2 * Short circuit background volume checks if a shutdown
3 or suspend request is received.
5 * Avoid contacting the file server in order to allocate
6 an empty buffer beyond the current length of the file.
8 * When a request for streams on a directory or mount
9 point object is received, do not offer a default stream.
10 This was affecting the saving of roaming profiles.
13 * Fix Windows 2000 compatibility. Do not call rand_s()
14 even if it is compatible with the C run time library
17 * Fix a data consistency error between the output of
18 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
19 workstation group name "AFS", and report server name as
20 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
22 * Enable executables to be run from \\AFS on Windows 7.
23 Return "Name not found" instead of "File not found" when
24 a directory or file name cannot be found.
26 * Prevent cache manager from marking file server "down" when
27 the data returned in response to either RXAFS_FetchData64
28 or RXAFS_StoreData64 is invalid.
30 * Add pioctl data validation to the AFS Explorer Shell extension.
33 * After 1.5.62 afslogon.dll would no longer crash but it
34 also would not load "domain" specific configurations.
37 * An uninitialized variable was present in the symlink
38 recursion detection loop.
40 * Reverse addition of unique per file GUID during
41 Create/Open responses. Inclusion of the GUID prevents
42 Cygwin applications (and others) from accessing //afs.
44 * Treat "filename::$DATA" as a synonym for "filename".
45 This addresses the recent report of roaming profiles
46 failing to save back to the file server.
48 * When given a choice between an error from rx_Write()
49 or RXAFS_EndStoreData(), use the rx_Write() error.
50 This ensures that over quota and access denied errors
53 * If an error occurs during RXAFS_StoreData operations,
54 apply that error to all of the buffers in the BIOD.
55 This ensures that failures will not be retried if they
58 * Do not perform a DNS lookup for a cellname if the
59 cellname prefix is _._AFS_IOCTL_._. This will avoid
60 DNS lookups when the query is for
61 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
63 * Fix Freelance root.afs volume object callback processing.
64 Object synchronization and status merging bugs.
66 * Fix Explorer Shell "invalid parameter" bug introduced
69 * Log RXKAD errors to the trace log.
71 * Prevent simultaneous pioctl calls from multiple processes
72 from stepping on each other's data.
74 * Prevent simultaneous pioctl calls from crashing the
77 * Add data validation to pioctl output processing in
78 ktc_GetToken and ktc_ListTokens.
80 * Add data validation to pioctl output processing in
81 fs.exe and symlink.exe.
84 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
85 No longer will browsing \\AFS produce truncated share names.
87 * Registry specified server preferences did not result in
88 the CM_SERVERFLAG_PREF_SET flag being applied to the
91 * Avoid unnecessary DNS lookups of share names as cells
92 when it is known that the name cannot be a cell name.
93 Any name that does not contain a dot is skipped.
95 * When processing Dfs Referral requests, do not return an
96 error if the server is down or busy; if the volume is
97 offline; the cell vldb cannot be reached; etc.
98 These paths are still \\AFS paths and so the client
99 should not be sent to search elsewhere for resolution.
101 * CRITICAL: Prevent the cm_Daemon thread from terminating when
102 the machine enters the suspend state. IF the cm_Daemon thread
103 dies, there is nothing to execute down server checks.
105 * Prevent use of smb_StartedLock before initialization which
106 can result in a panic in the lock package.
108 * Improve error checking in afslogon.dll in order to prevent
109 NULL pointer dereferences within GetLogonDomainOptions()
110 if the Lsa operations fail.
112 * Add support to the build system to automatically update
113 a Microsoft Symbol Store.
115 * Do not access the cm_conn_t after executing the cm_Analyze
116 loops. Doing so can result in an access to memory that has
119 * CRITICAL: If a StoreData request offset is not aligned on the
120 buffer module blockSize, (offset % blockSize) bytes of dirty
121 data will not be written to the file server.
123 * CRITICAL: If a StoreData64 request is sent to a file server
124 that does not support large files (>2GB) and the client is
125 unaware that the server has no large file support, the first
126 chunksize worth of data will not be written to the file server.
127 This bug was introduced in 1.5.3. It affects all IBM AFS file
128 servers and all OpenAFS file servers older than 1.4.0.
131 * If a file server becomes inaccessible while the cache manager has
132 dirty buffers to write, the afsd_service buf_IncrSync thread can
133 attempt to use 100% of the cpu.
135 * Fix the locking that protects the cm_cell_t lists
136 (all, name, id, free).
138 * Remove cm_cell_t objects from the name and id hash tables before
139 placing them on the free list.
141 * Fix "fs newcell" which was broken in 1.5.60.
143 * Do not attempt to synchronize dirty buffers if the associated
144 volume is known to be unavailable.
146 * Log to the Windows Event Log dirty buffers that cannot be written
147 during the shutdown process.
149 * Modify behavior of a Freelance mountpoint target that does not
150 specify a cell. Instead of assuming the target volume is in the
151 Freelance.Local cell, use the workstation "Cell" specified in the
152 registry. A mountpoint target of "#root.cell." will now mean the
153 root.cell volume in the workstation cell for the current session.
154 If the workstation cell changes from "athena.mit.edu" to
155 "andrew.cmu.edu", the referenced volume will also change without
156 requiring that the mount point targets be altered.
158 * Avoid false infinite loop errors when validating the cm_cell_t
159 free list during startup.
161 * Fix two more locations where xdr_free() should be used instead of
162 free() to permit a checked built of afsd_service.exe to work with
163 release builds of afsrpc.dll.
165 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
166 to permit the server Uuid to be used to lookup the server object
167 and from that determine the cell. This permits callbacks that are
168 received from alternate addresses to be processed with a known server
169 object. Previously a request from an unknown server would clear all
170 callbacks from all cells.
172 * Fix a bug that prevented optimal performance when using a non-zero
173 value for 'daemonCheckVolCBInterval'. As a reminder, when
174 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
175 volume callbacks are automatically renewed 90 minutes before their
178 * Fix automatic ranking of vldb servers whose values are obtained from
181 * Add logging for RX CALL DEAD errors which are generated if the remote
184 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
185 or the call is to a vldb server.
187 * Add registry based cell search functionality to NetIdMgr,
188 afs_config.exe, and klog.exe.
190 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
191 lookups on the host names in the CellServDB instead of using the
192 specified IP addresses. This provides aklog, pts, vos, etc. the same
193 CellServDB behavior that the Windows Cache Manager uses.
195 * Completely a behind the scenes change. The snprintf() routines have
196 been completely rewritten.
198 * When updating the stat cache entry callback of a .readonly object
199 from the volume group object, update the file server reference to
200 ensure it matches the most update to date callback.
202 * Add proper support for processing callbacks from multi-homed file
203 servers. Instead of comparing servers by cm_server_t pointer,
204 compare them by UUID when the UUID is known.
206 * During a shutdown short circuit the offline volume check daemon
209 * Properly track link counts during cm_Link and cm_Unlink operations.
211 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
212 preference to an error code reported by rx_EndCall.
214 * Add "PerFileAccessCheck" registry value to permit testing against
215 experimental file servers that include per-file acl support. This
216 value is intentionally undocumented. It is not to be used by
217 production environment deployments.
219 * Log a warning to trace log if a volume group is marked as a DFS
220 File Set. The Windows OpenAFS client does not include DFS support.
222 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
223 provider from obtaining tokens when referrals are in play.
225 * Add "fs chown" and "fs chgrp" commands to permit the owner and
226 group of objects stored in AFS to be set from Windows.
228 * Avoid performing background daemon operations when the machine is
229 going into suspend mode.
231 * Perform offline volume checks in most recently used order.
233 * Prevent crash when a data version for a cache object goes backwards.
235 * fs setquota, fs setcachesize, vos setfields, and vos create now
236 accept human readable orders of magnitude. (K, M, G)
238 * fs listquota fixed to permit large quota sizes to be displayed.
240 * rx packet resend and data packets sent counts were incorrect.
242 * Multi-thread safe library versions are now being generated and used.
243 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
245 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
246 values are now available on XP through Windows 7. Add functionality
247 to autodetect if such support is present on the machine. If so,
248 configure it if necessary and dynamically adjust the AFS Rx timeout
251 * Correct another race condition in the Rx library that could result
252 in an unexpected panic while freeing the Rx call iovq.
255 * A fix to the pioctl library to support drive substitution
256 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
258 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
259 that corrects a deadlock in the smb redirector and also adds new
260 functionality that permits the AFS SMB server to be given a longer
261 timeout than is normally the case. New functionality has been added
262 to configure these additional LanmanWorkstation\Parameter values.
263 (This functionality has been backported to XP SP3 and is scheduled
264 to be released on June 5th.)
266 * The BackConnectionHostNames registry value configuration was broken
267 when dynamic re-establishment of Netbios Name registrations was added.
270 * Hidden vos commands are revealed.
272 * If the "DisableLoopbackCheck" registry value is set, do not unset
273 it during the same service session.
275 * Reorganize code that prevents multiple Store operations to the same
278 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
279 permit broken symlinks to be treated as being in AFS.
281 * Fix vos commands that output 64-bit integer values.
283 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
284 Permits building cygwin applications against OpenAFS libraries.
286 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
287 as Windows HTML Help (.CHM) files.
289 * NSIS installer does a much better job of cleaning up files left
290 over from previous installs.
292 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
293 or "fs flushall" and on-going directory operations that can result in
294 afsd_service.exe crashing.
296 * Add support for DNS SRV records in place of AFSDB records.Â
297 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
298 Weight is currently ignored.
300 * Add a method of specifying Client CellServDB information within the
301 registry that can be used to either override the CellServDB file or
302 force the use of DNS lookups for a given cell.
304 The registry schema is as follows:
306 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
307 "LinkedCell" REG_SZ "[cellname]"
308 "Description" REG_SZ "[comment]"
309 "ForceDNS" DWORD {0,1}
311 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
312 "HostName" REG_SZ "[hostname]" Default: [servername]
313 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
314 "IPv6Address" REG_SZ "[address]" <future>
315 "Comment" REG_SZ "[comment]"
316 "Rank" DWORD "0..65535" Default: 0
317 "Clone" DWORD "{0,1}" <future: server only>
318 "vlserver" DWORD "7003" <future>
319 "ptserver" DWORD ... <future>
321 ForceDNS is implied non-zero if there are no [servername]
322 keys under the [cellname] key. Otherwise, ForceDNS is zero.
323 If [servername] keys are specified and none of them evaluate
324 to a valid server configuration, the return code is success.
325 This prevents failover to the CellServDB file or DNS.
327 Only one of "HostName" or "IPv4Address" is required. "HostName"
328 is optional if [servername] is the host name.
330 * Extend registry based CellServDB functionality to
331 afsconf_GetCellInfo() interface used by aklog and many
332 other command line utilities.
334 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
335 as it is now used by both client and server components.
338 * PriorityClass of afsd_service.exe process raised to
339 "High" to match the priority of the system services
340 that are waiting for afsd_service.exe to respond.
341 Configurable using [TransarcAFSDaemon\Parameters]
342 DWORD "PriorityClass" value. (use values from Win32
345 * Change error reporting of CM_ERROR_BADSHARENAME to
346 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
347 The error we were sending would cause the smb
348 redirector to drop the connection.
350 * Change the error reporting of CM_ERROR_ALLOFFLINE
351 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
352 Previous we would return Bad Network Path which
353 would cause the smb redirector to cancel all file
354 handles on the same "share".
356 * A debug version of afsd_service.exe can now be used
357 with a release version of afsrpc.dll (or vice versa).
358 afsrpc.dll now implements and exports xdr_free().
360 * Removed all unprotected OutputDebugString() calls as
361 those calls can block for up to 10 seconds if there
362 is contention for the globally shared memory or if
363 the debugger fails to read data in a timely manner.
365 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
366 supported for builds. The production releases still
370 * Discovered that XP/2003 supports a variant on recvmsg
371 called WSARecvMsg. The input parameters are mostly
372 compatible with the Posix recvmsg. This will save
373 multiple memcpy() calls per received packet.
375 * Discovered that Vista/2008 supports a variant on
376 sendmsg called WSASendMsg. The input parameters
377 are mostly compatible with the Posix sendmsg. This
378 will save multiple memcpy() calls per sent packet.
380 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
381 and utilizing rx_ReadProc32 from within the cache manager
382 as part of the FetchData operations. rx_ReadProc32 and
383 rx_WriteProc32 are optimized versions of rx_ReadProc and
384 rx_WriteProc for use when reading 32-bit integer values.
386 * The Network Identity Manager AFS provider now reports
387 progress events. These provide a better user experience
388 with Network Identity Manager 2.0.
390 * [RT 124293] A race condition exists which permits the
391 scp field of the an smb_fid_t object to become invalid
392 after a request on the smb_fid_t has begun. Remove the
393 race by protecting all access to the scp field with the
394 smb_fid_t mx mutex and obtaining a local reference on
395 the cm_scache_t object for the length of the request.
397 * [RT 124276] If the vldb is out of sync with the contents
398 of the file servers, afsd_service will retry too many
399 times when a file server reports a volume as not being
400 present. Now if the list reported by the vldb is the
401 same as the previously seen list, then the retry is
404 * [RT 124276] Read-only volume failover was broken in
405 1.5.53 whenever accessing a volume results in VNOVOL or
408 * [RT 124276] Prior to 1.3.70 the volume server reference
409 list was not reference counted and would be prematurely
410 freed while in use. When reference counting was added
411 in 1.3.70 a bug was introduced that could result in
412 service reference list corruption.
414 * Add Windows Application Event Log warning messages for
415 "Client SMB MPX value too large" and
416 "Client SMB Buffer Size too small".
418 * Renaming of files across directory boundaries would result
419 in an invalid handle error when attempting to access the
420 files after the move.
422 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
423 when a smb file descriptor cannot be found. The whole point
424 of using a Path Info function is because an smb file
425 descriptor wasn't allocated.
427 * More edge cases in which dynamic addition of Freelance
428 root.afs entries would get the wrong FID or where the
429 root.afs directory would not be refreshed.
431 * Buffer overflow could occur if the workstation cell name was
432 longer than 64 characters. Crashes could occur in
433 afscreds.exe, afslogon.dll, and afsd_service.exe.
435 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
438 * Improve consistency of the bad data version value used for
441 * Log file server uuid values as part of the cm_server object
442 when available. Dump the cm_server object list in response
445 * Optimize the performance of resetting access control lists
446 when tokens are set or removed.
448 * Fix vos.exe so that it can work with very large volume
451 * Remove symlink recursion tests and increase max symlink count
455 * Undo the "UAC manifest fix" applied to afs_config.exe.
457 * Fixed some additional race conditions in the Rx library
459 * Ensure that Freelance allocation of vnodes follow
460 the AFS convention of odd vnodes are directories
461 and everything else is an even vnode.
463 * Add Freelance logic to mount point and symlink
464 evaluation functions.
466 * Enhance smb_ParseASCIIBlock() so that it can handle
467 all of the STRING formats defined by the CIFS Technical
470 * Validate the output of smb_ParseASCIIBlock() in all
471 callers. Return CM_ERROR_BADSMB if the STRING field
472 cannot be parsed. CM_ERROR_BADSMB will cause the
473 contents of the packet to be logged.
475 * If multiple SMB Raw Write operations were taking
476 place at the same time, there could be data corruption
477 because unique event objects were not generated for
478 each Netbios receive operation.
482 * Fix a race in rxi_WriteProc which could result
483 in a freed or in use packet being passed to
484 rxi_PrepareSendPacket if the call is currently
485 in a transmit wait state. Passing a freed packet
486 will result in a panic. Passing an in use packet
487 might result in a call failure or perhaps data
490 * Kerberos referrals broke the NIM AFS provider
491 logic for automated configuration of the workstation
492 cell with a new identity. Now if a referrals
493 response is returned, assume the cell belongs to the
496 * The interface to buf_CleanAsync() and
497 buf_CleanAsyncLocked() has been changed to permit
498 errors to be propagated to the caller.
500 * During an enumeration of a directory, if the btree
501 has not yet been generated, return an error.
503 * When mapping CM_ERROR values to NTSTATUS values
504 do not treat 0 as an error.
506 * Avoid a cm_scache_t object reference count underflow
507 when renaming via the SMB interface.
509 * smb_IsMask() can be called with a non-8.3 input buffer.
510 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
511 input buffers and smb_IsMask() for non-8.3 input buffers.
513 * Add cm_buf_t error state to the "fs memdump" output.
515 * pioctl drive mapping checks need to take into account
516 Global Drive Mappings and SUBST assignments.
518 * Rename operations should not be permitted if the
519 destination name collides with an existing name that
520 differs only by case UNLESS the collision is the object
521 that is being renamed.
523 * Increase the maximum number of background daemon threads
524 to 64. Leave the default at 4.
526 * Set the FILE_UNICODE_ON_DISK flag when reporting the
527 file system characteristics.
529 * Report the maximum component name length as 255 even
530 for Unicode as part of the file system characteristics.
532 * Add RX debugging information to "fs memdump" output
535 * Refcount the cm_server_t objects returned by
538 * Fix the UAC manifest applied to afs_config.exe
540 * rxdebug now indicates if any RX_CONN flag is set
541 not just some of them
543 * Change the mapping of CM_ERROR_WOULDBLOCK from
544 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
548 * Includes #GCO Public CellServDB 07 Nov 2008
550 * Prevent separate cell entries from being created for
551 "foobar.org" and "foobar.org." which can happen if
552 an AFSDB record is found instead of a CellServDB
555 * Fix a panic when checking the SMB virtual circuit list.
556 If a VC is destroyed while the list is being checked
557 it is possible for the next object to become invalid.
559 * Increased logging to afsd_init.log of netbios name
562 * Fix a DNLC panic when a UTF-16 string that cannot be
563 converted to a utf-8 string is provided as a filename.
565 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
566 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
567 location to refer directly to the new location in
568 %programfiles%\openafs\client\program
570 * Be more aggressive about responding to Netbios
571 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
572 the LANA within the listener thread without performing a
575 * Use the WNet Enumeration of all connected drives as a filter
576 in the pioctl GetIoctlHandle operation. This should remove
577 the last situations in which there was a delay when opening
578 an Explorer Shell properties menu with the AFS Shell
581 * Fix the parsing of CellServDB in Windows to handle linked
582 cells as is done on Unix.
584 * Add support for linked cells to the cache manager,
585 netidmgr afs provider, and aklog. When two cells are
586 linked, a volume lookup that fails in one cell is looked
587 up in the other. This can be used for construction of
588 test cells that override specific volumes; cell renaming;
591 * Increase the length of cell and realm names that can
592 be entered in the NIM AFS provider.
594 * Actively prevent a 32-bit afsd_service.exe from
595 executing in the WOW64 environment.
597 * Protect against invalid IP address values in CellServDB
602 * Fix an Rx panic caused by an attempt to double free
603 an rx_packet when an error occurs while writing the
606 * Fix an exception cause by a race condition when
607 processing multiple file change notifications registered
608 by the Explorer Shell or other applications.
610 * Modify the Rx debug interface to use var_args and
611 permit Rx debugging to be used on 64-bit Windows
612 which previously had it disabled.
616 * Automated lock hierarchy enforcement added to the
617 osi base lock package. This permits detection of lock
618 violations that might but often does not result in
619 deadlocks. Enabled by default in the checked builds.
620 May be activated in release builds via the registry
621 value "LockOrderValidation" in the TransarcAFSDaemon
624 When a violation is detection an assertion is thrown
625 generating a dump file and a message indicating which
626 locks were obtained out of order is logged to the
627 Windows Application Event Log.
629 * Attempt to ensure that a dump file will be generated
630 whenever the service panics.
632 * Removed unused lock cm_bufGetMutex
634 * Many out of order lock acquisitions were detected
635 using the lock order verifier and were corrected.
637 * A race condition within the function path cm_GetSCache() ->
638 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
639 cm_scache_t object to be issued simultaneously to two
640 threads. This would eventually result in a panic due
641 to the resulting undercount.
643 * Fix interpretation of the empty string as the ioctl path
644 to mean the current directory. "fs lsm", "symlink list",
645 etc. now return a "not a ..." error instead of "does not
648 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
649 to indicate that the input string is a nul terminated string.
650 Assign ti when input strings are nul terminated instead of
651 length counted. (from 1.5.50)
653 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
654 does not match the observed behavor. The 'nameLength' is specified
655 in bytes, not in characters. Fix the implementation to match
656 the observed behavior. (1.5.50)
658 * The cm_HaveAccessRights() attempt at deadlock avoidance by
659 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
660 does not avoid the deadlock. Avoid the deadlock by enforcing the
661 lock order of the lowest vnode first. Then remove the infinite
662 loop avoidance in cm_SyncOp() that was returning an unwarranted
665 * File Server lock synchronization was not propertly enforced in
666 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
668 * Another change to the short name generator. Plus and equal are
669 not valid characters in a short name according to the
670 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
671 with left and right brace in order to minimize the impact.
673 * Fix a race condition that can result in a heap overwrite error if
674 the number of known file servers or vldb servers increases during
675 the preparation phase of a cm_CheckServers() operation.
677 * If the SMB Authentication would fail, a LSA allocated memory
678 buffer would be leaked.
680 * Rewrite the cm_cell_t allocation code to permit cell vldb server
681 lookups to fail without leaking cm_cell_t objects.
683 * Add "fs setcrypt auth" to permit the use of rxkad authentication
684 only mode. Not documented.
686 * Initialize Rx mutex and lock objects once and only once as part of
687 the DLL load process.
689 * When reporting the Volume Name "AFS" do not include the terminating
690 NUL in the reported length.
692 * Refresh Volume Location information upon receipt of a VNOVOL error
693 from the file server and avoid a race condition that prevented the
694 updated volume location information from being destroyed immediately
695 after it was acquired.
697 * In response to VNOVOL and VMOVED, remove the volume id from the server's
698 volume list so that the volume's up and down status is not affected
699 when the machine is shutdown.
701 * Many fixes to the Rx stack related to packet allocation and deallocation;
702 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
703 by overwriting the rx_call currentPacket pointer when it was already
704 referring to a packet.
706 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
707 in the TransarcAFSDeamon/Parameters key.
709 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
710 constructed directory B+ trees for flushed directory objects.
712 * When cached pages for directory objects have been locally modified, do
713 not permit the locally modified pages to be mixed with pages subsequently
714 obtained from the file server. Doing so can result in directory
717 * If Ubix errors are received in response to VL_xxx operations record
718 the error on the server instance and permit the operation request to
719 be retried in case another server is able to answer the query. We have
720 seen at least one instance of a cell in which a VL server returned
721 UNOQUORUM (possible database truncation) when the other servers were
722 responding successfully.
725 * Optimize the assignment of write errors to dirty
726 buffers by applying the first error to all buffers
727 belonging to the same object instead of trying each
730 * Add resource version info to xstat_fs_test and xstat_cm_test
732 * Fix a delete vs delete[] usage problem in the WiX custom
735 * When creating a mount point, validate the volume name to
738 * When processing a mount point, pay attention to VL_BADNAME
741 * Remove a race condition from cm_volume_t recycling that could
742 have resulted in two threads making use of the same object at
745 * Fix a reference count leak on smb_vc_t objects introduced in
746 1.5.50. Permit the objects to be freed once again.
748 * Fix the output of "fs memdump" for smb_vc_t objects that was
751 * Add routines for debugging smb refcount errors.
753 * When shutting down the cache manager, do not permit AFS
754 callback requests to be processed. Doing so can result in
755 invalid memory accesses.
757 * Do not recycle a deleted cm_scache_t object unless the
758 refCount is 0. Doing so can result in a deadlock and
759 invalid memory access.
761 * Minimize the number of RPCs sent to the file server for
762 directory bulk status requests. Only request status for
763 objects that we are actually reporting on and only if we
764 do not already have status for the object and then make
765 sure that we return as much as we can at once via the SMB
766 interface but not so much that it exceeds the RDR timeout
768 * Enable bos restricted operations
770 * Fix the creation of submounts used by the afscreds and afs_config
773 * Avoid repeated retries when attempting to obtain access rights.
774 If the access rights were obtained successfully but do not
775 satisfy the request, fail the request.
777 * Fix error code reported when attempting delete a file on a
778 readonly volume or one that is marked with the readonly DOS
781 * Fix a short name truncation bug.
783 * Fix a heap corruption error when reading the CellServDB file
786 * Fix directory listings when the directory has "lookup" permission
787 only and the directory has never been previously accessed.
789 * Fix several tracking issues for cm_scache_t rw locks.
791 * Ensure that all buffers used to store cell names are the
794 * Restore Windows 2000 support (broken in 1.5.50).
795 LOCALE_INVARIANT was introduced in XP.
797 * Add RxUdpBufSize registry value. Default is 256K.
799 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
801 * Do not include trailing NULs when returning file names in
802 a directory search response.
804 * Prevent RX from crashing the afsd_service by attempting
805 to allocate to many buffers at once when memory is fragmented.
807 * Avoid a deadlock when freeing cm_server_t objects
809 * Optimize search cell file. Do not search to end of file
810 if the cell has already been found.
812 * Remove explorer shell extension delay on local disks.
814 * Automatically pre-allocate 64 Rx packets per thread
816 * Add smb lock requests and stat cache lock allocations to
817 the output from fs memdump.
820 * Fix Large File Support (broken in 1.5.50)
822 * Fix "fs setacl" (broken in 1.5.50)
824 * Fix pioctl path evaluation of drive letter
825 mapped paths. (broken in 1.5.50)
827 * Fix installation of VC8 Runtime library in the EXE
828 Retail installer. (broken in 1.5.50)
830 * Avoid a potential deadlock during Volume Location
831 updates if all the file servers are unreachable at
832 the time the VL update succeeds.
834 * Ensure that rx connection object is references
837 * Fix propagation of over quota errors during cm_FSync()
840 Since 1.5.39 [1.5.50 released 16 July 2008]
841 * There were no releases between 1.5.39 and 1.5.50.
843 * FIRST RELEASE WITH UNICODE SUPPORT
845 * Pioctl interfaces to the cache manager refactored
846 to provide layering between the SMB specific code
847 and the general purpose ioctl operation.
849 * Garbage collect dead SMB virtual circuits as soon
850 as they are no longer being referenced. This avoids
851 problems with outstanding locks not being dropped
852 when the virtual circuit becomes invalid.
854 * Remove the IBM Administration Reference documentation
855 and replace it with the OpenAFS Command Reference
858 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
859 functions each time a connection is about to be used.
860 Do not hold a lock on the rx connection object while
861 it is being selected. This avoids a race between
862 threads attempting to set the timeout values and
863 removes a bottleneck that was hampering performance.
865 * Ensure that the smb directory attribute is set
866 for all directory objects.
868 * Add vs2008 support to the NSIS installer scripts
870 * Replace the VC Runtime EXE installer with the MSI
871 installer in the NSIS installer scripts
873 * Properly delete the VC7.1 C Runtime libraries.
875 Since 1.5.36 [1.5.39 released 23 June 2008]
876 * There were no 1.5.37 or 1.5.38 releases for Windows
878 * Use Visual Studio version of vsprintf() instead of
879 home grown version that resulted in linking conflicts
881 * Prevent buserver, ptserver and fileserver from crashing
882 due to non-portable strftime() format strings.
884 * Prevent server crashes caused by rx_InitMorePackets()
885 being called prior to allocation of all of the required
886 synchronization objects.
888 * Fixed a memory leak in the internal B+ tree directory.
889 For objects with non-8.3 names, the 'longname' memory
890 allocation would be leaked during object deletion or
893 * Fixed a memory leak in the background daemon cm_CheckServer()
894 routine. The memory allocated to store the server list
897 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
899 * Correct an abstraction layer violation. cm_scache_t objects
900 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
901 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
902 performed in cm_ReleaseSCache() when the reference count hits zero.
904 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
906 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
909 * Ensure that scache object 'nextp' pointers are set to NULL
910 when the objects are removed from the hash table linked list.
912 * If a dirty buffer cannot be written to the file server because
913 of an access denied error, mark the error in the cm_buf_t
914 structure so that it can be dealt with instead of retrying
918 * Update CellServDB file to GCO Public 23 Apr 2008
920 * Fix a cm_buf_t reference count leak when attempts to write
921 dirty buffers to the file server from within cm_IncrSyncer()
924 * Prevent udebug from crashing.
926 * Activate RX Idle Timeouts. If the file server is busy for more
927 then 30 seconds, attempt to failover to another server without
928 marking the busy server down.
930 * Another VNOVNODE issue fixed. When writing a dirty buffer
931 to the file server, if VNOVNODE is received, mark all buffers
932 as invalid without further attempts to contact the file server.
934 * Improved performance on high latency links.
937 * Fix a bug preventing the re-initialization of the Freelance
938 root.cell contents when a change is made.
940 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
941 or mountpoint that does not exist.
943 * aklog, afscreds, integrated logon, and the NIM AFS credential
944 provider will all attempt to first acquire an afs service ticket
945 using the service principal: afs/<cell>@<CLIENT_REALM>. This
946 permits the proper detection of the cell realm when multiple
947 realms are local realms for the cell.
949 * Fix fs commands that re-use file identifiers instead of forcing
950 a path evaluation each time. This was broken in 1.5.33 when
951 hashes were added to the cm_fid_t structure.
953 * Fix symlink and mountpoint removal and then recreation when
954 applied to the fake Freelance root.afs volume.
956 * Fix SMB change notifications for the Freelance root.afs volume
957 when creating or removing symlinks and mount points.
959 * Force DNS AFSDB cell name lookups to lower case if the DNS
960 response is not all lower case.
962 * 32-bit installers are now built with Visual Studio 2005
964 * All vos.exe commands now support the "-noresolve" option which
965 forces the display of IP addresses instead of DNS hostnames.
967 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
968 permits operations that do not require up to date volume
969 location data to continue while waiting for the RPC to complete.
970 Multiple threads waiting for a volume location update on the
971 same volume set no longer trigger multiple calls VL_xxx RPCs.
973 * Volume server references could be marked offline and never
974 reset to the not_busy state.
976 * Further optimizations to cm_Analyze() retry logic for
977 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
980 * Optimize the DNLC by applying Interlocked incrementing for
981 statistics gathering and enabling greater use of read
982 locks instead of write locks
984 * Further optimize osi_Log operations when logging is disabled
985 by removing the need for the function call overhead.
987 * Add a new lock operation, Convert Read to Write, and apply
988 it throughout the cache manager. This function provides a
989 fast transition from a read lock to a write lock when the
990 caller is the only reader.
992 * Fix a timeout problem when opening files or request locks
993 when the file in question is already write-locked by another
996 * Fix a deadlock that can occur if "fs flushall" is executed
997 while applications are writing to AFS.
999 * The cache manager would re-read data from the file server
1000 after file truncation when obtaining buffers to write to the
1001 same file until the first StoreData RPC completes.
1003 * Fixes error handling in cm_NTCheckDelete()
1005 * Replaces cm_scache_t mutex with a read-write lock permitting
1006 additional parallel access.
1008 * Prevent an cm_scache_t reference undercount when evaluating
1009 symlinks containing @sys when none of the active sysnames
1010 match any of the available targets.
1012 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1015 * Set RxMaxMTU registry value to 0 which is the equivalent of
1016 it not being set at all.
1018 * Reduce contention between BkgDaemon threads
1020 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1023 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1024 objects to use the faster Interlocked operations which avoid
1025 the need to hold write locks.
1027 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1028 non-fatal error that forces a retry against another service
1031 * Fix a crash in smb_WriteData that would occur if an application
1032 opened the file as read-only and then sent a write request anyway.
1035 * Respond to VL_NOENT errors by removing cm_volume_t object and
1036 setting it to be the first object to be recycled. This is not
1037 only a more efficient use of resources but it also prevents
1038 repeated attempts to query the VLDB server for the volume from
1039 the CheckOfflineVolumes routine.
1041 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1044 * A first cut at a cache manager statistics monitor that can be used
1045 to determine the necessary cache parameters to support the
1046 application working set.
1048 Off by default, the performance package can be activated by setting
1049 "daemonPerformanceTuningInterval" in the service Parameters key.
1050 As with the other daemon interval values the unit is in seconds.
1052 At service start and each succeeding interval the cache manager
1053 will write statistics to %TEMP%\afsd_performance.txt showing the
1054 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1055 The FID statistics keep track of all FIDs seen by the cache manager
1056 during the service session whether or not they are backed by any
1057 live object in the cache.
1059 These statistics are not stored in the cache file.
1061 * Fix windows event logging broken in 1.5.33.
1063 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1064 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1065 as well as when computing the server preference list.
1067 * When merging status after a StoreData operation prevent data
1068 buffers from being discarded that we would prefer to keep.
1069 (Bug introduced in 1.5.33.)
1071 * Modify installers the set RxMaxMTU to 0.
1073 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1074 dialog, the Background Daemons and Service Threads fields were
1077 * Increase default number of Background Daemons to 4.
1079 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1080 cm_volume_t objects to be recycled.
1082 * Windows Error Reporting has shown a number of previously unseen
1083 stack corruption errors in the Explorer Shell extension. The
1084 minidump appeared to indicate the problem was in ParseAcl().
1085 Reviewing the function it appears that numerous things could go
1086 wrong if invalid input was provided. Added error checking and
1087 data validation. Hopefully this will address the problem.
1089 * Windows Error Reporting has shown crashes in aklog as a result
1090 of krb5_cc_get_default() failures not being handled. They are
1094 * The Rx library used a 32-bit type for sockets which was
1095 truncating the socket value on 64-bit Windows. This
1096 was not an issue in general but became an issue when
1097 Rx "hot thread" support was enabled.
1099 * Enable "Rx hot thread" support which ensures that while
1100 a received message is being processed that a new thread
1101 is allocated to receive the next message to arrive.
1103 * Increased parallel access to smb_rctLock protected
1104 resources. Reduce the number of times that the
1105 lock is dropped and re-obtained.
1107 * Increased parallel access to cm_scache_t and cm_buf_t
1108 resources by replacing the previous global write lock
1109 protected reference counting mechanism with one built
1110 upon the use of InterlockedIncrement and InterlockedDecrement.
1112 * Implement a hash for cm_fid_t to reduce comparison costs
1113 and inline the cm_FidCmp() function.
1115 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1116 hash which has a better distribution
1118 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1119 the most recent read request instead of the next chunkSize.
1120 cm_chunkSize can be dozens or hundreds of buffers. As a result
1121 too much time is spent performing the evaluation.
1123 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1124 lock is to prevent the creation of new buffers while a truncation
1125 is being performed. All references to bufCreateLock have been
1126 removed except in two places:
1127 i. a write-lock surrounding the function that calls buf_Truncate()
1128 ii. a read-lock within buf_GetNewLocked() that actually allocates
1131 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1132 CM_CONFIG_DATA_VERSION number which value be used to force the
1133 replacement of the cache file contents when incompatible changes
1134 are made between releases.
1136 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1137 cm_SyncOpCheckContinue
1139 * Avoid calls to multi_Rx if nconns == 0
1141 * Modify smb_WriteData to perform background writes based upon the
1142 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1143 boundaries. This will slow down writes from the SMB interface but
1144 will avoid the risk of the CIFS client disconnecting from the AFS
1145 client SMB server. The default value for 'smb_AsyncStoreSize'
1148 * SMB AsyncStore functionality is on by default but can be
1149 disabled using the registry value:
1151 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1153 smb_AsyncStoreSize can be adjusted with
1155 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1157 The value must be a multiple of the buffer block size
1158 and cannot be larger than the chunk size.
1160 * Do not leak a read lock on the bufCreateLock
1162 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1165 * In MergeStatus, only remove clean buffers from the hash table.
1167 * Avoid the need to update the data version number on each buffer
1168 associated with a scache when MergeStatus is called after a
1169 StoreData by maintaining a range of valid data versions as part
1170 of the cm_scache_t object.
1172 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1173 from cm_HaveAccess... We can safely avoid the call so do so.
1175 * Add "ForceAnonVLDB" service parameters registry value. Default
1176 is 0. Set to 1 to force the use of anonymous connections to
1177 the VLDB servers. This is useful when using krb5 tokens and the
1178 VLDB servers have not been updated.
1181 * Do not attempt to search for a cell name that matches one
1182 of the special share names "svrsvc", "wkssvc", "ipc$"
1184 * Issue permission denied warnings from afscreds if attempts
1185 to modify the startup shortcut or registry values fail.
1186 This happens under Vista UAC.
1188 * Do not return BAD_NETWORK_ERROR in response to attempts
1189 to create directories or files in the Freelance root.afs
1190 volume. Instead return ACCESS_DENIED.
1192 BAD_NETWORK_ERROR was being sent because the Freelance cell
1193 has no vldb servers. Short circuit the error by testing for
1194 freelance in CreateFile and MakeDir
1196 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1197 to be opened and closed without actually existing. For now
1198 we treat them as equivalent to ioctls but this should
1201 * Fix RAP processing to not reject names that exist in the
1202 root.afs volume for GetInfo queries.
1204 * In afscreds, use GetUserNameEx to obtain the username instead
1205 of searching the registry. Try UserPrincipalName first and
1206 fallback to SamCompatible.
1208 * Stop linking to rpcns4.dll which is no longer supported on
1209 Vista and 2008. TaAfsAdmSvr is going to have to be
1212 * When processing dfs referral requests, if the sharename is
1213 a partial match and therefore will not be added automatically
1214 to the share list, make sure that we return an error.
1216 * Avoid a crash when logging VNOVOL errors returned from
1217 cm_CheckOfflineVolume.
1219 * When trace logging is disabled, do not permit osi_LogSaveString
1220 to perform the string copying. By disabling the string
1221 copies, throughput of the service more than doubles.
1223 * when checking offline volumes, update the cell vlserver info
1224 if required and don't perform a check if there are no servers
1227 * Add volumes in the vl_unknown state to the list of objects to be
1228 checked as offline or retried when an allbusy error is received
1231 * Add support for configuring rx_ExtraPackets
1233 * All calls to buf_Get() must be protected by a read lock on the
1234 cm_scache_t bufCreateLock
1236 * When a MergeStatus results in the invalidation of buffers, do
1237 not remove buffers from the hash tables that have active references.
1239 * Upon receiving a VNOVNODE error from a file server, be sure to
1240 invalidate any file locks. This prevents on-going attempts to
1241 extend the locks and will cause subsequent attempts to access
1242 the file to receive a invalid handle error.
1244 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1245 to directory entry names.
1247 * No longer discard tokens when RXKADBADTICKET error is received.
1248 Instead return STATUS_NO_KERB_KEY error.
1251 * Add "cmdebug <host> -cellservdb" option and client service
1252 implementation permits querying the vldb server for each
1253 cell known to the cache manager. The output is in a form
1254 that permits it to be used as a CellServDB file.
1256 * Add Symlinks->Show option to the AFS Explorer Shell Context
1259 * Prevent buffer overrun in the cmd parser used by command
1260 line tools such as fs, vos, pts, etc.
1262 * The path ioctl operations have several issues:
1264 (1) the specified path for "fs examine, whereis, whichcell,
1265 flush" and so always has the follow symlinks and mount
1266 points semantics. This makes it impossible to determine
1267 what the FID of a symlink or mount point is.
1269 (2) "fs examine" out is not the result of a single pioctl
1270 operation but is actually the combined output of half
1271 a dozen operations. Path evaluation is an expensive
1272 operation. It would be faster if the caller could
1273 evaluate the FID first and then perform all of the
1274 rest of the operations by specifying the FID instead
1277 (3) fs output reports all objects as files. By adding a
1278 GetFileType pioctl more informative output can be provided
1279 that indicates what type of object the path evaluates to.
1281 (4) the Windows fs command includes a number of commands that
1282 do nothing but exist only because the Unix cache manager
1285 A new extendible data structure cm_ioctl_query_opts_t has been
1286 added which can be optionally specified with pioctls that do not
1287 already require input data. The first two fields of this structure
1288 are 'literal' and 'fid'. The literal field is used to indicate
1289 whether the last component of the path should be evaluated following
1290 symlinks and mount points.
1292 The fid field permits a fid to be specified.
1294 * A new GetFileType pioctl has been added. The type of objects are
1297 * A new "-literal" option is available for "fs examine, flush, whereis,
1300 * Unimplemented fs commands have been removed.
1302 * There are circumstances where a volume object is being accessed and
1303 the volume is marked indicating that the volume location information
1304 is out of date but where it is also pointless and perhaps dangerous
1305 to block waiting for the rpc to complete. One example is when
1306 processing the cmdebug requests. If we know that we are not going
1307 to use the volume object to contact a server, then we can now set
1308 the CM_GETVOL_FLAG_NO_RESET flag.
1310 * when moving up in the directory tree we search the recorded fid list
1311 to find a matching fid that we have already crossed. we must also
1312 reset the fid count based upon what we discover.
1314 * Cell names are published as share names. As a result they are searched
1315 for as part of the Dfs Referral evaluation. If share "foo" can not be
1316 reached, the CIFS client will fallback to searching for "fo" as well.
1318 Since the freelance client automatically adds a symlink for prefixes
1319 of the cell name, this results in both "foo" and "fo" being added to
1320 the freelance root.afs volume. This patch prevents that negative side
1323 * when tracking Kerberos credential cache names be sure to include
1324 the cache type prefix.
1326 * Microsoft has assigned OpenAFS a network type value
1328 * The algorithms that were used to produce a Netbios name were broken.
1329 The name that was produced when the hostname was too long was in fact
1330 longer than the maximum permitted netbios name.
1332 Also, the "NetbiosName" value was not used as a suffix when the loopback
1333 adapter was not installed. The hardcoded string "AFS" was used instead.
1335 * avoid another deadlock during server probes initiated by the ipaddr
1336 change daemon thread.
1338 * prevent the afs client service from crashing when fs newcell is
1341 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1342 to \Program Files\OpenAFS\Client\Program in order to get them out
1343 of the %windir% tree
1345 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1346 end STATUS_OUT_OF_QUOTA can be returned to the application.
1348 * Modify the search order for determining the location of CellServDB
1349 and other client configuration files.
1351 1. AFSCONF environment variable
1353 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1354 4. use Program Files\OpenAFS\Client
1356 * More Vista Power Management tweaks
1358 * Close a socket handle leak in the kauth routines
1360 * Don't hold a cm_volume_t mutex across RPCs
1362 * Convert downTimes to use clock time instead of relative times.
1363 This permits correct comparisons with expiration times.
1365 * Update the Wix installer to properly install the prerequisites
1366 for translate_et.exe
1368 * Do not translate pioctl paths as they are not converted by the
1371 * Implement multi_rx probes for checking whether servers are up
1372 or down. Turns probes into a constant time activity instead
1373 of one dependent upon the number of known servers.
1375 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1376 update of the volume location data and retry. VNOVOL means
1377 that the volume could not be attached. It doesn't mean that
1378 the volume is not on this server.
1380 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1381 not attempt to validate the target path. There is no reliable
1384 * Improve performance of \\afs\<cell><mount-type><volume>\
1385 access by eliminating extraneous attempts to search the
1386 CellServDB file and DNS AFSDB records. Share names
1387 containing '%' or '#' are not valid cell names.
1389 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1390 to unlock a range that is not currently locked.
1392 * Enable the 32-bit tools to read the 64-bit registry keys
1393 when executing in the WOW64 environment.
1397 * Fix buffer refcount leak introduced in 1.5.29
1399 * Prevent the NIM AFS Provider from crashing when configured to use
1400 Kerberos v4 for token acquisition in situations where no Kerberos v4
1401 support is available. For example, 64-bit KFW or sites that delete
1402 the Kerberos v4 DLL.
1404 * In afsd_init.log, support logging cache sizes greater than 2GB.
1406 * The volume status output of "fs examine" was being determined based
1407 upon the wrong error value. Instead of using the pioctl() return
1408 code, the errno value must be used.
1410 * Prevent invalid pts auto-registration attempts by aklog and the
1411 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1413 * Add synchronization protection to all of the SMB Listener State
1414 variables. This prevents race conditions when the addition or
1415 loss of a network adapter takes place.
1417 * On Vista, there is a race condition between the restart of the
1418 afs client service and the network adapters. Ensure that the
1419 afs client service can handle switching between loopback and
1420 non-loopback modes of operation.
1422 * Fix FollowBackupPath to work with mount points that are not
1423 explicitly read/write.
1425 * Fix a deadlock that has been observed on Vista while resuming
1426 after at least four hours of sleep.
1428 * Add new HKLM registry configuration options that can be used
1429 to provide configuration hints to the NIM AFS Provider when
1430 creating new identities. See release notes for details on
1431 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1434 * Speed up write buffer operations by holding a global read-lock
1435 instead of a write-lock when updating the buffer data version
1436 number. The data version is protect by the associated stat
1439 * Speed up read buffer operations by removing out of date buffers
1440 from the buffer hash tables.
1442 * Protect against a null pointer dereference in afslogon.dll
1443 if QueryAdHomePathFromSid() fails.
1445 * Remove a race condition while accessing private freelance data
1448 * Update the buffer data versions when Freelance directory data
1451 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1454 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1455 instead of returning zero length.
1457 * Return ERROR_RETRY when a server sends VBUSY.
1459 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1460 a volstat plug-in is in use.
1462 * Enhance the volstat interface to support Dfs path mappings.
1464 * Fix "symlink list" to do the right thing for Dfs links
1466 * Fix callback expiration that was broken in 1.5.27.
1468 * Fix a crash at shutdown caused by the lack of synchronization
1469 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1471 * Hold the cm_server_t mutex while walking or modifying the
1472 cm_server_vols_t list in order to avoid a race condition.
1474 * Ensure that the cm_server_t refCount is safely modified.
1476 * The pthread library implementations of pthread_cond_wait
1477 and pthread_cond_timedwait were failing to return with
1478 the mutex held when error conditions occurred. This could
1479 result in race conditions wherever they are used. One case
1480 in which problems were seen was the Rx event_handler().
1482 * Prevent cm_GetBuffer() from locking all buffers associated with
1483 a stat cache object before deciding if the current buffer is up
1486 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1487 object for each offset that we will attempt to reserve.
1489 * Reduce the default number of background daemon threads used to
1490 prefetch or store to one in order to prevent two or more threads
1491 from stepping on each other.
1493 * Reduce the default chunk size to 256KB in order to prevent read
1494 timeouts on low bandwidth connections.
1496 * Fix a race condition that could result in threads becoming stranded
1497 when synchronizing multiple requests on the same stat cache object.
1500 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1501 has been disabled. It has been disabled because it turns out
1502 that the RPC is improperly implemented in all file servers
1503 previous to 1.4.6. The RPC is executed without the proper
1504 locks being held resulting in data corruption and eventual file
1505 server failure. The functionality will be re-enabled in a
1508 * a couple of .readonly Volume Callback optimizations were
1510 - Apply most recent volume callback to all stat cache objects
1511 in the volume. This avoids premature callback expirations
1512 and unnecessary FetchStatus calls.
1513 - Add option to permit attempts to renew .readonly Volume
1514 Callbacks on a periodic basis to avoid invalidation of stat
1515 cache data. [Requires daemonCheckVolCBInterval be set to
1516 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1518 * Adds support for 64-bit data version values. Previous releases
1519 only supported 32-bit values. Larger values sent by the file server
1520 would result in rollover.
1522 * aklog and asetkey no longer generate an exception if Kerberos for
1523 Windows is not installed.
1525 * If the LSA authentication functions report a temporary out of memory
1526 condition, fail the current authentication and do not cause the AFS
1527 client service to terminate unexpectedly.
1529 * When the AFS client service is shutting down, prevent the daemon check
1530 thread from executing additional checks which might attempt to access
1531 discarded locks or unmapped memory.
1533 * Correct a double free condition in the NIM AFS credentials provider.
1535 * Add "FollowBackupPath" service parameter registry value.
1536 This provides equivalent functionality to the UNIX afsd -backuptree
1537 option. When set to a non-zero value, normal mount points originating
1538 in a .backup volume will prefer a .backup volume.
1541 * Fix NSIS uninstall from the .MSI installer.
1543 * Add Stack Frame logging for AMD64
1545 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1546 server returns EWOULDBLOCK, retry the request every two
1547 seconds for up to the RDRtimeout.
1549 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1550 is not assigned the same value as EIO.
1552 * Update CellServDB to GCO Public 25 Oct 2007
1554 * Add additional validation and error handling code after
1555 each call to getSlot(). If an invalid slot is returned,
1556 return NONODE. If the invalid slot is returned when
1557 extracting a data node, invalidate the tree.
1559 * Modify compareKeys() to always perform a case-insensitive
1560 comparison and only perform a case sensistive comparison
1561 if the case-insensitive one matches. This ensures the
1562 ordering is consistently reported.
1564 * Add lock assertions to ensure that all calls are being
1565 performed with the correct locks being held. There have been
1566 some crash reports that provide stack data that does not appear
1567 to be possible unless there is a race. However, there are no
1568 obvious locations where the race is taking place and the test
1569 suite indicates that all of the correct locks are being held.
1570 We shall see what happens in the field.
1572 * For consistency replace all calls to findKey in which the range i
1573 s (1,numentries) with calls to getSlot().
1575 * Optimize the depth search loop by testing the slot value in the
1576 for statement instead of forcing the loop to be broken later.
1578 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1579 There are a number of locations where locks are obtained, dropped,
1580 and reobtained. This reorganization attempts to accomplish several
1583 (1) be optimistic for the most common case so it will be fast
1585 (2) add consistency checks after each location where locks are
1586 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1587 and the bplus tree is out of date, retry until we get to a
1588 consistent state that we can use.
1590 (3) Ensure that all operations take place with the correct locks.
1592 * One of the issues that has become a serious problem since the
1593 addition of the local directory updates is that although cm_SyncOp
1594 synchronizes operations, it does not preserve the order of requests.
1595 This has always been a problem in that it has been possible for a
1596 request to fail to complete due to its worker thread's bad luck.
1597 When a request takes longer than the Windows SMB Redirector's timeout,
1598 the SMB Redirector tears down the SMB virtual circuit.
1600 When using the local directory updates it is really important that
1601 the directory update operations complete in the order that they were
1602 sent to the file server. If they don't, then the local directory state
1603 and the file server state will not match and the local directory state
1604 must be discarded which in turn forces a new read of the entire
1605 directory contents over the network.
1607 This patch adds a new cm_scache_waiter_t object that is used to store
1608 the current thread, buffer, and syncop flags within a waiters queue
1609 on each cm_scache_t object. If a thread is forced to sleep in
1610 cm_SyncOp, upon waking it will check to see if there are any other
1611 threads waiting that are attempting to perform a similar task ahead
1612 of it in the queue. If yes, the thread goes back to sleep. If not,
1613 it goes ahead and enters the cm_SyncOp conflict resolution block.
1615 This patch has the additional side effect of reducing the number of
1616 competing threads that must obtain the cm_scache_t mutex and process
1617 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1618 utilization of the service and the clock time associated with processing
1619 requests will be reduced.
1621 * assert that the cm_scache_t mutex is held when calling
1622 buf_ForceDataVersion().
1624 * rename findNode to leafNode in bplus_Lookup
1626 * replace all OutputDebugString calls with osi_LogX calls
1628 * modify bestMatch to special case the return values for leaf nodes.
1629 If an entry is above or below the values available in the leaf node
1630 return BTLOWER or BTUPPER instead of BTERROR.
1632 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1633 if true convert to either slot 0 or Max and perform the insertion.
1634 This produces easier to read code when performing lookups.
1636 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1637 descriptive message.
1639 * the difference between osi_assert() and osi_assertx() is whether or not
1640 a description is specified as part of the call. When no message is
1641 specified afsd_notifier() is called with a NULL msgp parameter. This
1642 results in a null pointer dereference during the EventLog() call.
1644 As a result, none of the information describing the assertion is written
1647 This commit sets a default message for use when no other message is
1648 provided and it replaces all calls to osi_assert with osi_assertx and
1651 * Migrate B+ tree search key into thread local storage
1654 * Fix an uninitialized variable in lana_IsLoopback() which
1655 could result in a random determination that a loopback
1656 adapter is not a loopback adapter. This would result in
1657 a computed Netbios service name of %MACHINE%-AFS instead
1658 of AFS. This is a serious problem if the value computed
1659 by the AFS Client Service and the client process pioctl
1660 calls produce different results as it will result in a
1661 communication failure.
1663 * Fix a crash in the B+ Tree code that results when a file
1664 rename only results in a change of case.
1666 * Fix the pioctl path parsing. The previous algorithm
1667 broke the current directory of the client into two parts,
1668 the TIDpath which represented the device root and the
1669 relative path. The problem is that this is not sufficient
1670 from proper processing of symlinks and mount points.
1672 What is required is that the last component of the path
1673 must not have the symlink or mount point be evaluated
1674 whereas symlink and mount point evaluation must be performed
1675 for the rest of the relative path. Therefore, the path is
1676 now split into three parts: TIDpath, intermediate path,
1679 "fs examine" now reports the data of the symlink or mount
1680 point object instead of the target and "symlink list" and
1681 "symlink remove" now works. The "Explorer Shell Extension"
1682 menu now is displayed for symlink objects.
1684 * More informative symlink.exe error messages.
1686 * When the cmd.exe "del" command is used to remove a symlink,
1687 not only was the link object removed, but the stat cache
1688 object for the target was marked deleted. This resulted
1689 in cache corruption which would require the destruction
1690 of the cache to correct.
1692 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1693 which results in cached contents from read only volumes
1694 to be treated as valid when there is no callback registered
1695 and all servers are down or all instances are offline.
1697 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1699 * Add documented registry value, "AcceptDottedPrincipalNames"
1700 which is used to permit the acquisition of AFS tokens
1701 using Kerberos v5 principal names which include a dot in
1702 the first component.
1704 {HKLM\HKCU}\Software\OpenAFS\Client
1706 * "fs uuid" does not require "AFS Client Administrator" group
1707 membership to display the current UUID value.
1710 * Update NSIS installer to version 2.30
1712 * Update Wix installer to version 2.0.5325
1714 * 64-bit KFW support:
1715 - Network Identity Manager plug-in
1717 - 64-bit krb5 enable afscreds.exe, afs server manager,
1720 * Enable B+ tree Directory Search by default.
1722 * NIM plug-in is now compatible with FILE: ccaches
1724 * Install instloop.exe into Client/Programs/ even though
1725 it isn't required. Permit it to be used by organizations
1726 that clone Windows systems.
1728 * On a suspend, only drop callbacks for servers that
1729 are online. If the server is offline, maintain the callback
1730 until the server is once again available.
1732 * Prevent against acquisition of a directory write lock when
1733 there are no updates to be performed. This was resulting
1734 in a write-lock leak which later on would lead to a deadlock.
1736 * Add configurable file pre-fetch mode. For files that match
1737 the specified file extensions, pre-fetch the entire file into
1740 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1741 (MULTI_SZ) PrefetchExecutableExtensions
1743 * Fix the client_osi large integer support so that return values
1744 can be properly assigned to large integer variables.
1746 * Do not hold cm_scache_t mutex across cm_GetCell() calls
1747 since cm_GetCell() can block in a RPC.
1749 * Add short file name support to B+ tree searches.
1751 * Resolve a conflict between the desire to notify file servers
1752 that the client workstation is suspending and the desire to
1753 permit objects with valid callbacks when the client suspends
1754 to continue to use those objects from the cache if the client
1755 resumes without network connectivity.
1757 * Fix stack variables to ensure that all buffers storing paths
1758 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
1759 would cause a stack overflow condition.
1761 * On Vista, force afs_config.exe to run with the highest privileges
1762 available to the end user. For a user in the Administrators
1763 group running under UAC, this will cause afs_config.exe to be
1764 be run with full Administrators privileges after the user
1765 approves the privilege escalation.
1767 * More VNOVNODE handling issues. If a dirty buffer cannot be
1768 written due to VNOVNODE, clear the buffer and ignore the error.
1769 If a cm_scache_t object has been marked deleted due to a VNOVNODE
1770 error, the next time the smb_fid_t object that references it is
1771 used return either NOSUCHFILE or NOSUCHPATH and close the
1772 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
1773 result in an invalid handle error.
1775 * For Vista, set the RequestExecutionLevel of the NSIS installer
1778 * For the build environment, support the generation of manifests
1779 when using the CL1310 (VS .NET 2003) compiler/linker.
1781 * There was a bug in OutputDebugHexString() which could overflow
1782 a stack buffer resulting in unexpected process termination
1783 of afsd_service.exe if the input Security Token contained
1784 many bytes equivalent to the '%' char.
1786 * Add directory lookup statistics to "fs memdump" output.
1790 * Windows uses case-insensitive file name pattern matching
1791 but AFS is a case sensitive file system. The AFS3 directory
1792 format is block based, uses network byte order and
1793 includes a hash table for fast case sensitive lookups.
1794 This causes several problems for the Windows AFS client.
1795 (1) Traversing the directory blocks is cpu expensive
1796 (2) A hash table miss does not indicate that the desired
1797 entry does not exist.
1798 (3) Determining whether a non-ambiguous inexact match or
1799 the entry does not exist requires a linear traversal
1800 of the entire directory.
1801 These issues often result in 100% CPU utilization.
1803 These issues are addressed by building a modified B+ tree for
1804 each directory and then using the B+ tree for searches.
1806 This feature is disabled by default. To enable it set
1807 the "BPlusTrees" registry value to 0x01 in the AFS Client
1808 Service Parameters key.
1810 * Correct a null pointer dereference in the symlink recursion
1811 detection code. (Introduced in 1.5.22)
1813 * Add the Network Identity Manager AFS Provider to AMD64
1818 * A new registry value, "BlockSize", can now be used to adjust
1819 the size of the internal buffers used to store file data.
1820 The default is 4KB. When set, the values 1..1024 specify
1821 multiples of 4KB. Values > 1024 specify actual block sizes
1822 and must be multiples of 4KB. The actual block size in KB
1823 can not be larger than the chunk size which defaults to 1MB.
1825 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1828 * pts.exe and ptserver.exe built with -DSUPERGROUPS
1830 * Service Manager shutdown hints have been added. The hint
1831 specifies that it may take the AFS client service up to two
1832 minutes to shutdown.
1834 * Fix the recursion detection feature introduced in 1.5.22.
1835 The previous implementation broke the evaluation of paths
1836 using ../path or ./path. It also broke the evaluation of
1837 symlinks to absolute paths. For example, /afs/cell/path.
1839 * Fix an uninitialized variable in cm_MakeDir that could result
1840 in crash if the requested directory could not be created.
1842 * Fix a the VolumeStatusNotification log messages to save
1843 the input string in the circular log buffer so garbarge
1844 won't be output to afsd.log when the log is dumped.
1847 * Increase the SMB Maximum Path length to 260 characters
1848 from 256 as that is what CIFS supports.
1850 * On 64-bit Windows, use a unique description string for
1851 the 32-bit AFS Shell Extension.
1855 * In the AFS Admin Library, replaced all of the
1856 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
1857 allows parameter checking and proper type conversions to be
1858 performed. This is crucial when passing parameters that are
1859 64-bit wide on systems in which 'long' is a 32-bit value.
1860 Otherwise, the parameter list on the stack is interpretted
1861 incorrectly and the function not only doesn't succeed but it
1862 might corrupt memory and if lucky crash the application.
1864 The only program that we care about that uses this library
1865 is the AFS Server Manager. Now it doesn't crash.
1867 * Fixed smb directory searches for non-wildcard names. The
1868 optimized function should not have set the bulk in progress
1869 flag. This would adversely affect later directory searches
1870 that required a bulk search.
1872 * Detect infinite recursions (up to 512 stat cache objects
1873 in a chain.) If an infinite recursion is detected, return
1874 an error to stop the CIFS client from querying us forever.
1876 * Executables executed out of AFS name space can have their
1877 pages swapped out by Windows. Those pages will be reloaded
1878 when they are required. However, if the cache manager no
1879 longer has a callback and all of the file servers with
1880 the volume data are down, then the application will crash
1881 due to the inability to read the required page.
1883 We now track the down time of each server. If the callback
1884 expires after the server that issued it went down and all
1885 servers are down, we delay the expiration to permit the data
1886 in the cache to be continue to be used by Windows or an
1889 * The Windows AFS client has always had very poor performance
1890 with regards to directory management. In the case of
1891 directory updates caused by create, rename,and delete operations,
1892 the AFS cache manager will now perform local directory updates
1893 when possible in order to avoid re-reading the directory from
1894 the file server. This will be done in whenever both of the
1895 following conditions are true:
1897 - the data version on the directory increased by one as a
1898 result of the operation
1900 - all of the directory buffers necessary to make the change
1901 are present in the cache
1903 * The error message in the AFS Control Panel describing the
1904 need to be a Windows Administrator has been replaced with
1905 a message describing the need to be in the AFS Admins Group.
1907 * If the vldb server list is empty when attempting to evaluate
1908 a cell mountpoint it was possible for a crash to occur. This
1909 can only happen if there is an existing mount point and the
1910 is a failure attempting to load the server list from the
1911 CellServDB file or via AFSDB queries.
1913 * dirty buffers only have their dirty bytes stored to the file
1914 server instead of the entire chunksize
1916 * the default chunksize has been increased from 128KB to 1MB
1921 * changed the enum values for cm_serverRef_t state info to use a
1922 private name space to avoid collisions (srv_)
1924 * added a srv_deleted state for cm_serverRef_t objects. This
1925 state is set when cm_FreeServerList() is called with the
1926 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
1927 not always delete the cm_serverRef_t from the list if it is
1928 still in use by another thread. the srv_deleted state means
1929 the object's contents are no longer valid and it must be
1930 skipped. It will be deleted the next time the object is
1931 freed and the refcount hits zero.
1933 * the srv_deleted state is also used when a file server reports
1934 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
1935 as offline. This is done to prevent additional usage of the
1936 stale vldb data while waiting for the update volume request
1939 * added a state field to the cm_volume_t object (enum volstate
1940 vl_ name space) that maintains the state of the volume based
1941 upon the states of all of the cm_serverRef_t and cm_server_t
1944 * modified cm_UpdateVolume() to set the state of the cm_volume_t
1945 RW, RO, and BK to either vl_alldown or vl_online. There can't
1946 be any other states because cm_UpdateVolume() destroys any
1947 previous knowledge we might have had regarding busy or offline
1950 * modified cm_UpdateVolume() to update the volume name in the
1951 cm_volume_t to the volume base name if the previous value was
1954 * modified cm_FollowMountPoint() to check to see if the volume
1955 name is a volume ID and if so call cm_GetVolumeByID instead
1956 of cm_GetVolumeByName. This ensures that volume IDs are always
1957 looked up as numeric values. There is no longer a need to
1958 maintain a separate cm_volume_t containing the string representation
1961 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
1962 The first flag is a "CREATE" flag which is set by all existing
1963 calls. The flag is not set by calls to cm_GetVolumeByID() from
1964 the server probe code when volume status is being updated. We
1965 do not want the server probe operation to result in additional
1966 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
1967 which is set when the server probe code updates the volume status.
1968 This flag will be used to prevent the server probe operation from
1969 changing the order of the least recently used queue.
1971 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
1972 allocated for a given set of normal, readonly, and backup volumes
1973 regardless of whether or not the volume is accessed via name or
1974 ID number. The cm_volume_t namep field is always the base name
1977 * Added a new volume state, vl_unknown. This state is used as
1978 the initial state for all cm_volume_t when the cache manager starts,
1979 for each cm_volume_t at creation, and for each cm_volume_t when
1980 recycling. The cache manager does not know the state of all
1981 volumes in the world, only those that are in the cache and for
1982 which it has queried the VLDB and hosting file servers.
1984 * modified cm_GetVolumeByName() to initialize the state of a
1985 volume to vl_unknown. The actual state will be set when a
1986 cm_VolumeUpdate() call completes successfully.
1988 * changed name of scache hash table variables to avoid ambiguity
1989 when adding hash tables for volumes
1991 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
1993 * modified cm_UpdateVolume() to handle the case in which there is
1994 no RW volume but there is are RO volumes for a given base name.
1995 This is done by querying for the ".readonly" volume name if the
1996 base name does not exist in the VLDB. We never query for the
1997 .backup name because under the current usage model a .backup
1998 volume may only exist on the server that the read-write volume
1999 is located. If there is no RW volume, there can be no .backup.
2001 * Added four hash tables for cm_volume_t objects to improve the
2002 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2003 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2004 BKID. Three ID hash tables are necessary as long as it is
2005 desireable to maintain a single cm_volume_t containing all
2006 of the related RW, RO, and BK volume data. Having the RW and
2007 RO volume data in the same object is necessary for the
2008 implementation of cm_GetROVolumeID() which returns either the
2009 RO or RW ID depending upon the existence of RO volume instances.
2011 * Added a volume LRU queue so that volume reuse becomes fairer.
2012 This does not replace the all Volumes list which is used when
2013 it is desireable to walk a list of all the volumes whose order
2014 is not going to change out from underneath you which makes it
2015 safe to drop the cm_volumeLock.
2017 * handles volume hash table updates where volume name to
2018 volume ID number changes. The volume name remains
2019 constant in the cm_volume_t. if a vos rename is performed,
2020 the name of the volume will change and the volume IDs will be
2021 updated. Subsequent access to the old volume ID will create a
2022 new cm_volume_t with the new name.
2024 * Added a daemon thread operation to query the state of volumes
2025 listed as busy or offline. cm_CheckBusyVolumes() calls
2026 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2027 or vl_offline. If the volume is now online, the status on the
2028 volume is updated. The default period is 600 seconds. This can
2029 be configured with the BusyVolumeCheckInterval registry value.
2031 * Added prototype for smb_IoctlPrepareRead() which was missing a
2032 return type in the function definition.
2034 * Added volume id lists to the cm_server_t. These lists are
2035 allocated in blocks of ~32 IDs. When a cm_PingServer()
2036 detects a change in server state, the state of the cm_volume_t
2039 * Added volID to the cm_serverRef_t object. volID is used
2040 to identify the volume for which the object is a referral.
2041 cm_FreeServerList() uses the volID to remove the volume
2042 from the cm_server_t.
2044 * In cm_Analyze, when VNOVOL or VMOVED are received,
2045 call cm_ForceVolumeUpdate() to force a refresh of the volume
2048 * Added cm_VolumeStatusNotification() which is used at the moment
2049 to log volume status changes to the trace log. It will also
2050 be used as the access point to the File System Filter driver
2051 notification engine.
2053 * Added an all cm_scache_t list to cm_data. This replaces the use
2054 of the stat cache LRU queue when we need to enumerate all
2055 entries. The LRU list order is not static and when using it to
2056 enumerate all entries it can result in items being missed or
2057 items being processed more than once.
2059 * Modified cm_Analyze(). Instead of reseting the busy or offline
2060 state of a volume and forcing a retry of the operation
2061 cm_Analyze will defer to the background daemon thread that will
2062 update the state once every 600 seconds.
2064 * Added the automatic generation of a Freelance ".root" read-write
2065 mountpoint that refers to the root.afs volume of the workstation
2066 cellname at the time the mountpoint is created.
2068 * re-write cm_Analyze to make better use of the known volume
2069 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2072 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2073 busy volumes will be reset to srv_non_busy by the function
2074 but there is no mechanism for querying the busy state other
2075 than by attempting to access the resource.
2077 * cm_Analyze will query the state of an offline volume before
2078 deciding whether or not to retry when all volume instances
2081 * the algorithm used after an IP address list change detection
2082 has been modified. we now check the state of the VLDB servers,
2083 then reset the rx connections, and finally check the state of
2084 the file servers. Previously we flushed the rx connections
2085 and then checked the servers in whatever random order they
2086 were first discovered.
2088 * fix the volume status reported by "fs examine"
2090 * various improvements to the AFS Network Identity Manager Provider.
2091 - Don't access freed objects after the window has been destroyed.
2092 - Write back the AFS cell list in the identity configuration
2093 even when its empty.
2094 - Disable all use of the provider when AFS is disabled for a
2095 particular identity.
2097 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2098 alias instead of the full cellname, double check the full cell
2099 name and see if we have a matching cell before allocating a new
2102 * Dump more data with "fs memdump". All volumes and scaches and
2105 * Reverse the order of the all cells list. This ensures that
2106 the workstation cell is always first in the list.
2108 * If we recognize the server from which a callback revoke was
2109 received, then we can ensure that we only invalidate the status
2110 information for cm_scache_t objects belonging to that cell.
2112 * When building for amd64, use APPVER 5.02 as we do not support
2113 64-bit Windows 2000 and explicitly state the machine type in
2116 * no longer treat VMWare network adapters as equivalent to a
2117 loopback adapter. For some unknown reason, publishing the "AFS"
2118 netbios name on them frequently results in a name collision
2121 * Use read locks whenever possible instead of write locks when
2122 walking the allCellsp list. This permits better use of multiple
2125 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2126 is not present on the directory optimization. It would report
2127 mount points and directory objects for which there was no
2128 status information as a zero length file. Instead we can use
2129 the vnode number of the object to determine if it should be
2130 considered to be a directory or a file. If it is odd it is a
2131 directory; even it is a file.
2133 * Give Back Callbacks to the file servers in response to suspend
2134 and shutdown power management events. Only give up callbacks
2135 to UP servers and do so with a very short RX dead timeout.
2138 * Fix a race condition in the SMB FID cleanup code that could
2139 result in a stat cache mutex never being released.
2141 * Fix the graphics files that are installed as part of the
2142 IBM Administrators Guide.
2144 * Change the behavior of pioctl path parsing to not follow
2145 symlinks or mount points during evaluation. This results in
2146 the FID being the symlink or mount point and not the object
2147 they point to. This is consistent with the behavior of the
2150 * Enable AFSDB record lookup for command line tools such as
2153 * CellServDB update 2007-07-05
2155 * Added a Volume Status Plug-in Interface that permits third parties
2156 to receive notifications when the volume state changes occur.
2158 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2160 * Prevent the afsprocmgmt.dll from being unloaded while the
2161 background thread it starts is running.
2163 * Avoid a race condition when closing smb_fid_t objects.
2166 * fix install path of IBM Quick Start Guide when installed by
2169 * fix the client so that it can write into directories with
2170 only 'li' permissions.
2172 * fix installation of 32-bit Shell Extension for 64-bit systems.
2173 Install it under its own HKCR keys.
2175 * modify the wix installer to remove and upgrade install packages
2176 with the same version number.
2178 * when flushing volume location data every two hours or in
2179 response to "fs checkvolumes", also flush the mount point
2180 links in order to force their re-evaluation.
2182 * aklog was copying an uninitialized buffer as part of the
2183 foreign cell automatic pts registration
2185 * when obtaining directory listings, check for read permission
2186 on the directory. If read permission does not exist, fake the
2187 status information for the directory entries. This improves
2188 performance, prevents the client from pounding the file server,
2189 and prevents leakage of cached status information.
2191 * "fs getcacheparms" output was incorrect.
2193 * NetIdMgr AFS Provider should only automatically add token
2194 acquisition for cell foo.com if the realm of the cell matches
2195 the realm of the identity.
2197 * Fix a null pointer dereference in cm_MergeStatus()
2201 * network identity manager provider:
2203 - Correct for the fact that pr_CreateUser tampers with the
2204 contents of the username input buffer.
2206 - Take into account that the krb4 credential data structure
2207 does not have separate realm fields for the user and service.
2208 When krb524 is used with a cross-realm ticket, the realm
2209 describes the service not the user.
2211 * Many integrated logon changes:
2213 - Fix eventlog reporting. Do not attempt to log an event
2214 if the event source registration fails. Use DebugEvent0
2215 instead of DebugEvent when there are no parameters.
2217 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2220 - Fix the reading of Domain specific configuration for
2221 LogonScript and TheseCells. Previously the dwSize value was
2222 being overwritten so that subsequent RegQueryValueEx call
2225 - Fix a memory leak in the TheseCells reading code.
2227 - Add support for Domain specific "Realm" specification. The
2228 realm is the realm to be appended to the username. When
2229 logging in as a domain or to the local machine, the specified
2230 "Domain" name is not going to be a valid realm name.
2232 - Construct a proper principal name based upon the domain
2233 specified realm for use in obtaining tokens with KFW.
2235 - If the domain specified "TheseCells" list includes the
2236 default cell, do not obtain tokens twice.
2238 * An unprotected call to the Kerberos v5 function,
2239 krb5_free_context(), can result in an invalid memory access
2240 exception if the contents of the krb5 profile, krb5.ini, cannot
2241 be parsed correctly. The error is located within the afskfw
2242 library which is used by afscreds.exe, afssvrmgr.exe, and
2243 afslogon.dll. When the error is triggered within afslogon.dll
2244 a blue screen condition can occur when the winlogon.exe system
2245 process terminates unexpectedly.
2249 * Do not stop the afs client service with an assert
2250 if the cell name cannot be determined when responding
2251 to a "fs whereis" command.
2253 * Integrated Logon (afslogon.dll). No longer attempt
2254 to preserve Kerberos 5 tickets for the logon session
2255 as that job is now supported by KFW 3.1 and above.
2257 * The Network Identity Manager plug-in will attempt to
2258 remove expired service tickets from the cache.
2262 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2264 * AFS Server Manager and AFS Account Manager data cache
2265 moved from HKCU\Software\TransarcCorporation to
2266 HKCU\Software\OpenAFS to prevent crashes due to data
2269 * Reduce the default UP server probe time from 10 minutes
2270 to 5 minutes to match UNIX. This will not solve the NAT
2271 port mapping timeout issue because most NATs default to
2272 a value less that 5 minutes and this value ensures that
2273 the afs client server will start probes 5 minutes after
2274 the last probe set ended. The probe period for any one
2275 server is therefore longer than 5 minutes. The more
2276 servers in the list the longer the time between probes.
2278 * Before marking a server as down in cm_Analyze, retry the
2279 request using a new rx connection to the server. If the
2280 cache manager is behind a NAT, it is possible that the
2281 response from the server is being sent to the original
2282 address:port from which the rx connection was established.
2283 This will appear to the cache manager as the server being
2284 down when in fact it is not. Forcing the use of a new rx
2285 connection will result in the server responding to the
2286 most recent address:port assigned to the client by the NAT.
2288 If the server is in fact down, it will simply take a little
2289 longer to notice it.
2293 * Removing a directory that contained entries was failing with
2294 the wrong error code being returned to the CIFS client.
2298 * rxdebug exponential back-off. don't take down a server that
2299 is already struggling to respond.
2301 * Network Identity Manager plug-in registers help on the Help
2302 menu if it is supported by NIM. (KFW 3.2 and above)
2304 * Plug-n-play networking did not always work after coming back
2305 from a Windows Suspend. The AFS client service failed to
2306 re-compute the LAN adapter to bind to and instead attempted
2307 to bind to all LAN adapters. This would polute the public
2308 network with the name "AFS" which could lead to conflicts
2311 * Build installers using Microsoft Hotfixes provided at
2312 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2313 Daylight Savings Time issues.
2315 * "fs checkserver" does not require admin privileges unless
2316 the probe interval is being changed.
2320 * A reference leak was discovered in the mountpoint traversal
2321 functionality which resulted in cm_volume_t objects not becoming
2322 eligible for re-use. This will trigger a panic condition when
2323 more than the maximum number of volumes is accessed.
2325 * Fixed the success test for the "fs memdump" operation. Success
2326 is now reported as such. Added cm_volume_t object information
2329 * A data corruption error has been fixed. Write requests
2330 queued for background processing were not retried upon failure.
2331 This leads to inconsistencies between the contents of the file
2332 in the local cache and the contents of the file on the file
2335 The implemented fix is to re-queue any request that has failed
2336 due to timeout, volume busy, volume offline, server busy, or
2337 server offline. Requests are not processed while all servers
2338 for a volume are inaccessible. When servers become accessible
2339 the background request will be retried.
2341 In addition, if a background write fails, do not mark the buffers
2342 as clean. Failure to preserve the dirty flag prevents the buffers
2343 from being written to the file server in the future.
2345 This fix does not address the possibility that the credentials
2346 necessary to perform a write request might expire between the
2347 time the request is queued and when the request can be processed
2348 due to server accessibility.
2350 * The Freelance volume does not support per user ACLs. The
2351 cm_HaveAccessRights function did not handle this special case.
2352 Requests for privileges greater than read|lookup would result
2353 in afsd_service.exe entering an infinite loop. ISOBuster is
2354 known to cause this behavior.
2356 * When deleting files, the SMB Server would attempt to obtain
2357 the access permissions for the parent directory instead of
2358 for the file itself. This would result in an incorrect
2359 assessment of whether or not the file can be deleted.
2361 * The Network Identity Manager AFS plug-in has had its string
2362 tables updated for consistency with the language used by NIM.
2363 All credentials are referred to as credentials instead of tokens.
2364 Drop-down combo boxes now in fact drop down.
2366 * The AFS Salvager was not properly being built on Windows.
2367 This is relevant only if the AFS servers are in use.
2369 * The default signal handlers for the AFS servers incorrectly
2370 processed SIGQUIT. Instead of stopping the server, an
2371 exception was raised.
2373 * Fix another volume refcount leak. This one was triggered by
2374 an inability to contact the vlservers in order to update the
2377 * Fix checklist control used by the drive mappings tabs in
2378 afscreds.exe and afs_control.exe.
2380 * Improve output of "fs memdump" related to buffer management.
2382 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2383 already in the process of being fetched.
2385 * Improve cache performance by computing the buffer hash table
2386 size based upon the number of buffers so that the average
2387 number of entries in a bucket is seven.
2389 * Improve cache performance of read/write data by locally
2390 updating the dataVersion of all buffers associated with the
2391 file on each StoreData or StoreStatus operation provided that
2392 the dataVersion was incremented by one and the dataVersion of
2393 the buffer matched the original dataVersion of the cached
2394 status information for the file.
2396 * When recycling cm_scache_t objects, make sure that we
2397 clear all of the previous fields. It is no longer the
2398 case that we can assume that the stat cache object is being
2399 obtained only after a successful FetchStatus call.
2400 InlineBulkStatus will allocate stat cache objects for which
2401 the no status information is available.
2403 * When a file is being opened with the OPEN_ALWAYS disposition,
2404 do not require write permissions when the file already exists
2405 unless of course the file is being opened requesting write
2408 * When a file has been opened for read-only access, if the
2409 application requests an exclusive lock, obtain a read-lock
2410 instead of a write-lock. The purpose of the exclusive lock
2411 in this situation is to prevent changes to the file and the
2412 full file read lock will provide this guarrantee.
2417 * Enforce Unix Mode Write bit to prevent file deletion of read only
2420 * The CIFS READ and LIST permissions use the same value 0x0001.
2421 When a directory is being accessed, check for PRSFS_LOOKUP
2422 instead of PRSFS_READ.
2426 * Fix computations of short names of mount points. They were
2427 broken as part of the directory search optimization. (1.5.13a)
2429 * When all buffers are in use, periodically sleep the threads
2430 waiting for a buffer to become free while the global buffer
2431 lock is not being held in order to permit those threads that
2432 can free buffers to do so.
2434 * Power Management improvements. Maintain a global flag that
2435 specifies whether or not the service is in a suspend state.
2436 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2437 lana is no longer valid. Instead, stop the listener threads
2438 and if all listener threads are stopped, reset the lana_list.
2439 Allow the cm_Daemon() thread to periodically check the state
2440 of the smb listeners. If they are all stopped and the service
2441 is not suspended, attempt to restart them. If there are no valid
2442 lanas, return to the stopped state.
2444 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2445 the user is permitted to obtain read or write locks. This function
2446 would obtain the lock and then drop it returning whether or not
2447 the lock could be obtained. If the lock was in fact required,
2448 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2449 of course being that this pattern results in three RPCs to the
2450 file server (lock, unlock, lock). This is reduced to one RPC by
2451 implementing cm_CheckNTOpenDone() which frees the allocated byte
2452 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2455 * Optimize the SMB FindFirst processing for the case in which there
2456 are no wildcards. If a cm_Lookup succeeds, generate a response
2457 from the stat cache entry we have a callback for. Its cheaper
2458 than walking each buffer of directory data associated with the
2459 directory stat cache entry when we have thousands of directory
2462 * Increase the SMB Ioctl MaxData size to support larger token sizes
2464 * Disable autorun of the AFS Server Wizard when the AFS Server is
2467 * Add AFS Server service dependencies on tcpip and PNP_TDI
2469 * Remove the AFS Server Configuration Tool shortcut
2471 * Add Freelance mode awareness to AFS Server Configuration Tool
2475 * Unlike previous versions of the OS, Vista performs a shutdown on
2476 the Microsoft Loopback adapter just like it would on a real adapter.
2477 This causes the smb_Listener threads to trigger a panic during a
2478 suspend/hibernate power event.
2480 The fix is to unbind from the network adapters in response to a
2481 suspend/hibernate power event and then rebind to the adapters
2482 when a resume power event is received. Note that the resume events
2483 are not reliably delivered so it is possible the afs service will
2484 not be accessible. However, this is the best we can do.
2486 * don't allow environment variables we set in afslogon.dll to be
2487 inherited by child processes
2489 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2492 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2493 thread walking the all buffer list periodically searching for dirty
2494 buffers to write to the file server. This had several negative
2495 results. The algorithm ate up ever increasing amounts of CPU time
2496 even when AFS is idle as the size of the cache increases. Also,
2497 buffers were written to the file server in an order based upon the
2498 original buffer allocation which has nothing to do with the order
2499 in which the buffers became dirty.
2501 The new algorithm maintains a dirty buffer list. Items are added
2502 when the buffer is originally marked dirty. A buffer is only
2503 removed from the list by the buf_IncrSyncer when the buffer is no
2504 longer dirty. If the list is empty the thread goes back to sleep
2505 immediately without additional processing requirements.
2509 * Enable renaming of directories from the Explorer Shell when
2510 Offline Folders is enabled
2512 * An error is reported in response to the SMB ReceiveCoreFlush
2513 request when an error has not occurred.
2517 * optimize dirty buffer flushes in response to suspend operations
2518 do not flush if only loopback adapter is installed
2519 do no retry dirty buffer writes
2521 * do not reset netbios when performing loopback checks. doing
2522 so will break the existing lana bind
2524 * remove a deadlock in smb3.c QPathInfo
2528 * Add NetIDMgr AFS Plug-in
2530 * "Show Tray Icon" checkbox removed from afs_config.exe
2532 * Disable DNS Registrations for the Loopback Adapter and make sure
2533 that Netbios is turned on.
2535 * A reference count leak on stat cache objects was fixed
2537 * For debugging, tread ids of writers are now tracked on locks
2539 * Corrected an incorrect lock that was being held but not
2540 freed in a rare error case
2542 * Changed trace log end of line to CR-LF so it can be read
2545 * modify the buffer synchronization thread to write all dirty
2546 buffers and sleep only when there is no work to be done
2548 * improve behavior of cache manager when large numbers of
2549 temporary files are being created by multiple clients within
2552 * do not permit dirty buffers that belong to deleted files
2553 to be written to the file server. its a waste of bandwidth
2555 * do not leak stat cache references when the file server
2558 * afscreds: do not permit an attempt to obtain tokens if the
2559 afs service could not be started. this was causing stack
2560 overflows on Vista after resuming from a suspend
2564 * better handling of dot directory
2566 * fs sq and fs mkm now behave as on UNIX
2568 * add license to installers
2570 * Return Invalid Parameter error for unknown InfoLevels in
2571 SMB Tran2 Search Directory.
2573 * Correctly use WSAStartup and WSACleanup within the RX library
2575 * Check that afsd_service.exe is running before performing
2578 * Force AFS Shell Extension to load resource DLL using
2579 afs_shl_ext.dll as a base name.
2581 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2582 OpenAFS\Client\Program
2584 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2586 * Do not panic if the maximum number of volume entries are in use
2587 and one of them can be recycled.
2591 * Fix the code that reads the backconnnectionhostnames value
2592 so that there is not an uninitialized variable
2596 * integrated logon could cause winlogin.exe to blue screen windows
2597 if it was called with a domain that does not exist on the machine.
2598 this is possible during a sysprep.
2600 * corrections to cm_buf_t handling to ensure that buffers are
2601 removed from queues before they are freed and that holds are
2602 obtained on the next buffer in a chain before releasing the
2605 * do not generate MessageBox dialogs on network errors. they
2606 cause more trouble then they are worth.
2608 * move GlobalAutoMapper into a separate thread so that it does
2609 not prevent service startup if the WNetAddConnection2 call
2612 * restrict hard dead and connection timeouts to the CIFS Session
2613 Timeout value. This is being done to try to prevent virtual
2614 circuits from being destroyed prematurely.
2616 * Fix the code that writes the backconnectionhostnames value
2617 to ensure that the data buffer is written with the correct
2620 (there were no windows specific changes between 1.5.4 and 1.5.5)
2624 * Corrections to the Integrated Logon functionality that permits
2625 Kerberos 5 TGTs to be communicated into the user session. The
2626 transfer is now more secure and succeeds when the user does not
2627 have read permission to the %SystemRoot%\TEMP directory.
2629 * NSIS installer updated to version 2.18
2631 * WiX installer updated to version 2.0.4221.0
2633 * Updates to the IFS functionality for 64-bit correctness and
2634 permit compilation. The IFS functionality works for pioctl()
2635 operations but not much else at this point.
2637 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2638 query and set functions.
2640 * Ignore the Windows CIFS client session timeout entirely. Timeout
2641 only after the AFS Hard Dead Timeout period expires.
2645 * move headers, libraries, and samples from OpenAFS\Client to
2648 * remove the auto-registration of AFS IDs for foreign cells
2649 from integrated logon. this permits afslogon.dll to avoid
2650 initializing the RX library. afslogon.dll is a network
2651 provider library and as such it is loaded into all applications
2652 which perform a CreateProcessAsUser call. Initializing the
2653 RX library spawns background threads to listen for inbound
2654 packets. When the network provider is unloaded from the
2655 process, the RX library does not cleanup the threads. If they
2656 continue to execute after the library is unloaded, the process
2659 * implement inline bulk status calls. this will significantly
2660 improve the performance of directory listings.
2662 * prevent a crash when evaluating mountpoints to volumes that do not
2665 * new command: fs uuid [-generate]
2666 displays the current uuid and allows a new one to be generated
2667 on the fly without restarting the afs client service.
2669 * fix the 64-bit installers to properly document the install
2672 * move help files to the proper location so they can be located by
2673 the system tray tool and the control panel.
2675 * "over quota" and "disk full" errors were not being reported to the
2678 * largefile support has been added. Files greater than 31-bits can no
2679 longer be written to file servers which do not support large files.
2680 When file servers support large files, files up to 64-bits can be
2683 * cm_SyncOp could return with the input buffer unlocked upon error
2684 which would result in a panic condition. This bug was introduced
2687 * vlserver failover when freelance mode did not work. This prevented
2688 the root.afs volume from being loaded if the first vlserver could
2693 * A bug in the CIFS Standard Info response packet caused directories
2694 to appear as standard files. This prevented Microsoft Vista from
2695 being able to walk the AFS directory tree.
2697 * Mount points must be reported as directories in response to
2698 Query Path Info operations.
2700 * A deadlock was detected when performing "fs flushall" if the
2701 file server reports VNOVNODE. The scp->createBufferLock is
2702 already held by the current thread and there is no mechanism
2703 to propagate the knowledge. Therefore, an alternate mechanism
2704 for clearing the cache must be developed. A new function
2705 cm_RecycleSCache(scp,flags) has been extracted from
2706 cm_GetNewSCache(). This function performs the task of recycling
2707 an cm_scache_t object. When called from cm_GetNewSCache() with
2708 no flags the expectation is that there are no associated buffers
2709 that are queued to be read or written. When called from
2710 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2711 flag, any queued buffers will be de-queued and marked as if
2712 the operations were performed so the data can be discarded.
2714 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2716 * In response to VNOVNODE, the parent is only discarded if the current
2717 cm_scache_t is not a directory.
2719 * In the Ioctl FlushFile and FlushVolume functions, there were no
2720 checks to protect against flushing the Freelance SCache entries.
2722 * In FlushFile, the wrong cm_scache_t object was being released.
2724 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2725 recycled. Choose a new entry if cm_RecycleSCache fails.
2727 * Speed up the performance of the cache manager by not
2728 holding the smb_fid_t mutex across calls to cm_SyncOp
2731 * Ensure that all smb_fid_t flag references are protected.
2733 * Remove a deadlock between smb_fid_t mutex and
2734 smb_rctLock introduced in 1.4.1.
2736 * Fix aklog AFSID lookup to use not send the realm
2737 name when the realm is the local realm for the
2738 cell. The PT Server doesn't strip the realm
2739 name when it is the local realm.
2741 * Treat "\\afs\*." as an alias for "\\afs\all"
2743 * Dynamically adjust the priority of server threads
2744 based upon the age of the CIFS request that is
2745 being processed. The priority is bumped one level
2746 for every 15 seconds of processing time.
2748 * The original openafs contribution mis-used the
2749 cm_GetCallback function. This function should only
2750 be called as a side effect of a call to
2751 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
2752 was being called results in a FetchStatus RPC being
2753 performed even when the client already has a valid
2754 callback and does so in a manner that destroys the
2755 synchronization of the threads that are calling
2756 cm_SyncOp. If nothing else this results in a
2757 significant performance penalty.
2759 * More changes to cleanup of smb_vc_t objects in
2760 order to prevent race conditions.
2762 * Reduce the amount of Power Management event logging
2765 * If the client detects that the IP address has
2766 changed, force the use of new RX connections. This
2767 enables the file server to recognize that the client
2770 * Remove a race condition and deadlock associated
2771 with cleanup of smb_vc_t objects.
2773 * Fix the ANSI filename support. One function that
2774 required it was missing the translations.
2776 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
2779 * Add a DllMain() function to afspthread.dll. Perform
2780 cleanup of thread local storage, rx queues, critical
2781 sections, event handles, etc. This was implemented
2782 to avoid crashes in afspthread.dll when applications
2783 which load/unload the dll attempt to terminate.
2785 * Lock request behavior has been altered. As of this
2786 release a locally allocated lock will be used whenever
2787 the volume is RO or the requesting user's permissions
2788 are 'rl'. This will allow end users to execute programs
2789 from RW volumes when they do not have the 'k' privilege.
2791 Note that a recent discovery indicates that the 'k' bit
2792 is only supposed to control the ability to obtain read
2793 locks. Write locks are to be controlled by the 'w' and
2796 * Code has been added to enable the cache manager to be
2797 able to obtain the AFSID of the authenticated user.
2798 This is required to be able to determine when "Insert"
2799 privileges should apply. AFS considers a file to be
2800 a new file whenever the owner's AFSID matches the
2801 AFSID of the user. The current implementation relies
2802 on the name specified during the SetToken call to be
2803 correct. This will be replaced in the future with a
2804 new RPC that allows the AFSID of the authenticated
2805 user to be obtained.
2807 * A 32-bit-tools installer for use on 64-bit systems
2808 is now available. These tools are for use with 32-bit
2809 versions of Kerberos for Windows or JPSoftware's 4NT.
2811 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
2813 * Remove the use of the "active_vcp" variable. This variable
2814 no longer serves any useful purpose and was the cause of
2815 race conditions that produced under counts of the smb_vc_t
2818 * Change default minidump output to include data segments
2820 * restore access to cm_IoctlGetVolumeStatus and
2821 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
2822 Still need to figure out why they are being called
2823 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
2825 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
2827 * Update build system to include the "sys_name" in the
2830 * optimize the number of calls to obtain the smb_user_t for a
2831 smb virtual circuit.
2833 * add logging of server addresses, status and capabilities when
2836 * prevent multiple threads from pinging the same server simultaneously.
2837 instead, threads block and then use the result of the ping that was
2838 already taking place.
2840 * Add registry values to allow the defaults for daemon checks to be altered
2841 at startup. Log values to afsd_init.log.
2843 daemonCheckDownInterval
2844 daemonCheckUpInterval
2845 daemonCheckVolInterval
2846 daemonCheckCBInterval
2847 daemonCheckLockInterval
2848 daemonCheckTokenInterval
2850 * Add EventLog messages for transitions in fs crypt state.
2852 * Byte range locking (when locks will be obtained from the
2853 is still under discussion)
2855 * 64-bit Windows operating systems now supported
2856 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
2857 This is a native 64-bit implementation of OpenAFS for Windows.
2858 There is currently no 64-bit version of KFW and therefore
2859 there is no (easy) method for obtaining AFS credentials.
2860 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
2861 to obtain AFS tokens but there is not for the feint of
2864 * A faster implementation of the fcrypt algorithms improves
2867 * The afsd_service.exe now supplies Event Log Message Text.
2868 The event log no longer generates warning text due to lack
2869 of message configuration.
2871 * When afs servers are unreachable and network stack returns
2872 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
2873 mark the attempted server as being down. This allows for
2874 very fast response to the calling application. This change
2875 has a pronounced improvement when the Windows Explorer is
2876 used on disconnected mobile computers.
2878 * Added DebugOutputString debugging of the RX library
2880 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
2882 * Fix the rx-lwp implementation to randomly generate rx call
2883 identifiers which are supposed to be unique at a given point
2884 in time. This affects some of the command line tools used
2885 for debugging but not the AFS Client Service itself which
2886 uses the rx-pthreads implementation.
2888 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
2889 Windows although they are not distributed
2891 * don't generate an exception when logging unexpected SMB packets
2893 * the data structures used by cm_IoctlGetVolumeStatus and
2894 cm_IoctlSetVolumeStatus are not compatible with the CIFS
2895 Get/Set File Information calls so don't bother executing them.
2899 * Fix data written to registry as part of BackConnectionHostnames
2902 * Fix initialization bug when allocating cm_server_t objects that
2903 would cause servers to appear to be down after each allocation
2904 or renewal. Server would appear up after the next check servers
2909 * Windows will return WSAECONNRESET to the next receive packet
2910 operation on a socket if an ICMP packet is received in response
2911 to a packet that was sent and could not be delivered because
2912 the port was unreachable. We have to treat this as
2913 WSAEWOULDBLOCK so that we can retry the select() check and
2914 not mark the connection as bad.
2916 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
2918 * The client manager maintains a least recently used list of
2919 all the ACLs. If an ACL is already the most recently used
2920 do not bother to remove it from the list and place it on the
2923 * As part of the process of adding a new server, set the
2924 server is down flag and then perform a ping. If the server
2925 is up, the flag will be reset. If the server is down, the
2926 client will not timeout unnecessarily while attempting to
2929 * If a Mountpoint String is the empty string, return PATH NOT
2932 * Modify behavior of the Network Provider to avoid querying the
2933 profile location if integrated logon is known to be disabled.
2935 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
2936 will always be preserved regardless of whether or not integrated
2937 logon was used to obtain tokens. When unset, tokens will be
2938 destroyed regardless of whether integrated logon was used to
2939 obtain the tokens unless the profile is stored in AFS.
2941 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
2942 message is being delivered even when the user is not logging
2943 off of the computer.
2945 * Fix a deadlock condition introduced in rc6.
2947 * Cell names are case-insensitive. Since the founding of OpenAFS
2948 there has been a bug that treated "cell", "Cell", and "CELL" as
2949 different cells. This would result in what appeared to be
2950 random disappearance of user tokens or a failure to authenticate
2951 even when tokens were listed by the "tokens" command.
2953 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
2955 * Additional auditing of the smb code revealed that smb ioctl
2956 operations were failing to free smb_fid_t objects
2958 * The File Identifiers assigned to smb_fid_t objects could be
2959 given the value 65535 which is 0xFFFF which is also the
2960 INVALID_HANDLE value. Both 0 and 65535 are now treated as
2963 * Audited all assignments to fields and ensured that they are
2964 protected by obtaining the relevant mutex object.
2966 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
2967 if both the current cm_scache_t object and its parent are
2968 both returning VNOVNODE.
2970 * Removed all the logic associated with the 'dead_vcp' variable
2971 which must have been added for debugging because its used
2972 was never thread safe and it truly served no purpose.
2974 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
2976 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
2977 registry settings. Default is ON and 120 seconds. If set to
2978 OFF, tokens will not be destroyed.
2980 * Fixed a number of race conditions and reference count problems
2981 associated with global queues.
2983 * Added a method to detect abandoned smb virtual circuits so their
2984 associated file handles, locks, user tokens, and smb session
2985 resources can be released.
2987 * Added a mechanism to allow user tokens to be preserved
2988 during integrated login and be freed after logoff.
2990 * Check access permissions to the file before the file is
2993 * More improvements to the handling of SMB Virtual Circuits
2994 when SMB sessions are closed.
2996 * Correct a race condition introduced in rc4 that could result
2997 in a crash of afsd_service.exe.
2999 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3001 * Properly handle pre-mature SMB session closures.
3002 File handles, tree connections, and user sessions are now
3003 freed when the virtual circuit is closed.
3005 * smb_username_t and smb_user_t objects are now properly reference
3006 counted and freed as appropriate.
3008 * disable to use of AFS file locks. byte range locking is still
3009 implemented but all locks are managed locally. This prevents
3010 multiple processes from editing the same file on the same
3011 machine but it does prevent multiple processes on separate
3012 machines from modifying the file.
3014 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3016 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3017 the same server parameter would result in a panic condition
3018 if the server had not already been contacted by the AFS
3021 * Better warnings are written to the afsd.log file when
3022 the Windows RPC Protocols are not properly configured as
3023 is often the case on Windows 2000.
3025 * fix the procmgmt library to restore C RTL signal handlers when
3026 it is being detached from the process prior to process termination.
3028 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3030 * close a lock leak in the rx library that could be triggered while
3033 * prevent removal of drive mappings in response to afscreds.exe -M
3035 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3036 strings and flush the stat cache entry.
3038 * prevent server objects from being freed if user preferences are set.
3040 * Update WIX support to 2.0.2419.0
3042 * Release all locks on the server at service shutdown.
3044 * A down server should be marked up as long as it responds to the client.
3045 It doesn't matter if the server returned an error as long as the error
3046 is something other than VRESTARTING. If there is such an error, all
3047 connections on that server should be marked for replacement.
3049 * Change the check "up" servers period from once per hour to once every
3050 ten minutes. This matches the behavior of the Unix cache manager.
3052 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3053 * Byte Range Locking support has been added. Architectural documentation
3054 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3055 made to enforce Windows mandatory locking model using locally managed
3056 byte range locks with full file locks obtained from the file server.
3057 Read-only volumes provide the user with an implied locking privilege.
3058 Read-write volumes require that a user be granted the locking privilege
3059 in order for a lock to be obtained. As most Windows applications use
3060 Shared file access modes, they require the ability to obtain locks in
3061 order to access files.
3063 * Implement error translation functions to convert Universal Error Codes
3064 output by modern servers to local errno and winsock error values.
3065 This fixes the error message problems using pioctl() calls from fs.exe
3066 and the AFS Shell Extension.
3068 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3069 * Allow the AFS Client Service to set the firewall configuration by
3070 retrying until after Windows determines the boot period is over.
3072 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3075 * Detect IP address changes from within the AFS Client Service for the
3076 purpose of checking down servers sooner
3078 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3079 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3080 user that there is a Kerberos key problem. Previously the user would
3081 receive no feedback when there was something wrong.
3083 * Fix the large integer library support to properly handle values greater
3086 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3087 * Updated CellServDB
3089 * Fixed file name pattern matching. Question marks can match nothing
3090 if found at the end of a component.
3092 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3093 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3095 * The MSI installer now has a new Product ID for each build. This
3096 is to allow for automated upgrades. The Upgrade ID is remaining
3099 * Volumes that were cached and then subsequently were salvaged,
3100 moved or otherwise taken temporarily offline (as opposed to being
3101 busy) would become unreachable. A distinction is now made between
3102 a volume being offline and a server being down.
3104 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3106 * Fixed several AFS Shell Extension issues:
3107 - Checking Server Status would only work if a specific cell was
3109 - If servers were down, the Server Down dialog would write to
3110 invalid memory and cause explorer.exe to crash
3111 - Default Push Buttons were not specified for many dialogs
3112 - The extension would be unloaded by the Explorer Shell and
3113 never be reloaded under some circumstances.
3115 * Fixed the AFS System Tray menu to disappear if another Window
3116 was selected before a menu item was selected
3118 * Added volume owner and group information to "fs examine" output
3120 * Added fs and registry support for enabling/disabling/clearing
3121 rx statistics gathering.
3123 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3125 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3126 GUIDs for each platform.
3128 * Documentation directory structure and default installation rules
3131 * New OpenAFS for Windows HTML Release Notes installed
3132 as part of the documentation set.
3134 * Add thread id to rx debugging messages
3136 * When all servers are offline, return a bad network path error
3137 immediately. This is necessary to allow Explorer to do the
3138 right thing and not hang. The background daemon thread will
3139 check the down servers every 30 seconds to see if they are
3142 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3144 * The AFSCache file has been modified to store both the
3145 serial number of the volume on which the file resides
3146 and the SID of the local machine. If either change,
3147 a new UUID is generated.
3149 * add "fs flushall" command that forces all data buffers
3152 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3154 * add support for '/' instead of '\' in pioctl() calls
3156 * Apply AFS Client Admins group protect to AFS Shell Extension
3158 * Add support for \\afs\<name> to most AFS Shell Extension
3159 commands except the symlink methods.
3161 * when installing the Microsoft Loopback Adapter, enable
3162 MS Client for Networks and prevent an install failure
3163 by not calling CoInitialize twice in the same thread.
3165 * reload cell vldb values from the CellServDB every two
3166 hours in case it changes
3168 * When updating cell information from DNS, be sure to set
3171 * Add support to allow use of \\AFS\<foo> where <foo> is
3172 either a mount point or symlink. As <foo> is normally
3173 treated as a share name, we transform it into \\AFS\all\<foo>
3176 * Init 'code' to prevent false errors when integrated logon
3177 is disabled and the service is not running
3179 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3181 * Removed trace log messages that were hampering performance.
3183 * Fixed a deadlock that was being triggered by editing Word
3184 documents stored within AFS with WinWord 2003.
3186 * Bit 3 of the TraceOption registry value is now used to set the
3187 default for "fs trace" in the non-Debug builds. The new default
3188 is off for release builds and on for debug builds.
3190 * Bit 2 of the TraceOption registry value can be used to turn on
3191 real-time output of debug log entries to the Windows Debug
3192 Output monitor interface. This data can be viewed with tools
3193 such as Sysinternal's DbgView.
3195 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3196 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3197 a resource we should not be returning status codes such
3198 as Network Busy or Remote Host Down because these errors
3199 will imply that the AFS Client Service is Busy or Down
3200 and that is simply not the case. Instead we will return
3201 Bad Network Path as the path is temporarily not available.
3203 Instead of returning Sharing Paused when there is a
3204 timeout error, return Timeout. Once again, the AFS Client
3205 Service is not paused.
3207 * afscreds.exe would display an Obtain Creds dialog when
3208 the expired credentials reminder was triggered even
3209 if there was no network path to the KDC. This is prevented
3210 by adding KDC probe logic to the reminder thread.
3212 * afscreds.exe would display expired tokens no differently
3213 than unexpired ones. This would make it difficult for the
3214 user to distiguish when the tokens were expired. For the
3215 English build added a new resource string "(expired) that
3216 is displayed instead of the expiration time.
3219 * "fs wscell" when executed in freelance mode will return the
3220 name of the cell configured in the registry. The root.afs
3221 volume is not loaded from this cell, but it is used for the
3222 default for aklog and integrated login.
3224 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3225 no longer work on \\AFS\all when freelance is being used
3226 unless the user is a member of the "AFS Client Admins" Windows
3229 * some more checks performed during persistent cache validation.
3232 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3233 down servers for the current cell not all cells.
3235 * If the scache LRU list becomes corrupted, the AFS Client Service
3236 might enter an infinite loop at startup. Detect the corruption
3237 and discard the cache contents in this case.
3239 * Fixed the Explorer Shell's remove mountpoint functionality.
3241 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3242 obtain a mutex lock on a cm_scache_t object whose lock is held by
3243 a thread attempting to obtain the cm_aclLock.
3246 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3247 misinterpretted as "csail.mit.edu" when the resolver
3248 is configured to append ".mit.edu" to failed DNS queries.
3250 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3251 that can be used to force the preservation of user tokens upon logout.
3253 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3254 recommended installation categories: AFS Client, AFS Administrator,
3255 AFS Server, AFS Development Kit. Each category includes a different
3256 default subset of the OpenAFS components.
3258 The OpenAFS logo is now associated with the NSIS uninstall entry
3259 in the Add/Remove Programs control panel.
3261 * The user name associated with AFS tokens when obtained with
3262 integrated login, the afs systray tool, or aklog will always
3263 include the full Kerberos 5 user principal name regardless of
3264 whether or not the cell is local to the realm.
3266 * Modify integrated login so that it does not enter an infinite
3267 loop if the service is not set to auto start.
3269 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3272 * Added uninstall.exe to Wix installation
3274 * More modifications to algorithms used to wake sleeping threads.
3278 * Changes to the algorithms used to wake threads when they are
3279 put to sleep because of conflicts over buffer operations.
3282 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3283 the process of producing AFS tokens. The tokens are stored within
3284 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3285 New functionality has been added to temporarily copy the tickets to
3286 a file ccache during the short window between the NPLogonNotify and
3287 WinLogon Logon Event. The Logon Event handler starts a process as
3288 user, afscpcc.exe, that copies the tickets into the default ccache.
3289 The file ccache is then destroyed by the event handler.
3291 * During synchronization operations on stat cache entries, if threads
3292 are forced to sleep it was possible that the threads would not be
3293 woken under heavy load.
3295 * Enforce the delete on close flag specified by the NTCreateX()
3298 * A race condition in the rx library was fixed that resulted in
3299 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3300 was cleared when threads were still waiting to use the
3301 connection. Also, fixed a potential case in which the
3302 connection could be destroyed while threads were waiting to use it.
3304 * The DNLC freelist has been seen to become corrupted with still
3305 active entries being stored on the freelist. Changes were made
3306 to perform a better job of cleaning entries before placing them
3307 onto the freelist; marking them with the correct magic value when
3308 purging the list; and allowing the list to be purged on startup
3309 if the cache validation fails.
3311 * Windows Crash Reporting does not get triggered for afsd_service.exe
3312 because it provides its own top level exception handler. This patch
3313 places an explicit call to ReportCrash() on platforms which support
3316 If you configure Crash Reporting via AD Policy to capture crash
3317 reports within the domain, then you will receive this output.
3319 * Generate MiniDumps independent of the Windows Crash Reporting.
3321 (a) If an exception occurs,a minidump will be generated at the path
3322 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3323 can be set to any other type via use of the
3326 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3328 (b) MiniDumps can be generated at any time using the "fs minidump"
3329 command. This allows you to generate a minidump without
3330 stopping the process or even requiring the use of a debugger
3333 * an RX connection which reports bad ticket now treats it the same way
3334 the expired ticket is treated. The ticket is removed from the
3335 connection and a new anonymous connection is established.
3337 * fs memdump now requires that the caller be in the AFS Client Admins
3340 * additional debugging information is provided in cm_Analyze when the
3341 error code is ignored.
3343 * Fix a regression introduced into the pioctl() client code which
3344 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3345 MIT KFW is not installed. (1.3.82a)
3347 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3350 * Fix a race condition in cm_EndCallbackGrantingCall() that
3351 could leave 'nrevp' pointing to freed memory. In the process,
3352 optimize the processing to ensure that the applications monitoring
3353 the status of the stat cache entry are only notified at most
3356 * Conditionalize access to the Explorer Shell AFS->Mount Point
3357 ->Remove menu item. Disable the item if the selected item(s)
3358 are not mount points.
3360 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3361 Local Storage. This eliminates thread contention on the
3362 global RX free packet queue. This should improve performance
3363 on multi-CPU systems.
3365 * Fix the IP address filtering code to properly load/unload
3366 the IPHelper DLL so that Windows XP/2003 and beyond systems
3367 do not have to manually probe the registry.
3369 * CleanupACLEnt() was not being called consistently with
3370 the cm_scache_t object referred to by the back pointer
3371 mutex locked. This could in very rare conditions lead to
3372 an invalid memory access.
3374 * Added a script command to the msi installer "afs_replace"
3375 which can be used to replace one version of OpenAFS with
3376 another without requiring an uninstall and reinstall if
3377 the installed msi is no longer in its original location.
3379 * Apparently the problem with multi-domain forests with cross-
3380 realm trusts to non-Windows realms was not entirely solved.
3381 The authentication to the AFS SMB service failed because
3382 the wrong name was being used. Using ASU as an example,
3383 the authentication was being performed with the name
3384 "QAAD\user" (an account in the forest root) and not
3385 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3387 The solution was to add an additional dependency on KFW
3388 in order or to be able to easily obtain the client principal
3389 name stored in the MSLSA ccache TGT. This information is
3390 used in two locations:
3392 - the pioctl() function
3394 - a new WinLogon Event Handler for the "logon" event.
3396 The pioctl function will now be able to use the correct
3397 name when calling WNetAddConnection2() and the "logon"
3398 event handler will now be able to call WNetAddConnection2().
3399 The hope is that the "logon" event handler will be called
3400 before the profile is loaded but I have not guarrantee
3404 * Fixed a locking error in cm_TryBulkStat() which had the
3405 potential of deadlocking the system for the length of time
3406 it takes to perform a bulk status fetch operation.
3408 * Replaced time conversion code (UnixTime <-> FILETIME) to
3409 be completely arithmetic instead of relying on a bizarre
3410 algorithm involving a variety of C RTL time functions.
3411 This has the side effect that UnixTime and FILETIME which
3412 are both stored in UTC are interpretted as UTC throughout
3413 the year. Windows will apply the same localization to AFS
3414 as it does to NTFS. Applications which rely on the ability
3415 to sync files between the two file systems will no longer
3416 see the timestamps of files in AFS change an hour relative
3417 to the files stored in NTFS or Windows based backup devices.
3419 * Fixed a invalid memory access under a bizarre circumstance.
3420 Windows will allow a physical mass media device to be
3421 installed via Plug N Play to the system and will assign it
3422 the lowest available drive letter. This is true even if the
3423 drive letter is currently assigned in the user session to
3424 a network device via NET USE (or its equivalent.) When this
3425 happens, queries sent to the CIFS server will contain invalid
3426 data. This invalid data was not being caught by the AFS
3427 Client Service and was resulting in a crash when Freelance
3430 * Fixed a reference count error when registering callbacks
3431 on a stat cache object if the callback was already registered
3433 * Add a case to cm_Analyze() in order to handle the case where
3434 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3435 the server list for the volume is empty. In this case, force
3436 an update of the volume info and retry.
3438 * Insert a missing cm_EndCallbackGrantingCall() which could
3439 result in threads waiting for a callback to complete to never
3442 * In the persistent cache, there is a maximum number of volume
3443 entries. Allow volume entries to be reused if the maximum
3444 number have been allocated and their reference count is zero.
3446 * If we already have a dead virtual connection object, don't
3447 ignore the fact that additional dead vc objects must be taken
3450 * Removed a deadlock condition introduced in the summer of 2004.
3451 Do not hold mx locks around calls to RX functions including
3452 rx_NewCall(). That is what reference counts on the rx_connection
3455 * Fixed an initialization error in afslogon.dll which could
3456 result in random behavior including a failure to terminate
3457 the NPLogon function. (1.3.80b)
3459 * Fixed an error preventing the use of SMB authentication
3460 on Windows 2000. (1.3.80a)
3463 * Updated CellServDB to Public CellServDB 16 Mar 2005
3465 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3466 misinterpretted as "csail.mit.edu" when the resolver
3467 is configured to append ".mit.edu" to failed DNS queries.
3469 * Fixed another case in which the client would replace connections
3470 to the host when it was not necessary. In this case the
3471 situation would occur if cryptall was on and the connections
3472 were unauthenticated due to lack of an appropriate token.
3474 * OpenAFS for Windows has failed to work at sites which are
3475 utilizing a cross-realm trust between an MIT/Heimdal realm
3476 and a multi-domain Windows forest when the workstation being
3477 accessed is not located in the root domain. This is caused
3478 by a bug in the workstation which was triggered after the
3479 introduction of Windows 2003 Server. When the bug is triggered,
3480 the workstation attempts to authenticate users to afsd_service.exe
3481 by contacting the Domain Controller instead of using the
3482 LSA loopback authentication mechanism.
3484 One of the reasons this bug occurs is because the workstation
3485 does not have a reliable way of knowing that the service whose
3486 netbios name is "AFS" is located on the workstation. This will
3487 be fixed starting in Longhorn Beta 1 by Microsoft. The
3488 "BackConnectionHostNames" registry value will be used to
3489 indicate that the authentications to that service name should
3490 be performed using the loopback authentication mechanism.
3492 In the meantime, when Logon Caching is enabled, we can force
3493 afsd_service.exe to authenticate using the logon cache before
3494 contacting the Domain Controller. This will work with both
3495 password and smart card based logons.
3497 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3498 flag should not be cleared if a volume's server reference is
3499 marked as "offline". In the case where all of the volume's
3500 servers are either "down" or the volumes are "offline", we want
3501 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3502 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3503 should never occur in practice.
3505 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3506 5 seconds, clear the server down and volume busy flags, and
3507 then force an update from the VLDB. This allows the client to
3508 update the location of a volume if the reason for it being
3509 marked offline is because it is being moved. Calls to
3510 cm_ConnByMServers() will be retried either until success or
3511 the RDRtimeout period is reached.
3513 * Correct the Power Management code behavior in response to
3514 Standby, Suspend, and Shutdown events. Instead of flushing the
3515 buffers associated with the mounted SMB submounts, simply write
3516 all buffers which are marked dirty.
3518 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3519 provides a list of cells other than the default cell for which
3520 tokens should be obtained using the default Kerberos principal.
3521 This functionality is logon domain specific and is only available
3522 when using KFW for authentication.
3524 * Fixed FindNext Invalid Handle error caused by over agressive
3525 attempts at garbage collecting dirSearch entries when the
3526 dirSearch ID wraps from the maximum value to 1.
3528 * Add support for registry defined server preferences for VLDB and
3529 File servers. See registry.txt for details.
3531 * Increased default cache size to 96MB and default number of cache
3534 * Fixed refCount leaks related to directory lookups and pioctl calls.
3536 * Callbacks revoked during a race condition with an attempt to
3537 obtain the same callback no longer result in an inaccessible
3540 * New command line tool:
3542 afsdacl : Set or reset the DACL to allow starting or stopping
3543 the afsd service by any ordinary user.
3545 Usage : afsdacl [-set | -reset] [-show]
3546 -set : Sets the DACL
3547 -reset : Reset the DACL
3548 -show : Show current DACL (SDSF)
3550 * IP addresses are no longer queried once at startup. Instead IP
3551 addresses are obtained as needed. Loopback adapter addresses are
3552 no longer published to the server.
3554 * Pay attention to the MIT KFW registry configuration for automatic
3555 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3557 * Fix (once again) case-sensitive comparisons which was apparently
3560 * Activate support for vos listvol -format
3562 * Population of the "AFS Client Admins" group failed on non-English
3563 versions of Windows because the name of the "Administrators" group
3564 is localized. Now we lookup the name of the group by using the
3565 Administrators Group Alias Relative ID.
3567 * Multiple cell token acquisition within afscreds.exe was broken in the
3568 case of Kerberos 5 cross realm authentication. Instead of contacting
3569 the KDC associated with the cell's realm, afscreds would obtain a
3570 token from the KDC of the user's realm. This would result in
3571 "invalid kvno errors" while authenticating to the AFS servers.
3572 Unauthenticated access would work. The symptoms would vary based upon
3573 whether or not the VLDB servers had been contacted using unauthenticated
3574 connections prior to the user obtaining tokens.
3576 * The list of ACL entries was becoming corrupted. It appears as if
3577 an ACL when it expires was not being appropriately cleaned up. In
3578 fact, it was left in the list of ACLs associated with the scache
3581 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3582 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3583 and "amd64_win64" for amd 64-bit processors.
3585 * When executing executables, Windows provides a mechanisms by which
3586 the normal search for DLLs can be bypassed. If a file foo.exe is
3587 being executed and there exists a file or directory "foo.exe.local"
3588 (the contents are ignored if a file), then Windows will search for
3589 DLLs first in the "foo.exe.local" directory and second in the
3590 directory in which the "foo.exe" file is located.
3592 Previous releases of OAFW would improperly return
3593 STATUS_NOT_A_DIRECTORY instead of the expected
3594 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3595 to terminate the search for the DLL causing the execution of the
3596 application to fail.
3598 * Rework the reference counting of smb_vc_t objects. The references
3599 stored in the waiting locks were not counted. This could result in
3600 an assertion if the reference count drops to zero.
3602 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3603 when running in freelance mode.
3605 * Added support for persistent caching of file contents, stat entries,
3606 volume data, cell data, and name lookup data. When the registry
3607 value, "NonPersistentCaching", is not defined or set to 0 the file
3608 specified by "CachePath" becomes a persistent cache file. The
3609 size of the cache file is computed at run time. It has a maximum
3610 size of approximately 1.2GB. The contents of the file will be
3611 validated according to the rules specified by the "ValidateCache"
3614 * Added support for UUIDs. UUIDs are kept for as long as the
3615 cache file is intact.
3617 * Added cmdebug.exe and support cache manager debugging callback
3618 interfaces in afsd_service.exe.
3620 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3621 [-addrs] [-cache] [-help]
3622 Where: -long print all info
3623 -addrs print only host interfaces
3624 -cache print only cache configuration
3626 * Symlinks to \\AFS\[all\]... will now be treated the same as
3627 symlinks to /afs/.... However, please use /afs/... as otherwise
3628 the symlinks won't work on Unix.
3630 * Correct a problem with local CellID allocation for cells whose
3631 server list is obtained via DNS instead of a CellServDB file. If
3632 the DNS information expires the CellID assigned to the entry will
3633 be changed. This causes all of the cm_scache_t objects which refer
3634 to the old cellID number to become useless. Attempts to access
3635 files or directories with cm_scache_t objects using the old cellID
3636 will fail since the server list cannot be obtained.
3638 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3639 be triggered if the volume referenced by the cm_scache_t object is
3642 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3643 in addition to SERVICE_ACCEPT_STOP.
3645 * Move RPC shutdown until after the SMB and RX shutdown procedures
3646 complete. Block until RPC shutdown is complete.
3648 * Modify afslogon.dll (integrated logon) to wait for service if its
3649 state is START_PENDING. If the timeout period occurs, reset to
3650 the retry interval and not the sleep interval.
3652 * When renewing the server list for a cell obtained via DNS AFSDB
3653 records, the cm_cell_t entry must be removed from the list of all
3654 cells. Otherwise, the list of cells will be corrupted.
3656 * In the dcache and scache modules, use the cm_scache_t dataVersion
3657 instead of the cm_fakeDirVersion.
3659 * Synchronize fs.c with the unix version.
3661 * The variable used to determine whether a file or virtual memory
3662 mapped cache is used was not properly initialized to a default
3663 value. If the registry setting "NonPersistentCaching" was not
3664 set, the choice would be random. Properly initialized to be
3667 * The memory mapped view was never unmapped before closing the file
3668 at service shutdown. This is now properly cleaned up.
3670 * Default location of Cache file is now %TEMP%\AFSCache
3674 * A bug affecting new installations of 1.3.75/76 would result in
3675 the creating of incorrect mountpoints in the freelance root.afs
3676 volume for the default cell.
3678 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3679 of "root.cell", you have been affected by the bug. To correct the
3680 problem, execute the following commands:
3682 fs rmmount \\afs\all\<cellname>
3683 fs rmmount \\afs\all\.<cellname>
3685 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3686 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3690 * A bug has been fixed in the auto-generation of Freelance root.afs
3691 symlinks which produced random entries in the \\AFS\all directory.
3693 * Support has been added for multi-homed servers
3696 * Added a new registry value, "StoreAnsiFilenames", which can be used
3697 to force the use of ANSI character sets instead of OEM Code Pages.
3698 This feature is useful when users require the ability to create
3699 filenames with 8-bit characters and need to access the files from
3700 both Latin-1 based Unix systems as well as from Windows.
3702 Activation of this feature will prevent access to files stored with
3703 8-bit OEM characters.
3705 * Shutdown all SMB threads in a synchronized manner when stopping the
3708 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3709 by the largest contiguous block of unused memory within the 2GB process
3710 space which can be assigned to the memory mapped file. Unfortunately,
3711 when the executable digital signature verification code is activated
3712 Windows sees fit to further segment the process memory which in turn
3713 reduces the size of the maximum cache file to less then 800MB. If
3714 larger cache sizes are desired, a new registry value should be set:
3716 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3718 Setting this value will disable the runtime verification of digital
3719 signatures on afsd_service.exe and the afs dlls which it loads. It
3720 will not disable the the version number check on those same files.
3721 The signature verification is not a security messure and is only meant
3722 to enhance the ability to afsd_service.exe to detect potential
3723 destablizing mixtures of DLLs from incompatible distributions.
3725 Added code to auto-disable the signature verification check if
3726 the desired cache size is greater then 700MB.
3728 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3729 used to verify the validity of the afsd_service.exe binary
3730 as well as each of the AFS DLLs loaded by the service. Not only
3731 must the digital signature be valid but the signatures of the
3732 DLL must be signed by the same entity as the service.
3734 * Implement new functions: cm_freelanceMountPointExists and
3735 cm_freelanceSymlinkExists. Use them along with other validity
3736 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3737 ensure that name collisions do not occur and that empty strings
3738 are not valid file names.
3740 A symlink may not have a name which would resolve to a valid
3741 cell name. Doing so would prevent access to the cell.
3743 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3744 re-initialization code. The reference counts of the fake root.afs
3745 volume scache object(s) would become invalid when the mount point
3746 or symlink lists were altered.
3748 * Add registry entries to provide mappings from the afsdsbmt.ini
3749 to the new locations for applications which count on the use
3750 of the old Profile file APIs. These apps are likely to fail
3751 if the user does not have administrator privileges and the
3752 registry is locked down.
3754 * The afs_config.exe submounts dialog had two errors.
3755 First, attempts to remove entries failed because the registry
3756 key was being opened without KEY_WRITE privileges.
3757 Second, when editing a submount entry, changing the name
3758 would add a new key and leave the original one in place.
3759 Now the original submount will be removed if its name is
3762 * In recent months there have been several incidents in which
3763 users have experienced problems starting or accessing
3764 afsd_service.exe and after significant effort has been spent
3765 it has turned out that they have two versions of AFS on the
3766 machine or an inconsistent set of DLLs.
3768 Code has now been added to afsd_service.exe which will walk
3769 the list of modules loaded by afsd_service.exe and validate
3770 that the version of the AFS DLLs matches the version of the
3771 afsd_service.exe executable. If they do not match the service
3774 * When Freelance mode is enabled and there is no registry
3775 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
3776 will attempt to import the afs_freelance.ini file contents.
3777 If the file does not exist, it was creating a dummy file
3778 with a r/o and r/w entry for the default cell and then
3779 importing those values.
3781 This process has been changed. The temporary file is no
3782 longer created. Also, both the OpenAFS Client install
3783 directory as well as %WINDIR% are checked for previous
3784 afs_freelance.ini files.
3786 * Added support for VL_GetEntryByNameN(). Still need to add
3787 support for VL_GetEntryByNameU() for multi-homed support.
3789 * Fix a deadlock situation in afscreds.exe when canceling an
3790 auto-generated Obtain Tokens dialog
3793 * Fix the locking of objects during Directory Searches in the
3794 SMB/CIFS server. The failure to properly lock the reference
3795 counts was resulting in the premature freeing of smb_dirSearch_t
3796 objects while they were still in use by the SMB/CIFS client.
3797 This does not solve the "Invalid Handle" problem.
3799 * Fix Find Cell By Name pioctl call to return a valid cell
3800 name for the Freelance fake root.afs volume.
3801 "Freelance.Local.Root".
3803 * Fix the Explorer Shell Extension Symlinks->Add operation.
3804 The dialog template was missing and the link destination
3805 string was too short.
3807 * Add support for symlinks to Freelance root.afs volume
3808 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
3809 <number> = "<linkname>:<relative-path>."
3810 Use symlink.exe to create, list, or remove
3812 * Remove the fallback to the use of KFW's KRB4 library when obtaining
3813 tokens. We never obtain KRB4 tickets.
3815 * Fix AFS Client Configuration Control Panel to support new SysName
3818 * Fix a bug in afsd_service.exe which could result in the SysName
3819 not being read from the registry.
3822 * Add code to block the issuance of AFS tokens by aklog.exe or
3823 afscreds.exe when the Kerberos 5 principal name contains a dot.
3825 * Modify the IsAdmin() function to always treat the local SYSTEM
3826 account as an AFS client administrator. Affects fs.exe and
3829 * Modify the internal handling of Quota Exceeded errors
3831 * Upgrade all reference count fields in the Windows cache manager
3832 and the osi library to use unsigned long instead of signed short.
3833 A similar fix has been applied to the afs rpc (rx) library.
3835 * fix the Windows cache manager to prevent it from replacing the
3836 rx_connection object associated with the cm_conn_t object on each
3837 and every operation if "fs crypt" was set. This explains the
3838 dramatic performance difference when crypt is used vs clear.
3839 The problem: 'cryptall', a boolean flag indicating whether or not
3840 "fs crypt" is set, was being compared to the rx_connection
3841 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
3842 1 != 2 and therefore the rx_connection was always destroyed
3843 and replaced on each and every operation.
3845 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
3846 It is not safe for the cm_conn_t object to not be locked because
3847 rx_DestroyConnection might be called from another thread if:
3848 - the user's tokens have changed or expired
3849 - the crypt mode has changed
3851 This fix appears to have also taken care of the problems associated
3852 with Overlapped Writes resulting in Delayed Write errors.
3854 * fix NSIS installer's AdminGroup.exe to properly create and
3855 remove groups when given -create or -remove. The string comparison
3858 * fs sysname now accepts a list of sysname values
3860 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
3861 DWORD which when set to a non-zero value will cause error message
3862 text to be output to stderr from the pioctl() routine. Useful in
3863 debugging failures of fs.exe, tokens.exe, etc.
3865 * added a test to the power management code to only perform a
3866 flush operation if there is at least one network adapter which
3867 is not a loopback adapter.
3869 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
3870 "EnableKFW". This value will not be read if the key
3871 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
3872 value under that key does not.
3874 * provide mechanisms to force the use of krb524d for Kerberos 5
3875 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
3876 there is a new registry value "Use524" and for aklog.exe a new
3877 command line parameter "-m".
3879 * Fix the pattern matching algorithm to properly match patterns
3882 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
3883 which is used by both the original function and the new
3884 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
3885 creation of HardLinks in addition to Renaming. smb_Link()
3886 is a new function which creates HardLinks via cm_Link().
3887 cm_Link() is a new vnodeops function which creates links
3890 smb_ReceiveNTRename() does not support the File Copy and
3891 Move Cluster Information operations described in its interface.
3892 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
3894 * When opening files via symlinks, we should follow the symlinks
3895 until we reach the actual file stat cache entry. The stat cache
3896 entry of the file should then be stored in the FID instead of
3897 stat scache entry of the symlink.
3899 * return bad operation errors for all unimplemented functions
3900 even if we do not know the functions exist.
3902 * Log bad packets and unknown operation packets to the trace log
3904 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
3907 * Update list of known CIFS operations to include all those listed
3908 in CIFS-TR-1p00_FINAL.pdf.
3910 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
3911 to support the REG_EXPAND_SZ type.
3914 * A new Windows authorization group "AFS Client Admins" is now
3915 created and populated with the members of the "Administrators"
3916 group. The group is used to determine which accounts on the
3917 machine may be used to modify the AFS Client Configuration via
3918 the UI and command line tools. afs_config.exe, fs.exe,
3920 * Modify the WinLogon Logoff Event Handler to query NT4 domain
3921 controllers for the remote profile path if Active Directory
3922 services are not available.
3924 * Fix aklog.exe to not add the AFS ID to the username
3926 * PTS registration of new users to foreign cells has been added to
3929 * The cm_Daemon thread is used to perform checks for
3930 down servers, up servers, volumes, callback expirations,
3931 lock maintenance and token expiration. Due to a gaff in
3932 larger integer division the thread never performed any
3933 work. Instead the current time computation would always
3934 be less then the trigger times. This had an adverse affect
3935 on the client's ability to maintain communication with servers,
3936 keep volumes up to date, and flush user tokens and acls
3937 when they have expired. This was broken when the 1.3 branch
3938 was modified to support VC7 which no longer included
3941 * An initialization problem with the Freelance code was
3942 detected while fixing the callbackRequest. The cm_rootSCachep
3943 object is obtained during afsd_InitDaemons() but the callback
3944 information is incomplete. The callback information will not
3945 be obtained until cm_MergeStatus is called from within
3946 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
3947 test for the conditions under which the callback information
3950 * Reports have been filed indicating that callbacks were
3951 being lost. An examination of the code indicated that the
3952 cm_server_t objects were not being properly reference
3953 counted by the cm_scache_t and cm_callbackRequest_t objects.
3954 In particular, the cm_server_t objects may have been freed
3955 from beneath the cm_conn_t objects.
3957 All of the reference counting is now done via the functions:
3962 this improves the ability to track the referrals.
3964 Each cm_BeginCallbackGranting Call now allocates a reference
3965 to the cm_server_t. The cm_EndCallbackGrantingCall either
3966 frees the reference or transfers it to the cm_scache_t
3967 cbServerp field. These are then appropriately tracked
3968 through the cm_Analyze call.
3970 * Ensure that the dnlc hash table is the same size as the
3971 dir name hash table (as per original author's note).
3972 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
3973 for compatibility with 64-bit systems.
3975 * fix smb_ApplyV3DirListPatches to properly apply the hidden
3976 attribute to dotfiles when the infoLevel < 0x101 and
3977 cm_SyncOp has failed.
3979 * Fix the Freelance registry initialization code. There
3980 was a possibility that some systems could end up with
3981 garbage in the registry during a clean install.
3984 * file and directory names beginning with "." will now be given the
3985 hidden attribute when the volume access is anonymous. this matches
3986 the behavior when the volume access is via an authenticated user.
3988 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
3989 key. When a change occurs mark the root.afs data as invalid and
3990 for it to be reloaded on the next access. This allows administrators
3991 to modify the mount point list without restarting the service.
3993 The freelance client used to provide a fake modification time for
3994 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
3995 Added code to extract the last modification time of the Freelance
3996 registry key and use that instead. The time now represents the
3997 most recent mount point change.
3999 * PTS registration of new users to foreign cells has been added to
4002 * Additional Cache Control and Credential Manager options have been
4003 added to the WiX installer. See deployment guide for details.
4005 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4007 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4010 * The loopback installation code may have had a problem updating the
4011 %ETC%\HOSTS file which could have resulted in a premature failure.
4012 Work around code has been added for the case where the file cannot
4015 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4016 because Windows sends 64K blocks when using overlapped writes.
4018 * The default number of server threads was increased from 4 to 25 to
4019 better handle overlapped writes.
4021 * The "AfscredsShortcutParams" registry value was not being properly
4022 loaded by afscreds.exe. Therefore, the default value was always being
4023 used instead of the value set by the installer.
4025 * Windows XP provides downgrade attack detection to prevent an attacker
4026 from being able to force the use of NTLM simply by disrupting
4027 communication with the KDC. This attack cannot exist between the
4028 Windows CIFS client and the AFS Client Service. Therefore, when a
4029 downgrade has been detected the afs pioctl library will force the
4030 establishment of a new CIFS connection using NTLM.
4032 * A locking error was discovered surrounding all references to volume
4033 server lists within the cm_cell.c source file.
4035 * The logged into Windows username was incorrect on Terminal Server
4038 * A new registry value "NonPersistentCaching" was added to the service
4039 parameters key. When set to a non-zero value, the afs cache is stored
4040 in the Windows paging file. There are two limitations to choosing
4042 1. when persistent caching is implemented it won't work with
4043 this flag set since there will be nothing to persist.
4044 2. with this flag set the initial paging allocation cannot be
4045 changed while the service is running
4047 * An initialization bug was discovered in aklog.exe which affected users
4048 who have a domain name for their afs servers which could not be mapped
4052 * afs_config.exe now validates cell names against DNS in addition
4053 to the CellServDB file.
4055 * In order to allow the freelance client to connect to a volume with ID
4056 equal to 1 on the default cell we changed the fake root.afs volume ID
4057 once again. This time we choose 0xFFFFFFFF. In addition, we change
4058 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4059 It will now be impossible for a volume ID to match that of another
4060 cell unless the client is connected to 0xFFFFFFFD cells. That should
4061 be enough room for growth.
4063 * Fix "fs mkmount" command to work with UNC paths and when
4064 started from non-AFS drives. It is now possible to create a mount
4065 point in the freelance fake root.afs volume with the command
4067 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4071 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4072 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4074 * The algorithm used to re-attempt access to the servers associated with
4075 a volume has been altered to properly address the case in which all
4076 servers have been marked down. The previous algorithm did not reset
4077 the server's down flags so the servers were never actually retried.
4078 This caused a problem with active volumes if the network connectivity
4079 was lost as could be the case with a network cable removal, wireless
4080 drop, or laptop hibernation. With the fix volume access is restored
4081 almost instantenously when network connectivity becomes available.
4083 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4084 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4085 listing of AFS submounts and freelance mount points. Support for
4086 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4087 significantly enhances the behavior of AFS volumes within the Explorer
4088 Shell. For instance, "AFS" now shows up as server in the Explorer
4089 with each submount or freelance mount point visible as a share.
4090 The right click menu in each folder now works with full functionality
4091 on a consistent basis.
4093 * The network provider can be configured to have different behavior
4094 depending on the domain that the user logs into. These settings are
4095 only relevant when using integrated login. A domain refers to an
4096 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4097 local machine (i.e. local account logins). The domain name that is
4098 used for selecting the domain would be the domain that is passed into
4099 the NPLogonNotify function of the network provider. (see registry.txt
4102 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4103 "Authentication Cell" which may be used to specify a default
4104 authentication cell for afscreds.exe which is different from
4105 the default cell for the AFS Client Service daemon.
4107 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4108 afslogon.dll moved to %WINDIR%\System32\.
4109 New registry entries added to register the dll for Winlogon events.
4111 The logoff event will now force a call to ktc_ForgetAllTokens()
4112 using the context of the user being logged off as long as the
4113 user's profile is not loaded from within AFS. If the profile
4114 was loaded from AFS we can't release the tokens since the Logoff
4115 event is triggered prior to the profile being written back to
4116 the its source location. This is now performed in an XP SP2
4119 * Windows XP SP2 Internet Connection Firewall interoperability
4122 * The %WINDIR%\afsdsbmt.ini contains four sections:
4123 Submounts, Drive Mappings, Active Maps and CSC Policies.
4124 The Submounts and CSC policies are now stored in the registry under
4125 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4126 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4127 The Drive Mappings and Active Maps are stored in the registry under
4128 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4129 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4131 There is no automatic migration of this data as it would be impossible
4132 to consistently migrate data to user profiles which may not be active
4133 when the machine is updated.
4135 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4136 fake root.afs volume. For the same reasons as for the cellservdb file,
4137 this information should not be in %WINDIR%. This information is now
4138 kept under the registry key
4139 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4141 The data from the afs_freelance.ini file will be automatically
4142 migrated to the registry on first execution of afsd_service.exe
4144 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4145 troublesome for several reasons. One, it is confusing for those who
4146 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4147 Two, this file is not a Windows Profile formatted file. Three,
4148 applications should not be reading or writing to %WINDIR%. It causes
4149 problems for Windows Terminal Server.
4151 The new location of CellServDB will be the OpenAFS Client install
4152 directory which is by default C:\Program Files\OpenAFS\Client and can
4153 be determined by querying the registry for
4154 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4156 The existing afsdcell.ini will be migrated by the NSIS installer.
4157 The Wix installer must still be updated to do the same.
4159 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4160 High Security mode; and to add Terminal Services compatibility registry
4161 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4162 configuration files in %WINDIR%.
4164 * Add support for authenticated SMB connections. This will remove
4165 the need for high security mode in most situations. Both NTLM
4166 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4167 only NTLM can be used even though Kerberos is now supported. The
4168 reason is that it is not possible to construct a service principal
4169 which is unique to each individual machine.
4171 SMB Extended Auth does not work on XP SP2 unless one of two registry
4172 modifications are made:
4174 (1) To disable the check for matching host names on loopback connections
4175 set this key. This does not require a reboot:
4177 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4178 "DisableLoopbackCheck"=dword:00000001
4180 (2) To add the AFS SMB/CIFS service name to an approved list. This
4181 does require a reboot:
4183 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4184 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4186 afsd_service.exe will automatically add the current Netbios Name
4187 to the BackConnectionHostNames list and then temporarily disable
4188 the loopback check for one cycle of startup/shutdown of the service.
4189 We assume most folks do not start/stop without a reboot so this
4190 will be adequate in most cases.
4192 * Fix security hole in afslogon.dll which allowed passwords to be
4193 sent in clear text to the KDC in a misformed principal name.
4195 * Fix cm_GetCell() to properly handle expired dns entries
4198 * If Freelance mode is active and the afs_freelance.ini
4199 file does not exist, do not create an empty file.
4200 Instead create a file containing ro and rw mountpoints
4201 to the default cell using the standard conventions.
4203 * Modify the Freelance support to handle the ability
4204 to create rw mount points in the fake root.afs volume.
4206 * Changed the RPC mechanism used for token setting from
4207 named pipes to local. Use of named pipes can be restored
4208 by setting the environment variable AFS_RPC_PROTSEQ to
4211 Named pipes were required when a Windows 9x system was
4212 using a NT system in gateway mode which is incompatible
4213 with our use of local loopback adapters.
4215 * In afscreds.exe, if a username of the form user@REALM is
4216 specified and no password is specified, do not perform a
4217 kinit operation. Only perform the aklog functionality.
4219 * Add a new registry value which allows the number of processors
4220 on which afsd_service.exe executes to be restricted. Valid
4221 values are 1..numOfProcessors
4223 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4227 * A second MSI based installer option is now available.
4229 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4230 request tickets for user/instance@REALM instead of just user@REALM
4232 * Modify the Power Management Notify routine to wait for the Hard Dead
4233 timeout period instead of a fixed 19 seconds. With the longer timeout
4234 periods Hibernation and Standby could never succeed when network
4235 connectivity is not available.
4237 * The following fs.exe commands are now restricted to Administrator:
4238 - checkservers with a non-zero timer value
4241 - sysname with a new sysname list
4250 setting the default sysname for a machine should be done via the
4251 registry and not via "fs sysname".
4253 * NSIS installer adds options to install Debugging Symbols
4254 and the Microsoft Loopback Adapter; the user is now also
4255 given the ability to select the afscreds.exe startup options.
4257 * Build system modified to generate symbols for FREE (aka RELEASE)
4258 builds as well as CHECKED (aka DEBUG) builds
4260 * Sites which have a volume ID of 0x20000001 assigned to their
4261 root.cell volumes have been experiencing problems with accessing
4262 the root.cell volume of their cell when Freelance mode has been
4263 active. This was because 0x20000001 was assigned to the fake
4264 root.afs volume created by freelance. The fake volume id is
4265 now set to 0x00000001 to prevent conflicts.
4267 * The timeout logic in the AFS Client Service has been wrong
4268 for sometime. It is based on two different assumptions.
4269 First, the SMB client timeout is a fix value as was the case
4270 with OS/2 Lan Manager. This assumption is incorrect. The
4271 SMB timeout in Windows is a dynamic value computed based upon
4272 a fixed minimum timeout to which is added time based upon the
4273 size of the request and the performance characteristics of
4274 the connection. Second, it is the responsibility of the
4275 SMB Server to enforce the timeout requirements of the client.
4276 This is untrue. The SMB Server cannot be expected to know
4277 the requirements of the client. More importantly, if the
4278 SMB server uses the SMB client timeout as a value to restrict
4279 its behavior as an RX client, the performance characteristics
4280 of the local SMB session would be used to prematurely terminate
4281 WAN connections with significantly different performance
4284 The timeout logic has therefore been modified in the following
4286 . the Lan Manager Workstation (SMB) Session Timeout is used only
4287 as a basis for configuring the Connection Dead Timeout
4288 and Hard Dead Timeout values. The Connection Dead Timeout
4289 must be at least 15 seconds longer than the SMB Timeout
4290 and the Hard Dead Timeout must be at least double the
4291 Connection Dead Timeout.
4292 . New registry entries have been added to allow the Connection
4293 Dead Timeout and Hard Dead Timeout values independent of the
4294 Lan Manager Workstation Session Timeout
4295 . The test to enforce the SMB Client Timeout has been removed.
4297 One of the side-effects of removing the enforcement of the SMB
4298 Client Timeout is that regardless of whether or not the SMB client
4299 is available to receive the response (and how would the SMB server
4300 know) the RX protocol response can be used to update the AFS
4301 Client Service state for ready access by future SMB client
4304 This should be the end of the "Server paused or restarting messages"
4306 * Add support for arbitrary UNC paths to the pioctl() support.
4307 This enables the fs commands as well as the AFS Shell Extension
4308 to work correctly when UNC paths are being used.
4310 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4311 DNS if the cell configuration cannot be determined via CellServDB
4313 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4314 are ever set more than once at a time
4316 * Fix the management of lists of cm_cell_t structures when using
4317 DNS to lookup cell information. The previous code would fail to
4318 reuse the same cellID for a cell if DNS was used more than once
4319 for a given cell name. When the ttl expired, a single cm_cell_t
4320 could be inserted into the cm_allCellsp list more than once
4321 producing a loop. In addition, the vlServerp list belonging to
4322 the cell was not freed resulting in improper refCounting of the
4325 * Add DNS support to cm_IoctlNewCell() which previous only examined
4328 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4329 to properly garbage collect cm_server_t objects
4331 * Add numVCs variable to smb.c to track the number of smb_vc_t
4332 objects created and use it to initialize the vcID field which
4333 previously was set to 0 in all objects resulting in FindByID
4336 * Fixed DNS lookups to work consistently throughout the OpenAFS
4337 product instead of just from within the afsd_service.exe
4339 * Added a runtime check to ensure that AFS Client Service SMB
4340 Server is accepting connections before attempting to mount
4343 * Read IP addresses for volume servers out of the CellServDB
4344 file if gethostbyname() on the hostname fails.
4346 * Fix getcellconfig() to populate both the Hostnames as well
4347 as IP addresses when loading cell data via DNS
4349 * Increase the Connection Dead Time to 50 from 20 seconds
4350 Increase the Hard Dead Time to 120 from 40 seconds
4351 (matches the Unix values)
4353 * Fixed an assertion validating the number of allocated NCBs
4355 * Fixed the build environment to consistently build for
4356 Windows 2000 and above. (APPVER = 5.0)
4358 * Fixed rx_debug to properly validate the receipt in incoming
4359 data with select() and recvfrom(). Do not copy data out of
4360 the socket buffer unless success is indicated.
4363 * afsd_service.exe will now display a message box to the
4364 desktop when it terminates due to an IP Address Change.
4366 * installer no longer deletes AFS Server configuration data
4369 * installer generates a warning dialog if the RPC service
4370 is not properly configured
4372 * installer compressed with lzma instead of bzip2
4374 * afsd_service.exe shutdown crash solved once and for all
4376 * reference counting of smb_vc_t data structures improved
4378 * name space collision of smb_fid_t event objects corrected
4380 * the output of "fs memdump" is now written to
4381 %WINDIR%\TEMP\afsd_alloc.log
4383 * the file TaAfsApp_1033.dll is now properly installed allowing
4384 the User Manager to start
4386 * a new algorithm is used for computing filename pattern matches
4388 * afscreds.exe now accepts user names containing instance
4391 * Fix the Directory Name Lookup Cache to be case-sensitive.
4392 This is crucial in environments in which a Windows client
4393 is accessing a directory with more than one filename that
4394 differs only by case. If the directory contains "FOO"
4395 and "Foo". You want "DEL Foo" to delete the correct one.
4396 We still have a problem in that "DEL foo" will delete a
4397 random filename. This will be addressed in a future release.
4399 * Fix afscreds.exe -M option (renewMaps) to work when High
4400 Security mode is off. Also, remember to disable the ActiveMap
4401 flag in afsdsbmt.ini when a drive mapping is removed.
4403 * Updates to NSIS installer script. AFS Server configuration
4404 data will not be destroyed on un-install or re-install.
4405 Use a better compression algorithm.
4407 * afslogon.dll now uses KFW to obtain tokens when available
4409 * afslogon.dll when given an all uppercase username will
4410 attempt to authenticate with both the uppercase name
4411 and an all lowercase variation
4413 * DST modification removed. The fix appears to make things
4414 worse after a reboot of the machine.
4416 * fs.exe: added "cscpolicy" which is used to
4417 change client side caching policy for AFS shares
4419 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4420 [-documents] [-disable] [-help]
4422 * Several uninitialized variables have been initialized
4424 * It is now possible to obtain tokens using cross realm
4425 Kerberos within afscreds.exe:
4427 user: jaltman@ATHENA.MIT.EDU
4429 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4430 will will in turn be used to obtain afs@DEMENTIA.ORG.
4431 The resulting token will be stored with the display name
4432 jaltman@ATHENA.MIT.EDU@dementia.org
4434 * aklog.exe has been added to the client
4436 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4437 [[-p | -path] pathname]
4441 -d gives debugging information.
4442 krb_realm is the kerberos realm of a cell.
4443 pathname is the name of a directory to which you wish to authenticate.
4444 -noprdb means don't try to determine AFS ID.
4445 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4446 (default is Kerberos V)
4447 No commandline arguments means authenticate to the local cell.
4450 * All of the resource files have been restructured to adhere to
4451 a set of rules IBM implemented for loading string resources.
4452 These rules had either been forgotten or were not discovered
4453 by folks working on the OpenAFS sources. The end result was
4454 memory corruption. This is primary item which was preventing
4455 the AFS Server from working.
4457 * Increased the size of the maximum ticket size stored in a token
4458 from 344 bytes to 12,000. Increased the buffers used to convey
4459 messages between the pioctl() caller and the SMB Server from
4460 1000 bytes to 12,512. The code appeared to have been writing
4461 above the top of the stack by quite a few number of bytes.
4462 (The increased ticket size is necessary for the next item.)
4464 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4465 Instead the raw Kerberos 5 ticket is used in its entirety. This
4466 is extremely important as it allows us to use pure Kerberos 5 KDCs
4467 as the source of the AFS authentication. The use of up to 12,000 byte
4468 tickets will allow tickets produced by all versions of Microsoft
4469 Active Directory to be used.
4470 - create a user account.
4471 - designate it DES only
4473 - specify its UPN to be "afs@realm"
4474 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4476 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4477 is special when using long file names. (you must use "*.*" and not "*")
4478 Instead only enforce it when performing 8dot3 searches.
4480 * Fixed the DST problem with creation times being set one hour ahead
4482 * Fixed the problem when using \\afs\cell-alias. For example,
4483 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4484 for the alias name; instead simply expand the name. One of the
4485 symptoms of this problem was a loss of acquired tokens.
4487 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4488 of strings. (Only English strings provided.)
4490 * Fixed the installer to properly replace in use files.
4492 * Fixed the build system to cleanup generated component version files
4494 * The release build compiled with MSVC 6.0 compiler to avoid the
4495 afsd_service.exe shutdown crash. This does not solve the problem
4496 but simply avoids it for the time being.
4500 * fix afslogon.dll to not corrupt memory when High Security mode
4503 * fix afsd_service.exe to not attempt to restore the stack when
4504 an exception occurs. (not safe in multi-threaded programs)
4506 * fix uninstaller to properly remove the CRT and MFC DLLs
4508 * remove a Message Box from afscreds.exe when getcellconfig()
4509 fails on a kerberos realm which is not a cell
4511 The following is a list of changes to the OpenAFS for Window client
4514 * "fs setserverprefs" will leave afsd service deadlocked
4516 * "vos listaddrs" will core dump
4518 * installer sets the appropriate keys to support Integrated Logon
4520 * installer disables the "Find Lana by Name" functionality as it
4521 was causing headaches for many users
4523 * fix the intermittent crash of the power management thread when
4524 shutting down the AFS Client Service
4526 * optimizes the obtain drive mount list functionality which is
4527 executed every time the mount tab in afscreds.exe and afs_config.exe
4528 are refreshed. (this happens a lot)
4530 * fix the service shutdown logic. add the STOP_PENDING state
4531 and do not accept additional service events after we declare
4534 The following is a list of changes to the OpenAFS for Window client
4537 * flexelint was run against the source tree and hundreds (perhaps
4538 thousands) of corrections were applied to ensure prototypes
4539 were in use; types were used consistently; variables were
4540 initialized; unused variables were removed; etc.
4542 * A wide variety of instrumentation was added including the
4543 ability to produce a stack trace from within afsd_service.exe
4546 * Dynamic configuration of the RDRtimeout value based upon the
4547 LanMan Workstation Session Timeout
4549 * The mount root no longer needs to be called "/afs". This
4550 is now set by a registry value "MountRoot" within the key
4551 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4553 * The cell list is now only read out of afsdcell.ini when the
4554 file changes instead of each time a cell is resolved.
4556 * Thread synchronization was added to cm_server.c and ktc_nt.c
4558 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4559 calloc()/free(). The Global functions were needed on Windows 3.x
4560 but have caused a variety of problems on the Win32 platforms.
4561 Avoiding them is highly recommended by several Microsoft
4562 Knowledgebase articles
4564 * Support for Symbolic Links added to the AFS Shell Extension
4566 * Added a registry value "OverlayEnabled" to determine if
4567 Shell Extension Overlays should be enabled.
4568 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4570 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4571 separate trees for checked and free builds. Build system supports
4572 a custom directory src\WINNT\extra which can be used as a grafting
4573 location of organization specific additions to the build tree.
4575 * New installer built using NSIS 2.0.
4577 * Named all kernel objects in order to allow them to be monitored
4578 with tools such as SysInternals' ProcExp.exe.
4580 * Introduced new EventLog framework for AFSD
4582 * Introduced Power Management interface to AFSD for Standby and
4583 Hibernate modes to allow cache to be flushed prior to network
4586 * Utilize Win32 DNSQuery API instead of internal routines. This
4587 allows DNS SRV queries to be sent to all current domain name
4588 servers. Not just one specified in an INI file. DNS is now
4591 * "NetbiosName" registry value may be used to specify a fixed
4592 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4593 when the loopback adapter is in use. If you need to use the
4594 old notation with a loopback adapter installed specify a registry
4597 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4599 * Refactor all modules which depend on LAN Adapter and NetbiosName
4600 determination in a new library: lanahelper.lib. This allows for
4601 consistent behavior throughout the product.
4603 * Move the afsd.log and afsd_init.log files to the directory specified
4604 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4605 for services. Added the Date to the log entries.
4607 * New registry value "RxMaxMTU" used to limit the size of the RX
4608 packets sent by the AFS Client Service to the Server. In order
4609 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4610 size must be restricted to 1264 or smaller. The latest NSIS
4611 installer sets a value of 1260 by default.
4613 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4614 packets. This is not needed in order to work across the Cisco
4615 VPN but might be needed for other network environments. This
4616 value is not set by the NSIS installer.
4618 * New registry value "HideDotFiles" is used to apply the Hidden
4619 attribute to files whose names begin with a '.'. This value
4620 is set by the NSIS installer.
4622 * New registry value "MaxMpxRequests" allows the maximum number
4623 of multiplexed sessions to be configured at run time. This
4624 value is not set by the NSIS installer. The default value is
4627 * New registry value "MaxVCPerServer" allows the maxmimum number
4628 of VCs per server to be configured at run time. This value is
4629 not set by the NSIS installer. The default value is 100.
4631 * New registry value "AllSubmount" allows the "all" submount to
4632 be disabled by setting its value to 0x00.
4634 * Allow cells names to be valid mount points
4635 \\<netbiosName>\<cellname>
4637 * Store the active state of drive mappings in order for afscreds.exe
4638 to restore them upon startup
4640 * Add exception handling to generate a Stack Trace to the afsd_init.log
4641 file if one happens to occur.
4643 * Add lots of logging to help detect the cause of invalid SMB packets
4645 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4646 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4647 renewal of expiring tokens as long as afscreds.exe is running.
4649 * New afscreds.exe command line options:
4651 -M = renew drive maps
4652 -N = ip address change detection
4655 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4656 determines whether or not MIT Kerberos for Windows should be used
4657 to obtain tokens via Kerberos 5 tickets.
4659 * New registry value "AfscredsShortcutParams" in
4660 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4661 determines the command line parameters to be specified when "fixing"
4662 the AFS Shortcut in the user's startup folder.
4664 * The "ShowTrayIcon" registry value has been moved from
4665 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4666 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4668 * The <cell name> registry values used to store the token expiration
4669 reminders have been moved from
4670 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4671 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4673 * Obtain the Logon User Name from the Explorer key when available
4675 * new text document doc\txt\winnotes\registry.txt lists all registry
4676 values used by OpenAFS (excluding the AFS Server)
4678 * BUG: rx_securityClass objects were not properly reference
4679 counted and were never freed.
4681 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4682 would be generated. The existence of a server does not imply
4683 that it is not down. If all of the servers for a cell are down
4684 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4687 * BUG: the directory name lookup cache failed to free the entries
4688 in the cache when the name cache entries cycled. The entries
4689 in the cache would become dereferenced without being freed.
4691 * BUG: fs setserverprefs could be executed without Administrator
4694 * BUG: the number of allocated NCB objects (100) exceeded the number
4695 which could actually be waited upon by the kernel (64). Any objects
4696 which were utilized above the limit could never have event completions
4699 * BUG: smb_username_t objects were not being reference counted and
4700 were not properly freed.
4702 * BUG: smb_tid_t objects could under unusual circumstances be freed
4703 before they were no longer referenced.
4705 * BUG: smb_fid_t object pointer were frequently used even when
4706 their value could be NULL. They were not properly released and
4707 therefore they were never freed.
4709 * BUG: smb_packet_t data structures were not completely initialized
4712 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4713 Denied" because that causes the Explorer Shell to try again until
4714 access is obtained. Instead return "Remote Resources" which allows
4715 the shell to move on and treat the error as transient.
4717 * BUG: when initializing the NCBreturns structure, separate Event objects
4718 were created for each NCB although a single Event object was supposed
4719 to be shared by all.
4721 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4724 * BUG: smb_tran2Packet_t objects were not being properly referenced
4727 * BUG: directory path creation did not handle the case of multiple
4728 directories requiring creation in one attempt
4730 * BUG: SMB requests which required an Extended Response were ignored.
4731 This prevented some files from being written to AFS volumes.
4733 * BUG: character strings were being freed even after they were
4734 inserted into in use data structures
4736 * BUG: inconsistent usernames were used when High Security mode was
4737 enabled. (there is still much to do in this area)
4739 * BUG: pioctl() calls which require out of band RPC operations were
4740 susceptible to race conditions when performed by multiple processes
4742 * BUG: memory allocation and deallocation crossed instances of the
4743 C Runtime Library producing memory leakage and corruption in
4744 afscreds and the client configurator.