all cells.
1413 * Fix a bug that prevented optimal performance when using a non-zero
1414 value for 'daemonCheckVolCBInterval'. As a reminder, when
1415 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
1416 volume callbacks are automatically renewed 90 minutes before their
1419 * Fix automatic ranking of vldb servers whose values are obtained from
1420 the CellServDB file.
1422 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
1423 or the call is to a vldb server.
1425 * Add registry based cell search functionality to NetIdMgr,
1426 afs_config.exe, and klog.exe.
1428 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
1429 lookups on the host names in the CellServDB instead of using the
1430 specified IP addresses. This provides aklog, pts, vos, etc. the
1431 same CellServDB behavior that the Windows Cache Manager uses.
1433 * When updating the stat cache entry callback of a .readonly object
1434 from the volume group object, update the file server reference to
1435 ensure it matches the most update to date callback.
1437 * Add proper support for processing callbacks from multi-homed file
1438 servers. Instead of comparing servers by cm_server_t pointer,
1439 compare them by UUID when the UUID is known.
1441 * During a shutdown short circuit the offline volume check daemon
1444 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
1445 preference to an error code reported by rx_EndCall.
1447 * Add "PerFileAccessCheck" registry value to permit testing against
1448 experimental file servers that include per-file acl support. This
1449 value is intentionally undocumented. It is not to be used by
1450 production environment deployments.
1452 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
1453 provider from obtaining tokens when referrals are in play.
1455 * Add "fs chown" and "fs chgrp" commands to permit the owner and group
1456 of objects stored in AFS to be set from Windows.
1458 * Avoid performing background daemon operations when the machine is
1459 going into suspend mode.
1461 * Perform offline volume checks in most recently used order.
1463 * Prevent crash when a data version for a cache object goes backwards.
1465 * Multi-thread safe library versions are now being generated and used.
1466 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
1468 * Microsoft SMB Redirector (mrxsmb.sys) support for
1469 ExtendedSessTimeout values are now available on XP through Windows
1470 7. Add functionality to autodetect if such support is present on
1471 the machine. If so, configure it if necessary and dynamically
1472 adjust the AFS Rx timeout values accordingly.
1474 All UNIX client platforms
1476 * Fix out-of-tree source builds.
1480 * GUI installer now asks for local cell information.
1482 * AFS Commander preferences pane is now installed by default.
1486 * Avoid kernel panics due to null pointer dereferences in the network
1487 interface poller kernel thread.
1490 OpenAFS 1.5.60 (2009-05-31)
1494 * Retry volserver transaction creation on failure.
1496 * Allow building HTML and PDF documentation from included XML copies
1497 of User Guide, Admin Guide and Quick Start Guide for Unix.
1499 * Documentation updates and additional documentation.
1501 * Add -encrypt support to pts client.
1503 * Convert MR-AFS fs commands to OSD commands.
1505 All server platforms
1507 * Updated background sync process in fileserver to avoid a race which
1508 could result in a volume being taken offline.
1512 * On April 9th Microsoft released a Hot Fix for Windows Server 2003
1513 SP2 that corrects a deadlock in the smb redirector and also adds new
1514 functionality that permits the AFS SMB server to be given a longer
1515 timeout than is normally the case. New functionality has been added
1516 to configure these additional LanmanWorkstation\Parameter values.
1517 (This functionality has been backported to XP SP3 and is scheduled
1518 to be released on June 5th.)
1520 * Fix RT#124787, a race condition between "fs flush <dir>", "fs
1521 flushvolume", or "fs flushall" and on-going directory operations
1522 that can result in afsd_service.exe crashing.
1524 * Release Notes, User and Administrator guides are now shipped as
1525 indexed Windows HtmlHelp Files. (.chm). Shortcuts are provided from
1528 * A method of specifying Client CellServDB information within the
1529 registry has been added that can be used to either override the
1530 CellServDB file or force the use of DNS lookups for a given cell.
1531 See the release notes for details.
1533 * The pioctl interface now properly handles drive letter substitution
1534 to UNC paths. (SUBST <d:> <\\afs\cell\path>)
1536 * The BackConnectionHostNames registry value configuration was broken
1537 when dynamic re-establishment of Netbios Name registrations was
1538 added. This release restores the functionality.
1540 * All hidden vos.exe commands are now revealed.
1542 * Attempts to store the same dirty file chunk from multiple threads
1545 * The IsPathInAfs test in Explorer Shell Extension and fs.exe now
1546 permits broken symlinks to be treated as being in AFS.
1548 * vos.exe commands that output 64-bit integer values once again do so.
1549 This was broken in 1.5.59.
1551 * Cygwin Import Libraries are provided in the SDK for all OpenAFS
1552 DLLs. This permits building cygwin applications against OpenAFS
1555 * NSIS installer does a much better job of cleaning up files left over
1556 from previous installs.
1558 * libafsconf.dll moved from Client\Program to Common directory as is
1559 is now used by all modules for CellServDB processing.
1561 All UNIX client platforms
1563 * Write back changes on last store for memcache to avoid discarding
1566 * Abstract disk cache support to allow for path, fh, inode based
1567 caches with no need for messy ifdef structures each time a new type
1572 * Support as a userland port.
1576 * Corrected structure definition for userspace cache manager to allow
1581 * Corrected client locking support.
1583 * Updated patch to stop deadlocking in the kernel during mmap.
1585 * Avoid oops when setting up groups for PAGs to match keyrings.
1587 * Use Linux fh-based cache in cases where possible by default.
1591 * Corrected structure definition for userspace cache manager to allow
1596 * Support for OpenBSD 4.5.
1600 * Corrected support for server-side vos split interface.
1603 OpenAFS 1.5.59 (2009-04-06)
1607 * Increased service priority class to "High" to match the priority of
1608 system components that are dependent upon the a timely response.
1610 * SMB error responses avoid returning errors that could confuse the
1611 Microsoft SMB redirector into disconnecting the connection to \\AFS.
1613 All UNIX client platforms (except MacOS X 10.4 and 10.5)
1615 * OpenAFS 1.5.59 contains fixes for the client issues addressed by the
1616 security advisories OPENAFS-SA-2009-001 and OPENAFS-SA-2009-002.
1620 * Support for prerelease Linux 2.6.30 kernels.
1623 OpenAFS 1.5.58 (2009-03-30)
1627 * Code cleanup and prototyping.
1629 * Avoid unnecessary blocking in Rx periodic cleanup code.
1631 All server platforms
1633 * Fileserver CopyOnWrite routine optimized for performance.
1635 * Make fileserver callback dumps 64 bit safe.
1637 * Fix byte order issues with fileserver host hashing.
1639 * Fix buffer size issues with butc.
1641 * Fix several Ubik recovery issues.
1643 * Avoid leaking file references in the fileserver.
1645 * Fix a race in DAFS while closing vnodes, and another offlining
1648 * volserver interfaces for volume splitting client.
1652 * [RT 124293] A race condition exists which can result in a crash.
1654 * [RT 124276] If the vldb is out of sync with the contents of the file
1655 servers, afsd_service will retry too many times when a file server
1656 reports a volume as not being present. Now if the list reported by
1657 the vldb is the same as the previously seen list, then the retry is
1660 * [RT 124276] Read-only volume failover was broken in 1.5.53 whenever
1661 accessing a volume results in VNOVOL or VMOVED.
1663 * [RT 124276] Prior to 1.3.70 the volume server reference list was not
1664 reference counted and would be prematurely freed while in use. When
1665 reference counting was added in 1.3.70 a bug was introduced that
1666 could result in service reference list corruption.
1668 * Add Windows Application Event Log warning messages for "Client SMB
1669 MPX value too large" and "Client SMB Buffer Size too small".
1671 * Renaming of files across directory boundaries would result in an
1672 invalid handle error when attempting to access the files after the
1675 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail when a
1676 smb file descriptor cannot be found. The whole point of using a
1677 Path Info function is because an smb file descriptor wasn't
1680 * More edge cases in which dynamic addition of Freelance root.afs
1681 entries would get the wrong FID or where the root.afs directory
1682 would not be refreshed.
1684 * Buffer overflow could occur if the workstation cell name was longer
1685 than 64 characters. Crashes could occur in afscreds.exe,
1686 afslogon.dll, and afsd_service.exe.
1688 * VNOSERVICE and VOFFLINE errors were leaking and were exposed to the
1691 * Log file server uuid values as part of the cm_server object when
1692 available. Dump the cm_server object list in response to "fs
1695 * Optimize the performance of resetting access control lists when
1696 tokens are set or removed.
1698 * Remove symlink recursion tests and increase max symlink count to 64
1701 * Windows specific Rx performance improvements.
1703 * Support for Network Identity Manager 2.0
1705 All UNIX client platforms
1707 * Avoid issues with freeing resources at shutdown.
1709 * Numerous fixes to disconnected AFS.
1711 * Disconnected AFS fixes for replaying changes without double-freeing.
1713 * Attempt to use krb524 principal conversion in aklog if available.
1717 * Kerberos configuration at build time corrected.
1721 * Default to dynamic allocation of AFS kernel cache entries to allow
1722 growth for inotify()-pinned entries. (beagle, famd, etc)
1724 * Change client truncation routines to avoid locking issues.
1726 * IA64 port clients fixed on Linux 2.6.
1728 * RPMs now install fstrace message catalog.
1730 * Support through kernel 2.6.29 tested.
1732 * Fix locking issues on CellServDB file in /proc.
1736 * Support OpenBSD 4.4
1739 OpenAFS 1.5.57 (2009-01-23)
1743 * Conditional compilation of rxdebug support is now possible.
1745 * Documentation updates.
1747 * Further race connditions in Rx have been corrected.
1749 All server platforms
1751 * Salvager no longer attempts to recreate headers in the wrong
1754 * Volumes are properly marked in use on creation and subsequently on
1755 examination with vos.
1759 * Undo the "UAC manifest fix" applied to afs_config.exe.
1761 * Ensure that Freelance allocation of vnodes follow the AFS convention
1762 of odd vnodes are directories and everything else is an even vnode.
1764 * Add Freelance logic to mount point and symlink evaluation functions.
1766 * Enhance smb_ParseASCIIBlock() so that it can handle all of the
1767 STRING formats defined by the CIFS Technical Report 1.0.
1769 * Validate the output of smb_ParseASCIIBlock() in all callers. Return
1770 CM_ERROR_BADSMB if the STRING field cannot be parsed.
1771 CM_ERROR_BADSMB will cause the contents of the packet to be logged.
1773 * If multiple SMB Raw Write operations were taking place at the same
1774 time, there could be data corruption because unique event objects
1775 were not generated for each Netbios receive operation.
1777 All UNIX client platforms
1779 * Userspace AFS library can now deal with large files when supported
1782 * Numerous updates to disconnected AFS support, including changes to
1783 allow reconnection to work in more circumstances.
1787 * FreeBSD unstrategy code has been updated.
1791 * A race during file truncation has been corrected.
1793 * System call probing routines have been updated.
1795 * 2.6.29 is now supported.
1799 * 10.3 support has been corrected.
1803 * Initial OpenBSD 4.4 support.
1807 * Updates to allow compiling on newer OpenSolaris are now included.
1810 OpenAFS 1.5.56 (2008-12-30)
1814 * libuafs (userspace cache manager) updated to correct several errors.
1816 * Additional rx debugging support is available as a conditional
1819 * A race condition in Rx leading to a panic has been corrected.
1821 * Rx idle time tracking has been corrected.
1823 * ubik clone server support has been corrected.
1825 All server platforms
1827 * Salvager no longer leaves cores in vice partitions.
1829 * The vol-dump tool now supports dumps larger than 2gb where possible.
1831 * Operations on multiple files now report all FIDs in the audit log.
1833 * butc XBSA support now works correctly on amd64.
1837 * The NetIDMgr AFS Provider automated configuration logic was broken
1838 by the introduction of Kerberos referrals. If the realm of the
1839 identity cannot be determined, the workstation cell is now assumed
1840 to belong to the newly created identity.
1842 * Avoid a reference count under flow during rename operations.
1844 * Avoid a crash caused by treating an arbitrary length directory
1845 search mask as an 8.3 mask.
1847 * Prevent rename operations if a case insensitive match for the target
1848 name already exists and does not refer to the object being renamed.
1850 * Increase the maximum number of background daemons to 64.
1852 * Fix the UAC manifest applied to afs_config.exe
1854 All UNIX client platforms
1856 * Updates to disconnected AFS support.
1860 * FreeBSD 7.1 is now supported.
1862 * amd64 FreeBSD is now supported.
1866 * Generic fh (exportfs API) cache type is now available.
1868 * Avoid some oopses due to backing_dev_info inode fields not being
1871 * 2.6.28 is now supported.
1875 * 10.3 support has been corrected.
1879 * Large partition support has been corrected.
1881 * Filesystem-agnostic cache is now available on Solaris 10 and 11.
1884 OpenAFS 1.5.55 (2009-11-10)
1888 * Salvager avoids leaving core files in vice partitions.
1890 * NFS translator fixes.
1892 * Unresponsive server handling fixes.
1894 * A volserver race which could result in duplicate transactions is
1899 * Fixes a panic caused by corruption of the SMB virtual circuit list.
1902 * Fixes a panic caused by receipt of a UTF-16 string that cannot be
1905 * Implements a more aggressive recovery algorithm for Netbios errors
1906 that result in loss of communication to the AFS SMB server.
1908 * Improve pioctl response time when testing whether or not a PATH is
1911 * Adds support for linked cells.
1913 * Increases the length of the cell and realm names that can be input
1914 into the Network Identity Manage AFS provider configuration dialog.
1916 All UNIX client platforms
1918 * Disconnected AFS avoids infinite recursion during rmdir.
1922 * Support for 2.6.28 (not complete for NFS translator modules).
1924 * Support for using exportfs API for filesystem-agnostic cache.
1926 * Disable backing store readahead.
1928 * Avoid deadlocks when writing back mmapped files larger than the
1931 * Avoid Oops when doing PAG garbage collection.
1934 OpenAFS 1.5.54 (2008-10-08)
1938 * Updates for new Tivoli X/Open API finding.
1940 * A double-free is corrected in Rx.
1942 All server platforms
1944 * Ubik cleans up file descriptor cache correctly when doing recovery.
1946 * Enhanced vldb error checker included.
1950 * Prevent a crash that could occur when multiple file / directory
1951 change notifications are processed simultaneously.
1955 * AFS claims more free space so Finder will attempt file copies.
1959 * Avoid spurious ENOENT when calling gwtcwd() on a volume root.
1961 * Avoid spurious ENOTDIR during fakestat.
1964 OpenAFS 1.5.53 (2008-09-26)
1968 * rx avoids many packet leaks.
1970 * rx jumbogram disabling now works (and is the default).
1972 All server platforms
1974 * Demand Attach fileserver tries to avoid issues tracking offline
1979 * Many potential deadlock conditions due to out of order lock
1980 acquisitions have been corrected.
1982 * A race condition resulting in an undercount on the cm_scache_t
1983 reference counts has been corrected.
1985 * An empty string when sent as an ioctl path is now properly
1986 interpreted as meaning the current directory. This affects "fs
1987 lsm", "symlink list", etc.
1989 * Fix smb string parsing differences where the smb protocol
1990 documentation does not match the actual Windows implementation.
1992 * Random access denied errors fixed.
1994 * A file server lock synchronization issue was corrected in SMB
1995 NTCreateX and NTTranCreate operations. This bug prevented properly
1996 operation when loading roaming profiles.
1998 * Fix a heap overwrite error during server probe operations if a new
1999 server is added while a probe operation is in progress.
2001 * Fix an LSA memory leak that was the result of an LSA error.
2003 * Do not leak cm_cell_t objects if the VLDB server lookup fails.
2005 * Only initialize rx mutex/lock objects once.
2007 * Do not nul terminate the AFS volume name when reported to Windows.
2009 * Improve VNOVOL error handling. Prevent updated vl information from
2010 being destroyed immediately after it was acquired. This bug
2011 prevented proper fail over when volumes are moved or removed from a
2014 * Remove volume id from the server volume list in response to VMOVED
2017 * "fs flushXXX" commands now destroy locally built B+ directory trees.
2019 * Prevent mixture of locally modified directory pages and file server
2022 * Fail over to alternate vl servers if a ubik error is returned.
2024 All UNIX client platforms
2026 * Disconnected AFS now supports read-write mode.
2028 * volserver now builds correctly.
2032 * AIX 6 is now supported.
2036 * FreeBSD 7 is now supported.
2040 * cache bypass is now supported.
2042 * 2.6.x kmod compilation now uses kernel compile options always.
2044 * Support through 2.6.27.
2048 * Show more space free so Finder doesn't get confused.
2052 * Default to namei rather than inode.
2055 OpenAFS 1.5.52 (2008-08-18)
2059 * Initialize volume updateDate at volume creation.
2061 * Avoid potential corruption of directories during salvage.
2063 * Check for out of memory condition during allocation of additional Rx
2068 * Restore support for Windows 2000 (broken in 1.5.50).
2070 * Perform additional validation on volume names in mount points during
2071 creation and evaluation.
2073 * Fix several deadlocks, races, and reference count issues.
2075 * Further optimize SMB Directory Search processing and minimize the
2076 number of InlineBulkStatus RPCs sent to the file server.
2078 * Enable "bos restricted" operations.
2080 * Fix the create of submounts used by the AFSCreds and afs_config
2083 * Fix a short name truncation bug. (1.5.50)
2085 * Fix the error code reported when attempting to delete a file on a
2086 readonly volume or one that is marked with the readonly DOS
2089 * Fix a heap corruption error when reading the CellServDB file
2092 * Add the "RxUdpBufSize" registry value. The new default is 256KB.
2094 * Do not include trailing NULs in the directory search output.
2097 * Pre-allocate 64 Rx Packet buffers per thread in order to improve
2100 * For debugging: add smb lock requests and stat cache lock allocations
2101 to the output from "fs memdump".
2105 * Workaround broken sigwait() to allow fileserver to shut down
2106 correctly pre NetBSD 5.0.
2110 * Default to namei fileserver; Allow inode fileserver at configure
2115 * Try harder to avoid kernel panics for malformed requests.
2118 OpenAFS 1.5.51 (2008-07-29)
2122 * salvager tries harder to arrange for clients to get VBUSY while
2123 salvaging single volumes.
2125 * salvager avoids certain corruption when salvaging directories.
2127 * Rx connection clones disabled.
2131 * The 32-bit EXE 1.5.50 installer failed to properly install the C
2132 Runtime library. When used as an upgrade OpenAFS would continue to
2133 work but when used as a new installation, OpenAFS binaries would
2136 * Fixes the "fs" and "symlink" commands to properly parse Unicode path
2137 prefixes during the pioctl remote procedure call. This bug would
2138 result in file not found errors for files and directories that
2139 clearly exist. (Bug introduced in 1.5.50)
2141 * Large File support is disabled. (Bug introduced in 1.5.50)
2143 * Removes the possibility of a deadlock during volume location update
2144 operation if all of the reported file servers are unreachable at the
2147 * Ensures that reference counts are properly incremented/decremented
2148 on Rx connection objects used for volume location RPCs.
2150 * Over Quota errors during cm_FSync() calls would lead to an infinite
2151 loop as the error was never propagated to the caller.
2153 All UNIX client platforms
2155 * Bugfixes to disconnected AFS support in the cache manager.
2158 OpenAFS 1.5.50 (2008-07-16)
2162 * volserver puts recloned volumes back online before returning the
2163 volume to the fileserver, avoiding spurious VNOVOL errors.
2165 * Updated TSM X/Open API support available.
2167 * Demand Attach fileserver will not crash due to accesses during
2170 * Substantial documentation updates.
2172 * Demand Attach fileserver state tracking and analyzer tool
2175 * UAFS userspace cachemanager updates.
2177 * Corrected support for anti-meltdown protection in the client.
2181 * UNICODE Character Set Support.
2183 * Pioctl interfaces to the cache manager have been refactored to
2184 provide layering between the SMB specific code and the general
2185 purpose ioctl operation.
2187 * Garbage collect dead SMB virtual circuits as soon as they are no
2188 longer being referenced. This avoids problems with outstanding
2189 locks not being dropped when the virtual circuit becomes invalid.
2191 * Remove the IBM Administration Reference documentation and replace it
2192 with the OpenAFS Command Reference Manual.
2194 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime functions each
2195 time a connection is about to be used. Do not hold a lock on the rx
2196 connection object while it is being selected. This avoids a race
2197 between threads attempting to set the timeout values and removes a
2198 bottleneck that was hampering performance.
2200 * Ensure that the smb directory attribute is set for all directory
2203 * Replace the VC Runtime EXE installer with the MSI installer in the
2204 NSIS installer scripts.
2208 * Support for updates to OpenSolaris.
2212 * Correct dentry revalidation for cross-directory renames.
2214 * Updated rpm packaging materials for 1.5 release series and 2.4
2217 * Corrected syscall table probing.
2219 * NFS translator updates for current kernels.
2222 OpenAFS 1.5.39 (2008-06-24)
2226 * Updates for Demand Attach fileserver.
2230 * Fix two memory leaks.
2232 * Fix one missing lock.
2234 * Handle access denied errors when writing dirty buffers.
2236 * Fix two errors that would cause the *experimental* AFS Servers
2240 OpenAFS 1.5.38 (2008-05-24)
2244 * Add read-only disconnected support.
2247 OpenAFS 1.5.37 (2008-05-21)
2251 * Includes a number of optimizations for testing.
2254 OpenAFS 1.5.36 (2008-05-09)
2258 * Rx optimizations now attempt to deal with high latency WANs.
2260 * Client will not wait infinitely for a server which is not providing
2261 data. Additional servers will be polled without marking the server
2262 which is not providing data down.
2264 * vos move will not erroneously unlock locked vldb entries on failure.
2266 All server platforms
2268 * Fileserver avoids a potential infinite loop when a client
2269 relinquishes an address.
2271 * Fileserver large setting now configures more threads.
2273 * Fileserver properly registers uuids of new clients.
2275 * Ubik servers do not improperly hide updates from clients.
2277 * Fileservers reserve enough file descriptors such that each thread
2278 can cache one to avoid spurious errors.
2282 * Fix a cm_buf_t reference count leak when attempts to write dirty
2283 buffers to the file server from within cm_IncrSyncer() fail.
2285 * Prevent udebug from crashing.
2287 * Another VNOVNODE issue fixed. When writing a dirty buffer to the
2288 file server, if VNOVNODE is received, mark all buffers as invalid
2289 without further attempts to contact the file server.
2292 =======> Changes back to 1.3 have not yet been incorporated here <=======
2297 * -nosettime is now the default for afsd. Use "-settime" to get the
2300 * OpenBSD is now supported.
2302 * Mountpoint directory information is now only faked for cross-cell
2303 mountpoints when using the -fakestat flag (e.g. for the directories
2304 under /afs, but not for most other volumes mounted inside the cell).
2305 The -fakestat-all switch can be used to fake information for all
2308 * When fakestat is enabled on MacOSX, the Finder can be used to browse
2309 a fully-populated /afs directory. However, this precludes reliable
2310 use of entire volumes as MacOS bundles (i.e. containing a Contents
2311 directory in the root of the volume).
2313 * Mountpoint directory information can be faked by the cache manager,
2314 making operations such as stat'ing all cells under /afs much faster.
2315 This is enabled by passing -fakestat to afsd, but might not be stable
2321 * The kaserver now defaults to not allowing interrealm authentication,
2322 due to security vulnerabilities in the krb4 protocol. The new
2323 "-crossrealm" flag to the kaserver is provided to reenable interrealm
2324 authentication if desired.
2326 * RedHat Linux 9.0 is now supported.
2328 * Solaris 9 12/02 is now supported. Solaris 7 and 8 x86 should now
2331 * On Linux machines using 2.2 series kernels, 2.2.19 or higher is now
2334 * An OpenAFS 1.2.9 afsd will not work with kernel modules built from
2335 an earlier OpenAFS release. In general, using a mismatched afsd and
2336 kernel modules set is unsupported; it is not recommended that you use
2337 such a configuration on a regular basis.
2342 * Mountpoint directory information is now only faked for cross-cell
2343 mountpoits when using the -fakestat flag (e.g. for the directories
2344 under /afs, but not for most other volumes mounted inside the cell).
2345 The -fakestat-all switch can be used to fake information for all
2348 * HPUX 11.0 is now supported.
2350 * It is now possible for AFS to use Kerberos 5 directly, via rxkad 2b.
2351 See the OpenAFS 1.2.8 Release Notes for more information on using this
2354 * An NFS translator kernel module is now included and compiled by
2355 default for Solaris only.
2360 * MacOS X 10.2 is now supported. FreeBSD 4.3 and later support is
2361 included in this release, but is still under active development and
2362 should only be used by those doing active development on the OpenAFS
2365 * When fakestat is enabled on MacOSX, the Finder can be used to browse a
2366 fully-populated /afs directory. However, this precludes reliable use
2367 of entire volumes as MacOS bundles (i.e. containing a Contents
2368 directory in the root of the volume).
2370 * The fileserver will now use Rx pings to determine if clients are
2371 reachable prior to allocating resources to them, to prevent asymmetric
2372 clients from consuming all fileserver resources.
2377 * Mountpoint directory information can be faked by the cache manager,
2378 making operations such as stat'ing all cells under /afs much faster.
2379 This is enabled by passing -fakestat to afsd.
2381 * Solaris 9 FCS and Solaris 7 and 8 x86 are now supported.
2386 * A remote denial of service attack in the AIX and IRIX clients has been
2387 fixed. Users of those platforms are strongly encouraged to upgrade.
2389 * Fixed race conditions in fileserver that could result in crash.
2394 * Server logfiles now more consistant about format in which hosts are
2397 * vfsck on Solaris will now allow force runs (using -y flag) even if old
2403 * Cell aliases for dynroot can be specified in the CellAlias file in
2404 /usr/vice/etc or /usr/local/etc/openafs, in format "realname alias",
2405 one per line. They can also be managed at runtime with "fs newalias"
2406 and "fs listaliases".
2411 * Solaris 9 and Linux PA-RISC are now supported.
2413 * fileserver will not erroneously delay legitimate errors for 3 seconds
2414 after 10 errors are returned (e.g. stat() on a directory you can't
2417 * Rx MTU calculation now works for Irix, Solaris and Linux
2419 * If afsd is started with the -dynroot flag, /afs will be locally
2420 generated from the CellServDB. AFSDB cells will be mounted
2421 automatically upon access.
2423 * The namei fileserver allows vice "partitions" to be directories
2424 instead of partitions and will attach and display accordingly.
2425 Creating the file "AlwaysAttach" in the /vicepX directory is used as
2426 the trigger to attach it.
2428 * TSM support for butc no longer requires editing a Makefile, simply
2429 specify the --enable-tivoli-tsm configure option.
2431 * Linux builds no longer require source changes every time the kernel
2432 inode structure changes; the OpenAFS sources will now configure itself
2433 to the actual inode structure as defined in the kernel sources.
2438 * vfsck on Digital UNIX and Solaris will now refuse to fsck mounted
2444 * AFS now supports --prefix and the other directory options of
2445 configure. By default AFS builds assuming it will be installed in
2446 /usr/local. In order to get traditional AFS directory paths (/usr/afs
2447 and /usr/vice/etc) use the --enable-transarc-paths option to
2448 configure. More details on the new directory layout are found in
2454 * Windows 95/98/ME/NT/2000 - Consistent versioning: Installation, AFS
2455 Control Center, Client dialog boxes and properties pages for
2456 executables display a consistent OpenAFS version number. Installation
2457 detects previous installation and prompts the user for upgrade
2460 * Windows 95/98/ME/NT/2000 - Installation features: During installation
2461 the user can select the source of the CellservDB file, AFS home cell,
2462 and drive mappings. During installation a drive path mapping can
2463 include a variable that will be substituted with the current UserName
2466 * Windows 2000/NT - Integrated logon: The Integrated Logon feature
2469 * Windows 95/98/ME - Logon script features: The Windows 95/98/ME client
2470 now offers a command-line option for starting up the AFS client
2471 without authenication. It is now possilbe to start the AFS client
2472 first and obtain tokens, and map drives all through Windows scripts.
2473 This helps using Windows 95/98/ME client in Kerberos 5 environment.
2475 * Windows 2000/NT - LANA numbers: AFS client now scans the LANA numbers
2476 to establish the correct NETBIOS connection. NetBEUI is no longer
2477 needed. The user no longer needs to find the correct LANA number.
2479 * Windows 2000/NT - OpenAFS naming consistancy: Further progress has
2480 been made to remove references to "Transarc AFS" and replace with
2486 * AFS now builds with configure. The README for building has been
2487 updated and includes full details.
2489 * A client system can now have multiple sysname values for @sys. They
2490 will be searched in order when looking up files in AFS. The
2491 -newsysname argument to fs sysname can be repeated to set multiple
2494 * A new system group is created for new cells (system:ptsviewers with id
2495 -203). If this group exists, members of this group can examine and
2496 read the entire protection database. They can examine all users and
2497 groups and can get the membership of any group.
2499 * A new program, pt_util has been added to the distribution. This
2500 program allows users to print the contents of the protection database
2501 or to edit the protection database without running a ptserver. It can
2502 be used to set up a new cell without ever running in noauth mode. Run
2503 pt_util -h for help.
2505 * The fs setcrypt and fs getcrypt commands have been added. These
2506 commands allow the system administrator to require that the client
2507 encrypt all authenticated traffic between the client workstation and
2508 AFS. The encryption used is weak, but is likely better than sending
2509 unencrypted traffic in most environments. Some functions, such as
2510 looking for a volume may not be encrypted, but data transfer certainly
2511 is. By default data is not encrypted. At this time no significant
2512 experimentation with server performance has been conducted.
2514 * By default AFS is compiled with AFS_AFSDB_ENV, enabling the -afsdb
2515 option to be given to afsd on startup. If this option is used, then
2516 new cells will be looked up using AFSDB records stored in DNS if they
2517 are not found in CellServDB. This means that users can create
2518 cross-cell mountpoints in directories they control to access cells not
2519 in root.afs, and that cells in root.afs need not be in the client's
2522 * AFS database servers can be marked as read-only clones. Surround the
2523 hostname in square brackets on the bos addhost command and the
2524 database server will never be elected sync site. This is useful for
2525 cells distributed over a wide region.
2527 * The AFS servers now support the -syslog flag. This flag causes them
2528 to log to syslog rather than to files. This flag is not supported on
2529 NT. For all servers besides the salvager, the flag can also be
2530 specified as -syslog=facility, where facility is an integer facility
2531 code from syslog.h. A -syslogfacility option is provided for the
2532 salvager to accomplish the same goal.
2534 * If the --enable-fast-restart flag is given when configuring AFS, then
2535 the salvager supports the -dontsalvage flag which causes it to exit
2536 without salvaging any volumes. If this is configured into the third
2537 command of a fs process, then the fileserver will start without
2538 salvaging. It will fail to attach volumes that need salvaging and
2539 they can be salvaged manually. This provides significantly better
2540 server startup performance at the cost of administrative complexity.
2542 * If the --enable-bitmap-later flag is given when configuring AFS, then
2543 the fileserver creates bitmaps for free vnodes on demand, allowing
2546 * If bosserver finds a BosConfig.new file at startup, it reads this file
2547 and renames it to BosConfig. This allows bosserver to be reconfigured
2550 * The bosserver can be placed in a restricted mode in which AFS
2551 superusers are only granted limited access to the server host. The
2552 following functionality is disabled when restricted mode is in use:
2555 bos getlog (except for files with no '/'s in their name)*
2561 specific exceptions are made for functionality that "bos salvage"
2564 A cron bnode who's name is "salvage-tmp", time is now, and command
2565 begins with "/usr/afs/bin/salvager" may be created. This bnode deletes
2566 itself when complete, so no special "delete" support is needed. This
2567 functionality may be removed in the future if a "Salvage" RPC is
2570 The file with the exact path /usr/afs/logs/SalvageLog may be fetched,
2571 since that is how bos salvage [...] -showlog is implimented.
2573 Restricted mode is enabled using a new bos command (bos setrestricted)
2574 or bossever command line switch (bosserver -restricted). Restricted
2575 mode can be disabled by a) sending the bosserver process a SIGFPE
2576 (which will then allow restricted operations until the next restart or
2577 setrestricted command) or b) editing /usr/afs/local/BosConfig (or
2578 BosConfig.new), and restarting the bosserver.
2580 * The bos UserList of trusted administrators can now contain cross-realm
2581 Kerberos principals.
2583 * udebug now takes --server not --servers.
2585 * Several error messages have been improved to include volume numbers.
2587 * Several new ports have been included for UNIX platforms: Darwin
2588 (ppc_darwin_12 and ppc_darwin_13), Linux 2.4 (i386_linux24), Linux on
2589 the Powerpc (ppc_linux22 and ppc_linux24), Linux on the Sparc
2590 (sparc_linux22, sparc64_linux22 and sparc64_linux24).
2592 * Incomplete FreeBSD and Alpha Linux ports are included. The FreeBSD
2593 port has a working server and the Alpha Linux port has a partially
2596 * A native client for Windows 95/98/ME has been added to the
2597 distribution. With this program, a gateway machine is no longer
2598 required for Windows 9x to access AFS files. One drive letter will be
2599 created on your machine by default - Z:. The Z: drive will be the
2600 root of the AFS tree, allowing you to browse all sites that have AFS
2601 servers available. Additional drive letters can be defined for other
2602 AFS directories. A Windows Explorer shell extension is included that
2603 allows you to right click on items within an AFS tree to bring up an
2604 "AFS" menu item and perform various operations on a file or directory.
2605 The most useful item is "Access Control Lists", which allows you to
2606 view and edit the permissions of a particular directory. Command line
2607 tools are also available in the install directory. These commands
2608 include klog, unlog, tokens, kpasswd, symlink, fs and pts. The
2609 installable includes a readme file that contains more information on
2610 how to use the client program and known issues.
2612 * Support for large caches in afsd. Cachefiles are stored in
2613 subdirectories. The default is 2048 files per subdirectory, which
2614 should work fine in most situations. You can use the new afsd option
2615 -files_per_subdir to change this number. Note that the first time you
2616 run afsd with this patch, your cachefiles will get moved into
2617 subdirectories. If you subsequently run an older version of afsd, you
2618 will lose all your cached files.