2 * Fix Windows 2000 compatibility. Do not call rand_s()
3 even if it is compatible with the C run time library
6 * Fix a data consistency error between the output of
7 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
8 workstation group name "AFS", and report server name as
9 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
11 * Enable executables to be run from \\AFS on Windows 7.
12 Return "Name not found" instead of "File not found" when
13 a directory or file name cannot be found.
15 * Prevent cache manager from marking file server "down" when
16 the data returned in response to either RXAFS_FetchData64
17 or RXAFS_StoreData64 is invalid.
19 * Add pioctl data validation to the AFS Explorer Shell extension.
22 * After 1.5.62 afslogon.dll would no longer crash but it
23 also would not load "domain" specific configurations.
26 * An uninitialized variable was present in the symlink
27 recursion detection loop.
29 * Reverse addition of unique per file GUID during
30 Create/Open responses. Inclusion of the GUID prevents
31 Cygwin applications (and others) from accessing //afs.
33 * Treat "filename::$DATA" as a synonym for "filename".
34 This addresses the recent report of roaming profiles
35 failing to save back to the file server.
37 * When given a choice between an error from rx_Write()
38 or RXAFS_EndStoreData(), use the rx_Write() error.
39 This ensures that over quota and access denied errors
42 * If an error occurs during RXAFS_StoreData operations,
43 apply that error to all of the buffers in the BIOD.
44 This ensures that failures will not be retried if they
47 * Do not perform a DNS lookup for a cellname if the
48 cellname prefix is _._AFS_IOCTL_._. This will avoid
49 DNS lookups when the query is for
50 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
52 * Fix Freelance root.afs volume object callback processing.
53 Object synchronization and status merging bugs.
55 * Fix Explorer Shell "invalid parameter" bug introduced
58 * Log RXKAD errors to the trace log.
60 * Prevent simultaneous pioctl calls from multiple processes
61 from stepping on each other's data.
63 * Prevent simultaneous pioctl calls from crashing the
66 * Add data validation to pioctl output processing in
67 ktc_GetToken and ktc_ListTokens.
69 * Add data validation to pioctl output processing in
70 fs.exe and symlink.exe.
73 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
74 No longer will browsing \\AFS produce truncated share names.
76 * Registry specified server preferences did not result in
77 the CM_SERVERFLAG_PREF_SET flag being applied to the
80 * Avoid unnecessary DNS lookups of share names as cells
81 when it is known that the name cannot be a cell name.
82 Any name that does not contain a dot is skipped.
84 * When processing Dfs Referral requests, do not return an
85 error if the server is down or busy; if the volume is
86 offline; the cell vldb cannot be reached; etc.
87 These paths are still \\AFS paths and so the client
88 should not be sent to search elsewhere for resolution.
90 * CRITICAL: Prevent the cm_Daemon thread from terminating when
91 the machine enters the suspend state. IF the cm_Daemon thread
92 dies, there is nothing to execute down server checks.
94 * Prevent use of smb_StartedLock before initialization which
95 can result in a panic in the lock package.
97 * Improve error checking in afslogon.dll in order to prevent
98 NULL pointer dereferences within GetLogonDomainOptions()
99 if the Lsa operations fail.
101 * Add support to the build system to automatically update
102 a Microsoft Symbol Store.
104 * Do not access the cm_conn_t after executing the cm_Analyze
105 loops. Doing so can result in an access to memory that has
108 * CRITICAL: If a StoreData request offset is not aligned on the
109 buffer module blockSize, (offset % blockSize) bytes of dirty
110 data will not be written to the file server.
112 * CRITICAL: If a StoreData64 request is sent to a file server
113 that does not support large files (>2GB) and the client is
114 unaware that the server has no large file support, the first
115 chunksize worth of data will not be written to the file server.
116 This bug was introduced in 1.5.3. It affects all IBM AFS file
117 servers and all OpenAFS file servers older than 1.4.0.
120 * If a file server becomes inaccessible while the cache manager has
121 dirty buffers to write, the afsd_service buf_IncrSync thread can
122 attempt to use 100% of the cpu.
124 * Fix the locking that protects the cm_cell_t lists
125 (all, name, id, free).
127 * Remove cm_cell_t objects from the name and id hash tables before
128 placing them on the free list.
130 * Fix "fs newcell" which was broken in 1.5.60.
132 * Do not attempt to synchronize dirty buffers if the associated
133 volume is known to be unavailable.
135 * Log to the Windows Event Log dirty buffers that cannot be written
136 during the shutdown process.
138 * Modify behavior of a Freelance mountpoint target that does not
139 specify a cell. Instead of assuming the target volume is in the
140 Freelance.Local cell, use the workstation "Cell" specified in the
141 registry. A mountpoint target of "#root.cell." will now mean the
142 root.cell volume in the workstation cell for the current session.
143 If the workstation cell changes from "athena.mit.edu" to
144 "andrew.cmu.edu", the referenced volume will also change without
145 requiring that the mount point targets be altered.
147 * Avoid false infinite loop errors when validating the cm_cell_t
148 free list during startup.
150 * Fix two more locations where xdr_free() should be used instead of
151 free() to permit a checked built of afsd_service.exe to work with
152 release builds of afsrpc.dll.
154 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
155 to permit the server Uuid to be used to lookup the server object
156 and from that determine the cell. This permits callbacks that are
157 received from alternate addresses to be processed with a known server
158 object. Previously a request from an unknown server would clear all
159 callbacks from all cells.
161 * Fix a bug that prevented optimal performance when using a non-zero
162 value for 'daemonCheckVolCBInterval'. As a reminder, when
163 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
164 volume callbacks are automatically renewed 90 minutes before their
167 * Fix automatic ranking of vldb servers whose values are obtained from
170 * Add logging for RX CALL DEAD errors which are generated if the remote
173 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
174 or the call is to a vldb server.
176 * Add registry based cell search functionality to NetIdMgr,
177 afs_config.exe, and klog.exe.
179 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
180 lookups on the host names in the CellServDB instead of using the
181 specified IP addresses. This provides aklog, pts, vos, etc. the same
182 CellServDB behavior that the Windows Cache Manager uses.
184 * Completely a behind the scenes change. The snprintf() routines have
185 been completely rewritten.
187 * When updating the stat cache entry callback of a .readonly object
188 from the volume group object, update the file server reference to
189 ensure it matches the most update to date callback.
191 * Add proper support for processing callbacks from multi-homed file
192 servers. Instead of comparing servers by cm_server_t pointer,
193 compare them by UUID when the UUID is known.
195 * During a shutdown short circuit the offline volume check daemon
198 * Properly track link counts during cm_Link and cm_Unlink operations.
200 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
201 preference to an error code reported by rx_EndCall.
203 * Add "PerFileAccessCheck" registry value to permit testing against
204 experimental file servers that include per-file acl support. This
205 value is intentionally undocumented. It is not to be used by
206 production environment deployments.
208 * Log a warning to trace log if a volume group is marked as a DFS
209 File Set. The Windows OpenAFS client does not include DFS support.
211 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
212 provider from obtaining tokens when referrals are in play.
214 * Add "fs chown" and "fs chgrp" commands to permit the owner and
215 group of objects stored in AFS to be set from Windows.
217 * Avoid performing background daemon operations when the machine is
218 going into suspend mode.
220 * Perform offline volume checks in most recently used order.
222 * Prevent crash when a data version for a cache object goes backwards.
224 * fs setquota, fs setcachesize, vos setfields, and vos create now
225 accept human readable orders of magnitude. (K, M, G)
227 * fs listquota fixed to permit large quota sizes to be displayed.
229 * rx packet resend and data packets sent counts were incorrect.
231 * Multi-thread safe library versions are now being generated and used.
232 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
234 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
235 values are now available on XP through Windows 7. Add functionality
236 to autodetect if such support is present on the machine. If so,
237 configure it if necessary and dynamically adjust the AFS Rx timeout
240 * Correct another race condition in the Rx library that could result
241 in an unexpected panic while freeing the Rx call iovq.
244 * A fix to the pioctl library to support drive substitution
245 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
247 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
248 that corrects a deadlock in the smb redirector and also adds new
249 functionality that permits the AFS SMB server to be given a longer
250 timeout than is normally the case. New functionality has been added
251 to configure these additional LanmanWorkstation\Parameter values.
252 (This functionality has been backported to XP SP3 and is scheduled
253 to be released on June 5th.)
255 * The BackConnectionHostNames registry value configuration was broken
256 when dynamic re-establishment of Netbios Name registrations was added.
259 * Hidden vos commands are revealed.
261 * If the "DisableLoopbackCheck" registry value is set, do not unset
262 it during the same service session.
264 * Reorganize code that prevents multiple Store operations to the same
267 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
268 permit broken symlinks to be treated as being in AFS.
270 * Fix vos commands that output 64-bit integer values.
272 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
273 Permits building cygwin applications against OpenAFS libraries.
275 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
276 as Windows HTML Help (.CHM) files.
278 * NSIS installer does a much better job of cleaning up files left
279 over from previous installs.
281 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
282 or "fs flushall" and on-going directory operations that can result in
283 afsd_service.exe crashing.
285 * Add support for DNS SRV records in place of AFSDB records.Â
286 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
287 Weight is currently ignored.
289 * Add a method of specifying Client CellServDB information within the
290 registry that can be used to either override the CellServDB file or
291 force the use of DNS lookups for a given cell.
293 The registry schema is as follows:
295 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
296 "LinkedCell" REG_SZ "[cellname]"
297 "Description" REG_SZ "[comment]"
298 "ForceDNS" DWORD {0,1}
300 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
301 "HostName" REG_SZ "[hostname]" Default: [servername]
302 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
303 "IPv6Address" REG_SZ "[address]" <future>
304 "Comment" REG_SZ "[comment]"
305 "Rank" DWORD "0..65535" Default: 0
306 "Clone" DWORD "{0,1}" <future: server only>
307 "vlserver" DWORD "7003" <future>
308 "ptserver" DWORD ... <future>
310 ForceDNS is implied non-zero if there are no [servername]
311 keys under the [cellname] key. Otherwise, ForceDNS is zero.
312 If [servername] keys are specified and none of them evaluate
313 to a valid server configuration, the return code is success.
314 This prevents failover to the CellServDB file or DNS.
316 Only one of "HostName" or "IPv4Address" is required. "HostName"
317 is optional if [servername] is the host name.
319 * Extend registry based CellServDB functionality to
320 afsconf_GetCellInfo() interface used by aklog and many
321 other command line utilities.
323 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
324 as it is now used by both client and server components.
327 * PriorityClass of afsd_service.exe process raised to
328 "High" to match the priority of the system services
329 that are waiting for afsd_service.exe to respond.
330 Configurable using [TransarcAFSDaemon\Parameters]
331 DWORD "PriorityClass" value. (use values from Win32
334 * Change error reporting of CM_ERROR_BADSHARENAME to
335 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
336 The error we were sending would cause the smb
337 redirector to drop the connection.
339 * Change the error reporting of CM_ERROR_ALLOFFLINE
340 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
341 Previous we would return Bad Network Path which
342 would cause the smb redirector to cancel all file
343 handles on the same "share".
345 * A debug version of afsd_service.exe can now be used
346 with a release version of afsrpc.dll (or vice versa).
347 afsrpc.dll now implements and exports xdr_free().
349 * Removed all unprotected OutputDebugString() calls as
350 those calls can block for up to 10 seconds if there
351 is contention for the globally shared memory or if
352 the debugger fails to read data in a timely manner.
354 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
355 supported for builds. The production releases still
359 * Discovered that XP/2003 supports a variant on recvmsg
360 called WSARecvMsg. The input parameters are mostly
361 compatible with the Posix recvmsg. This will save
362 multiple memcpy() calls per received packet.
364 * Discovered that Vista/2008 supports a variant on
365 sendmsg called WSASendMsg. The input parameters
366 are mostly compatible with the Posix sendmsg. This
367 will save multiple memcpy() calls per sent packet.
369 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
370 and utilizing rx_ReadProc32 from within the cache manager
371 as part of the FetchData operations. rx_ReadProc32 and
372 rx_WriteProc32 are optimized versions of rx_ReadProc and
373 rx_WriteProc for use when reading 32-bit integer values.
375 * The Network Identity Manager AFS provider now reports
376 progress events. These provide a better user experience
377 with Network Identity Manager 2.0.
379 * [RT 124293] A race condition exists which permits the
380 scp field of the an smb_fid_t object to become invalid
381 after a request on the smb_fid_t has begun. Remove the
382 race by protecting all access to the scp field with the
383 smb_fid_t mx mutex and obtaining a local reference on
384 the cm_scache_t object for the length of the request.
386 * [RT 124276] If the vldb is out of sync with the contents
387 of the file servers, afsd_service will retry too many
388 times when a file server reports a volume as not being
389 present. Now if the list reported by the vldb is the
390 same as the previously seen list, then the retry is
393 * [RT 124276] Read-only volume failover was broken in
394 1.5.53 whenever accessing a volume results in VNOVOL or
397 * [RT 124276] Prior to 1.3.70 the volume server reference
398 list was not reference counted and would be prematurely
399 freed while in use. When reference counting was added
400 in 1.3.70 a bug was introduced that could result in
401 service reference list corruption.
403 * Add Windows Application Event Log warning messages for
404 "Client SMB MPX value too large" and
405 "Client SMB Buffer Size too small".
407 * Renaming of files across directory boundaries would result
408 in an invalid handle error when attempting to access the
409 files after the move.
411 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
412 when a smb file descriptor cannot be found. The whole point
413 of using a Path Info function is because an smb file
414 descriptor wasn't allocated.
416 * More edge cases in which dynamic addition of Freelance
417 root.afs entries would get the wrong FID or where the
418 root.afs directory would not be refreshed.
420 * Buffer overflow could occur if the workstation cell name was
421 longer than 64 characters. Crashes could occur in
422 afscreds.exe, afslogon.dll, and afsd_service.exe.
424 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
427 * Improve consistency of the bad data version value used for
430 * Log file server uuid values as part of the cm_server object
431 when available. Dump the cm_server object list in response
434 * Optimize the performance of resetting access control lists
435 when tokens are set or removed.
437 * Fix vos.exe so that it can work with very large volume
440 * Remove symlink recursion tests and increase max symlink count
444 * Undo the "UAC manifest fix" applied to afs_config.exe.
446 * Fixed some additional race conditions in the Rx library
448 * Ensure that Freelance allocation of vnodes follow
449 the AFS convention of odd vnodes are directories
450 and everything else is an even vnode.
452 * Add Freelance logic to mount point and symlink
453 evaluation functions.
455 * Enhance smb_ParseASCIIBlock() so that it can handle
456 all of the STRING formats defined by the CIFS Technical
459 * Validate the output of smb_ParseASCIIBlock() in all
460 callers. Return CM_ERROR_BADSMB if the STRING field
461 cannot be parsed. CM_ERROR_BADSMB will cause the
462 contents of the packet to be logged.
464 * If multiple SMB Raw Write operations were taking
465 place at the same time, there could be data corruption
466 because unique event objects were not generated for
467 each Netbios receive operation.
471 * Fix a race in rxi_WriteProc which could result
472 in a freed or in use packet being passed to
473 rxi_PrepareSendPacket if the call is currently
474 in a transmit wait state. Passing a freed packet
475 will result in a panic. Passing an in use packet
476 might result in a call failure or perhaps data
479 * Kerberos referrals broke the NIM AFS provider
480 logic for automated configuration of the workstation
481 cell with a new identity. Now if a referrals
482 response is returned, assume the cell belongs to the
485 * The interface to buf_CleanAsync() and
486 buf_CleanAsyncLocked() has been changed to permit
487 errors to be propagated to the caller.
489 * During an enumeration of a directory, if the btree
490 has not yet been generated, return an error.
492 * When mapping CM_ERROR values to NTSTATUS values
493 do not treat 0 as an error.
495 * Avoid a cm_scache_t object reference count underflow
496 when renaming via the SMB interface.
498 * smb_IsMask() can be called with a non-8.3 input buffer.
499 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
500 input buffers and smb_IsMask() for non-8.3 input buffers.
502 * Add cm_buf_t error state to the "fs memdump" output.
504 * pioctl drive mapping checks need to take into account
505 Global Drive Mappings and SUBST assignments.
507 * Rename operations should not be permitted if the
508 destination name collides with an existing name that
509 differs only by case UNLESS the collision is the object
510 that is being renamed.
512 * Increase the maximum number of background daemon threads
513 to 64. Leave the default at 4.
515 * Set the FILE_UNICODE_ON_DISK flag when reporting the
516 file system characteristics.
518 * Report the maximum component name length as 255 even
519 for Unicode as part of the file system characteristics.
521 * Add RX debugging information to "fs memdump" output
524 * Refcount the cm_server_t objects returned by
527 * Fix the UAC manifest applied to afs_config.exe
529 * rxdebug now indicates if any RX_CONN flag is set
530 not just some of them
532 * Change the mapping of CM_ERROR_WOULDBLOCK from
533 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
537 * Includes #GCO Public CellServDB 07 Nov 2008
539 * Prevent separate cell entries from being created for
540 "foobar.org" and "foobar.org." which can happen if
541 an AFSDB record is found instead of a CellServDB
544 * Fix a panic when checking the SMB virtual circuit list.
545 If a VC is destroyed while the list is being checked
546 it is possible for the next object to become invalid.
548 * Increased logging to afsd_init.log of netbios name
551 * Fix a DNLC panic when a UTF-16 string that cannot be
552 converted to a utf-8 string is provided as a filename.
554 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
555 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
556 location to refer directly to the new location in
557 %programfiles%\openafs\client\program
559 * Be more aggressive about responding to Netbios
560 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
561 the LANA within the listener thread without performing a
564 * Use the WNet Enumeration of all connected drives as a filter
565 in the pioctl GetIoctlHandle operation. This should remove
566 the last situations in which there was a delay when opening
567 an Explorer Shell properties menu with the AFS Shell
570 * Fix the parsing of CellServDB in Windows to handle linked
571 cells as is done on Unix.
573 * Add support for linked cells to the cache manager,
574 netidmgr afs provider, and aklog. When two cells are
575 linked, a volume lookup that fails in one cell is looked
576 up in the other. This can be used for construction of
577 test cells that override specific volumes; cell renaming;
580 * Increase the length of cell and realm names that can
581 be entered in the NIM AFS provider.
583 * Actively prevent a 32-bit afsd_service.exe from
584 executing in the WOW64 environment.
586 * Protect against invalid IP address values in CellServDB
591 * Fix an Rx panic caused by an attempt to double free
592 an rx_packet when an error occurs while writing the
595 * Fix an exception cause by a race condition when
596 processing multiple file change notifications registered
597 by the Explorer Shell or other applications.
599 * Modify the Rx debug interface to use var_args and
600 permit Rx debugging to be used on 64-bit Windows
601 which previously had it disabled.
605 * Automated lock hierarchy enforcement added to the
606 osi base lock package. This permits detection of lock
607 violations that might but often does not result in
608 deadlocks. Enabled by default in the checked builds.
609 May be activated in release builds via the registry
610 value "LockOrderValidation" in the TransarcAFSDaemon
613 When a violation is detection an assertion is thrown
614 generating a dump file and a message indicating which
615 locks were obtained out of order is logged to the
616 Windows Application Event Log.
618 * Attempt to ensure that a dump file will be generated
619 whenever the service panics.
621 * Removed unused lock cm_bufGetMutex
623 * Many out of order lock acquisitions were detected
624 using the lock order verifier and were corrected.
626 * A race condition within the function path cm_GetSCache() ->
627 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
628 cm_scache_t object to be issued simultaneously to two
629 threads. This would eventually result in a panic due
630 to the resulting undercount.
632 * Fix interpretation of the empty string as the ioctl path
633 to mean the current directory. "fs lsm", "symlink list",
634 etc. now return a "not a ..." error instead of "does not
637 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
638 to indicate that the input string is a nul terminated string.
639 Assign ti when input strings are nul terminated instead of
640 length counted. (from 1.5.50)
642 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
643 does not match the observed behavor. The 'nameLength' is specified
644 in bytes, not in characters. Fix the implementation to match
645 the observed behavior. (1.5.50)
647 * The cm_HaveAccessRights() attempt at deadlock avoidance by
648 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
649 does not avoid the deadlock. Avoid the deadlock by enforcing the
650 lock order of the lowest vnode first. Then remove the infinite
651 loop avoidance in cm_SyncOp() that was returning an unwarranted
654 * File Server lock synchronization was not propertly enforced in
655 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
657 * Another change to the short name generator. Plus and equal are
658 not valid characters in a short name according to the
659 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
660 with left and right brace in order to minimize the impact.
662 * Fix a race condition that can result in a heap overwrite error if
663 the number of known file servers or vldb servers increases during
664 the preparation phase of a cm_CheckServers() operation.
666 * If the SMB Authentication would fail, a LSA allocated memory
667 buffer would be leaked.
669 * Rewrite the cm_cell_t allocation code to permit cell vldb server
670 lookups to fail without leaking cm_cell_t objects.
672 * Add "fs setcrypt auth" to permit the use of rxkad authentication
673 only mode. Not documented.
675 * Initialize Rx mutex and lock objects once and only once as part of
676 the DLL load process.
678 * When reporting the Volume Name "AFS" do not include the terminating
679 NUL in the reported length.
681 * Refresh Volume Location information upon receipt of a VNOVOL error
682 from the file server and avoid a race condition that prevented the
683 updated volume location information from being destroyed immediately
684 after it was acquired.
686 * In response to VNOVOL and VMOVED, remove the volume id from the server's
687 volume list so that the volume's up and down status is not affected
688 when the machine is shutdown.
690 * Many fixes to the Rx stack related to packet allocation and deallocation;
691 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
692 by overwriting the rx_call currentPacket pointer when it was already
693 referring to a packet.
695 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
696 in the TransarcAFSDeamon/Parameters key.
698 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
699 constructed directory B+ trees for flushed directory objects.
701 * When cached pages for directory objects have been locally modified, do
702 not permit the locally modified pages to be mixed with pages subsequently
703 obtained from the file server. Doing so can result in directory
706 * If Ubix errors are received in response to VL_xxx operations record
707 the error on the server instance and permit the operation request to
708 be retried in case another server is able to answer the query. We have
709 seen at least one instance of a cell in which a VL server returned
710 UNOQUORUM (possible database truncation) when the other servers were
711 responding successfully.
714 * Optimize the assignment of write errors to dirty
715 buffers by applying the first error to all buffers
716 belonging to the same object instead of trying each
719 * Add resource version info to xstat_fs_test and xstat_cm_test
721 * Fix a delete vs delete[] usage problem in the WiX custom
724 * When creating a mount point, validate the volume name to
727 * When processing a mount point, pay attention to VL_BADNAME
730 * Remove a race condition from cm_volume_t recycling that could
731 have resulted in two threads making use of the same object at
734 * Fix a reference count leak on smb_vc_t objects introduced in
735 1.5.50. Permit the objects to be freed once again.
737 * Fix the output of "fs memdump" for smb_vc_t objects that was
740 * Add routines for debugging smb refcount errors.
742 * When shutting down the cache manager, do not permit AFS
743 callback requests to be processed. Doing so can result in
744 invalid memory accesses.
746 * Do not recycle a deleted cm_scache_t object unless the
747 refCount is 0. Doing so can result in a deadlock and
748 invalid memory access.
750 * Minimize the number of RPCs sent to the file server for
751 directory bulk status requests. Only request status for
752 objects that we are actually reporting on and only if we
753 do not already have status for the object and then make
754 sure that we return as much as we can at once via the SMB
755 interface but not so much that it exceeds the RDR timeout
757 * Enable bos restricted operations
759 * Fix the creation of submounts used by the afscreds and afs_config
762 * Avoid repeated retries when attempting to obtain access rights.
763 If the access rights were obtained successfully but do not
764 satisfy the request, fail the request.
766 * Fix error code reported when attempting delete a file on a
767 readonly volume or one that is marked with the readonly DOS
770 * Fix a short name truncation bug.
772 * Fix a heap corruption error when reading the CellServDB file
775 * Fix directory listings when the directory has "lookup" permission
776 only and the directory has never been previously accessed.
778 * Fix several tracking issues for cm_scache_t rw locks.
780 * Ensure that all buffers used to store cell names are the
783 * Restore Windows 2000 support (broken in 1.5.50).
784 LOCALE_INVARIANT was introduced in XP.
786 * Add RxUdpBufSize registry value. Default is 256K.
788 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
790 * Do not include trailing NULs when returning file names in
791 a directory search response.
793 * Prevent RX from crashing the afsd_service by attempting
794 to allocate to many buffers at once when memory is fragmented.
796 * Avoid a deadlock when freeing cm_server_t objects
798 * Optimize search cell file. Do not search to end of file
799 if the cell has already been found.
801 * Remove explorer shell extension delay on local disks.
803 * Automatically pre-allocate 64 Rx packets per thread
805 * Add smb lock requests and stat cache lock allocations to
806 the output from fs memdump.
809 * Fix Large File Support (broken in 1.5.50)
811 * Fix "fs setacl" (broken in 1.5.50)
813 * Fix pioctl path evaluation of drive letter
814 mapped paths. (broken in 1.5.50)
816 * Fix installation of VC8 Runtime library in the EXE
817 Retail installer. (broken in 1.5.50)
819 * Avoid a potential deadlock during Volume Location
820 updates if all the file servers are unreachable at
821 the time the VL update succeeds.
823 * Ensure that rx connection object is references
826 * Fix propagation of over quota errors during cm_FSync()
829 Since 1.5.39 [1.5.50 released 16 July 2008]
830 * There were no releases between 1.5.39 and 1.5.50.
832 * FIRST RELEASE WITH UNICODE SUPPORT
834 * Pioctl interfaces to the cache manager refactored
835 to provide layering between the SMB specific code
836 and the general purpose ioctl operation.
838 * Garbage collect dead SMB virtual circuits as soon
839 as they are no longer being referenced. This avoids
840 problems with outstanding locks not being dropped
841 when the virtual circuit becomes invalid.
843 * Remove the IBM Administration Reference documentation
844 and replace it with the OpenAFS Command Reference
847 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
848 functions each time a connection is about to be used.
849 Do not hold a lock on the rx connection object while
850 it is being selected. This avoids a race between
851 threads attempting to set the timeout values and
852 removes a bottleneck that was hampering performance.
854 * Ensure that the smb directory attribute is set
855 for all directory objects.
857 * Add vs2008 support to the NSIS installer scripts
859 * Replace the VC Runtime EXE installer with the MSI
860 installer in the NSIS installer scripts
862 * Properly delete the VC7.1 C Runtime libraries.
864 Since 1.5.36 [1.5.39 released 23 June 2008]
865 * There were no 1.5.37 or 1.5.38 releases for Windows
867 * Use Visual Studio version of vsprintf() instead of
868 home grown version that resulted in linking conflicts
870 * Prevent buserver, ptserver and fileserver from crashing
871 due to non-portable strftime() format strings.
873 * Prevent server crashes caused by rx_InitMorePackets()
874 being called prior to allocation of all of the required
875 synchronization objects.
877 * Fixed a memory leak in the internal B+ tree directory.
878 For objects with non-8.3 names, the 'longname' memory
879 allocation would be leaked during object deletion or
882 * Fixed a memory leak in the background daemon cm_CheckServer()
883 routine. The memory allocated to store the server list
886 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
888 * Correct an abstraction layer violation. cm_scache_t objects
889 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
890 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
891 performed in cm_ReleaseSCache() when the reference count hits zero.
893 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
895 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
898 * Ensure that scache object 'nextp' pointers are set to NULL
899 when the objects are removed from the hash table linked list.
901 * If a dirty buffer cannot be written to the file server because
902 of an access denied error, mark the error in the cm_buf_t
903 structure so that it can be dealt with instead of retrying
907 * Update CellServDB file to GCO Public 23 Apr 2008
909 * Fix a cm_buf_t reference count leak when attempts to write
910 dirty buffers to the file server from within cm_IncrSyncer()
913 * Prevent udebug from crashing.
915 * Activate RX Idle Timeouts. If the file server is busy for more
916 then 30 seconds, attempt to failover to another server without
917 marking the busy server down.
919 * Another VNOVNODE issue fixed. When writing a dirty buffer
920 to the file server, if VNOVNODE is received, mark all buffers
921 as invalid without further attempts to contact the file server.
923 * Improved performance on high latency links.
926 * Fix a bug preventing the re-initialization of the Freelance
927 root.cell contents when a change is made.
929 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
930 or mountpoint that does not exist.
932 * aklog, afscreds, integrated logon, and the NIM AFS credential
933 provider will all attempt to first acquire an afs service ticket
934 using the service principal: afs/<cell>@<CLIENT_REALM>. This
935 permits the proper detection of the cell realm when multiple
936 realms are local realms for the cell.
938 * Fix fs commands that re-use file identifiers instead of forcing
939 a path evaluation each time. This was broken in 1.5.33 when
940 hashes were added to the cm_fid_t structure.
942 * Fix symlink and mountpoint removal and then recreation when
943 applied to the fake Freelance root.afs volume.
945 * Fix SMB change notifications for the Freelance root.afs volume
946 when creating or removing symlinks and mount points.
948 * Force DNS AFSDB cell name lookups to lower case if the DNS
949 response is not all lower case.
951 * 32-bit installers are now built with Visual Studio 2005
953 * All vos.exe commands now support the "-noresolve" option which
954 forces the display of IP addresses instead of DNS hostnames.
956 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
957 permits operations that do not require up to date volume
958 location data to continue while waiting for the RPC to complete.
959 Multiple threads waiting for a volume location update on the
960 same volume set no longer trigger multiple calls VL_xxx RPCs.
962 * Volume server references could be marked offline and never
963 reset to the not_busy state.
965 * Further optimizations to cm_Analyze() retry logic for
966 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
969 * Optimize the DNLC by applying Interlocked incrementing for
970 statistics gathering and enabling greater use of read
971 locks instead of write locks
973 * Further optimize osi_Log operations when logging is disabled
974 by removing the need for the function call overhead.
976 * Add a new lock operation, Convert Read to Write, and apply
977 it throughout the cache manager. This function provides a
978 fast transition from a read lock to a write lock when the
979 caller is the only reader.
981 * Fix a timeout problem when opening files or request locks
982 when the file in question is already write-locked by another
985 * Fix a deadlock that can occur if "fs flushall" is executed
986 while applications are writing to AFS.
988 * The cache manager would re-read data from the file server
989 after file truncation when obtaining buffers to write to the
990 same file until the first StoreData RPC completes.
992 * Fixes error handling in cm_NTCheckDelete()
994 * Replaces cm_scache_t mutex with a read-write lock permitting
995 additional parallel access.
997 * Prevent an cm_scache_t reference undercount when evaluating
998 symlinks containing @sys when none of the active sysnames
999 match any of the available targets.
1001 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1004 * Set RxMaxMTU registry value to 0 which is the equivalent of
1005 it not being set at all.
1007 * Reduce contention between BkgDaemon threads
1009 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1012 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1013 objects to use the faster Interlocked operations which avoid
1014 the need to hold write locks.
1016 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1017 non-fatal error that forces a retry against another service
1020 * Fix a crash in smb_WriteData that would occur if an application
1021 opened the file as read-only and then sent a write request anyway.
1024 * Respond to VL_NOENT errors by removing cm_volume_t object and
1025 setting it to be the first object to be recycled. This is not
1026 only a more efficient use of resources but it also prevents
1027 repeated attempts to query the VLDB server for the volume from
1028 the CheckOfflineVolumes routine.
1030 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1033 * A first cut at a cache manager statistics monitor that can be used
1034 to determine the necessary cache parameters to support the
1035 application working set.
1037 Off by default, the performance package can be activated by setting
1038 "daemonPerformanceTuningInterval" in the service Parameters key.
1039 As with the other daemon interval values the unit is in seconds.
1041 At service start and each succeeding interval the cache manager
1042 will write statistics to %TEMP%\afsd_performance.txt showing the
1043 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1044 The FID statistics keep track of all FIDs seen by the cache manager
1045 during the service session whether or not they are backed by any
1046 live object in the cache.
1048 These statistics are not stored in the cache file.
1050 * Fix windows event logging broken in 1.5.33.
1052 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1053 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1054 as well as when computing the server preference list.
1056 * When merging status after a StoreData operation prevent data
1057 buffers from being discarded that we would prefer to keep.
1058 (Bug introduced in 1.5.33.)
1060 * Modify installers the set RxMaxMTU to 0.
1062 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1063 dialog, the Background Daemons and Service Threads fields were
1066 * Increase default number of Background Daemons to 4.
1068 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1069 cm_volume_t objects to be recycled.
1071 * Windows Error Reporting has shown a number of previously unseen
1072 stack corruption errors in the Explorer Shell extension. The
1073 minidump appeared to indicate the problem was in ParseAcl().
1074 Reviewing the function it appears that numerous things could go
1075 wrong if invalid input was provided. Added error checking and
1076 data validation. Hopefully this will address the problem.
1078 * Windows Error Reporting has shown crashes in aklog as a result
1079 of krb5_cc_get_default() failures not being handled. They are
1083 * The Rx library used a 32-bit type for sockets which was
1084 truncating the socket value on 64-bit Windows. This
1085 was not an issue in general but became an issue when
1086 Rx "hot thread" support was enabled.
1088 * Enable "Rx hot thread" support which ensures that while
1089 a received message is being processed that a new thread
1090 is allocated to receive the next message to arrive.
1092 * Increased parallel access to smb_rctLock protected
1093 resources. Reduce the number of times that the
1094 lock is dropped and re-obtained.
1096 * Increased parallel access to cm_scache_t and cm_buf_t
1097 resources by replacing the previous global write lock
1098 protected reference counting mechanism with one built
1099 upon the use of InterlockedIncrement and InterlockedDecrement.
1101 * Implement a hash for cm_fid_t to reduce comparison costs
1102 and inline the cm_FidCmp() function.
1104 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1105 hash which has a better distribution
1107 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1108 the most recent read request instead of the next chunkSize.
1109 cm_chunkSize can be dozens or hundreds of buffers. As a result
1110 too much time is spent performing the evaluation.
1112 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1113 lock is to prevent the creation of new buffers while a truncation
1114 is being performed. All references to bufCreateLock have been
1115 removed except in two places:
1116 i. a write-lock surrounding the function that calls buf_Truncate()
1117 ii. a read-lock within buf_GetNewLocked() that actually allocates
1120 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1121 CM_CONFIG_DATA_VERSION number which value be used to force the
1122 replacement of the cache file contents when incompatible changes
1123 are made between releases.
1125 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1126 cm_SyncOpCheckContinue
1128 * Avoid calls to multi_Rx if nconns == 0
1130 * Modify smb_WriteData to perform background writes based upon the
1131 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1132 boundaries. This will slow down writes from the SMB interface but
1133 will avoid the risk of the CIFS client disconnecting from the AFS
1134 client SMB server. The default value for 'smb_AsyncStoreSize'
1137 * SMB AsyncStore functionality is on by default but can be
1138 disabled using the registry value:
1140 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1142 smb_AsyncStoreSize can be adjusted with
1144 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1146 The value must be a multiple of the buffer block size
1147 and cannot be larger than the chunk size.
1149 * Do not leak a read lock on the bufCreateLock
1151 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1154 * In MergeStatus, only remove clean buffers from the hash table.
1156 * Avoid the need to update the data version number on each buffer
1157 associated with a scache when MergeStatus is called after a
1158 StoreData by maintaining a range of valid data versions as part
1159 of the cm_scache_t object.
1161 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1162 from cm_HaveAccess... We can safely avoid the call so do so.
1164 * Add "ForceAnonVLDB" service parameters registry value. Default
1165 is 0. Set to 1 to force the use of anonymous connections to
1166 the VLDB servers. This is useful when using krb5 tokens and the
1167 VLDB servers have not been updated.
1170 * Do not attempt to search for a cell name that matches one
1171 of the special share names "svrsvc", "wkssvc", "ipc$"
1173 * Issue permission denied warnings from afscreds if attempts
1174 to modify the startup shortcut or registry values fail.
1175 This happens under Vista UAC.
1177 * Do not return BAD_NETWORK_ERROR in response to attempts
1178 to create directories or files in the Freelance root.afs
1179 volume. Instead return ACCESS_DENIED.
1181 BAD_NETWORK_ERROR was being sent because the Freelance cell
1182 has no vldb servers. Short circuit the error by testing for
1183 freelance in CreateFile and MakeDir
1185 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1186 to be opened and closed without actually existing. For now
1187 we treat them as equivalent to ioctls but this should
1190 * Fix RAP processing to not reject names that exist in the
1191 root.afs volume for GetInfo queries.
1193 * In afscreds, use GetUserNameEx to obtain the username instead
1194 of searching the registry. Try UserPrincipalName first and
1195 fallback to SamCompatible.
1197 * Stop linking to rpcns4.dll which is no longer supported on
1198 Vista and 2008. TaAfsAdmSvr is going to have to be
1201 * When processing dfs referral requests, if the sharename is
1202 a partial match and therefore will not be added automatically
1203 to the share list, make sure that we return an error.
1205 * Avoid a crash when logging VNOVOL errors returned from
1206 cm_CheckOfflineVolume.
1208 * When trace logging is disabled, do not permit osi_LogSaveString
1209 to perform the string copying. By disabling the string
1210 copies, throughput of the service more than doubles.
1212 * when checking offline volumes, update the cell vlserver info
1213 if required and don't perform a check if there are no servers
1216 * Add volumes in the vl_unknown state to the list of objects to be
1217 checked as offline or retried when an allbusy error is received
1220 * Add support for configuring rx_ExtraPackets
1222 * All calls to buf_Get() must be protected by a read lock on the
1223 cm_scache_t bufCreateLock
1225 * When a MergeStatus results in the invalidation of buffers, do
1226 not remove buffers from the hash tables that have active references.
1228 * Upon receiving a VNOVNODE error from a file server, be sure to
1229 invalidate any file locks. This prevents on-going attempts to
1230 extend the locks and will cause subsequent attempts to access
1231 the file to receive a invalid handle error.
1233 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1234 to directory entry names.
1236 * No longer discard tokens when RXKADBADTICKET error is received.
1237 Instead return STATUS_NO_KERB_KEY error.
1240 * Add "cmdebug <host> -cellservdb" option and client service
1241 implementation permits querying the vldb server for each
1242 cell known to the cache manager. The output is in a form
1243 that permits it to be used as a CellServDB file.
1245 * Add Symlinks->Show option to the AFS Explorer Shell Context
1248 * Prevent buffer overrun in the cmd parser used by command
1249 line tools such as fs, vos, pts, etc.
1251 * The path ioctl operations have several issues:
1253 (1) the specified path for "fs examine, whereis, whichcell,
1254 flush" and so always has the follow symlinks and mount
1255 points semantics. This makes it impossible to determine
1256 what the FID of a symlink or mount point is.
1258 (2) "fs examine" out is not the result of a single pioctl
1259 operation but is actually the combined output of half
1260 a dozen operations. Path evaluation is an expensive
1261 operation. It would be faster if the caller could
1262 evaluate the FID first and then perform all of the
1263 rest of the operations by specifying the FID instead
1266 (3) fs output reports all objects as files. By adding a
1267 GetFileType pioctl more informative output can be provided
1268 that indicates what type of object the path evaluates to.
1270 (4) the Windows fs command includes a number of commands that
1271 do nothing but exist only because the Unix cache manager
1274 A new extendible data structure cm_ioctl_query_opts_t has been
1275 added which can be optionally specified with pioctls that do not
1276 already require input data. The first two fields of this structure
1277 are 'literal' and 'fid'. The literal field is used to indicate
1278 whether the last component of the path should be evaluated following
1279 symlinks and mount points.
1281 The fid field permits a fid to be specified.
1283 * A new GetFileType pioctl has been added. The type of objects are
1286 * A new "-literal" option is available for "fs examine, flush, whereis,
1289 * Unimplemented fs commands have been removed.
1291 * There are circumstances where a volume object is being accessed and
1292 the volume is marked indicating that the volume location information
1293 is out of date but where it is also pointless and perhaps dangerous
1294 to block waiting for the rpc to complete. One example is when
1295 processing the cmdebug requests. If we know that we are not going
1296 to use the volume object to contact a server, then we can now set
1297 the CM_GETVOL_FLAG_NO_RESET flag.
1299 * when moving up in the directory tree we search the recorded fid list
1300 to find a matching fid that we have already crossed. we must also
1301 reset the fid count based upon what we discover.
1303 * Cell names are published as share names. As a result they are searched
1304 for as part of the Dfs Referral evaluation. If share "foo" can not be
1305 reached, the CIFS client will fallback to searching for "fo" as well.
1307 Since the freelance client automatically adds a symlink for prefixes
1308 of the cell name, this results in both "foo" and "fo" being added to
1309 the freelance root.afs volume. This patch prevents that negative side
1312 * when tracking Kerberos credential cache names be sure to include
1313 the cache type prefix.
1315 * Microsoft has assigned OpenAFS a network type value
1317 * The algorithms that were used to produce a Netbios name were broken.
1318 The name that was produced when the hostname was too long was in fact
1319 longer than the maximum permitted netbios name.
1321 Also, the "NetbiosName" value was not used as a suffix when the loopback
1322 adapter was not installed. The hardcoded string "AFS" was used instead.
1324 * avoid another deadlock during server probes initiated by the ipaddr
1325 change daemon thread.
1327 * prevent the afs client service from crashing when fs newcell is
1330 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1331 to \Program Files\OpenAFS\Client\Program in order to get them out
1332 of the %windir% tree
1334 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1335 end STATUS_OUT_OF_QUOTA can be returned to the application.
1337 * Modify the search order for determining the location of CellServDB
1338 and other client configuration files.
1340 1. AFSCONF environment variable
1342 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1343 4. use Program Files\OpenAFS\Client
1345 * More Vista Power Management tweaks
1347 * Close a socket handle leak in the kauth routines
1349 * Don't hold a cm_volume_t mutex across RPCs
1351 * Convert downTimes to use clock time instead of relative times.
1352 This permits correct comparisons with expiration times.
1354 * Update the Wix installer to properly install the prerequisites
1355 for translate_et.exe
1357 * Do not translate pioctl paths as they are not converted by the
1360 * Implement multi_rx probes for checking whether servers are up
1361 or down. Turns probes into a constant time activity instead
1362 of one dependent upon the number of known servers.
1364 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1365 update of the volume location data and retry. VNOVOL means
1366 that the volume could not be attached. It doesn't mean that
1367 the volume is not on this server.
1369 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1370 not attempt to validate the target path. There is no reliable
1373 * Improve performance of \\afs\<cell><mount-type><volume>\
1374 access by eliminating extraneous attempts to search the
1375 CellServDB file and DNS AFSDB records. Share names
1376 containing '%' or '#' are not valid cell names.
1378 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1379 to unlock a range that is not currently locked.
1381 * Enable the 32-bit tools to read the 64-bit registry keys
1382 when executing in the WOW64 environment.
1386 * Fix buffer refcount leak introduced in 1.5.29
1388 * Prevent the NIM AFS Provider from crashing when configured to use
1389 Kerberos v4 for token acquisition in situations where no Kerberos v4
1390 support is available. For example, 64-bit KFW or sites that delete
1391 the Kerberos v4 DLL.
1393 * In afsd_init.log, support logging cache sizes greater than 2GB.
1395 * The volume status output of "fs examine" was being determined based
1396 upon the wrong error value. Instead of using the pioctl() return
1397 code, the errno value must be used.
1399 * Prevent invalid pts auto-registration attempts by aklog and the
1400 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1402 * Add synchronization protection to all of the SMB Listener State
1403 variables. This prevents race conditions when the addition or
1404 loss of a network adapter takes place.
1406 * On Vista, there is a race condition between the restart of the
1407 afs client service and the network adapters. Ensure that the
1408 afs client service can handle switching between loopback and
1409 non-loopback modes of operation.
1411 * Fix FollowBackupPath to work with mount points that are not
1412 explicitly read/write.
1414 * Fix a deadlock that has been observed on Vista while resuming
1415 after at least four hours of sleep.
1417 * Add new HKLM registry configuration options that can be used
1418 to provide configuration hints to the NIM AFS Provider when
1419 creating new identities. See release notes for details on
1420 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1423 * Speed up write buffer operations by holding a global read-lock
1424 instead of a write-lock when updating the buffer data version
1425 number. The data version is protect by the associated stat
1428 * Speed up read buffer operations by removing out of date buffers
1429 from the buffer hash tables.
1431 * Protect against a null pointer dereference in afslogon.dll
1432 if QueryAdHomePathFromSid() fails.
1434 * Remove a race condition while accessing private freelance data
1437 * Update the buffer data versions when Freelance directory data
1440 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1443 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1444 instead of returning zero length.
1446 * Return ERROR_RETRY when a server sends VBUSY.
1448 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1449 a volstat plug-in is in use.
1451 * Enhance the volstat interface to support Dfs path mappings.
1453 * Fix "symlink list" to do the right thing for Dfs links
1455 * Fix callback expiration that was broken in 1.5.27.
1457 * Fix a crash at shutdown caused by the lack of synchronization
1458 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1460 * Hold the cm_server_t mutex while walking or modifying the
1461 cm_server_vols_t list in order to avoid a race condition.
1463 * Ensure that the cm_server_t refCount is safely modified.
1465 * The pthread library implementations of pthread_cond_wait
1466 and pthread_cond_timedwait were failing to return with
1467 the mutex held when error conditions occurred. This could
1468 result in race conditions wherever they are used. One case
1469 in which problems were seen was the Rx event_handler().
1471 * Prevent cm_GetBuffer() from locking all buffers associated with
1472 a stat cache object before deciding if the current buffer is up
1475 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1476 object for each offset that we will attempt to reserve.
1478 * Reduce the default number of background daemon threads used to
1479 prefetch or store to one in order to prevent two or more threads
1480 from stepping on each other.
1482 * Reduce the default chunk size to 256KB in order to prevent read
1483 timeouts on low bandwidth connections.
1485 * Fix a race condition that could result in threads becoming stranded
1486 when synchronizing multiple requests on the same stat cache object.
1489 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1490 has been disabled. It has been disabled because it turns out
1491 that the RPC is improperly implemented in all file servers
1492 previous to 1.4.6. The RPC is executed without the proper
1493 locks being held resulting in data corruption and eventual file
1494 server failure. The functionality will be re-enabled in a
1497 * a couple of .readonly Volume Callback optimizations were
1499 - Apply most recent volume callback to all stat cache objects
1500 in the volume. This avoids premature callback expirations
1501 and unnecessary FetchStatus calls.
1502 - Add option to permit attempts to renew .readonly Volume
1503 Callbacks on a periodic basis to avoid invalidation of stat
1504 cache data. [Requires daemonCheckVolCBInterval be set to
1505 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1507 * Adds support for 64-bit data version values. Previous releases
1508 only supported 32-bit values. Larger values sent by the file server
1509 would result in rollover.
1511 * aklog and asetkey no longer generate an exception if Kerberos for
1512 Windows is not installed.
1514 * If the LSA authentication functions report a temporary out of memory
1515 condition, fail the current authentication and do not cause the AFS
1516 client service to terminate unexpectedly.
1518 * When the AFS client service is shutting down, prevent the daemon check
1519 thread from executing additional checks which might attempt to access
1520 discarded locks or unmapped memory.
1522 * Correct a double free condition in the NIM AFS credentials provider.
1524 * Add "FollowBackupPath" service parameter registry value.
1525 This provides equivalent functionality to the UNIX afsd -backuptree
1526 option. When set to a non-zero value, normal mount points originating
1527 in a .backup volume will prefer a .backup volume.
1530 * Fix NSIS uninstall from the .MSI installer.
1532 * Add Stack Frame logging for AMD64
1534 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1535 server returns EWOULDBLOCK, retry the request every two
1536 seconds for up to the RDRtimeout.
1538 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1539 is not assigned the same value as EIO.
1541 * Update CellServDB to GCO Public 25 Oct 2007
1543 * Add additional validation and error handling code after
1544 each call to getSlot(). If an invalid slot is returned,
1545 return NONODE. If the invalid slot is returned when
1546 extracting a data node, invalidate the tree.
1548 * Modify compareKeys() to always perform a case-insensitive
1549 comparison and only perform a case sensistive comparison
1550 if the case-insensitive one matches. This ensures the
1551 ordering is consistently reported.
1553 * Add lock assertions to ensure that all calls are being
1554 performed with the correct locks being held. There have been
1555 some crash reports that provide stack data that does not appear
1556 to be possible unless there is a race. However, there are no
1557 obvious locations where the race is taking place and the test
1558 suite indicates that all of the correct locks are being held.
1559 We shall see what happens in the field.
1561 * For consistency replace all calls to findKey in which the range i
1562 s (1,numentries) with calls to getSlot().
1564 * Optimize the depth search loop by testing the slot value in the
1565 for statement instead of forcing the loop to be broken later.
1567 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1568 There are a number of locations where locks are obtained, dropped,
1569 and reobtained. This reorganization attempts to accomplish several
1572 (1) be optimistic for the most common case so it will be fast
1574 (2) add consistency checks after each location where locks are
1575 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1576 and the bplus tree is out of date, retry until we get to a
1577 consistent state that we can use.
1579 (3) Ensure that all operations take place with the correct locks.
1581 * One of the issues that has become a serious problem since the
1582 addition of the local directory updates is that although cm_SyncOp
1583 synchronizes operations, it does not preserve the order of requests.
1584 This has always been a problem in that it has been possible for a
1585 request to fail to complete due to its worker thread's bad luck.
1586 When a request takes longer than the Windows SMB Redirector's timeout,
1587 the SMB Redirector tears down the SMB virtual circuit.
1589 When using the local directory updates it is really important that
1590 the directory update operations complete in the order that they were
1591 sent to the file server. If they don't, then the local directory state
1592 and the file server state will not match and the local directory state
1593 must be discarded which in turn forces a new read of the entire
1594 directory contents over the network.
1596 This patch adds a new cm_scache_waiter_t object that is used to store
1597 the current thread, buffer, and syncop flags within a waiters queue
1598 on each cm_scache_t object. If a thread is forced to sleep in
1599 cm_SyncOp, upon waking it will check to see if there are any other
1600 threads waiting that are attempting to perform a similar task ahead
1601 of it in the queue. If yes, the thread goes back to sleep. If not,
1602 it goes ahead and enters the cm_SyncOp conflict resolution block.
1604 This patch has the additional side effect of reducing the number of
1605 competing threads that must obtain the cm_scache_t mutex and process
1606 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1607 utilization of the service and the clock time associated with processing
1608 requests will be reduced.
1610 * assert that the cm_scache_t mutex is held when calling
1611 buf_ForceDataVersion().
1613 * rename findNode to leafNode in bplus_Lookup
1615 * replace all OutputDebugString calls with osi_LogX calls
1617 * modify bestMatch to special case the return values for leaf nodes.
1618 If an entry is above or below the values available in the leaf node
1619 return BTLOWER or BTUPPER instead of BTERROR.
1621 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1622 if true convert to either slot 0 or Max and perform the insertion.
1623 This produces easier to read code when performing lookups.
1625 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1626 descriptive message.
1628 * the difference between osi_assert() and osi_assertx() is whether or not
1629 a description is specified as part of the call. When no message is
1630 specified afsd_notifier() is called with a NULL msgp parameter. This
1631 results in a null pointer dereference during the EventLog() call.
1633 As a result, none of the information describing the assertion is written
1636 This commit sets a default message for use when no other message is
1637 provided and it replaces all calls to osi_assert with osi_assertx and
1640 * Migrate B+ tree search key into thread local storage
1643 * Fix an uninitialized variable in lana_IsLoopback() which
1644 could result in a random determination that a loopback
1645 adapter is not a loopback adapter. This would result in
1646 a computed Netbios service name of %MACHINE%-AFS instead
1647 of AFS. This is a serious problem if the value computed
1648 by the AFS Client Service and the client process pioctl
1649 calls produce different results as it will result in a
1650 communication failure.
1652 * Fix a crash in the B+ Tree code that results when a file
1653 rename only results in a change of case.
1655 * Fix the pioctl path parsing. The previous algorithm
1656 broke the current directory of the client into two parts,
1657 the TIDpath which represented the device root and the
1658 relative path. The problem is that this is not sufficient
1659 from proper processing of symlinks and mount points.
1661 What is required is that the last component of the path
1662 must not have the symlink or mount point be evaluated
1663 whereas symlink and mount point evaluation must be performed
1664 for the rest of the relative path. Therefore, the path is
1665 now split into three parts: TIDpath, intermediate path,
1668 "fs examine" now reports the data of the symlink or mount
1669 point object instead of the target and "symlink list" and
1670 "symlink remove" now works. The "Explorer Shell Extension"
1671 menu now is displayed for symlink objects.
1673 * More informative symlink.exe error messages.
1675 * When the cmd.exe "del" command is used to remove a symlink,
1676 not only was the link object removed, but the stat cache
1677 object for the target was marked deleted. This resulted
1678 in cache corruption which would require the destruction
1679 of the cache to correct.
1681 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1682 which results in cached contents from read only volumes
1683 to be treated as valid when there is no callback registered
1684 and all servers are down or all instances are offline.
1686 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1688 * Add documented registry value, "AcceptDottedPrincipalNames"
1689 which is used to permit the acquisition of AFS tokens
1690 using Kerberos v5 principal names which include a dot in
1691 the first component.
1693 {HKLM\HKCU}\Software\OpenAFS\Client
1695 * "fs uuid" does not require "AFS Client Administrator" group
1696 membership to display the current UUID value.
1699 * Update NSIS installer to version 2.30
1701 * Update Wix installer to version 2.0.5325
1703 * 64-bit KFW support:
1704 - Network Identity Manager plug-in
1706 - 64-bit krb5 enable afscreds.exe, afs server manager,
1709 * Enable B+ tree Directory Search by default.
1711 * NIM plug-in is now compatible with FILE: ccaches
1713 * Install instloop.exe into Client/Programs/ even though
1714 it isn't required. Permit it to be used by organizations
1715 that clone Windows systems.
1717 * On a suspend, only drop callbacks for servers that
1718 are online. If the server is offline, maintain the callback
1719 until the server is once again available.
1721 * Prevent against acquisition of a directory write lock when
1722 there are no updates to be performed. This was resulting
1723 in a write-lock leak which later on would lead to a deadlock.
1725 * Add configurable file pre-fetch mode. For files that match
1726 the specified file extensions, pre-fetch the entire file into
1729 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1730 (MULTI_SZ) PrefetchExecutableExtensions
1732 * Fix the client_osi large integer support so that return values
1733 can be properly assigned to large integer variables.
1735 * Do not hold cm_scache_t mutex across cm_GetCell() calls
1736 since cm_GetCell() can block in a RPC.
1738 * Add short file name support to B+ tree searches.
1740 * Resolve a conflict between the desire to notify file servers
1741 that the client workstation is suspending and the desire to
1742 permit objects with valid callbacks when the client suspends
1743 to continue to use those objects from the cache if the client
1744 resumes without network connectivity.
1746 * Fix stack variables to ensure that all buffers storing paths
1747 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
1748 would cause a stack overflow condition.
1750 * On Vista, force afs_config.exe to run with the highest privileges
1751 available to the end user. For a user in the Administrators
1752 group running under UAC, this will cause afs_config.exe to be
1753 be run with full Administrators privileges after the user
1754 approves the privilege escalation.
1756 * More VNOVNODE handling issues. If a dirty buffer cannot be
1757 written due to VNOVNODE, clear the buffer and ignore the error.
1758 If a cm_scache_t object has been marked deleted due to a VNOVNODE
1759 error, the next time the smb_fid_t object that references it is
1760 used return either NOSUCHFILE or NOSUCHPATH and close the
1761 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
1762 result in an invalid handle error.
1764 * For Vista, set the RequestExecutionLevel of the NSIS installer
1767 * For the build environment, support the generation of manifests
1768 when using the CL1310 (VS .NET 2003) compiler/linker.
1770 * There was a bug in OutputDebugHexString() which could overflow
1771 a stack buffer resulting in unexpected process termination
1772 of afsd_service.exe if the input Security Token contained
1773 many bytes equivalent to the '%' char.
1775 * Add directory lookup statistics to "fs memdump" output.
1779 * Windows uses case-insensitive file name pattern matching
1780 but AFS is a case sensitive file system. The AFS3 directory
1781 format is block based, uses network byte order and
1782 includes a hash table for fast case sensitive lookups.
1783 This causes several problems for the Windows AFS client.
1784 (1) Traversing the directory blocks is cpu expensive
1785 (2) A hash table miss does not indicate that the desired
1786 entry does not exist.
1787 (3) Determining whether a non-ambiguous inexact match or
1788 the entry does not exist requires a linear traversal
1789 of the entire directory.
1790 These issues often result in 100% CPU utilization.
1792 These issues are addressed by building a modified B+ tree for
1793 each directory and then using the B+ tree for searches.
1795 This feature is disabled by default. To enable it set
1796 the "BPlusTrees" registry value to 0x01 in the AFS Client
1797 Service Parameters key.
1799 * Correct a null pointer dereference in the symlink recursion
1800 detection code. (Introduced in 1.5.22)
1802 * Add the Network Identity Manager AFS Provider to AMD64
1807 * A new registry value, "BlockSize", can now be used to adjust
1808 the size of the internal buffers used to store file data.
1809 The default is 4KB. When set, the values 1..1024 specify
1810 multiples of 4KB. Values > 1024 specify actual block sizes
1811 and must be multiples of 4KB. The actual block size in KB
1812 can not be larger than the chunk size which defaults to 1MB.
1814 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1817 * pts.exe and ptserver.exe built with -DSUPERGROUPS
1819 * Service Manager shutdown hints have been added. The hint
1820 specifies that it may take the AFS client service up to two
1821 minutes to shutdown.
1823 * Fix the recursion detection feature introduced in 1.5.22.
1824 The previous implementation broke the evaluation of paths
1825 using ../path or ./path. It also broke the evaluation of
1826 symlinks to absolute paths. For example, /afs/cell/path.
1828 * Fix an uninitialized variable in cm_MakeDir that could result
1829 in crash if the requested directory could not be created.
1831 * Fix a the VolumeStatusNotification log messages to save
1832 the input string in the circular log buffer so garbarge
1833 won't be output to afsd.log when the log is dumped.
1836 * Increase the SMB Maximum Path length to 260 characters
1837 from 256 as that is what CIFS supports.
1839 * On 64-bit Windows, use a unique description string for
1840 the 32-bit AFS Shell Extension.
1844 * In the AFS Admin Library, replaced all of the
1845 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
1846 allows parameter checking and proper type conversions to be
1847 performed. This is crucial when passing parameters that are
1848 64-bit wide on systems in which 'long' is a 32-bit value.
1849 Otherwise, the parameter list on the stack is interpretted
1850 incorrectly and the function not only doesn't succeed but it
1851 might corrupt memory and if lucky crash the application.
1853 The only program that we care about that uses this library
1854 is the AFS Server Manager. Now it doesn't crash.
1856 * Fixed smb directory searches for non-wildcard names. The
1857 optimized function should not have set the bulk in progress
1858 flag. This would adversely affect later directory searches
1859 that required a bulk search.
1861 * Detect infinite recursions (up to 512 stat cache objects
1862 in a chain.) If an infinite recursion is detected, return
1863 an error to stop the CIFS client from querying us forever.
1865 * Executables executed out of AFS name space can have their
1866 pages swapped out by Windows. Those pages will be reloaded
1867 when they are required. However, if the cache manager no
1868 longer has a callback and all of the file servers with
1869 the volume data are down, then the application will crash
1870 due to the inability to read the required page.
1872 We now track the down time of each server. If the callback
1873 expires after the server that issued it went down and all
1874 servers are down, we delay the expiration to permit the data
1875 in the cache to be continue to be used by Windows or an
1878 * The Windows AFS client has always had very poor performance
1879 with regards to directory management. In the case of
1880 directory updates caused by create, rename,and delete operations,
1881 the AFS cache manager will now perform local directory updates
1882 when possible in order to avoid re-reading the directory from
1883 the file server. This will be done in whenever both of the
1884 following conditions are true:
1886 - the data version on the directory increased by one as a
1887 result of the operation
1889 - all of the directory buffers necessary to make the change
1890 are present in the cache
1892 * The error message in the AFS Control Panel describing the
1893 need to be a Windows Administrator has been replaced with
1894 a message describing the need to be in the AFS Admins Group.
1896 * If the vldb server list is empty when attempting to evaluate
1897 a cell mountpoint it was possible for a crash to occur. This
1898 can only happen if there is an existing mount point and the
1899 is a failure attempting to load the server list from the
1900 CellServDB file or via AFSDB queries.
1902 * dirty buffers only have their dirty bytes stored to the file
1903 server instead of the entire chunksize
1905 * the default chunksize has been increased from 128KB to 1MB
1910 * changed the enum values for cm_serverRef_t state info to use a
1911 private name space to avoid collisions (srv_)
1913 * added a srv_deleted state for cm_serverRef_t objects. This
1914 state is set when cm_FreeServerList() is called with the
1915 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
1916 not always delete the cm_serverRef_t from the list if it is
1917 still in use by another thread. the srv_deleted state means
1918 the object's contents are no longer valid and it must be
1919 skipped. It will be deleted the next time the object is
1920 freed and the refcount hits zero.
1922 * the srv_deleted state is also used when a file server reports
1923 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
1924 as offline. This is done to prevent additional usage of the
1925 stale vldb data while waiting for the update volume request
1928 * added a state field to the cm_volume_t object (enum volstate
1929 vl_ name space) that maintains the state of the volume based
1930 upon the states of all of the cm_serverRef_t and cm_server_t
1933 * modified cm_UpdateVolume() to set the state of the cm_volume_t
1934 RW, RO, and BK to either vl_alldown or vl_online. There can't
1935 be any other states because cm_UpdateVolume() destroys any
1936 previous knowledge we might have had regarding busy or offline
1939 * modified cm_UpdateVolume() to update the volume name in the
1940 cm_volume_t to the volume base name if the previous value was
1943 * modified cm_FollowMountPoint() to check to see if the volume
1944 name is a volume ID and if so call cm_GetVolumeByID instead
1945 of cm_GetVolumeByName. This ensures that volume IDs are always
1946 looked up as numeric values. There is no longer a need to
1947 maintain a separate cm_volume_t containing the string representation
1950 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
1951 The first flag is a "CREATE" flag which is set by all existing
1952 calls. The flag is not set by calls to cm_GetVolumeByID() from
1953 the server probe code when volume status is being updated. We
1954 do not want the server probe operation to result in additional
1955 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
1956 which is set when the server probe code updates the volume status.
1957 This flag will be used to prevent the server probe operation from
1958 changing the order of the least recently used queue.
1960 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
1961 allocated for a given set of normal, readonly, and backup volumes
1962 regardless of whether or not the volume is accessed via name or
1963 ID number. The cm_volume_t namep field is always the base name
1966 * Added a new volume state, vl_unknown. This state is used as
1967 the initial state for all cm_volume_t when the cache manager starts,
1968 for each cm_volume_t at creation, and for each cm_volume_t when
1969 recycling. The cache manager does not know the state of all
1970 volumes in the world, only those that are in the cache and for
1971 which it has queried the VLDB and hosting file servers.
1973 * modified cm_GetVolumeByName() to initialize the state of a
1974 volume to vl_unknown. The actual state will be set when a
1975 cm_VolumeUpdate() call completes successfully.
1977 * changed name of scache hash table variables to avoid ambiguity
1978 when adding hash tables for volumes
1980 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
1982 * modified cm_UpdateVolume() to handle the case in which there is
1983 no RW volume but there is are RO volumes for a given base name.
1984 This is done by querying for the ".readonly" volume name if the
1985 base name does not exist in the VLDB. We never query for the
1986 .backup name because under the current usage model a .backup
1987 volume may only exist on the server that the read-write volume
1988 is located. If there is no RW volume, there can be no .backup.
1990 * Added four hash tables for cm_volume_t objects to improve the
1991 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
1992 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
1993 BKID. Three ID hash tables are necessary as long as it is
1994 desireable to maintain a single cm_volume_t containing all
1995 of the related RW, RO, and BK volume data. Having the RW and
1996 RO volume data in the same object is necessary for the
1997 implementation of cm_GetROVolumeID() which returns either the
1998 RO or RW ID depending upon the existence of RO volume instances.
2000 * Added a volume LRU queue so that volume reuse becomes fairer.
2001 This does not replace the all Volumes list which is used when
2002 it is desireable to walk a list of all the volumes whose order
2003 is not going to change out from underneath you which makes it
2004 safe to drop the cm_volumeLock.
2006 * handles volume hash table updates where volume name to
2007 volume ID number changes. The volume name remains
2008 constant in the cm_volume_t. if a vos rename is performed,
2009 the name of the volume will change and the volume IDs will be
2010 updated. Subsequent access to the old volume ID will create a
2011 new cm_volume_t with the new name.
2013 * Added a daemon thread operation to query the state of volumes
2014 listed as busy or offline. cm_CheckBusyVolumes() calls
2015 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2016 or vl_offline. If the volume is now online, the status on the
2017 volume is updated. The default period is 600 seconds. This can
2018 be configured with the BusyVolumeCheckInterval registry value.
2020 * Added prototype for smb_IoctlPrepareRead() which was missing a
2021 return type in the function definition.
2023 * Added volume id lists to the cm_server_t. These lists are
2024 allocated in blocks of ~32 IDs. When a cm_PingServer()
2025 detects a change in server state, the state of the cm_volume_t
2028 * Added volID to the cm_serverRef_t object. volID is used
2029 to identify the volume for which the object is a referral.
2030 cm_FreeServerList() uses the volID to remove the volume
2031 from the cm_server_t.
2033 * In cm_Analyze, when VNOVOL or VMOVED are received,
2034 call cm_ForceVolumeUpdate() to force a refresh of the volume
2037 * Added cm_VolumeStatusNotification() which is used at the moment
2038 to log volume status changes to the trace log. It will also
2039 be used as the access point to the File System Filter driver
2040 notification engine.
2042 * Added an all cm_scache_t list to cm_data. This replaces the use
2043 of the stat cache LRU queue when we need to enumerate all
2044 entries. The LRU list order is not static and when using it to
2045 enumerate all entries it can result in items being missed or
2046 items being processed more than once.
2048 * Modified cm_Analyze(). Instead of reseting the busy or offline
2049 state of a volume and forcing a retry of the operation
2050 cm_Analyze will defer to the background daemon thread that will
2051 update the state once every 600 seconds.
2053 * Added the automatic generation of a Freelance ".root" read-write
2054 mountpoint that refers to the root.afs volume of the workstation
2055 cellname at the time the mountpoint is created.
2057 * re-write cm_Analyze to make better use of the known volume
2058 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2061 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2062 busy volumes will be reset to srv_non_busy by the function
2063 but there is no mechanism for querying the busy state other
2064 than by attempting to access the resource.
2066 * cm_Analyze will query the state of an offline volume before
2067 deciding whether or not to retry when all volume instances
2070 * the algorithm used after an IP address list change detection
2071 has been modified. we now check the state of the VLDB servers,
2072 then reset the rx connections, and finally check the state of
2073 the file servers. Previously we flushed the rx connections
2074 and then checked the servers in whatever random order they
2075 were first discovered.
2077 * fix the volume status reported by "fs examine"
2079 * various improvements to the AFS Network Identity Manager Provider.
2080 - Don't access freed objects after the window has been destroyed.
2081 - Write back the AFS cell list in the identity configuration
2082 even when its empty.
2083 - Disable all use of the provider when AFS is disabled for a
2084 particular identity.
2086 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2087 alias instead of the full cellname, double check the full cell
2088 name and see if we have a matching cell before allocating a new
2091 * Dump more data with "fs memdump". All volumes and scaches and
2094 * Reverse the order of the all cells list. This ensures that
2095 the workstation cell is always first in the list.
2097 * If we recognize the server from which a callback revoke was
2098 received, then we can ensure that we only invalidate the status
2099 information for cm_scache_t objects belonging to that cell.
2101 * When building for amd64, use APPVER 5.02 as we do not support
2102 64-bit Windows 2000 and explicitly state the machine type in
2105 * no longer treat VMWare network adapters as equivalent to a
2106 loopback adapter. For some unknown reason, publishing the "AFS"
2107 netbios name on them frequently results in a name collision
2110 * Use read locks whenever possible instead of write locks when
2111 walking the allCellsp list. This permits better use of multiple
2114 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2115 is not present on the directory optimization. It would report
2116 mount points and directory objects for which there was no
2117 status information as a zero length file. Instead we can use
2118 the vnode number of the object to determine if it should be
2119 considered to be a directory or a file. If it is odd it is a
2120 directory; even it is a file.
2122 * Give Back Callbacks to the file servers in response to suspend
2123 and shutdown power management events. Only give up callbacks
2124 to UP servers and do so with a very short RX dead timeout.
2127 * Fix a race condition in the SMB FID cleanup code that could
2128 result in a stat cache mutex never being released.
2130 * Fix the graphics files that are installed as part of the
2131 IBM Administrators Guide.
2133 * Change the behavior of pioctl path parsing to not follow
2134 symlinks or mount points during evaluation. This results in
2135 the FID being the symlink or mount point and not the object
2136 they point to. This is consistent with the behavior of the
2139 * Enable AFSDB record lookup for command line tools such as
2142 * CellServDB update 2007-07-05
2144 * Added a Volume Status Plug-in Interface that permits third parties
2145 to receive notifications when the volume state changes occur.
2147 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2149 * Prevent the afsprocmgmt.dll from being unloaded while the
2150 background thread it starts is running.
2152 * Avoid a race condition when closing smb_fid_t objects.
2155 * fix install path of IBM Quick Start Guide when installed by
2158 * fix the client so that it can write into directories with
2159 only 'li' permissions.
2161 * fix installation of 32-bit Shell Extension for 64-bit systems.
2162 Install it under its own HKCR keys.
2164 * modify the wix installer to remove and upgrade install packages
2165 with the same version number.
2167 * when flushing volume location data every two hours or in
2168 response to "fs checkvolumes", also flush the mount point
2169 links in order to force their re-evaluation.
2171 * aklog was copying an uninitialized buffer as part of the
2172 foreign cell automatic pts registration
2174 * when obtaining directory listings, check for read permission
2175 on the directory. If read permission does not exist, fake the
2176 status information for the directory entries. This improves
2177 performance, prevents the client from pounding the file server,
2178 and prevents leakage of cached status information.
2180 * "fs getcacheparms" output was incorrect.
2182 * NetIdMgr AFS Provider should only automatically add token
2183 acquisition for cell foo.com if the realm of the cell matches
2184 the realm of the identity.
2186 * Fix a null pointer dereference in cm_MergeStatus()
2190 * network identity manager provider:
2192 - Correct for the fact that pr_CreateUser tampers with the
2193 contents of the username input buffer.
2195 - Take into account that the krb4 credential data structure
2196 does not have separate realm fields for the user and service.
2197 When krb524 is used with a cross-realm ticket, the realm
2198 describes the service not the user.
2200 * Many integrated logon changes:
2202 - Fix eventlog reporting. Do not attempt to log an event
2203 if the event source registration fails. Use DebugEvent0
2204 instead of DebugEvent when there are no parameters.
2206 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2209 - Fix the reading of Domain specific configuration for
2210 LogonScript and TheseCells. Previously the dwSize value was
2211 being overwritten so that subsequent RegQueryValueEx call
2214 - Fix a memory leak in the TheseCells reading code.
2216 - Add support for Domain specific "Realm" specification. The
2217 realm is the realm to be appended to the username. When
2218 logging in as a domain or to the local machine, the specified
2219 "Domain" name is not going to be a valid realm name.
2221 - Construct a proper principal name based upon the domain
2222 specified realm for use in obtaining tokens with KFW.
2224 - If the domain specified "TheseCells" list includes the
2225 default cell, do not obtain tokens twice.
2227 * An unprotected call to the Kerberos v5 function,
2228 krb5_free_context(), can result in an invalid memory access
2229 exception if the contents of the krb5 profile, krb5.ini, cannot
2230 be parsed correctly. The error is located within the afskfw
2231 library which is used by afscreds.exe, afssvrmgr.exe, and
2232 afslogon.dll. When the error is triggered within afslogon.dll
2233 a blue screen condition can occur when the winlogon.exe system
2234 process terminates unexpectedly.
2238 * Do not stop the afs client service with an assert
2239 if the cell name cannot be determined when responding
2240 to a "fs whereis" command.
2242 * Integrated Logon (afslogon.dll). No longer attempt
2243 to preserve Kerberos 5 tickets for the logon session
2244 as that job is now supported by KFW 3.1 and above.
2246 * The Network Identity Manager plug-in will attempt to
2247 remove expired service tickets from the cache.
2251 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2253 * AFS Server Manager and AFS Account Manager data cache
2254 moved from HKCU\Software\TransarcCorporation to
2255 HKCU\Software\OpenAFS to prevent crashes due to data
2258 * Reduce the default UP server probe time from 10 minutes
2259 to 5 minutes to match UNIX. This will not solve the NAT
2260 port mapping timeout issue because most NATs default to
2261 a value less that 5 minutes and this value ensures that
2262 the afs client server will start probes 5 minutes after
2263 the last probe set ended. The probe period for any one
2264 server is therefore longer than 5 minutes. The more
2265 servers in the list the longer the time between probes.
2267 * Before marking a server as down in cm_Analyze, retry the
2268 request using a new rx connection to the server. If the
2269 cache manager is behind a NAT, it is possible that the
2270 response from the server is being sent to the original
2271 address:port from which the rx connection was established.
2272 This will appear to the cache manager as the server being
2273 down when in fact it is not. Forcing the use of a new rx
2274 connection will result in the server responding to the
2275 most recent address:port assigned to the client by the NAT.
2277 If the server is in fact down, it will simply take a little
2278 longer to notice it.
2282 * Removing a directory that contained entries was failing with
2283 the wrong error code being returned to the CIFS client.
2287 * rxdebug exponential back-off. don't take down a server that
2288 is already struggling to respond.
2290 * Network Identity Manager plug-in registers help on the Help
2291 menu if it is supported by NIM. (KFW 3.2 and above)
2293 * Plug-n-play networking did not always work after coming back
2294 from a Windows Suspend. The AFS client service failed to
2295 re-compute the LAN adapter to bind to and instead attempted
2296 to bind to all LAN adapters. This would polute the public
2297 network with the name "AFS" which could lead to conflicts
2300 * Build installers using Microsoft Hotfixes provided at
2301 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2302 Daylight Savings Time issues.
2304 * "fs checkserver" does not require admin privileges unless
2305 the probe interval is being changed.
2309 * A reference leak was discovered in the mountpoint traversal
2310 functionality which resulted in cm_volume_t objects not becoming
2311 eligible for re-use. This will trigger a panic condition when
2312 more than the maximum number of volumes is accessed.
2314 * Fixed the success test for the "fs memdump" operation. Success
2315 is now reported as such. Added cm_volume_t object information
2318 * A data corruption error has been fixed. Write requests
2319 queued for background processing were not retried upon failure.
2320 This leads to inconsistencies between the contents of the file
2321 in the local cache and the contents of the file on the file
2324 The implemented fix is to re-queue any request that has failed
2325 due to timeout, volume busy, volume offline, server busy, or
2326 server offline. Requests are not processed while all servers
2327 for a volume are inaccessible. When servers become accessible
2328 the background request will be retried.
2330 In addition, if a background write fails, do not mark the buffers
2331 as clean. Failure to preserve the dirty flag prevents the buffers
2332 from being written to the file server in the future.
2334 This fix does not address the possibility that the credentials
2335 necessary to perform a write request might expire between the
2336 time the request is queued and when the request can be processed
2337 due to server accessibility.
2339 * The Freelance volume does not support per user ACLs. The
2340 cm_HaveAccessRights function did not handle this special case.
2341 Requests for privileges greater than read|lookup would result
2342 in afsd_service.exe entering an infinite loop. ISOBuster is
2343 known to cause this behavior.
2345 * When deleting files, the SMB Server would attempt to obtain
2346 the access permissions for the parent directory instead of
2347 for the file itself. This would result in an incorrect
2348 assessment of whether or not the file can be deleted.
2350 * The Network Identity Manager AFS plug-in has had its string
2351 tables updated for consistency with the language used by NIM.
2352 All credentials are referred to as credentials instead of tokens.
2353 Drop-down combo boxes now in fact drop down.
2355 * The AFS Salvager was not properly being built on Windows.
2356 This is relevant only if the AFS servers are in use.
2358 * The default signal handlers for the AFS servers incorrectly
2359 processed SIGQUIT. Instead of stopping the server, an
2360 exception was raised.
2362 * Fix another volume refcount leak. This one was triggered by
2363 an inability to contact the vlservers in order to update the
2366 * Fix checklist control used by the drive mappings tabs in
2367 afscreds.exe and afs_control.exe.
2369 * Improve output of "fs memdump" related to buffer management.
2371 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2372 already in the process of being fetched.
2374 * Improve cache performance by computing the buffer hash table
2375 size based upon the number of buffers so that the average
2376 number of entries in a bucket is seven.
2378 * Improve cache performance of read/write data by locally
2379 updating the dataVersion of all buffers associated with the
2380 file on each StoreData or StoreStatus operation provided that
2381 the dataVersion was incremented by one and the dataVersion of
2382 the buffer matched the original dataVersion of the cached
2383 status information for the file.
2385 * When recycling cm_scache_t objects, make sure that we
2386 clear all of the previous fields. It is no longer the
2387 case that we can assume that the stat cache object is being
2388 obtained only after a successful FetchStatus call.
2389 InlineBulkStatus will allocate stat cache objects for which
2390 the no status information is available.
2392 * When a file is being opened with the OPEN_ALWAYS disposition,
2393 do not require write permissions when the file already exists
2394 unless of course the file is being opened requesting write
2397 * When a file has been opened for read-only access, if the
2398 application requests an exclusive lock, obtain a read-lock
2399 instead of a write-lock. The purpose of the exclusive lock
2400 in this situation is to prevent changes to the file and the
2401 full file read lock will provide this guarrantee.
2406 * Enforce Unix Mode Write bit to prevent file deletion of read only
2409 * The CIFS READ and LIST permissions use the same value 0x0001.
2410 When a directory is being accessed, check for PRSFS_LOOKUP
2411 instead of PRSFS_READ.
2415 * Fix computations of short names of mount points. They were
2416 broken as part of the directory search optimization. (1.5.13a)
2418 * When all buffers are in use, periodically sleep the threads
2419 waiting for a buffer to become free while the global buffer
2420 lock is not being held in order to permit those threads that
2421 can free buffers to do so.
2423 * Power Management improvements. Maintain a global flag that
2424 specifies whether or not the service is in a suspend state.
2425 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2426 lana is no longer valid. Instead, stop the listener threads
2427 and if all listener threads are stopped, reset the lana_list.
2428 Allow the cm_Daemon() thread to periodically check the state
2429 of the smb listeners. If they are all stopped and the service
2430 is not suspended, attempt to restart them. If there are no valid
2431 lanas, return to the stopped state.
2433 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2434 the user is permitted to obtain read or write locks. This function
2435 would obtain the lock and then drop it returning whether or not
2436 the lock could be obtained. If the lock was in fact required,
2437 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2438 of course being that this pattern results in three RPCs to the
2439 file server (lock, unlock, lock). This is reduced to one RPC by
2440 implementing cm_CheckNTOpenDone() which frees the allocated byte
2441 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2444 * Optimize the SMB FindFirst processing for the case in which there
2445 are no wildcards. If a cm_Lookup succeeds, generate a response
2446 from the stat cache entry we have a callback for. Its cheaper
2447 than walking each buffer of directory data associated with the
2448 directory stat cache entry when we have thousands of directory
2451 * Increase the SMB Ioctl MaxData size to support larger token sizes
2453 * Disable autorun of the AFS Server Wizard when the AFS Server is
2456 * Add AFS Server service dependencies on tcpip and PNP_TDI
2458 * Remove the AFS Server Configuration Tool shortcut
2460 * Add Freelance mode awareness to AFS Server Configuration Tool
2464 * Unlike previous versions of the OS, Vista performs a shutdown on
2465 the Microsoft Loopback adapter just like it would on a real adapter.
2466 This causes the smb_Listener threads to trigger a panic during a
2467 suspend/hibernate power event.
2469 The fix is to unbind from the network adapters in response to a
2470 suspend/hibernate power event and then rebind to the adapters
2471 when a resume power event is received. Note that the resume events
2472 are not reliably delivered so it is possible the afs service will
2473 not be accessible. However, this is the best we can do.
2475 * don't allow environment variables we set in afslogon.dll to be
2476 inherited by child processes
2478 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2481 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2482 thread walking the all buffer list periodically searching for dirty
2483 buffers to write to the file server. This had several negative
2484 results. The algorithm ate up ever increasing amounts of CPU time
2485 even when AFS is idle as the size of the cache increases. Also,
2486 buffers were written to the file server in an order based upon the
2487 original buffer allocation which has nothing to do with the order
2488 in which the buffers became dirty.
2490 The new algorithm maintains a dirty buffer list. Items are added
2491 when the buffer is originally marked dirty. A buffer is only
2492 removed from the list by the buf_IncrSyncer when the buffer is no
2493 longer dirty. If the list is empty the thread goes back to sleep
2494 immediately without additional processing requirements.
2498 * Enable renaming of directories from the Explorer Shell when
2499 Offline Folders is enabled
2501 * An error is reported in response to the SMB ReceiveCoreFlush
2502 request when an error has not occurred.
2506 * optimize dirty buffer flushes in response to suspend operations
2507 do not flush if only loopback adapter is installed
2508 do no retry dirty buffer writes
2510 * do not reset netbios when performing loopback checks. doing
2511 so will break the existing lana bind
2513 * remove a deadlock in smb3.c QPathInfo
2517 * Add NetIDMgr AFS Plug-in
2519 * "Show Tray Icon" checkbox removed from afs_config.exe
2521 * Disable DNS Registrations for the Loopback Adapter and make sure
2522 that Netbios is turned on.
2524 * A reference count leak on stat cache objects was fixed
2526 * For debugging, tread ids of writers are now tracked on locks
2528 * Corrected an incorrect lock that was being held but not
2529 freed in a rare error case
2531 * Changed trace log end of line to CR-LF so it can be read
2534 * modify the buffer synchronization thread to write all dirty
2535 buffers and sleep only when there is no work to be done
2537 * improve behavior of cache manager when large numbers of
2538 temporary files are being created by multiple clients within
2541 * do not permit dirty buffers that belong to deleted files
2542 to be written to the file server. its a waste of bandwidth
2544 * do not leak stat cache references when the file server
2547 * afscreds: do not permit an attempt to obtain tokens if the
2548 afs service could not be started. this was causing stack
2549 overflows on Vista after resuming from a suspend
2553 * better handling of dot directory
2555 * fs sq and fs mkm now behave as on UNIX
2557 * add license to installers
2559 * Return Invalid Parameter error for unknown InfoLevels in
2560 SMB Tran2 Search Directory.
2562 * Correctly use WSAStartup and WSACleanup within the RX library
2564 * Check that afsd_service.exe is running before performing
2567 * Force AFS Shell Extension to load resource DLL using
2568 afs_shl_ext.dll as a base name.
2570 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2571 OpenAFS\Client\Program
2573 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2575 * Do not panic if the maximum number of volume entries are in use
2576 and one of them can be recycled.
2580 * Fix the code that reads the backconnnectionhostnames value
2581 so that there is not an uninitialized variable
2585 * integrated logon could cause winlogin.exe to blue screen windows
2586 if it was called with a domain that does not exist on the machine.
2587 this is possible during a sysprep.
2589 * corrections to cm_buf_t handling to ensure that buffers are
2590 removed from queues before they are freed and that holds are
2591 obtained on the next buffer in a chain before releasing the
2594 * do not generate MessageBox dialogs on network errors. they
2595 cause more trouble then they are worth.
2597 * move GlobalAutoMapper into a separate thread so that it does
2598 not prevent service startup if the WNetAddConnection2 call
2601 * restrict hard dead and connection timeouts to the CIFS Session
2602 Timeout value. This is being done to try to prevent virtual
2603 circuits from being destroyed prematurely.
2605 * Fix the code that writes the backconnectionhostnames value
2606 to ensure that the data buffer is written with the correct
2609 (there were no windows specific changes between 1.5.4 and 1.5.5)
2613 * Corrections to the Integrated Logon functionality that permits
2614 Kerberos 5 TGTs to be communicated into the user session. The
2615 transfer is now more secure and succeeds when the user does not
2616 have read permission to the %SystemRoot%\TEMP directory.
2618 * NSIS installer updated to version 2.18
2620 * WiX installer updated to version 2.0.4221.0
2622 * Updates to the IFS functionality for 64-bit correctness and
2623 permit compilation. The IFS functionality works for pioctl()
2624 operations but not much else at this point.
2626 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2627 query and set functions.
2629 * Ignore the Windows CIFS client session timeout entirely. Timeout
2630 only after the AFS Hard Dead Timeout period expires.
2634 * move headers, libraries, and samples from OpenAFS\Client to
2637 * remove the auto-registration of AFS IDs for foreign cells
2638 from integrated logon. this permits afslogon.dll to avoid
2639 initializing the RX library. afslogon.dll is a network
2640 provider library and as such it is loaded into all applications
2641 which perform a CreateProcessAsUser call. Initializing the
2642 RX library spawns background threads to listen for inbound
2643 packets. When the network provider is unloaded from the
2644 process, the RX library does not cleanup the threads. If they
2645 continue to execute after the library is unloaded, the process
2648 * implement inline bulk status calls. this will significantly
2649 improve the performance of directory listings.
2651 * prevent a crash when evaluating mountpoints to volumes that do not
2654 * new command: fs uuid [-generate]
2655 displays the current uuid and allows a new one to be generated
2656 on the fly without restarting the afs client service.
2658 * fix the 64-bit installers to properly document the install
2661 * move help files to the proper location so they can be located by
2662 the system tray tool and the control panel.
2664 * "over quota" and "disk full" errors were not being reported to the
2667 * largefile support has been added. Files greater than 31-bits can no
2668 longer be written to file servers which do not support large files.
2669 When file servers support large files, files up to 64-bits can be
2672 * cm_SyncOp could return with the input buffer unlocked upon error
2673 which would result in a panic condition. This bug was introduced
2676 * vlserver failover when freelance mode did not work. This prevented
2677 the root.afs volume from being loaded if the first vlserver could
2682 * A bug in the CIFS Standard Info response packet caused directories
2683 to appear as standard files. This prevented Microsoft Vista from
2684 being able to walk the AFS directory tree.
2686 * Mount points must be reported as directories in response to
2687 Query Path Info operations.
2689 * A deadlock was detected when performing "fs flushall" if the
2690 file server reports VNOVNODE. The scp->createBufferLock is
2691 already held by the current thread and there is no mechanism
2692 to propagate the knowledge. Therefore, an alternate mechanism
2693 for clearing the cache must be developed. A new function
2694 cm_RecycleSCache(scp,flags) has been extracted from
2695 cm_GetNewSCache(). This function performs the task of recycling
2696 an cm_scache_t object. When called from cm_GetNewSCache() with
2697 no flags the expectation is that there are no associated buffers
2698 that are queued to be read or written. When called from
2699 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2700 flag, any queued buffers will be de-queued and marked as if
2701 the operations were performed so the data can be discarded.
2703 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2705 * In response to VNOVNODE, the parent is only discarded if the current
2706 cm_scache_t is not a directory.
2708 * In the Ioctl FlushFile and FlushVolume functions, there were no
2709 checks to protect against flushing the Freelance SCache entries.
2711 * In FlushFile, the wrong cm_scache_t object was being released.
2713 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2714 recycled. Choose a new entry if cm_RecycleSCache fails.
2716 * Speed up the performance of the cache manager by not
2717 holding the smb_fid_t mutex across calls to cm_SyncOp
2720 * Ensure that all smb_fid_t flag references are protected.
2722 * Remove a deadlock between smb_fid_t mutex and
2723 smb_rctLock introduced in 1.4.1.
2725 * Fix aklog AFSID lookup to use not send the realm
2726 name when the realm is the local realm for the
2727 cell. The PT Server doesn't strip the realm
2728 name when it is the local realm.
2730 * Treat "\\afs\*." as an alias for "\\afs\all"
2732 * Dynamically adjust the priority of server threads
2733 based upon the age of the CIFS request that is
2734 being processed. The priority is bumped one level
2735 for every 15 seconds of processing time.
2737 * The original openafs contribution mis-used the
2738 cm_GetCallback function. This function should only
2739 be called as a side effect of a call to
2740 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
2741 was being called results in a FetchStatus RPC being
2742 performed even when the client already has a valid
2743 callback and does so in a manner that destroys the
2744 synchronization of the threads that are calling
2745 cm_SyncOp. If nothing else this results in a
2746 significant performance penalty.
2748 * More changes to cleanup of smb_vc_t objects in
2749 order to prevent race conditions.
2751 * Reduce the amount of Power Management event logging
2754 * If the client detects that the IP address has
2755 changed, force the use of new RX connections. This
2756 enables the file server to recognize that the client
2759 * Remove a race condition and deadlock associated
2760 with cleanup of smb_vc_t objects.
2762 * Fix the ANSI filename support. One function that
2763 required it was missing the translations.
2765 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
2768 * Add a DllMain() function to afspthread.dll. Perform
2769 cleanup of thread local storage, rx queues, critical
2770 sections, event handles, etc. This was implemented
2771 to avoid crashes in afspthread.dll when applications
2772 which load/unload the dll attempt to terminate.
2774 * Lock request behavior has been altered. As of this
2775 release a locally allocated lock will be used whenever
2776 the volume is RO or the requesting user's permissions
2777 are 'rl'. This will allow end users to execute programs
2778 from RW volumes when they do not have the 'k' privilege.
2780 Note that a recent discovery indicates that the 'k' bit
2781 is only supposed to control the ability to obtain read
2782 locks. Write locks are to be controlled by the 'w' and
2785 * Code has been added to enable the cache manager to be
2786 able to obtain the AFSID of the authenticated user.
2787 This is required to be able to determine when "Insert"
2788 privileges should apply. AFS considers a file to be
2789 a new file whenever the owner's AFSID matches the
2790 AFSID of the user. The current implementation relies
2791 on the name specified during the SetToken call to be
2792 correct. This will be replaced in the future with a
2793 new RPC that allows the AFSID of the authenticated
2794 user to be obtained.
2796 * A 32-bit-tools installer for use on 64-bit systems
2797 is now available. These tools are for use with 32-bit
2798 versions of Kerberos for Windows or JPSoftware's 4NT.
2800 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
2802 * Remove the use of the "active_vcp" variable. This variable
2803 no longer serves any useful purpose and was the cause of
2804 race conditions that produced under counts of the smb_vc_t
2807 * Change default minidump output to include data segments
2809 * restore access to cm_IoctlGetVolumeStatus and
2810 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
2811 Still need to figure out why they are being called
2812 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
2814 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
2816 * Update build system to include the "sys_name" in the
2819 * optimize the number of calls to obtain the smb_user_t for a
2820 smb virtual circuit.
2822 * add logging of server addresses, status and capabilities when
2825 * prevent multiple threads from pinging the same server simultaneously.
2826 instead, threads block and then use the result of the ping that was
2827 already taking place.
2829 * Add registry values to allow the defaults for daemon checks to be altered
2830 at startup. Log values to afsd_init.log.
2832 daemonCheckDownInterval
2833 daemonCheckUpInterval
2834 daemonCheckVolInterval
2835 daemonCheckCBInterval
2836 daemonCheckLockInterval
2837 daemonCheckTokenInterval
2839 * Add EventLog messages for transitions in fs crypt state.
2841 * Byte range locking (when locks will be obtained from the
2842 is still under discussion)
2844 * 64-bit Windows operating systems now supported
2845 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
2846 This is a native 64-bit implementation of OpenAFS for Windows.
2847 There is currently no 64-bit version of KFW and therefore
2848 there is no (easy) method for obtaining AFS credentials.
2849 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
2850 to obtain AFS tokens but there is not for the feint of
2853 * A faster implementation of the fcrypt algorithms improves
2856 * The afsd_service.exe now supplies Event Log Message Text.
2857 The event log no longer generates warning text due to lack
2858 of message configuration.
2860 * When afs servers are unreachable and network stack returns
2861 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
2862 mark the attempted server as being down. This allows for
2863 very fast response to the calling application. This change
2864 has a pronounced improvement when the Windows Explorer is
2865 used on disconnected mobile computers.
2867 * Added DebugOutputString debugging of the RX library
2869 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
2871 * Fix the rx-lwp implementation to randomly generate rx call
2872 identifiers which are supposed to be unique at a given point
2873 in time. This affects some of the command line tools used
2874 for debugging but not the AFS Client Service itself which
2875 uses the rx-pthreads implementation.
2877 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
2878 Windows although they are not distributed
2880 * don't generate an exception when logging unexpected SMB packets
2882 * the data structures used by cm_IoctlGetVolumeStatus and
2883 cm_IoctlSetVolumeStatus are not compatible with the CIFS
2884 Get/Set File Information calls so don't bother executing them.
2888 * Fix data written to registry as part of BackConnectionHostnames
2891 * Fix initialization bug when allocating cm_server_t objects that
2892 would cause servers to appear to be down after each allocation
2893 or renewal. Server would appear up after the next check servers
2898 * Windows will return WSAECONNRESET to the next receive packet
2899 operation on a socket if an ICMP packet is received in response
2900 to a packet that was sent and could not be delivered because
2901 the port was unreachable. We have to treat this as
2902 WSAEWOULDBLOCK so that we can retry the select() check and
2903 not mark the connection as bad.
2905 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
2907 * The client manager maintains a least recently used list of
2908 all the ACLs. If an ACL is already the most recently used
2909 do not bother to remove it from the list and place it on the
2912 * As part of the process of adding a new server, set the
2913 server is down flag and then perform a ping. If the server
2914 is up, the flag will be reset. If the server is down, the
2915 client will not timeout unnecessarily while attempting to
2918 * If a Mountpoint String is the empty string, return PATH NOT
2921 * Modify behavior of the Network Provider to avoid querying the
2922 profile location if integrated logon is known to be disabled.
2924 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
2925 will always be preserved regardless of whether or not integrated
2926 logon was used to obtain tokens. When unset, tokens will be
2927 destroyed regardless of whether integrated logon was used to
2928 obtain the tokens unless the profile is stored in AFS.
2930 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
2931 message is being delivered even when the user is not logging
2932 off of the computer.
2934 * Fix a deadlock condition introduced in rc6.
2936 * Cell names are case-insensitive. Since the founding of OpenAFS
2937 there has been a bug that treated "cell", "Cell", and "CELL" as
2938 different cells. This would result in what appeared to be
2939 random disappearance of user tokens or a failure to authenticate
2940 even when tokens were listed by the "tokens" command.
2942 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
2944 * Additional auditing of the smb code revealed that smb ioctl
2945 operations were failing to free smb_fid_t objects
2947 * The File Identifiers assigned to smb_fid_t objects could be
2948 given the value 65535 which is 0xFFFF which is also the
2949 INVALID_HANDLE value. Both 0 and 65535 are now treated as
2952 * Audited all assignments to fields and ensured that they are
2953 protected by obtaining the relevant mutex object.
2955 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
2956 if both the current cm_scache_t object and its parent are
2957 both returning VNOVNODE.
2959 * Removed all the logic associated with the 'dead_vcp' variable
2960 which must have been added for debugging because its used
2961 was never thread safe and it truly served no purpose.
2963 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
2965 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
2966 registry settings. Default is ON and 120 seconds. If set to
2967 OFF, tokens will not be destroyed.
2969 * Fixed a number of race conditions and reference count problems
2970 associated with global queues.
2972 * Added a method to detect abandoned smb virtual circuits so their
2973 associated file handles, locks, user tokens, and smb session
2974 resources can be released.
2976 * Added a mechanism to allow user tokens to be preserved
2977 during integrated login and be freed after logoff.
2979 * Check access permissions to the file before the file is
2982 * More improvements to the handling of SMB Virtual Circuits
2983 when SMB sessions are closed.
2985 * Correct a race condition introduced in rc4 that could result
2986 in a crash of afsd_service.exe.
2988 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
2990 * Properly handle pre-mature SMB session closures.
2991 File handles, tree connections, and user sessions are now
2992 freed when the virtual circuit is closed.
2994 * smb_username_t and smb_user_t objects are now properly reference
2995 counted and freed as appropriate.
2997 * disable to use of AFS file locks. byte range locking is still
2998 implemented but all locks are managed locally. This prevents
2999 multiple processes from editing the same file on the same
3000 machine but it does prevent multiple processes on separate
3001 machines from modifying the file.
3003 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3005 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3006 the same server parameter would result in a panic condition
3007 if the server had not already been contacted by the AFS
3010 * Better warnings are written to the afsd.log file when
3011 the Windows RPC Protocols are not properly configured as
3012 is often the case on Windows 2000.
3014 * fix the procmgmt library to restore C RTL signal handlers when
3015 it is being detached from the process prior to process termination.
3017 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3019 * close a lock leak in the rx library that could be triggered while
3022 * prevent removal of drive mappings in response to afscreds.exe -M
3024 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3025 strings and flush the stat cache entry.
3027 * prevent server objects from being freed if user preferences are set.
3029 * Update WIX support to 2.0.2419.0
3031 * Release all locks on the server at service shutdown.
3033 * A down server should be marked up as long as it responds to the client.
3034 It doesn't matter if the server returned an error as long as the error
3035 is something other than VRESTARTING. If there is such an error, all
3036 connections on that server should be marked for replacement.
3038 * Change the check "up" servers period from once per hour to once every
3039 ten minutes. This matches the behavior of the Unix cache manager.
3041 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3042 * Byte Range Locking support has been added. Architectural documentation
3043 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3044 made to enforce Windows mandatory locking model using locally managed
3045 byte range locks with full file locks obtained from the file server.
3046 Read-only volumes provide the user with an implied locking privilege.
3047 Read-write volumes require that a user be granted the locking privilege
3048 in order for a lock to be obtained. As most Windows applications use
3049 Shared file access modes, they require the ability to obtain locks in
3050 order to access files.
3052 * Implement error translation functions to convert Universal Error Codes
3053 output by modern servers to local errno and winsock error values.
3054 This fixes the error message problems using pioctl() calls from fs.exe
3055 and the AFS Shell Extension.
3057 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3058 * Allow the AFS Client Service to set the firewall configuration by
3059 retrying until after Windows determines the boot period is over.
3061 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3064 * Detect IP address changes from within the AFS Client Service for the
3065 purpose of checking down servers sooner
3067 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3068 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3069 user that there is a Kerberos key problem. Previously the user would
3070 receive no feedback when there was something wrong.
3072 * Fix the large integer library support to properly handle values greater
3075 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3076 * Updated CellServDB
3078 * Fixed file name pattern matching. Question marks can match nothing
3079 if found at the end of a component.
3081 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3082 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3084 * The MSI installer now has a new Product ID for each build. This
3085 is to allow for automated upgrades. The Upgrade ID is remaining
3088 * Volumes that were cached and then subsequently were salvaged,
3089 moved or otherwise taken temporarily offline (as opposed to being
3090 busy) would become unreachable. A distinction is now made between
3091 a volume being offline and a server being down.
3093 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3095 * Fixed several AFS Shell Extension issues:
3096 - Checking Server Status would only work if a specific cell was
3098 - If servers were down, the Server Down dialog would write to
3099 invalid memory and cause explorer.exe to crash
3100 - Default Push Buttons were not specified for many dialogs
3101 - The extension would be unloaded by the Explorer Shell and
3102 never be reloaded under some circumstances.
3104 * Fixed the AFS System Tray menu to disappear if another Window
3105 was selected before a menu item was selected
3107 * Added volume owner and group information to "fs examine" output
3109 * Added fs and registry support for enabling/disabling/clearing
3110 rx statistics gathering.
3112 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3114 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3115 GUIDs for each platform.
3117 * Documentation directory structure and default installation rules
3120 * New OpenAFS for Windows HTML Release Notes installed
3121 as part of the documentation set.
3123 * Add thread id to rx debugging messages
3125 * When all servers are offline, return a bad network path error
3126 immediately. This is necessary to allow Explorer to do the
3127 right thing and not hang. The background daemon thread will
3128 check the down servers every 30 seconds to see if they are
3131 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3133 * The AFSCache file has been modified to store both the
3134 serial number of the volume on which the file resides
3135 and the SID of the local machine. If either change,
3136 a new UUID is generated.
3138 * add "fs flushall" command that forces all data buffers
3141 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3143 * add support for '/' instead of '\' in pioctl() calls
3145 * Apply AFS Client Admins group protect to AFS Shell Extension
3147 * Add support for \\afs\<name> to most AFS Shell Extension
3148 commands except the symlink methods.
3150 * when installing the Microsoft Loopback Adapter, enable
3151 MS Client for Networks and prevent an install failure
3152 by not calling CoInitialize twice in the same thread.
3154 * reload cell vldb values from the CellServDB every two
3155 hours in case it changes
3157 * When updating cell information from DNS, be sure to set
3160 * Add support to allow use of \\AFS\<foo> where <foo> is
3161 either a mount point or symlink. As <foo> is normally
3162 treated as a share name, we transform it into \\AFS\all\<foo>
3165 * Init 'code' to prevent false errors when integrated logon
3166 is disabled and the service is not running
3168 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3170 * Removed trace log messages that were hampering performance.
3172 * Fixed a deadlock that was being triggered by editing Word
3173 documents stored within AFS with WinWord 2003.
3175 * Bit 3 of the TraceOption registry value is now used to set the
3176 default for "fs trace" in the non-Debug builds. The new default
3177 is off for release builds and on for debug builds.
3179 * Bit 2 of the TraceOption registry value can be used to turn on
3180 real-time output of debug log entries to the Windows Debug
3181 Output monitor interface. This data can be viewed with tools
3182 such as Sysinternal's DbgView.
3184 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3185 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3186 a resource we should not be returning status codes such
3187 as Network Busy or Remote Host Down because these errors
3188 will imply that the AFS Client Service is Busy or Down
3189 and that is simply not the case. Instead we will return
3190 Bad Network Path as the path is temporarily not available.
3192 Instead of returning Sharing Paused when there is a
3193 timeout error, return Timeout. Once again, the AFS Client
3194 Service is not paused.
3196 * afscreds.exe would display an Obtain Creds dialog when
3197 the expired credentials reminder was triggered even
3198 if there was no network path to the KDC. This is prevented
3199 by adding KDC probe logic to the reminder thread.
3201 * afscreds.exe would display expired tokens no differently
3202 than unexpired ones. This would make it difficult for the
3203 user to distiguish when the tokens were expired. For the
3204 English build added a new resource string "(expired) that
3205 is displayed instead of the expiration time.
3208 * "fs wscell" when executed in freelance mode will return the
3209 name of the cell configured in the registry. The root.afs
3210 volume is not loaded from this cell, but it is used for the
3211 default for aklog and integrated login.
3213 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3214 no longer work on \\AFS\all when freelance is being used
3215 unless the user is a member of the "AFS Client Admins" Windows
3218 * some more checks performed during persistent cache validation.
3221 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3222 down servers for the current cell not all cells.
3224 * If the scache LRU list becomes corrupted, the AFS Client Service
3225 might enter an infinite loop at startup. Detect the corruption
3226 and discard the cache contents in this case.
3228 * Fixed the Explorer Shell's remove mountpoint functionality.
3230 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3231 obtain a mutex lock on a cm_scache_t object whose lock is held by
3232 a thread attempting to obtain the cm_aclLock.
3235 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3236 misinterpretted as "csail.mit.edu" when the resolver
3237 is configured to append ".mit.edu" to failed DNS queries.
3239 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3240 that can be used to force the preservation of user tokens upon logout.
3242 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3243 recommended installation categories: AFS Client, AFS Administrator,
3244 AFS Server, AFS Development Kit. Each category includes a different
3245 default subset of the OpenAFS components.
3247 The OpenAFS logo is now associated with the NSIS uninstall entry
3248 in the Add/Remove Programs control panel.
3250 * The user name associated with AFS tokens when obtained with
3251 integrated login, the afs systray tool, or aklog will always
3252 include the full Kerberos 5 user principal name regardless of
3253 whether or not the cell is local to the realm.
3255 * Modify integrated login so that it does not enter an infinite
3256 loop if the service is not set to auto start.
3258 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3261 * Added uninstall.exe to Wix installation
3263 * More modifications to algorithms used to wake sleeping threads.
3267 * Changes to the algorithms used to wake threads when they are
3268 put to sleep because of conflicts over buffer operations.
3271 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3272 the process of producing AFS tokens. The tokens are stored within
3273 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3274 New functionality has been added to temporarily copy the tickets to
3275 a file ccache during the short window between the NPLogonNotify and
3276 WinLogon Logon Event. The Logon Event handler starts a process as
3277 user, afscpcc.exe, that copies the tickets into the default ccache.
3278 The file ccache is then destroyed by the event handler.
3280 * During synchronization operations on stat cache entries, if threads
3281 are forced to sleep it was possible that the threads would not be
3282 woken under heavy load.
3284 * Enforce the delete on close flag specified by the NTCreateX()
3287 * A race condition in the rx library was fixed that resulted in
3288 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3289 was cleared when threads were still waiting to use the
3290 connection. Also, fixed a potential case in which the
3291 connection could be destroyed while threads were waiting to use it.
3293 * The DNLC freelist has been seen to become corrupted with still
3294 active entries being stored on the freelist. Changes were made
3295 to perform a better job of cleaning entries before placing them
3296 onto the freelist; marking them with the correct magic value when
3297 purging the list; and allowing the list to be purged on startup
3298 if the cache validation fails.
3300 * Windows Crash Reporting does not get triggered for afsd_service.exe
3301 because it provides its own top level exception handler. This patch
3302 places an explicit call to ReportCrash() on platforms which support
3305 If you configure Crash Reporting via AD Policy to capture crash
3306 reports within the domain, then you will receive this output.
3308 * Generate MiniDumps independent of the Windows Crash Reporting.
3310 (a) If an exception occurs,a minidump will be generated at the path
3311 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3312 can be set to any other type via use of the
3315 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3317 (b) MiniDumps can be generated at any time using the "fs minidump"
3318 command. This allows you to generate a minidump without
3319 stopping the process or even requiring the use of a debugger
3322 * an RX connection which reports bad ticket now treats it the same way
3323 the expired ticket is treated. The ticket is removed from the
3324 connection and a new anonymous connection is established.
3326 * fs memdump now requires that the caller be in the AFS Client Admins
3329 * additional debugging information is provided in cm_Analyze when the
3330 error code is ignored.
3332 * Fix a regression introduced into the pioctl() client code which
3333 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3334 MIT KFW is not installed. (1.3.82a)
3336 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3339 * Fix a race condition in cm_EndCallbackGrantingCall() that
3340 could leave 'nrevp' pointing to freed memory. In the process,
3341 optimize the processing to ensure that the applications monitoring
3342 the status of the stat cache entry are only notified at most
3345 * Conditionalize access to the Explorer Shell AFS->Mount Point
3346 ->Remove menu item. Disable the item if the selected item(s)
3347 are not mount points.
3349 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3350 Local Storage. This eliminates thread contention on the
3351 global RX free packet queue. This should improve performance
3352 on multi-CPU systems.
3354 * Fix the IP address filtering code to properly load/unload
3355 the IPHelper DLL so that Windows XP/2003 and beyond systems
3356 do not have to manually probe the registry.
3358 * CleanupACLEnt() was not being called consistently with
3359 the cm_scache_t object referred to by the back pointer
3360 mutex locked. This could in very rare conditions lead to
3361 an invalid memory access.
3363 * Added a script command to the msi installer "afs_replace"
3364 which can be used to replace one version of OpenAFS with
3365 another without requiring an uninstall and reinstall if
3366 the installed msi is no longer in its original location.
3368 * Apparently the problem with multi-domain forests with cross-
3369 realm trusts to non-Windows realms was not entirely solved.
3370 The authentication to the AFS SMB service failed because
3371 the wrong name was being used. Using ASU as an example,
3372 the authentication was being performed with the name
3373 "QAAD\user" (an account in the forest root) and not
3374 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3376 The solution was to add an additional dependency on KFW
3377 in order or to be able to easily obtain the client principal
3378 name stored in the MSLSA ccache TGT. This information is
3379 used in two locations:
3381 - the pioctl() function
3383 - a new WinLogon Event Handler for the "logon" event.
3385 The pioctl function will now be able to use the correct
3386 name when calling WNetAddConnection2() and the "logon"
3387 event handler will now be able to call WNetAddConnection2().
3388 The hope is that the "logon" event handler will be called
3389 before the profile is loaded but I have not guarrantee
3393 * Fixed a locking error in cm_TryBulkStat() which had the
3394 potential of deadlocking the system for the length of time
3395 it takes to perform a bulk status fetch operation.
3397 * Replaced time conversion code (UnixTime <-> FILETIME) to
3398 be completely arithmetic instead of relying on a bizarre
3399 algorithm involving a variety of C RTL time functions.
3400 This has the side effect that UnixTime and FILETIME which
3401 are both stored in UTC are interpretted as UTC throughout
3402 the year. Windows will apply the same localization to AFS
3403 as it does to NTFS. Applications which rely on the ability
3404 to sync files between the two file systems will no longer
3405 see the timestamps of files in AFS change an hour relative
3406 to the files stored in NTFS or Windows based backup devices.
3408 * Fixed a invalid memory access under a bizarre circumstance.
3409 Windows will allow a physical mass media device to be
3410 installed via Plug N Play to the system and will assign it
3411 the lowest available drive letter. This is true even if the
3412 drive letter is currently assigned in the user session to
3413 a network device via NET USE (or its equivalent.) When this
3414 happens, queries sent to the CIFS server will contain invalid
3415 data. This invalid data was not being caught by the AFS
3416 Client Service and was resulting in a crash when Freelance
3419 * Fixed a reference count error when registering callbacks
3420 on a stat cache object if the callback was already registered
3422 * Add a case to cm_Analyze() in order to handle the case where
3423 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3424 the server list for the volume is empty. In this case, force
3425 an update of the volume info and retry.
3427 * Insert a missing cm_EndCallbackGrantingCall() which could
3428 result in threads waiting for a callback to complete to never
3431 * In the persistent cache, there is a maximum number of volume
3432 entries. Allow volume entries to be reused if the maximum
3433 number have been allocated and their reference count is zero.
3435 * If we already have a dead virtual connection object, don't
3436 ignore the fact that additional dead vc objects must be taken
3439 * Removed a deadlock condition introduced in the summer of 2004.
3440 Do not hold mx locks around calls to RX functions including
3441 rx_NewCall(). That is what reference counts on the rx_connection
3444 * Fixed an initialization error in afslogon.dll which could
3445 result in random behavior including a failure to terminate
3446 the NPLogon function. (1.3.80b)
3448 * Fixed an error preventing the use of SMB authentication
3449 on Windows 2000. (1.3.80a)
3452 * Updated CellServDB to Public CellServDB 16 Mar 2005
3454 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3455 misinterpretted as "csail.mit.edu" when the resolver
3456 is configured to append ".mit.edu" to failed DNS queries.
3458 * Fixed another case in which the client would replace connections
3459 to the host when it was not necessary. In this case the
3460 situation would occur if cryptall was on and the connections
3461 were unauthenticated due to lack of an appropriate token.
3463 * OpenAFS for Windows has failed to work at sites which are
3464 utilizing a cross-realm trust between an MIT/Heimdal realm
3465 and a multi-domain Windows forest when the workstation being
3466 accessed is not located in the root domain. This is caused
3467 by a bug in the workstation which was triggered after the
3468 introduction of Windows 2003 Server. When the bug is triggered,
3469 the workstation attempts to authenticate users to afsd_service.exe
3470 by contacting the Domain Controller instead of using the
3471 LSA loopback authentication mechanism.
3473 One of the reasons this bug occurs is because the workstation
3474 does not have a reliable way of knowing that the service whose
3475 netbios name is "AFS" is located on the workstation. This will
3476 be fixed starting in Longhorn Beta 1 by Microsoft. The
3477 "BackConnectionHostNames" registry value will be used to
3478 indicate that the authentications to that service name should
3479 be performed using the loopback authentication mechanism.
3481 In the meantime, when Logon Caching is enabled, we can force
3482 afsd_service.exe to authenticate using the logon cache before
3483 contacting the Domain Controller. This will work with both
3484 password and smart card based logons.
3486 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3487 flag should not be cleared if a volume's server reference is
3488 marked as "offline". In the case where all of the volume's
3489 servers are either "down" or the volumes are "offline", we want
3490 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3491 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3492 should never occur in practice.
3494 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3495 5 seconds, clear the server down and volume busy flags, and
3496 then force an update from the VLDB. This allows the client to
3497 update the location of a volume if the reason for it being
3498 marked offline is because it is being moved. Calls to
3499 cm_ConnByMServers() will be retried either until success or
3500 the RDRtimeout period is reached.
3502 * Correct the Power Management code behavior in response to
3503 Standby, Suspend, and Shutdown events. Instead of flushing the
3504 buffers associated with the mounted SMB submounts, simply write
3505 all buffers which are marked dirty.
3507 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3508 provides a list of cells other than the default cell for which
3509 tokens should be obtained using the default Kerberos principal.
3510 This functionality is logon domain specific and is only available
3511 when using KFW for authentication.
3513 * Fixed FindNext Invalid Handle error caused by over agressive
3514 attempts at garbage collecting dirSearch entries when the
3515 dirSearch ID wraps from the maximum value to 1.
3517 * Add support for registry defined server preferences for VLDB and
3518 File servers. See registry.txt for details.
3520 * Increased default cache size to 96MB and default number of cache
3523 * Fixed refCount leaks related to directory lookups and pioctl calls.
3525 * Callbacks revoked during a race condition with an attempt to
3526 obtain the same callback no longer result in an inaccessible
3529 * New command line tool:
3531 afsdacl : Set or reset the DACL to allow starting or stopping
3532 the afsd service by any ordinary user.
3534 Usage : afsdacl [-set | -reset] [-show]
3535 -set : Sets the DACL
3536 -reset : Reset the DACL
3537 -show : Show current DACL (SDSF)
3539 * IP addresses are no longer queried once at startup. Instead IP
3540 addresses are obtained as needed. Loopback adapter addresses are
3541 no longer published to the server.
3543 * Pay attention to the MIT KFW registry configuration for automatic
3544 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3546 * Fix (once again) case-sensitive comparisons which was apparently
3549 * Activate support for vos listvol -format
3551 * Population of the "AFS Client Admins" group failed on non-English
3552 versions of Windows because the name of the "Administrators" group
3553 is localized. Now we lookup the name of the group by using the
3554 Administrators Group Alias Relative ID.
3556 * Multiple cell token acquisition within afscreds.exe was broken in the
3557 case of Kerberos 5 cross realm authentication. Instead of contacting
3558 the KDC associated with the cell's realm, afscreds would obtain a
3559 token from the KDC of the user's realm. This would result in
3560 "invalid kvno errors" while authenticating to the AFS servers.
3561 Unauthenticated access would work. The symptoms would vary based upon
3562 whether or not the VLDB servers had been contacted using unauthenticated
3563 connections prior to the user obtaining tokens.
3565 * The list of ACL entries was becoming corrupted. It appears as if
3566 an ACL when it expires was not being appropriately cleaned up. In
3567 fact, it was left in the list of ACLs associated with the scache
3570 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3571 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3572 and "amd64_win64" for amd 64-bit processors.
3574 * When executing executables, Windows provides a mechanisms by which
3575 the normal search for DLLs can be bypassed. If a file foo.exe is
3576 being executed and there exists a file or directory "foo.exe.local"
3577 (the contents are ignored if a file), then Windows will search for
3578 DLLs first in the "foo.exe.local" directory and second in the
3579 directory in which the "foo.exe" file is located.
3581 Previous releases of OAFW would improperly return
3582 STATUS_NOT_A_DIRECTORY instead of the expected
3583 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3584 to terminate the search for the DLL causing the execution of the
3585 application to fail.
3587 * Rework the reference counting of smb_vc_t objects. The references
3588 stored in the waiting locks were not counted. This could result in
3589 an assertion if the reference count drops to zero.
3591 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3592 when running in freelance mode.
3594 * Added support for persistent caching of file contents, stat entries,
3595 volume data, cell data, and name lookup data. When the registry
3596 value, "NonPersistentCaching", is not defined or set to 0 the file
3597 specified by "CachePath" becomes a persistent cache file. The
3598 size of the cache file is computed at run time. It has a maximum
3599 size of approximately 1.2GB. The contents of the file will be
3600 validated according to the rules specified by the "ValidateCache"
3603 * Added support for UUIDs. UUIDs are kept for as long as the
3604 cache file is intact.
3606 * Added cmdebug.exe and support cache manager debugging callback
3607 interfaces in afsd_service.exe.
3609 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3610 [-addrs] [-cache] [-help]
3611 Where: -long print all info
3612 -addrs print only host interfaces
3613 -cache print only cache configuration
3615 * Symlinks to \\AFS\[all\]... will now be treated the same as
3616 symlinks to /afs/.... However, please use /afs/... as otherwise
3617 the symlinks won't work on Unix.
3619 * Correct a problem with local CellID allocation for cells whose
3620 server list is obtained via DNS instead of a CellServDB file. If
3621 the DNS information expires the CellID assigned to the entry will
3622 be changed. This causes all of the cm_scache_t objects which refer
3623 to the old cellID number to become useless. Attempts to access
3624 files or directories with cm_scache_t objects using the old cellID
3625 will fail since the server list cannot be obtained.
3627 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3628 be triggered if the volume referenced by the cm_scache_t object is
3631 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3632 in addition to SERVICE_ACCEPT_STOP.
3634 * Move RPC shutdown until after the SMB and RX shutdown procedures
3635 complete. Block until RPC shutdown is complete.
3637 * Modify afslogon.dll (integrated logon) to wait for service if its
3638 state is START_PENDING. If the timeout period occurs, reset to
3639 the retry interval and not the sleep interval.
3641 * When renewing the server list for a cell obtained via DNS AFSDB
3642 records, the cm_cell_t entry must be removed from the list of all
3643 cells. Otherwise, the list of cells will be corrupted.
3645 * In the dcache and scache modules, use the cm_scache_t dataVersion
3646 instead of the cm_fakeDirVersion.
3648 * Synchronize fs.c with the unix version.
3650 * The variable used to determine whether a file or virtual memory
3651 mapped cache is used was not properly initialized to a default
3652 value. If the registry setting "NonPersistentCaching" was not
3653 set, the choice would be random. Properly initialized to be
3656 * The memory mapped view was never unmapped before closing the file
3657 at service shutdown. This is now properly cleaned up.
3659 * Default location of Cache file is now %TEMP%\AFSCache
3663 * A bug affecting new installations of 1.3.75/76 would result in
3664 the creating of incorrect mountpoints in the freelance root.afs
3665 volume for the default cell.
3667 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3668 of "root.cell", you have been affected by the bug. To correct the
3669 problem, execute the following commands:
3671 fs rmmount \\afs\all\<cellname>
3672 fs rmmount \\afs\all\.<cellname>
3674 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3675 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3679 * A bug has been fixed in the auto-generation of Freelance root.afs
3680 symlinks which produced random entries in the \\AFS\all directory.
3682 * Support has been added for multi-homed servers
3685 * Added a new registry value, "StoreAnsiFilenames", which can be used
3686 to force the use of ANSI character sets instead of OEM Code Pages.
3687 This feature is useful when users require the ability to create
3688 filenames with 8-bit characters and need to access the files from
3689 both Latin-1 based Unix systems as well as from Windows.
3691 Activation of this feature will prevent access to files stored with
3692 8-bit OEM characters.
3694 * Shutdown all SMB threads in a synchronized manner when stopping the
3697 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3698 by the largest contiguous block of unused memory within the 2GB process
3699 space which can be assigned to the memory mapped file. Unfortunately,
3700 when the executable digital signature verification code is activated
3701 Windows sees fit to further segment the process memory which in turn
3702 reduces the size of the maximum cache file to less then 800MB. If
3703 larger cache sizes are desired, a new registry value should be set:
3705 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3707 Setting this value will disable the runtime verification of digital
3708 signatures on afsd_service.exe and the afs dlls which it loads. It
3709 will not disable the the version number check on those same files.
3710 The signature verification is not a security messure and is only meant
3711 to enhance the ability to afsd_service.exe to detect potential
3712 destablizing mixtures of DLLs from incompatible distributions.
3714 Added code to auto-disable the signature verification check if
3715 the desired cache size is greater then 700MB.
3717 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3718 used to verify the validity of the afsd_service.exe binary
3719 as well as each of the AFS DLLs loaded by the service. Not only
3720 must the digital signature be valid but the signatures of the
3721 DLL must be signed by the same entity as the service.
3723 * Implement new functions: cm_freelanceMountPointExists and
3724 cm_freelanceSymlinkExists. Use them along with other validity
3725 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3726 ensure that name collisions do not occur and that empty strings
3727 are not valid file names.
3729 A symlink may not have a name which would resolve to a valid
3730 cell name. Doing so would prevent access to the cell.
3732 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3733 re-initialization code. The reference counts of the fake root.afs
3734 volume scache object(s) would become invalid when the mount point
3735 or symlink lists were altered.
3737 * Add registry entries to provide mappings from the afsdsbmt.ini
3738 to the new locations for applications which count on the use
3739 of the old Profile file APIs. These apps are likely to fail
3740 if the user does not have administrator privileges and the
3741 registry is locked down.
3743 * The afs_config.exe submounts dialog had two errors.
3744 First, attempts to remove entries failed because the registry
3745 key was being opened without KEY_WRITE privileges.
3746 Second, when editing a submount entry, changing the name
3747 would add a new key and leave the original one in place.
3748 Now the original submount will be removed if its name is
3751 * In recent months there have been several incidents in which
3752 users have experienced problems starting or accessing
3753 afsd_service.exe and after significant effort has been spent
3754 it has turned out that they have two versions of AFS on the
3755 machine or an inconsistent set of DLLs.
3757 Code has now been added to afsd_service.exe which will walk
3758 the list of modules loaded by afsd_service.exe and validate
3759 that the version of the AFS DLLs matches the version of the
3760 afsd_service.exe executable. If they do not match the service
3763 * When Freelance mode is enabled and there is no registry
3764 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
3765 will attempt to import the afs_freelance.ini file contents.
3766 If the file does not exist, it was creating a dummy file
3767 with a r/o and r/w entry for the default cell and then
3768 importing those values.
3770 This process has been changed. The temporary file is no
3771 longer created. Also, both the OpenAFS Client install
3772 directory as well as %WINDIR% are checked for previous
3773 afs_freelance.ini files.
3775 * Added support for VL_GetEntryByNameN(). Still need to add
3776 support for VL_GetEntryByNameU() for multi-homed support.
3778 * Fix a deadlock situation in afscreds.exe when canceling an
3779 auto-generated Obtain Tokens dialog
3782 * Fix the locking of objects during Directory Searches in the
3783 SMB/CIFS server. The failure to properly lock the reference
3784 counts was resulting in the premature freeing of smb_dirSearch_t
3785 objects while they were still in use by the SMB/CIFS client.
3786 This does not solve the "Invalid Handle" problem.
3788 * Fix Find Cell By Name pioctl call to return a valid cell
3789 name for the Freelance fake root.afs volume.
3790 "Freelance.Local.Root".
3792 * Fix the Explorer Shell Extension Symlinks->Add operation.
3793 The dialog template was missing and the link destination
3794 string was too short.
3796 * Add support for symlinks to Freelance root.afs volume
3797 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
3798 <number> = "<linkname>:<relative-path>."
3799 Use symlink.exe to create, list, or remove
3801 * Remove the fallback to the use of KFW's KRB4 library when obtaining
3802 tokens. We never obtain KRB4 tickets.
3804 * Fix AFS Client Configuration Control Panel to support new SysName
3807 * Fix a bug in afsd_service.exe which could result in the SysName
3808 not being read from the registry.
3811 * Add code to block the issuance of AFS tokens by aklog.exe or
3812 afscreds.exe when the Kerberos 5 principal name contains a dot.
3814 * Modify the IsAdmin() function to always treat the local SYSTEM
3815 account as an AFS client administrator. Affects fs.exe and
3818 * Modify the internal handling of Quota Exceeded errors
3820 * Upgrade all reference count fields in the Windows cache manager
3821 and the osi library to use unsigned long instead of signed short.
3822 A similar fix has been applied to the afs rpc (rx) library.
3824 * fix the Windows cache manager to prevent it from replacing the
3825 rx_connection object associated with the cm_conn_t object on each
3826 and every operation if "fs crypt" was set. This explains the
3827 dramatic performance difference when crypt is used vs clear.
3828 The problem: 'cryptall', a boolean flag indicating whether or not
3829 "fs crypt" is set, was being compared to the rx_connection
3830 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
3831 1 != 2 and therefore the rx_connection was always destroyed
3832 and replaced on each and every operation.
3834 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
3835 It is not safe for the cm_conn_t object to not be locked because
3836 rx_DestroyConnection might be called from another thread if:
3837 - the user's tokens have changed or expired
3838 - the crypt mode has changed
3840 This fix appears to have also taken care of the problems associated
3841 with Overlapped Writes resulting in Delayed Write errors.
3843 * fix NSIS installer's AdminGroup.exe to properly create and
3844 remove groups when given -create or -remove. The string comparison
3847 * fs sysname now accepts a list of sysname values
3849 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
3850 DWORD which when set to a non-zero value will cause error message
3851 text to be output to stderr from the pioctl() routine. Useful in
3852 debugging failures of fs.exe, tokens.exe, etc.
3854 * added a test to the power management code to only perform a
3855 flush operation if there is at least one network adapter which
3856 is not a loopback adapter.
3858 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
3859 "EnableKFW". This value will not be read if the key
3860 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
3861 value under that key does not.
3863 * provide mechanisms to force the use of krb524d for Kerberos 5
3864 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
3865 there is a new registry value "Use524" and for aklog.exe a new
3866 command line parameter "-m".
3868 * Fix the pattern matching algorithm to properly match patterns
3871 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
3872 which is used by both the original function and the new
3873 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
3874 creation of HardLinks in addition to Renaming. smb_Link()
3875 is a new function which creates HardLinks via cm_Link().
3876 cm_Link() is a new vnodeops function which creates links
3879 smb_ReceiveNTRename() does not support the File Copy and
3880 Move Cluster Information operations described in its interface.
3881 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
3883 * When opening files via symlinks, we should follow the symlinks
3884 until we reach the actual file stat cache entry. The stat cache
3885 entry of the file should then be stored in the FID instead of
3886 stat scache entry of the symlink.
3888 * return bad operation errors for all unimplemented functions
3889 even if we do not know the functions exist.
3891 * Log bad packets and unknown operation packets to the trace log
3893 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
3896 * Update list of known CIFS operations to include all those listed
3897 in CIFS-TR-1p00_FINAL.pdf.
3899 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
3900 to support the REG_EXPAND_SZ type.
3903 * A new Windows authorization group "AFS Client Admins" is now
3904 created and populated with the members of the "Administrators"
3905 group. The group is used to determine which accounts on the
3906 machine may be used to modify the AFS Client Configuration via
3907 the UI and command line tools. afs_config.exe, fs.exe,
3909 * Modify the WinLogon Logoff Event Handler to query NT4 domain
3910 controllers for the remote profile path if Active Directory
3911 services are not available.
3913 * Fix aklog.exe to not add the AFS ID to the username
3915 * PTS registration of new users to foreign cells has been added to
3918 * The cm_Daemon thread is used to perform checks for
3919 down servers, up servers, volumes, callback expirations,
3920 lock maintenance and token expiration. Due to a gaff in
3921 larger integer division the thread never performed any
3922 work. Instead the current time computation would always
3923 be less then the trigger times. This had an adverse affect
3924 on the client's ability to maintain communication with servers,
3925 keep volumes up to date, and flush user tokens and acls
3926 when they have expired. This was broken when the 1.3 branch
3927 was modified to support VC7 which no longer included
3930 * An initialization problem with the Freelance code was
3931 detected while fixing the callbackRequest. The cm_rootSCachep
3932 object is obtained during afsd_InitDaemons() but the callback
3933 information is incomplete. The callback information will not
3934 be obtained until cm_MergeStatus is called from within
3935 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
3936 test for the conditions under which the callback information
3939 * Reports have been filed indicating that callbacks were
3940 being lost. An examination of the code indicated that the
3941 cm_server_t objects were not being properly reference
3942 counted by the cm_scache_t and cm_callbackRequest_t objects.
3943 In particular, the cm_server_t objects may have been freed
3944 from beneath the cm_conn_t objects.
3946 All of the reference counting is now done via the functions:
3951 this improves the ability to track the referrals.
3953 Each cm_BeginCallbackGranting Call now allocates a reference
3954 to the cm_server_t. The cm_EndCallbackGrantingCall either
3955 frees the reference or transfers it to the cm_scache_t
3956 cbServerp field. These are then appropriately tracked
3957 through the cm_Analyze call.
3959 * Ensure that the dnlc hash table is the same size as the
3960 dir name hash table (as per original author's note).
3961 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
3962 for compatibility with 64-bit systems.
3964 * fix smb_ApplyV3DirListPatches to properly apply the hidden
3965 attribute to dotfiles when the infoLevel < 0x101 and
3966 cm_SyncOp has failed.
3968 * Fix the Freelance registry initialization code. There
3969 was a possibility that some systems could end up with
3970 garbage in the registry during a clean install.
3973 * file and directory names beginning with "." will now be given the
3974 hidden attribute when the volume access is anonymous. this matches
3975 the behavior when the volume access is via an authenticated user.
3977 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
3978 key. When a change occurs mark the root.afs data as invalid and
3979 for it to be reloaded on the next access. This allows administrators
3980 to modify the mount point list without restarting the service.
3982 The freelance client used to provide a fake modification time for
3983 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
3984 Added code to extract the last modification time of the Freelance
3985 registry key and use that instead. The time now represents the
3986 most recent mount point change.
3988 * PTS registration of new users to foreign cells has been added to
3991 * Additional Cache Control and Credential Manager options have been
3992 added to the WiX installer. See deployment guide for details.
3994 * The CachePath setting is now optionally a REG_EXPAND_SZ type
3996 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
3999 * The loopback installation code may have had a problem updating the
4000 %ETC%\HOSTS file which could have resulted in a premature failure.
4001 Work around code has been added for the case where the file cannot
4004 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4005 because Windows sends 64K blocks when using overlapped writes.
4007 * The default number of server threads was increased from 4 to 25 to
4008 better handle overlapped writes.
4010 * The "AfscredsShortcutParams" registry value was not being properly
4011 loaded by afscreds.exe. Therefore, the default value was always being
4012 used instead of the value set by the installer.
4014 * Windows XP provides downgrade attack detection to prevent an attacker
4015 from being able to force the use of NTLM simply by disrupting
4016 communication with the KDC. This attack cannot exist between the
4017 Windows CIFS client and the AFS Client Service. Therefore, when a
4018 downgrade has been detected the afs pioctl library will force the
4019 establishment of a new CIFS connection using NTLM.
4021 * A locking error was discovered surrounding all references to volume
4022 server lists within the cm_cell.c source file.
4024 * The logged into Windows username was incorrect on Terminal Server
4027 * A new registry value "NonPersistentCaching" was added to the service
4028 parameters key. When set to a non-zero value, the afs cache is stored
4029 in the Windows paging file. There are two limitations to choosing
4031 1. when persistent caching is implemented it won't work with
4032 this flag set since there will be nothing to persist.
4033 2. with this flag set the initial paging allocation cannot be
4034 changed while the service is running
4036 * An initialization bug was discovered in aklog.exe which affected users
4037 who have a domain name for their afs servers which could not be mapped
4041 * afs_config.exe now validates cell names against DNS in addition
4042 to the CellServDB file.
4044 * In order to allow the freelance client to connect to a volume with ID
4045 equal to 1 on the default cell we changed the fake root.afs volume ID
4046 once again. This time we choose 0xFFFFFFFF. In addition, we change
4047 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4048 It will now be impossible for a volume ID to match that of another
4049 cell unless the client is connected to 0xFFFFFFFD cells. That should
4050 be enough room for growth.
4052 * Fix "fs mkmount" command to work with UNC paths and when
4053 started from non-AFS drives. It is now possible to create a mount
4054 point in the freelance fake root.afs volume with the command
4056 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4060 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4061 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4063 * The algorithm used to re-attempt access to the servers associated with
4064 a volume has been altered to properly address the case in which all
4065 servers have been marked down. The previous algorithm did not reset
4066 the server's down flags so the servers were never actually retried.
4067 This caused a problem with active volumes if the network connectivity
4068 was lost as could be the case with a network cable removal, wireless
4069 drop, or laptop hibernation. With the fix volume access is restored
4070 almost instantenously when network connectivity becomes available.
4072 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4073 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4074 listing of AFS submounts and freelance mount points. Support for
4075 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4076 significantly enhances the behavior of AFS volumes within the Explorer
4077 Shell. For instance, "AFS" now shows up as server in the Explorer
4078 with each submount or freelance mount point visible as a share.
4079 The right click menu in each folder now works with full functionality
4080 on a consistent basis.
4082 * The network provider can be configured to have different behavior
4083 depending on the domain that the user logs into. These settings are
4084 only relevant when using integrated login. A domain refers to an
4085 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4086 local machine (i.e. local account logins). The domain name that is
4087 used for selecting the domain would be the domain that is passed into
4088 the NPLogonNotify function of the network provider. (see registry.txt
4091 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4092 "Authentication Cell" which may be used to specify a default
4093 authentication cell for afscreds.exe which is different from
4094 the default cell for the AFS Client Service daemon.
4096 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4097 afslogon.dll moved to %WINDIR%\System32\.
4098 New registry entries added to register the dll for Winlogon events.
4100 The logoff event will now force a call to ktc_ForgetAllTokens()
4101 using the context of the user being logged off as long as the
4102 user's profile is not loaded from within AFS. If the profile
4103 was loaded from AFS we can't release the tokens since the Logoff
4104 event is triggered prior to the profile being written back to
4105 the its source location. This is now performed in an XP SP2
4108 * Windows XP SP2 Internet Connection Firewall interoperability
4111 * The %WINDIR%\afsdsbmt.ini contains four sections:
4112 Submounts, Drive Mappings, Active Maps and CSC Policies.
4113 The Submounts and CSC policies are now stored in the registry under
4114 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4115 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4116 The Drive Mappings and Active Maps are stored in the registry under
4117 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4118 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4120 There is no automatic migration of this data as it would be impossible
4121 to consistently migrate data to user profiles which may not be active
4122 when the machine is updated.
4124 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4125 fake root.afs volume. For the same reasons as for the cellservdb file,
4126 this information should not be in %WINDIR%. This information is now
4127 kept under the registry key
4128 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4130 The data from the afs_freelance.ini file will be automatically
4131 migrated to the registry on first execution of afsd_service.exe
4133 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4134 troublesome for several reasons. One, it is confusing for those who
4135 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4136 Two, this file is not a Windows Profile formatted file. Three,
4137 applications should not be reading or writing to %WINDIR%. It causes
4138 problems for Windows Terminal Server.
4140 The new location of CellServDB will be the OpenAFS Client install
4141 directory which is by default C:\Program Files\OpenAFS\Client and can
4142 be determined by querying the registry for
4143 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4145 The existing afsdcell.ini will be migrated by the NSIS installer.
4146 The Wix installer must still be updated to do the same.
4148 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4149 High Security mode; and to add Terminal Services compatibility registry
4150 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4151 configuration files in %WINDIR%.
4153 * Add support for authenticated SMB connections. This will remove
4154 the need for high security mode in most situations. Both NTLM
4155 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4156 only NTLM can be used even though Kerberos is now supported. The
4157 reason is that it is not possible to construct a service principal
4158 which is unique to each individual machine.
4160 SMB Extended Auth does not work on XP SP2 unless one of two registry
4161 modifications are made:
4163 (1) To disable the check for matching host names on loopback connections
4164 set this key. This does not require a reboot:
4166 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4167 "DisableLoopbackCheck"=dword:00000001
4169 (2) To add the AFS SMB/CIFS service name to an approved list. This
4170 does require a reboot:
4172 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4173 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4175 afsd_service.exe will automatically add the current Netbios Name
4176 to the BackConnectionHostNames list and then temporarily disable
4177 the loopback check for one cycle of startup/shutdown of the service.
4178 We assume most folks do not start/stop without a reboot so this
4179 will be adequate in most cases.
4181 * Fix security hole in afslogon.dll which allowed passwords to be
4182 sent in clear text to the KDC in a misformed principal name.
4184 * Fix cm_GetCell() to properly handle expired dns entries
4187 * If Freelance mode is active and the afs_freelance.ini
4188 file does not exist, do not create an empty file.
4189 Instead create a file containing ro and rw mountpoints
4190 to the default cell using the standard conventions.
4192 * Modify the Freelance support to handle the ability
4193 to create rw mount points in the fake root.afs volume.
4195 * Changed the RPC mechanism used for token setting from
4196 named pipes to local. Use of named pipes can be restored
4197 by setting the environment variable AFS_RPC_PROTSEQ to
4200 Named pipes were required when a Windows 9x system was
4201 using a NT system in gateway mode which is incompatible
4202 with our use of local loopback adapters.
4204 * In afscreds.exe, if a username of the form user@REALM is
4205 specified and no password is specified, do not perform a
4206 kinit operation. Only perform the aklog functionality.
4208 * Add a new registry value which allows the number of processors
4209 on which afsd_service.exe executes to be restricted. Valid
4210 values are 1..numOfProcessors
4212 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4216 * A second MSI based installer option is now available.
4218 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4219 request tickets for user/instance@REALM instead of just user@REALM
4221 * Modify the Power Management Notify routine to wait for the Hard Dead
4222 timeout period instead of a fixed 19 seconds. With the longer timeout
4223 periods Hibernation and Standby could never succeed when network
4224 connectivity is not available.
4226 * The following fs.exe commands are now restricted to Administrator:
4227 - checkservers with a non-zero timer value
4230 - sysname with a new sysname list
4239 setting the default sysname for a machine should be done via the
4240 registry and not via "fs sysname".
4242 * NSIS installer adds options to install Debugging Symbols
4243 and the Microsoft Loopback Adapter; the user is now also
4244 given the ability to select the afscreds.exe startup options.
4246 * Build system modified to generate symbols for FREE (aka RELEASE)
4247 builds as well as CHECKED (aka DEBUG) builds
4249 * Sites which have a volume ID of 0x20000001 assigned to their
4250 root.cell volumes have been experiencing problems with accessing
4251 the root.cell volume of their cell when Freelance mode has been
4252 active. This was because 0x20000001 was assigned to the fake
4253 root.afs volume created by freelance. The fake volume id is
4254 now set to 0x00000001 to prevent conflicts.
4256 * The timeout logic in the AFS Client Service has been wrong
4257 for sometime. It is based on two different assumptions.
4258 First, the SMB client timeout is a fix value as was the case
4259 with OS/2 Lan Manager. This assumption is incorrect. The
4260 SMB timeout in Windows is a dynamic value computed based upon
4261 a fixed minimum timeout to which is added time based upon the
4262 size of the request and the performance characteristics of
4263 the connection. Second, it is the responsibility of the
4264 SMB Server to enforce the timeout requirements of the client.
4265 This is untrue. The SMB Server cannot be expected to know
4266 the requirements of the client. More importantly, if the
4267 SMB server uses the SMB client timeout as a value to restrict
4268 its behavior as an RX client, the performance characteristics
4269 of the local SMB session would be used to prematurely terminate
4270 WAN connections with significantly different performance
4273 The timeout logic has therefore been modified in the following
4275 . the Lan Manager Workstation (SMB) Session Timeout is used only
4276 as a basis for configuring the Connection Dead Timeout
4277 and Hard Dead Timeout values. The Connection Dead Timeout
4278 must be at least 15 seconds longer than the SMB Timeout
4279 and the Hard Dead Timeout must be at least double the
4280 Connection Dead Timeout.
4281 . New registry entries have been added to allow the Connection
4282 Dead Timeout and Hard Dead Timeout values independent of the
4283 Lan Manager Workstation Session Timeout
4284 . The test to enforce the SMB Client Timeout has been removed.
4286 One of the side-effects of removing the enforcement of the SMB
4287 Client Timeout is that regardless of whether or not the SMB client
4288 is available to receive the response (and how would the SMB server
4289 know) the RX protocol response can be used to update the AFS
4290 Client Service state for ready access by future SMB client
4293 This should be the end of the "Server paused or restarting messages"
4295 * Add support for arbitrary UNC paths to the pioctl() support.
4296 This enables the fs commands as well as the AFS Shell Extension
4297 to work correctly when UNC paths are being used.
4299 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4300 DNS if the cell configuration cannot be determined via CellServDB
4302 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4303 are ever set more than once at a time
4305 * Fix the management of lists of cm_cell_t structures when using
4306 DNS to lookup cell information. The previous code would fail to
4307 reuse the same cellID for a cell if DNS was used more than once
4308 for a given cell name. When the ttl expired, a single cm_cell_t
4309 could be inserted into the cm_allCellsp list more than once
4310 producing a loop. In addition, the vlServerp list belonging to
4311 the cell was not freed resulting in improper refCounting of the
4314 * Add DNS support to cm_IoctlNewCell() which previous only examined
4317 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4318 to properly garbage collect cm_server_t objects
4320 * Add numVCs variable to smb.c to track the number of smb_vc_t
4321 objects created and use it to initialize the vcID field which
4322 previously was set to 0 in all objects resulting in FindByID
4325 * Fixed DNS lookups to work consistently throughout the OpenAFS
4326 product instead of just from within the afsd_service.exe
4328 * Added a runtime check to ensure that AFS Client Service SMB
4329 Server is accepting connections before attempting to mount
4332 * Read IP addresses for volume servers out of the CellServDB
4333 file if gethostbyname() on the hostname fails.
4335 * Fix getcellconfig() to populate both the Hostnames as well
4336 as IP addresses when loading cell data via DNS
4338 * Increase the Connection Dead Time to 50 from 20 seconds
4339 Increase the Hard Dead Time to 120 from 40 seconds
4340 (matches the Unix values)
4342 * Fixed an assertion validating the number of allocated NCBs
4344 * Fixed the build environment to consistently build for
4345 Windows 2000 and above. (APPVER = 5.0)
4347 * Fixed rx_debug to properly validate the receipt in incoming
4348 data with select() and recvfrom(). Do not copy data out of
4349 the socket buffer unless success is indicated.
4352 * afsd_service.exe will now display a message box to the
4353 desktop when it terminates due to an IP Address Change.
4355 * installer no longer deletes AFS Server configuration data
4358 * installer generates a warning dialog if the RPC service
4359 is not properly configured
4361 * installer compressed with lzma instead of bzip2
4363 * afsd_service.exe shutdown crash solved once and for all
4365 * reference counting of smb_vc_t data structures improved
4367 * name space collision of smb_fid_t event objects corrected
4369 * the output of "fs memdump" is now written to
4370 %WINDIR%\TEMP\afsd_alloc.log
4372 * the file TaAfsApp_1033.dll is now properly installed allowing
4373 the User Manager to start
4375 * a new algorithm is used for computing filename pattern matches
4377 * afscreds.exe now accepts user names containing instance
4380 * Fix the Directory Name Lookup Cache to be case-sensitive.
4381 This is crucial in environments in which a Windows client
4382 is accessing a directory with more than one filename that
4383 differs only by case. If the directory contains "FOO"
4384 and "Foo". You want "DEL Foo" to delete the correct one.
4385 We still have a problem in that "DEL foo" will delete a
4386 random filename. This will be addressed in a future release.
4388 * Fix afscreds.exe -M option (renewMaps) to work when High
4389 Security mode is off. Also, remember to disable the ActiveMap
4390 flag in afsdsbmt.ini when a drive mapping is removed.
4392 * Updates to NSIS installer script. AFS Server configuration
4393 data will not be destroyed on un-install or re-install.
4394 Use a better compression algorithm.
4396 * afslogon.dll now uses KFW to obtain tokens when available
4398 * afslogon.dll when given an all uppercase username will
4399 attempt to authenticate with both the uppercase name
4400 and an all lowercase variation
4402 * DST modification removed. The fix appears to make things
4403 worse after a reboot of the machine.
4405 * fs.exe: added "cscpolicy" which is used to
4406 change client side caching policy for AFS shares
4408 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4409 [-documents] [-disable] [-help]
4411 * Several uninitialized variables have been initialized
4413 * It is now possible to obtain tokens using cross realm
4414 Kerberos within afscreds.exe:
4416 user: jaltman@ATHENA.MIT.EDU
4418 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4419 will will in turn be used to obtain afs@DEMENTIA.ORG.
4420 The resulting token will be stored with the display name
4421 jaltman@ATHENA.MIT.EDU@dementia.org
4423 * aklog.exe has been added to the client
4425 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4426 [[-p | -path] pathname]
4430 -d gives debugging information.
4431 krb_realm is the kerberos realm of a cell.
4432 pathname is the name of a directory to which you wish to authenticate.
4433 -noprdb means don't try to determine AFS ID.
4434 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4435 (default is Kerberos V)
4436 No commandline arguments means authenticate to the local cell.
4439 * All of the resource files have been restructured to adhere to
4440 a set of rules IBM implemented for loading string resources.
4441 These rules had either been forgotten or were not discovered
4442 by folks working on the OpenAFS sources. The end result was
4443 memory corruption. This is primary item which was preventing
4444 the AFS Server from working.
4446 * Increased the size of the maximum ticket size stored in a token
4447 from 344 bytes to 12,000. Increased the buffers used to convey
4448 messages between the pioctl() caller and the SMB Server from
4449 1000 bytes to 12,512. The code appeared to have been writing
4450 above the top of the stack by quite a few number of bytes.
4451 (The increased ticket size is necessary for the next item.)
4453 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4454 Instead the raw Kerberos 5 ticket is used in its entirety. This
4455 is extremely important as it allows us to use pure Kerberos 5 KDCs
4456 as the source of the AFS authentication. The use of up to 12,000 byte
4457 tickets will allow tickets produced by all versions of Microsoft
4458 Active Directory to be used.
4459 - create a user account.
4460 - designate it DES only
4462 - specify its UPN to be "afs@realm"
4463 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4465 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4466 is special when using long file names. (you must use "*.*" and not "*")
4467 Instead only enforce it when performing 8dot3 searches.
4469 * Fixed the DST problem with creation times being set one hour ahead
4471 * Fixed the problem when using \\afs\cell-alias. For example,
4472 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4473 for the alias name; instead simply expand the name. One of the
4474 symptoms of this problem was a loss of acquired tokens.
4476 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4477 of strings. (Only English strings provided.)
4479 * Fixed the installer to properly replace in use files.
4481 * Fixed the build system to cleanup generated component version files
4483 * The release build compiled with MSVC 6.0 compiler to avoid the
4484 afsd_service.exe shutdown crash. This does not solve the problem
4485 but simply avoids it for the time being.
4489 * fix afslogon.dll to not corrupt memory when High Security mode
4492 * fix afsd_service.exe to not attempt to restore the stack when
4493 an exception occurs. (not safe in multi-threaded programs)
4495 * fix uninstaller to properly remove the CRT and MFC DLLs
4497 * remove a Message Box from afscreds.exe when getcellconfig()
4498 fails on a kerberos realm which is not a cell
4500 The following is a list of changes to the OpenAFS for Window client
4503 * "fs setserverprefs" will leave afsd service deadlocked
4505 * "vos listaddrs" will core dump
4507 * installer sets the appropriate keys to support Integrated Logon
4509 * installer disables the "Find Lana by Name" functionality as it
4510 was causing headaches for many users
4512 * fix the intermittent crash of the power management thread when
4513 shutting down the AFS Client Service
4515 * optimizes the obtain drive mount list functionality which is
4516 executed every time the mount tab in afscreds.exe and afs_config.exe
4517 are refreshed. (this happens a lot)
4519 * fix the service shutdown logic. add the STOP_PENDING state
4520 and do not accept additional service events after we declare
4523 The following is a list of changes to the OpenAFS for Window client
4526 * flexelint was run against the source tree and hundreds (perhaps
4527 thousands) of corrections were applied to ensure prototypes
4528 were in use; types were used consistently; variables were
4529 initialized; unused variables were removed; etc.
4531 * A wide variety of instrumentation was added including the
4532 ability to produce a stack trace from within afsd_service.exe
4535 * Dynamic configuration of the RDRtimeout value based upon the
4536 LanMan Workstation Session Timeout
4538 * The mount root no longer needs to be called "/afs". This
4539 is now set by a registry value "MountRoot" within the key
4540 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4542 * The cell list is now only read out of afsdcell.ini when the
4543 file changes instead of each time a cell is resolved.
4545 * Thread synchronization was added to cm_server.c and ktc_nt.c
4547 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4548 calloc()/free(). The Global functions were needed on Windows 3.x
4549 but have caused a variety of problems on the Win32 platforms.
4550 Avoiding them is highly recommended by several Microsoft
4551 Knowledgebase articles
4553 * Support for Symbolic Links added to the AFS Shell Extension
4555 * Added a registry value "OverlayEnabled" to determine if
4556 Shell Extension Overlays should be enabled.
4557 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4559 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4560 separate trees for checked and free builds. Build system supports
4561 a custom directory src\WINNT\extra which can be used as a grafting
4562 location of organization specific additions to the build tree.
4564 * New installer built using NSIS 2.0.
4566 * Named all kernel objects in order to allow them to be monitored
4567 with tools such as SysInternals' ProcExp.exe.
4569 * Introduced new EventLog framework for AFSD
4571 * Introduced Power Management interface to AFSD for Standby and
4572 Hibernate modes to allow cache to be flushed prior to network
4575 * Utilize Win32 DNSQuery API instead of internal routines. This
4576 allows DNS SRV queries to be sent to all current domain name
4577 servers. Not just one specified in an INI file. DNS is now
4580 * "NetbiosName" registry value may be used to specify a fixed
4581 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4582 when the loopback adapter is in use. If you need to use the
4583 old notation with a loopback adapter installed specify a registry
4586 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4588 * Refactor all modules which depend on LAN Adapter and NetbiosName
4589 determination in a new library: lanahelper.lib. This allows for
4590 consistent behavior throughout the product.
4592 * Move the afsd.log and afsd_init.log files to the directory specified
4593 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4594 for services. Added the Date to the log entries.
4596 * New registry value "RxMaxMTU" used to limit the size of the RX
4597 packets sent by the AFS Client Service to the Server. In order
4598 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4599 size must be restricted to 1264 or smaller. The latest NSIS
4600 installer sets a value of 1260 by default.
4602 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4603 packets. This is not needed in order to work across the Cisco
4604 VPN but might be needed for other network environments. This
4605 value is not set by the NSIS installer.
4607 * New registry value "HideDotFiles" is used to apply the Hidden
4608 attribute to files whose names begin with a '.'. This value
4609 is set by the NSIS installer.
4611 * New registry value "MaxMpxRequests" allows the maximum number
4612 of multiplexed sessions to be configured at run time. This
4613 value is not set by the NSIS installer. The default value is
4616 * New registry value "MaxVCPerServer" allows the maxmimum number
4617 of VCs per server to be configured at run time. This value is
4618 not set by the NSIS installer. The default value is 100.
4620 * New registry value "AllSubmount" allows the "all" submount to
4621 be disabled by setting its value to 0x00.
4623 * Allow cells names to be valid mount points
4624 \\<netbiosName>\<cellname>
4626 * Store the active state of drive mappings in order for afscreds.exe
4627 to restore them upon startup
4629 * Add exception handling to generate a Stack Trace to the afsd_init.log
4630 file if one happens to occur.
4632 * Add lots of logging to help detect the cause of invalid SMB packets
4634 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4635 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4636 renewal of expiring tokens as long as afscreds.exe is running.
4638 * New afscreds.exe command line options:
4640 -M = renew drive maps
4641 -N = ip address change detection
4644 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4645 determines whether or not MIT Kerberos for Windows should be used
4646 to obtain tokens via Kerberos 5 tickets.
4648 * New registry value "AfscredsShortcutParams" in
4649 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4650 determines the command line parameters to be specified when "fixing"
4651 the AFS Shortcut in the user's startup folder.
4653 * The "ShowTrayIcon" registry value has been moved from
4654 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4655 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4657 * The <cell name> registry values used to store the token expiration
4658 reminders have been moved from
4659 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4660 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4662 * Obtain the Logon User Name from the Explorer key when available
4664 * new text document doc\txt\winnotes\registry.txt lists all registry
4665 values used by OpenAFS (excluding the AFS Server)
4667 * BUG: rx_securityClass objects were not properly reference
4668 counted and were never freed.
4670 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4671 would be generated. The existence of a server does not imply
4672 that it is not down. If all of the servers for a cell are down
4673 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4676 * BUG: the directory name lookup cache failed to free the entries
4677 in the cache when the name cache entries cycled. The entries
4678 in the cache would become dereferenced without being freed.
4680 * BUG: fs setserverprefs could be executed without Administrator
4683 * BUG: the number of allocated NCB objects (100) exceeded the number
4684 which could actually be waited upon by the kernel (64). Any objects
4685 which were utilized above the limit could never have event completions
4688 * BUG: smb_username_t objects were not being reference counted and
4689 were not properly freed.
4691 * BUG: smb_tid_t objects could under unusual circumstances be freed
4692 before they were no longer referenced.
4694 * BUG: smb_fid_t object pointer were frequently used even when
4695 their value could be NULL. They were not properly released and
4696 therefore they were never freed.
4698 * BUG: smb_packet_t data structures were not completely initialized
4701 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4702 Denied" because that causes the Explorer Shell to try again until
4703 access is obtained. Instead return "Remote Resources" which allows
4704 the shell to move on and treat the error as transient.
4706 * BUG: when initializing the NCBreturns structure, separate Event objects
4707 were created for each NCB although a single Event object was supposed
4708 to be shared by all.
4710 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4713 * BUG: smb_tran2Packet_t objects were not being properly referenced
4716 * BUG: directory path creation did not handle the case of multiple
4717 directories requiring creation in one attempt
4719 * BUG: SMB requests which required an Extended Response were ignored.
4720 This prevented some files from being written to AFS volumes.
4722 * BUG: character strings were being freed even after they were
4723 inserted into in use data structures
4725 * BUG: inconsistent usernames were used when High Security mode was
4726 enabled. (there is still much to do in this area)
4728 * BUG: pioctl() calls which require out of band RPC operations were
4729 susceptible to race conditions when performed by multiple processes
4731 * BUG: memory allocation and deallocation crossed instances of the
4732 C Runtime Library producing memory leakage and corruption in
4733 afscreds and the client configurator.