2 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
3 No longer will browsing \\AFS produce truncated share names.
5 * Registry specified server preferences did not result in
6 the CM_SERVERFLAG_PREF_SET flag being applied to the
9 * Avoid unnecessary DNS lookups of share names as cells
10 when it is known that the name cannot be a cell name.
11 Any name that does not contain a dot is skipped.
13 * When processing Dfs Referral requests, do not return an
14 error if the server is down or busy; if the volume is
15 offline; the cell vldb cannot be reached; etc.
16 These paths are still \\AFS paths and so the client
17 should not be sent to search elsewhere for resolution.
19 * CRITICAL: Prevent the cm_Daemon thread from terminating when
20 the machine enters the suspend state. IF the cm_Daemon thread
21 dies, there is nothing to execute down server checks.
23 * Prevent use of smb_StartedLock before initialization which
24 can result in a panic in the lock package.
26 * Improve error checking in afslogon.dll in order to prevent
27 NULL pointer dereferences within GetLogonDomainOptions()
28 if the Lsa operations fail.
30 * Add support to the build system to automatically update
31 a Microsoft Symbol Store.
33 * Do not access the cm_conn_t after executing the cm_Analyze
34 loops. Doing so can result in an access to memory that has
37 * CRITICAL: If a StoreData request offset is not aligned on the
38 buffer module blockSize, (offset % blockSize) bytes of dirty
39 data will not be written to the file server.
41 * CRITICAL: If a StoreData64 request is sent to a file server
42 that does not support large files (>2GB) and the client is
43 unaware that the server has no large file support, the first
44 chunksize worth of data will not be written to the file server.
45 This bug was introduced in 1.5.3. It affects all IBM AFS file
46 servers and all OpenAFS file servers older than 1.4.0.
49 * If a file server becomes inaccessible while the cache manager has
50 dirty buffers to write, the afsd_service buf_IncrSync thread can
51 attempt to use 100% of the cpu.
53 * Fix the locking that protects the cm_cell_t lists
54 (all, name, id, free).
56 * Remove cm_cell_t objects from the name and id hash tables before
57 placing them on the free list.
59 * Fix "fs newcell" which was broken in 1.5.60.
61 * Do not attempt to synchronize dirty buffers if the associated
62 volume is known to be unavailable.
64 * Log to the Windows Event Log dirty buffers that cannot be written
65 during the shutdown process.
67 * Modify behavior of a Freelance mountpoint target that does not
68 specify a cell. Instead of assuming the target volume is in the
69 Freelance.Local cell, use the workstation "Cell" specified in the
70 registry. A mountpoint target of "#root.cell." will now mean the
71 root.cell volume in the workstation cell for the current session.
72 If the workstation cell changes from "athena.mit.edu" to
73 "andrew.cmu.edu", the referenced volume will also change without
74 requiring that the mount point targets be altered.
76 * Avoid false infinite loop errors when validating the cm_cell_t
77 free list during startup.
79 * Fix two more locations where xdr_free() should be used instead of
80 free() to permit a checked built of afsd_service.exe to work with
81 release builds of afsrpc.dll.
83 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
84 to permit the server Uuid to be used to lookup the server object
85 and from that determine the cell. This permits callbacks that are
86 received from alternate addresses to be processed with a known server
87 object. Previously a request from an unknown server would clear all
88 callbacks from all cells.
90 * Fix a bug that prevented optimal performance when using a non-zero
91 value for 'daemonCheckVolCBInterval'. As a reminder, when
92 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
93 volume callbacks are automatically renewed 90 minutes before their
96 * Fix automatic ranking of vldb servers whose values are obtained from
99 * Add logging for RX CALL DEAD errors which are generated if the remote
102 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
103 or the call is to a vldb server.
105 * Add registry based cell search functionality to NetIdMgr,
106 afs_config.exe, and klog.exe.
108 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
109 lookups on the host names in the CellServDB instead of using the
110 specified IP addresses. This provides aklog, pts, vos, etc. the same
111 CellServDB behavior that the Windows Cache Manager uses.
113 * Completely a behind the scenes change. The snprintf() routines have
114 been completely rewritten.
116 * When updating the stat cache entry callback of a .readonly object
117 from the volume group object, update the file server reference to
118 ensure it matches the most update to date callback.
120 * Add proper support for processing callbacks from multi-homed file
121 servers. Instead of comparing servers by cm_server_t pointer,
122 compare them by UUID when the UUID is known.
124 * During a shutdown short circuit the offline volume check daemon
127 * Properly track link counts during cm_Link and cm_Unlink operations.
129 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
130 preference to an error code reported by rx_EndCall.
132 * Add "PerFileAccessCheck" registry value to permit testing against
133 experimental file servers that include per-file acl support. This
134 value is intentionally undocumented. It is not to be used by
135 production environment deployments.
137 * Log a warning to trace log if a volume group is marked as a DFS
138 File Set. The Windows OpenAFS client does not include DFS support.
140 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
141 provider from obtaining tokens when referrals are in play.
143 * Add "fs chown" and "fs chgrp" commands to permit the owner and
144 group of objects stored in AFS to be set from Windows.
146 * Avoid performing background daemon operations when the machine is
147 going into suspend mode.
149 * Perform offline volume checks in most recently used order.
151 * Prevent crash when a data version for a cache object goes backwards.
153 * fs setquota, fs setcachesize, vos setfields, and vos create now
154 accept human readable orders of magnitude. (K, M, G)
156 * fs listquota fixed to permit large quota sizes to be displayed.
158 * rx packet resend and data packets sent counts were incorrect.
160 * Multi-thread safe library versions are now being generated and used.
161 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
163 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
164 values are now available on XP through Windows 7. Add functionality
165 to autodetect if such support is present on the machine. If so,
166 configure it if necessary and dynamically adjust the AFS Rx timeout
169 * Correct another race condition in the Rx library that could result
170 in an unexpected panic while freeing the Rx call iovq.
173 * A fix to the pioctl library to support drive substitution
174 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
176 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
177 that corrects a deadlock in the smb redirector and also adds new
178 functionality that permits the AFS SMB server to be given a longer
179 timeout than is normally the case. New functionality has been added
180 to configure these additional LanmanWorkstation\Parameter values.
181 (This functionality has been backported to XP SP3 and is scheduled
182 to be released on June 5th.)
184 * The BackConnectionHostNames registry value configuration was broken
185 when dynamic re-establishment of Netbios Name registrations was added.
188 * Hidden vos commands are revealed.
190 * If the "DisableLoopbackCheck" registry value is set, do not unset
191 it during the same service session.
193 * Reorganize code that prevents multiple Store operations to the same
196 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
197 permit broken symlinks to be treated as being in AFS.
199 * Fix vos commands that output 64-bit integer values.
201 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
202 Permits building cygwin applications against OpenAFS libraries.
204 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
205 as Windows HTML Help (.CHM) files.
207 * NSIS installer does a much better job of cleaning up files left
208 over from previous installs.
210 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
211 or "fs flushall" and on-going directory operations that can result in
212 afsd_service.exe crashing.
214 * Add support for DNS SRV records in place of AFSDB records.
215 _afs3-vlserver._udp.<cellname>. Priority field is used.
216 Weight is currently ignored.
218 * Add a method of specifying Client CellServDB information within the
219 registry that can be used to either override the CellServDB file or
220 force the use of DNS lookups for a given cell.
222 The registry schema is as follows:
224 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
225 "LinkedCell" REG_SZ "[cellname]"
226 "Description" REG_SZ "[comment]"
227 "ForceDNS" DWORD {0,1}
229 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
230 "HostName" REG_SZ "[hostname]" Default: [servername]
231 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
232 "IPv6Address" REG_SZ "[address]" <future>
233 "Comment" REG_SZ "[comment]"
234 "Rank" DWORD "0..65535" Default: 0
235 "Clone" DWORD "{0,1}" <future: server only>
236 "vlserver" DWORD "7003" <future>
237 "ptserver" DWORD ... <future>
239 ForceDNS is implied non-zero if there are no [servername]
240 keys under the [cellname] key. Otherwise, ForceDNS is zero.
241 If [servername] keys are specified and none of them evaluate
242 to a valid server configuration, the return code is success.
243 This prevents failover to the CellServDB file or DNS.
245 Only one of "HostName" or "IPv4Address" is required. "HostName"
246 is optional if [servername] is the host name.
248 * Extend registry based CellServDB functionality to
249 afsconf_GetCellInfo() interface used by aklog and many
250 other command line utilities.
252 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
253 as it is now used by both client and server components.
256 * PriorityClass of afsd_service.exe process raised to
257 "High" to match the priority of the system services
258 that are waiting for afsd_service.exe to respond.
259 Configurable using [TransarcAFSDaemon\Parameters]
260 DWORD "PriorityClass" value. (use values from Win32
263 * Change error reporting of CM_ERROR_BADSHARENAME to
264 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
265 The error we were sending would cause the smb
266 redirector to drop the connection.
268 * Change the error reporting of CM_ERROR_ALLOFFLINE
269 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
270 Previous we would return Bad Network Path which
271 would cause the smb redirector to cancel all file
272 handles on the same "share".
274 * A debug version of afsd_service.exe can now be used
275 with a release version of afsrpc.dll (or vice versa).
276 afsrpc.dll now implements and exports xdr_free().
278 * Removed all unprotected OutputDebugString() calls as
279 those calls can block for up to 10 seconds if there
280 is contention for the globally shared memory or if
281 the debugger fails to read data in a timely manner.
283 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
284 supported for builds. The production releases still
288 * Discovered that XP/2003 supports a variant on recvmsg
289 called WSARecvMsg. The input parameters are mostly
290 compatible with the Posix recvmsg. This will save
291 multiple memcpy() calls per received packet.
293 * Discovered that Vista/2008 supports a variant on
294 sendmsg called WSASendMsg. The input parameters
295 are mostly compatible with the Posix sendmsg. This
296 will save multiple memcpy() calls per sent packet.
298 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
299 and utilizing rx_ReadProc32 from within the cache manager
300 as part of the FetchData operations. rx_ReadProc32 and
301 rx_WriteProc32 are optimized versions of rx_ReadProc and
302 rx_WriteProc for use when reading 32-bit integer values.
304 * The Network Identity Manager AFS provider now reports
305 progress events. These provide a better user experience
306 with Network Identity Manager 2.0.
308 * [RT 124293] A race condition exists which permits the
309 scp field of the an smb_fid_t object to become invalid
310 after a request on the smb_fid_t has begun. Remove the
311 race by protecting all access to the scp field with the
312 smb_fid_t mx mutex and obtaining a local reference on
313 the cm_scache_t object for the length of the request.
315 * [RT 124276] If the vldb is out of sync with the contents
316 of the file servers, afsd_service will retry too many
317 times when a file server reports a volume as not being
318 present. Now if the list reported by the vldb is the
319 same as the previously seen list, then the retry is
322 * [RT 124276] Read-only volume failover was broken in
323 1.5.53 whenever accessing a volume results in VNOVOL or
326 * [RT 124276] Prior to 1.3.70 the volume server reference
327 list was not reference counted and would be prematurely
328 freed while in use. When reference counting was added
329 in 1.3.70 a bug was introduced that could result in
330 service reference list corruption.
332 * Add Windows Application Event Log warning messages for
333 "Client SMB MPX value too large" and
334 "Client SMB Buffer Size too small".
336 * Renaming of files across directory boundaries would result
337 in an invalid handle error when attempting to access the
338 files after the move.
340 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
341 when a smb file descriptor cannot be found. The whole point
342 of using a Path Info function is because an smb file
343 descriptor wasn't allocated.
345 * More edge cases in which dynamic addition of Freelance
346 root.afs entries would get the wrong FID or where the
347 root.afs directory would not be refreshed.
349 * Buffer overflow could occur if the workstation cell name was
350 longer than 64 characters. Crashes could occur in
351 afscreds.exe, afslogon.dll, and afsd_service.exe.
353 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
356 * Improve consistency of the bad data version value used for
359 * Log file server uuid values as part of the cm_server object
360 when available. Dump the cm_server object list in response
363 * Optimize the performance of resetting access control lists
364 when tokens are set or removed.
366 * Fix vos.exe so that it can work with very large volume
369 * Remove symlink recursion tests and increase max symlink count
373 * Undo the "UAC manifest fix" applied to afs_config.exe.
375 * Fixed some additional race conditions in the Rx library
377 * Ensure that Freelance allocation of vnodes follow
378 the AFS convention of odd vnodes are directories
379 and everything else is an even vnode.
381 * Add Freelance logic to mount point and symlink
382 evaluation functions.
384 * Enhance smb_ParseASCIIBlock() so that it can handle
385 all of the STRING formats defined by the CIFS Technical
388 * Validate the output of smb_ParseASCIIBlock() in all
389 callers. Return CM_ERROR_BADSMB if the STRING field
390 cannot be parsed. CM_ERROR_BADSMB will cause the
391 contents of the packet to be logged.
393 * If multiple SMB Raw Write operations were taking
394 place at the same time, there could be data corruption
395 because unique event objects were not generated for
396 each Netbios receive operation.
400 * Fix a race in rxi_WriteProc which could result
401 in a freed or in use packet being passed to
402 rxi_PrepareSendPacket if the call is currently
403 in a transmit wait state. Passing a freed packet
404 will result in a panic. Passing an in use packet
405 might result in a call failure or perhaps data
408 * Kerberos referrals broke the NIM AFS provider
409 logic for automated configuration of the workstation
410 cell with a new identity. Now if a referrals
411 response is returned, assume the cell belongs to the
414 * The interface to buf_CleanAsync() and
415 buf_CleanAsyncLocked() has been changed to permit
416 errors to be propagated to the caller.
418 * During an enumeration of a directory, if the btree
419 has not yet been generated, return an error.
421 * When mapping CM_ERROR values to NTSTATUS values
422 do not treat 0 as an error.
424 * Avoid a cm_scache_t object reference count underflow
425 when renaming via the SMB interface.
427 * smb_IsMask() can be called with a non-8.3 input buffer.
428 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
429 input buffers and smb_IsMask() for non-8.3 input buffers.
431 * Add cm_buf_t error state to the "fs memdump" output.
433 * pioctl drive mapping checks need to take into account
434 Global Drive Mappings and SUBST assignments.
436 * Rename operations should not be permitted if the
437 destination name collides with an existing name that
438 differs only by case UNLESS the collision is the object
439 that is being renamed.
441 * Increase the maximum number of background daemon threads
442 to 64. Leave the default at 4.
444 * Set the FILE_UNICODE_ON_DISK flag when reporting the
445 file system characteristics.
447 * Report the maximum component name length as 255 even
448 for Unicode as part of the file system characteristics.
450 * Add RX debugging information to "fs memdump" output
453 * Refcount the cm_server_t objects returned by
456 * Fix the UAC manifest applied to afs_config.exe
458 * rxdebug now indicates if any RX_CONN flag is set
459 not just some of them
461 * Change the mapping of CM_ERROR_WOULDBLOCK from
462 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
466 * Includes #GCO Public CellServDB 07 Nov 2008
468 * Prevent separate cell entries from being created for
469 "foobar.org" and "foobar.org." which can happen if
470 an AFSDB record is found instead of a CellServDB
473 * Fix a panic when checking the SMB virtual circuit list.
474 If a VC is destroyed while the list is being checked
475 it is possible for the next object to become invalid.
477 * Increased logging to afsd_init.log of netbios name
480 * Fix a DNLC panic when a UTF-16 string that cannot be
481 converted to a utf-8 string is provided as a filename.
483 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
484 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
485 location to refer directly to the new location in
486 %programfiles%\openafs\client\program
488 * Be more aggressive about responding to Netbios
489 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
490 the LANA within the listener thread without performing a
493 * Use the WNet Enumeration of all connected drives as a filter
494 in the pioctl GetIoctlHandle operation. This should remove
495 the last situations in which there was a delay when opening
496 an Explorer Shell properties menu with the AFS Shell
499 * Fix the parsing of CellServDB in Windows to handle linked
500 cells as is done on Unix.
502 * Add support for linked cells to the cache manager,
503 netidmgr afs provider, and aklog. When two cells are
504 linked, a volume lookup that fails in one cell is looked
505 up in the other. This can be used for construction of
506 test cells that override specific volumes; cell renaming;
509 * Increase the length of cell and realm names that can
510 be entered in the NIM AFS provider.
512 * Actively prevent a 32-bit afsd_service.exe from
513 executing in the WOW64 environment.
515 * Protect against invalid IP address values in CellServDB
520 * Fix an Rx panic caused by an attempt to double free
521 an rx_packet when an error occurs while writing the
524 * Fix an exception cause by a race condition when
525 processing multiple file change notifications registered
526 by the Explorer Shell or other applications.
528 * Modify the Rx debug interface to use var_args and
529 permit Rx debugging to be used on 64-bit Windows
530 which previously had it disabled.
534 * Automated lock hierarchy enforcement added to the
535 osi base lock package. This permits detection of lock
536 violations that might but often does not result in
537 deadlocks. Enabled by default in the checked builds.
538 May be activated in release builds via the registry
539 value "LockOrderValidation" in the TransarcAFSDaemon
542 When a violation is detection an assertion is thrown
543 generating a dump file and a message indicating which
544 locks were obtained out of order is logged to the
545 Windows Application Event Log.
547 * Attempt to ensure that a dump file will be generated
548 whenever the service panics.
550 * Removed unused lock cm_bufGetMutex
552 * Many out of order lock acquisitions were detected
553 using the lock order verifier and were corrected.
555 * A race condition within the function path cm_GetSCache() ->
556 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
557 cm_scache_t object to be issued simultaneously to two
558 threads. This would eventually result in a panic due
559 to the resulting undercount.
561 * Fix interpretation of the empty string as the ioctl path
562 to mean the current directory. "fs lsm", "symlink list",
563 etc. now return a "not a ..." error instead of "does not
566 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
567 to indicate that the input string is a nul terminated string.
568 Assign ti when input strings are nul terminated instead of
569 length counted. (from 1.5.50)
571 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
572 does not match the observed behavor. The 'nameLength' is specified
573 in bytes, not in characters. Fix the implementation to match
574 the observed behavior. (1.5.50)
576 * The cm_HaveAccessRights() attempt at deadlock avoidance by
577 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
578 does not avoid the deadlock. Avoid the deadlock by enforcing the
579 lock order of the lowest vnode first. Then remove the infinite
580 loop avoidance in cm_SyncOp() that was returning an unwarranted
583 * File Server lock synchronization was not propertly enforced in
584 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
586 * Another change to the short name generator. Plus and equal are
587 not valid characters in a short name according to the
588 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
589 with left and right brace in order to minimize the impact.
591 * Fix a race condition that can result in a heap overwrite error if
592 the number of known file servers or vldb servers increases during
593 the preparation phase of a cm_CheckServers() operation.
595 * If the SMB Authentication would fail, a LSA allocated memory
596 buffer would be leaked.
598 * Rewrite the cm_cell_t allocation code to permit cell vldb server
599 lookups to fail without leaking cm_cell_t objects.
601 * Add "fs setcrypt auth" to permit the use of rxkad authentication
602 only mode. Not documented.
604 * Initialize Rx mutex and lock objects once and only once as part of
605 the DLL load process.
607 * When reporting the Volume Name "AFS" do not include the terminating
608 NUL in the reported length.
610 * Refresh Volume Location information upon receipt of a VNOVOL error
611 from the file server and avoid a race condition that prevented the
612 updated volume location information from being destroyed immediately
613 after it was acquired.
615 * In response to VNOVOL and VMOVED, remove the volume id from the server's
616 volume list so that the volume's up and down status is not affected
617 when the machine is shutdown.
619 * Many fixes to the Rx stack related to packet allocation and deallocation;
620 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
621 by overwriting the rx_call currentPacket pointer when it was already
622 referring to a packet.
624 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
625 in the TransarcAFSDeamon/Parameters key.
627 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
628 constructed directory B+ trees for flushed directory objects.
630 * When cached pages for directory objects have been locally modified, do
631 not permit the locally modified pages to be mixed with pages subsequently
632 obtained from the file server. Doing so can result in directory
635 * If Ubix errors are received in response to VL_xxx operations record
636 the error on the server instance and permit the operation request to
637 be retried in case another server is able to answer the query. We have
638 seen at least one instance of a cell in which a VL server returned
639 UNOQUORUM (possible database truncation) when the other servers were
640 responding successfully.
643 * Optimize the assignment of write errors to dirty
644 buffers by applying the first error to all buffers
645 belonging to the same object instead of trying each
648 * Add resource version info to xstat_fs_test and xstat_cm_test
650 * Fix a delete vs delete[] usage problem in the WiX custom
653 * When creating a mount point, validate the volume name to
656 * When processing a mount point, pay attention to VL_BADNAME
659 * Remove a race condition from cm_volume_t recycling that could
660 have resulted in two threads making use of the same object at
663 * Fix a reference count leak on smb_vc_t objects introduced in
664 1.5.50. Permit the objects to be freed once again.
666 * Fix the output of "fs memdump" for smb_vc_t objects that was
669 * Add routines for debugging smb refcount errors.
671 * When shutting down the cache manager, do not permit AFS
672 callback requests to be processed. Doing so can result in
673 invalid memory accesses.
675 * Do not recycle a deleted cm_scache_t object unless the
676 refCount is 0. Doing so can result in a deadlock and
677 invalid memory access.
679 * Minimize the number of RPCs sent to the file server for
680 directory bulk status requests. Only request status for
681 objects that we are actually reporting on and only if we
682 do not already have status for the object and then make
683 sure that we return as much as we can at once via the SMB
684 interface but not so much that it exceeds the RDR timeout
686 * Enable bos restricted operations
688 * Fix the creation of submounts used by the afscreds and afs_config
691 * Avoid repeated retries when attempting to obtain access rights.
692 If the access rights were obtained successfully but do not
693 satisfy the request, fail the request.
695 * Fix error code reported when attempting delete a file on a
696 readonly volume or one that is marked with the readonly DOS
699 * Fix a short name truncation bug.
701 * Fix a heap corruption error when reading the CellServDB file
704 * Fix directory listings when the directory has "lookup" permission
705 only and the directory has never been previously accessed.
707 * Fix several tracking issues for cm_scache_t rw locks.
709 * Ensure that all buffers used to store cell names are the
712 * Restore Windows 2000 support (broken in 1.5.50).
713 LOCALE_INVARIANT was introduced in XP.
715 * Add RxUdpBufSize registry value. Default is 256K.
717 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
719 * Do not include trailing NULs when returning file names in
720 a directory search response.
722 * Prevent RX from crashing the afsd_service by attempting
723 to allocate to many buffers at once when memory is fragmented.
725 * Avoid a deadlock when freeing cm_server_t objects
727 * Optimize search cell file. Do not search to end of file
728 if the cell has already been found.
730 * Remove explorer shell extension delay on local disks.
732 * Automatically pre-allocate 64 Rx packets per thread
734 * Add smb lock requests and stat cache lock allocations to
735 the output from fs memdump.
738 * Fix Large File Support (broken in 1.5.50)
740 * Fix "fs setacl" (broken in 1.5.50)
742 * Fix pioctl path evaluation of drive letter
743 mapped paths. (broken in 1.5.50)
745 * Fix installation of VC8 Runtime library in the EXE
746 Retail installer. (broken in 1.5.50)
748 * Avoid a potential deadlock during Volume Location
749 updates if all the file servers are unreachable at
750 the time the VL update succeeds.
752 * Ensure that rx connection object is references
755 * Fix propagation of over quota errors during cm_FSync()
758 Since 1.5.39 [1.5.50 released 16 July 2008]
759 * There were no releases between 1.5.39 and 1.5.50.
761 * FIRST RELEASE WITH UNICODE SUPPORT
763 * Pioctl interfaces to the cache manager refactored
764 to provide layering between the SMB specific code
765 and the general purpose ioctl operation.
767 * Garbage collect dead SMB virtual circuits as soon
768 as they are no longer being referenced. This avoids
769 problems with outstanding locks not being dropped
770 when the virtual circuit becomes invalid.
772 * Remove the IBM Administration Reference documentation
773 and replace it with the OpenAFS Command Reference
776 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
777 functions each time a connection is about to be used.
778 Do not hold a lock on the rx connection object while
779 it is being selected. This avoids a race between
780 threads attempting to set the timeout values and
781 removes a bottleneck that was hampering performance.
783 * Ensure that the smb directory attribute is set
784 for all directory objects.
786 * Add vs2008 support to the NSIS installer scripts
788 * Replace the VC Runtime EXE installer with the MSI
789 installer in the NSIS installer scripts
791 * Properly delete the VC7.1 C Runtime libraries.
793 Since 1.5.36 [1.5.39 released 23 June 2008]
794 * There were no 1.5.37 or 1.5.38 releases for Windows
796 * Use Visual Studio version of vsprintf() instead of
797 home grown version that resulted in linking conflicts
799 * Prevent buserver, ptserver and fileserver from crashing
800 due to non-portable strftime() format strings.
802 * Prevent server crashes caused by rx_InitMorePackets()
803 being called prior to allocation of all of the required
804 synchronization objects.
806 * Fixed a memory leak in the internal B+ tree directory.
807 For objects with non-8.3 names, the 'longname' memory
808 allocation would be leaked during object deletion or
811 * Fixed a memory leak in the background daemon cm_CheckServer()
812 routine. The memory allocated to store the server list
815 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
817 * Correct an abstraction layer violation. cm_scache_t objects
818 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
819 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
820 performed in cm_ReleaseSCache() when the reference count hits zero.
822 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
824 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
827 * Ensure that scache object 'nextp' pointers are set to NULL
828 when the objects are removed from the hash table linked list.
830 * If a dirty buffer cannot be written to the file server because
831 of an access denied error, mark the error in the cm_buf_t
832 structure so that it can be dealt with instead of retrying
836 * Update CellServDB file to GCO Public 23 Apr 2008
838 * Fix a cm_buf_t reference count leak when attempts to write
839 dirty buffers to the file server from within cm_IncrSyncer()
842 * Prevent udebug from crashing.
844 * Activate RX Idle Timeouts. If the file server is busy for more
845 then 30 seconds, attempt to failover to another server without
846 marking the busy server down.
848 * Another VNOVNODE issue fixed. When writing a dirty buffer
849 to the file server, if VNOVNODE is received, mark all buffers
850 as invalid without further attempts to contact the file server.
852 * Improved performance on high latency links.
855 * Fix a bug preventing the re-initialization of the Freelance
856 root.cell contents when a change is made.
858 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
859 or mountpoint that does not exist.
861 * aklog, afscreds, integrated logon, and the NIM AFS credential
862 provider will all attempt to first acquire an afs service ticket
863 using the service principal: afs/<cell>@<CLIENT_REALM>. This
864 permits the proper detection of the cell realm when multiple
865 realms are local realms for the cell.
867 * Fix fs commands that re-use file identifiers instead of forcing
868 a path evaluation each time. This was broken in 1.5.33 when
869 hashes were added to the cm_fid_t structure.
871 * Fix symlink and mountpoint removal and then recreation when
872 applied to the fake Freelance root.afs volume.
874 * Fix SMB change notifications for the Freelance root.afs volume
875 when creating or removing symlinks and mount points.
877 * Force DNS AFSDB cell name lookups to lower case if the DNS
878 response is not all lower case.
880 * 32-bit installers are now built with Visual Studio 2005
882 * All vos.exe commands now support the "-noresolve" option which
883 forces the display of IP addresses instead of DNS hostnames.
885 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
886 permits operations that do not require up to date volume
887 location data to continue while waiting for the RPC to complete.
888 Multiple threads waiting for a volume location update on the
889 same volume set no longer trigger multiple calls VL_xxx RPCs.
891 * Volume server references could be marked offline and never
892 reset to the not_busy state.
894 * Further optimizations to cm_Analyze() retry logic for
895 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
898 * Optimize the DNLC by applying Interlocked incrementing for
899 statistics gathering and enabling greater use of read
900 locks instead of write locks
902 * Further optimize osi_Log operations when logging is disabled
903 by removing the need for the function call overhead.
905 * Add a new lock operation, Convert Read to Write, and apply
906 it throughout the cache manager. This function provides a
907 fast transition from a read lock to a write lock when the
908 caller is the only reader.
910 * Fix a timeout problem when opening files or request locks
911 when the file in question is already write-locked by another
914 * Fix a deadlock that can occur if "fs flushall" is executed
915 while applications are writing to AFS.
917 * The cache manager would re-read data from the file server
918 after file truncation when obtaining buffers to write to the
919 same file until the first StoreData RPC completes.
921 * Fixes error handling in cm_NTCheckDelete()
923 * Replaces cm_scache_t mutex with a read-write lock permitting
924 additional parallel access.
926 * Prevent an cm_scache_t reference undercount when evaluating
927 symlinks containing @sys when none of the active sysnames
928 match any of the available targets.
930 * Prevent the leak of a cm_scache_t rwlock if associated symlink
933 * Set RxMaxMTU registry value to 0 which is the equivalent of
934 it not being set at all.
936 * Reduce contention between BkgDaemon threads
938 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
941 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
942 objects to use the faster Interlocked operations which avoid
943 the need to hold write locks.
945 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
946 non-fatal error that forces a retry against another service
949 * Fix a crash in smb_WriteData that would occur if an application
950 opened the file as read-only and then sent a write request anyway.
953 * Respond to VL_NOENT errors by removing cm_volume_t object and
954 setting it to be the first object to be recycled. This is not
955 only a more efficient use of resources but it also prevents
956 repeated attempts to query the VLDB server for the volume from
957 the CheckOfflineVolumes routine.
959 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
962 * A first cut at a cache manager statistics monitor that can be used
963 to determine the necessary cache parameters to support the
964 application working set.
966 Off by default, the performance package can be activated by setting
967 "daemonPerformanceTuningInterval" in the service Parameters key.
968 As with the other daemon interval values the unit is in seconds.
970 At service start and each succeeding interval the cache manager
971 will write statistics to %TEMP%\afsd_performance.txt showing the
972 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
973 The FID statistics keep track of all FIDs seen by the cache manager
974 during the service session whether or not they are backed by any
975 live object in the cache.
977 These statistics are not stored in the cache file.
979 * Fix windows event logging broken in 1.5.33.
981 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
982 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
983 as well as when computing the server preference list.
985 * When merging status after a StoreData operation prevent data
986 buffers from being discarded that we would prefer to keep.
987 (Bug introduced in 1.5.33.)
989 * Modify installers the set RxMaxMTU to 0.
991 * In the AFS Configuration Panel on the Advanced->Miscellaneous
992 dialog, the Background Daemons and Service Threads fields were
995 * Increase default number of Background Daemons to 4.
997 * Remove cm_volume_t reference from the cm_scache_t object. Permit
998 cm_volume_t objects to be recycled.
1000 * Windows Error Reporting has shown a number of previously unseen
1001 stack corruption errors in the Explorer Shell extension. The
1002 minidump appeared to indicate the problem was in ParseAcl().
1003 Reviewing the function it appears that numerous things could go
1004 wrong if invalid input was provided. Added error checking and
1005 data validation. Hopefully this will address the problem.
1007 * Windows Error Reporting has shown crashes in aklog as a result
1008 of krb5_cc_get_default() failures not being handled. They are
1012 * The Rx library used a 32-bit type for sockets which was
1013 truncating the socket value on 64-bit Windows. This
1014 was not an issue in general but became an issue when
1015 Rx "hot thread" support was enabled.
1017 * Enable "Rx hot thread" support which ensures that while
1018 a received message is being processed that a new thread
1019 is allocated to receive the next message to arrive.
1021 * Increased parallel access to smb_rctLock protected
1022 resources. Reduce the number of times that the
1023 lock is dropped and re-obtained.
1025 * Increased parallel access to cm_scache_t and cm_buf_t
1026 resources by replacing the previous global write lock
1027 protected reference counting mechanism with one built
1028 upon the use of InterlockedIncrement and InterlockedDecrement.
1030 * Implement a hash for cm_fid_t to reduce comparison costs
1031 and inline the cm_FidCmp() function.
1033 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1034 hash which has a better distribution
1036 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1037 the most recent read request instead of the next chunkSize.
1038 cm_chunkSize can be dozens or hundreds of buffers. As a result
1039 too much time is spent performing the evaluation.
1041 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1042 lock is to prevent the creation of new buffers while a truncation
1043 is being performed. All references to bufCreateLock have been
1044 removed except in two places:
1045 i. a write-lock surrounding the function that calls buf_Truncate()
1046 ii. a read-lock within buf_GetNewLocked() that actually allocates
1049 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1050 CM_CONFIG_DATA_VERSION number which value be used to force the
1051 replacement of the cache file contents when incompatible changes
1052 are made between releases.
1054 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1055 cm_SyncOpCheckContinue
1057 * Avoid calls to multi_Rx if nconns == 0
1059 * Modify smb_WriteData to perform background writes based upon the
1060 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1061 boundaries. This will slow down writes from the SMB interface but
1062 will avoid the risk of the CIFS client disconnecting from the AFS
1063 client SMB server. The default value for 'smb_AsyncStoreSize'
1066 * SMB AsyncStore functionality is on by default but can be
1067 disabled using the registry value:
1069 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1071 smb_AsyncStoreSize can be adjusted with
1073 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1075 The value must be a multiple of the buffer block size
1076 and cannot be larger than the chunk size.
1078 * Do not leak a read lock on the bufCreateLock
1080 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1083 * In MergeStatus, only remove clean buffers from the hash table.
1085 * Avoid the need to update the data version number on each buffer
1086 associated with a scache when MergeStatus is called after a
1087 StoreData by maintaining a range of valid data versions as part
1088 of the cm_scache_t object.
1090 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1091 from cm_HaveAccess... We can safely avoid the call so do so.
1093 * Add "ForceAnonVLDB" service parameters registry value. Default
1094 is 0. Set to 1 to force the use of anonymous connections to
1095 the VLDB servers. This is useful when using krb5 tokens and the
1096 VLDB servers have not been updated.
1099 * Do not attempt to search for a cell name that matches one
1100 of the special share names "svrsvc", "wkssvc", "ipc$"
1102 * Issue permission denied warnings from afscreds if attempts
1103 to modify the startup shortcut or registry values fail.
1104 This happens under Vista UAC.
1106 * Do not return BAD_NETWORK_ERROR in response to attempts
1107 to create directories or files in the Freelance root.afs
1108 volume. Instead return ACCESS_DENIED.
1110 BAD_NETWORK_ERROR was being sent because the Freelance cell
1111 has no vldb servers. Short circuit the error by testing for
1112 freelance in CreateFile and MakeDir
1114 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1115 to be opened and closed without actually existing. For now
1116 we treat them as equivalent to ioctls but this should
1119 * Fix RAP processing to not reject names that exist in the
1120 root.afs volume for GetInfo queries.
1122 * In afscreds, use GetUserNameEx to obtain the username instead
1123 of searching the registry. Try UserPrincipalName first and
1124 fallback to SamCompatible.
1126 * Stop linking to rpcns4.dll which is no longer supported on
1127 Vista and 2008. TaAfsAdmSvr is going to have to be
1130 * When processing dfs referral requests, if the sharename is
1131 a partial match and therefore will not be added automatically
1132 to the share list, make sure that we return an error.
1134 * Avoid a crash when logging VNOVOL errors returned from
1135 cm_CheckOfflineVolume.
1137 * When trace logging is disabled, do not permit osi_LogSaveString
1138 to perform the string copying. By disabling the string
1139 copies, throughput of the service more than doubles.
1141 * when checking offline volumes, update the cell vlserver info
1142 if required and don't perform a check if there are no servers
1145 * Add volumes in the vl_unknown state to the list of objects to be
1146 checked as offline or retried when an allbusy error is received
1149 * Add support for configuring rx_ExtraPackets
1151 * All calls to buf_Get() must be protected by a read lock on the
1152 cm_scache_t bufCreateLock
1154 * When a MergeStatus results in the invalidation of buffers, do
1155 not remove buffers from the hash tables that have active references.
1157 * Upon receiving a VNOVNODE error from a file server, be sure to
1158 invalidate any file locks. This prevents on-going attempts to
1159 extend the locks and will cause subsequent attempts to access
1160 the file to receive a invalid handle error.
1162 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1163 to directory entry names.
1165 * No longer discard tokens when RXKADBADTICKET error is received.
1166 Instead return STATUS_NO_KERB_KEY error.
1169 * Add "cmdebug <host> -cellservdb" option and client service
1170 implementation permits querying the vldb server for each
1171 cell known to the cache manager. The output is in a form
1172 that permits it to be used as a CellServDB file.
1174 * Add Symlinks->Show option to the AFS Explorer Shell Context
1177 * Prevent buffer overrun in the cmd parser used by command
1178 line tools such as fs, vos, pts, etc.
1180 * The path ioctl operations have several issues:
1182 (1) the specified path for "fs examine, whereis, whichcell,
1183 flush" and so always has the follow symlinks and mount
1184 points semantics. This makes it impossible to determine
1185 what the FID of a symlink or mount point is.
1187 (2) "fs examine" out is not the result of a single pioctl
1188 operation but is actually the combined output of half
1189 a dozen operations. Path evaluation is an expensive
1190 operation. It would be faster if the caller could
1191 evaluate the FID first and then perform all of the
1192 rest of the operations by specifying the FID instead
1195 (3) fs output reports all objects as files. By adding a
1196 GetFileType pioctl more informative output can be provided
1197 that indicates what type of object the path evaluates to.
1199 (4) the Windows fs command includes a number of commands that
1200 do nothing but exist only because the Unix cache manager
1203 A new extendible data structure cm_ioctl_query_opts_t has been
1204 added which can be optionally specified with pioctls that do not
1205 already require input data. The first two fields of this structure
1206 are 'literal' and 'fid'. The literal field is used to indicate
1207 whether the last component of the path should be evaluated following
1208 symlinks and mount points.
1210 The fid field permits a fid to be specified.
1212 * A new GetFileType pioctl has been added. The type of objects are
1215 * A new "-literal" option is available for "fs examine, flush, whereis,
1218 * Unimplemented fs commands have been removed.
1220 * There are circumstances where a volume object is being accessed and
1221 the volume is marked indicating that the volume location information
1222 is out of date but where it is also pointless and perhaps dangerous
1223 to block waiting for the rpc to complete. One example is when
1224 processing the cmdebug requests. If we know that we are not going
1225 to use the volume object to contact a server, then we can now set
1226 the CM_GETVOL_FLAG_NO_RESET flag.
1228 * when moving up in the directory tree we search the recorded fid list
1229 to find a matching fid that we have already crossed. we must also
1230 reset the fid count based upon what we discover.
1232 * Cell names are published as share names. As a result they are searched
1233 for as part of the Dfs Referral evaluation. If share "foo" can not be
1234 reached, the CIFS client will fallback to searching for "fo" as well.
1236 Since the freelance client automatically adds a symlink for prefixes
1237 of the cell name, this results in both "foo" and "fo" being added to
1238 the freelance root.afs volume. This patch prevents that negative side
1241 * when tracking Kerberos credential cache names be sure to include
1242 the cache type prefix.
1244 * Microsoft has assigned OpenAFS a network type value
1246 * The algorithms that were used to produce a Netbios name were broken.
1247 The name that was produced when the hostname was too long was in fact
1248 longer than the maximum permitted netbios name.
1250 Also, the "NetbiosName" value was not used as a suffix when the loopback
1251 adapter was not installed. The hardcoded string "AFS" was used instead.
1253 * avoid another deadlock during server probes initiated by the ipaddr
1254 change daemon thread.
1256 * prevent the afs client service from crashing when fs newcell is
1259 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1260 to \Program Files\OpenAFS\Client\Program in order to get them out
1261 of the %windir% tree
1263 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1264 end STATUS_OUT_OF_QUOTA can be returned to the application.
1266 * Modify the search order for determining the location of CellServDB
1267 and other client configuration files.
1269 1. AFSCONF environment variable
1271 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1272 4. use Program Files\OpenAFS\Client
1274 * More Vista Power Management tweaks
1276 * Close a socket handle leak in the kauth routines
1278 * Don't hold a cm_volume_t mutex across RPCs
1280 * Convert downTimes to use clock time instead of relative times.
1281 This permits correct comparisons with expiration times.
1283 * Update the Wix installer to properly install the prerequisites
1284 for translate_et.exe
1286 * Do not translate pioctl paths as they are not converted by the
1289 * Implement multi_rx probes for checking whether servers are up
1290 or down. Turns probes into a constant time activity instead
1291 of one dependent upon the number of known servers.
1293 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1294 update of the volume location data and retry. VNOVOL means
1295 that the volume could not be attached. It doesn't mean that
1296 the volume is not on this server.
1298 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1299 not attempt to validate the target path. There is no reliable
1302 * Improve performance of \\afs\<cell><mount-type><volume>\
1303 access by eliminating extraneous attempts to search the
1304 CellServDB file and DNS AFSDB records. Share names
1305 containing '%' or '#' are not valid cell names.
1307 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1308 to unlock a range that is not currently locked.
1310 * Enable the 32-bit tools to read the 64-bit registry keys
1311 when executing in the WOW64 environment.
1315 * Fix buffer refcount leak introduced in 1.5.29
1317 * Prevent the NIM AFS Provider from crashing when configured to use
1318 Kerberos v4 for token acquisition in situations where no Kerberos v4
1319 support is available. For example, 64-bit KFW or sites that delete
1320 the Kerberos v4 DLL.
1322 * In afsd_init.log, support logging cache sizes greater than 2GB.
1324 * The volume status output of "fs examine" was being determined based
1325 upon the wrong error value. Instead of using the pioctl() return
1326 code, the errno value must be used.
1328 * Prevent invalid pts auto-registration attempts by aklog and the
1329 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1331 * Add synchronization protection to all of the SMB Listener State
1332 variables. This prevents race conditions when the addition or
1333 loss of a network adapter takes place.
1335 * On Vista, there is a race condition between the restart of the
1336 afs client service and the network adapters. Ensure that the
1337 afs client service can handle switching between loopback and
1338 non-loopback modes of operation.
1340 * Fix FollowBackupPath to work with mount points that are not
1341 explicitly read/write.
1343 * Fix a deadlock that has been observed on Vista while resuming
1344 after at least four hours of sleep.
1346 * Add new HKLM registry configuration options that can be used
1347 to provide configuration hints to the NIM AFS Provider when
1348 creating new identities. See release notes for details on
1349 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1352 * Speed up write buffer operations by holding a global read-lock
1353 instead of a write-lock when updating the buffer data version
1354 number. The data version is protect by the associated stat
1357 * Speed up read buffer operations by removing out of date buffers
1358 from the buffer hash tables.
1360 * Protect against a null pointer dereference in afslogon.dll
1361 if QueryAdHomePathFromSid() fails.
1363 * Remove a race condition while accessing private freelance data
1366 * Update the buffer data versions when Freelance directory data
1369 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1372 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1373 instead of returning zero length.
1375 * Return ERROR_RETRY when a server sends VBUSY.
1377 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1378 a volstat plug-in is in use.
1380 * Enhance the volstat interface to support Dfs path mappings.
1382 * Fix "symlink list" to do the right thing for Dfs links
1384 * Fix callback expiration that was broken in 1.5.27.
1386 * Fix a crash at shutdown caused by the lack of synchronization
1387 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1389 * Hold the cm_server_t mutex while walking or modifying the
1390 cm_server_vols_t list in order to avoid a race condition.
1392 * Ensure that the cm_server_t refCount is safely modified.
1394 * The pthread library implementations of pthread_cond_wait
1395 and pthread_cond_timedwait were failing to return with
1396 the mutex held when error conditions occurred. This could
1397 result in race conditions wherever they are used. One case
1398 in which problems were seen was the Rx event_handler().
1400 * Prevent cm_GetBuffer() from locking all buffers associated with
1401 a stat cache object before deciding if the current buffer is up
1404 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1405 object for each offset that we will attempt to reserve.
1407 * Reduce the default number of background daemon threads used to
1408 prefetch or store to one in order to prevent two or more threads
1409 from stepping on each other.
1411 * Reduce the default chunk size to 256KB in order to prevent read
1412 timeouts on low bandwidth connections.
1414 * Fix a race condition that could result in threads becoming stranded
1415 when synchronizing multiple requests on the same stat cache object.
1418 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1419 has been disabled. It has been disabled because it turns out
1420 that the RPC is improperly implemented in all file servers
1421 previous to 1.4.6. The RPC is executed without the proper
1422 locks being held resulting in data corruption and eventual file
1423 server failure. The functionality will be re-enabled in a
1426 * a couple of .readonly Volume Callback optimizations were
1428 - Apply most recent volume callback to all stat cache objects
1429 in the volume. This avoids premature callback expirations
1430 and unnecessary FetchStatus calls.
1431 - Add option to permit attempts to renew .readonly Volume
1432 Callbacks on a periodic basis to avoid invalidation of stat
1433 cache data. [Requires daemonCheckVolCBInterval be set to
1434 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1436 * Adds support for 64-bit data version values. Previous releases
1437 only supported 32-bit values. Larger values sent by the file server
1438 would result in rollover.
1440 * aklog and asetkey no longer generate an exception if Kerberos for
1441 Windows is not installed.
1443 * If the LSA authentication functions report a temporary out of memory
1444 condition, fail the current authentication and do not cause the AFS
1445 client service to terminate unexpectedly.
1447 * When the AFS client service is shutting down, prevent the daemon check
1448 thread from executing additional checks which might attempt to access
1449 discarded locks or unmapped memory.
1451 * Correct a double free condition in the NIM AFS credentials provider.
1453 * Add "FollowBackupPath" service parameter registry value.
1454 This provides equivalent functionality to the UNIX afsd -backuptree
1455 option. When set to a non-zero value, normal mount points originating
1456 in a .backup volume will prefer a .backup volume.
1459 * Fix NSIS uninstall from the .MSI installer.
1461 * Add Stack Frame logging for AMD64
1463 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1464 server returns EWOULDBLOCK, retry the request every two
1465 seconds for up to the RDRtimeout.
1467 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1468 is not assigned the same value as EIO.
1470 * Update CellServDB to GCO Public 25 Oct 2007
1472 * Add additional validation and error handling code after
1473 each call to getSlot(). If an invalid slot is returned,
1474 return NONODE. If the invalid slot is returned when
1475 extracting a data node, invalidate the tree.
1477 * Modify compareKeys() to always perform a case-insensitive
1478 comparison and only perform a case sensistive comparison
1479 if the case-insensitive one matches. This ensures the
1480 ordering is consistently reported.
1482 * Add lock assertions to ensure that all calls are being
1483 performed with the correct locks being held. There have been
1484 some crash reports that provide stack data that does not appear
1485 to be possible unless there is a race. However, there are no
1486 obvious locations where the race is taking place and the test
1487 suite indicates that all of the correct locks are being held.
1488 We shall see what happens in the field.
1490 * For consistency replace all calls to findKey in which the range i
1491 s (1,numentries) with calls to getSlot().
1493 * Optimize the depth search loop by testing the slot value in the
1494 for statement instead of forcing the loop to be broken later.
1496 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1497 There are a number of locations where locks are obtained, dropped,
1498 and reobtained. This reorganization attempts to accomplish several
1501 (1) be optimistic for the most common case so it will be fast
1503 (2) add consistency checks after each location where locks are
1504 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1505 and the bplus tree is out of date, retry until we get to a
1506 consistent state that we can use.
1508 (3) Ensure that all operations take place with the correct locks.
1510 * One of the issues that has become a serious problem since the
1511 addition of the local directory updates is that although cm_SyncOp
1512 synchronizes operations, it does not preserve the order of requests.
1513 This has always been a problem in that it has been possible for a
1514 request to fail to complete due to its worker thread's bad luck.
1515 When a request takes longer than the Windows SMB Redirector's timeout,
1516 the SMB Redirector tears down the SMB virtual circuit.
1518 When using the local directory updates it is really important that
1519 the directory update operations complete in the order that they were
1520 sent to the file server. If they don't, then the local directory state
1521 and the file server state will not match and the local directory state
1522 must be discarded which in turn forces a new read of the entire
1523 directory contents over the network.
1525 This patch adds a new cm_scache_waiter_t object that is used to store
1526 the current thread, buffer, and syncop flags within a waiters queue
1527 on each cm_scache_t object. If a thread is forced to sleep in
1528 cm_SyncOp, upon waking it will check to see if there are any other
1529 threads waiting that are attempting to perform a similar task ahead
1530 of it in the queue. If yes, the thread goes back to sleep. If not,
1531 it goes ahead and enters the cm_SyncOp conflict resolution block.
1533 This patch has the additional side effect of reducing the number of
1534 competing threads that must obtain the cm_scache_t mutex and process
1535 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1536 utilization of the service and the clock time associated with processing
1537 requests will be reduced.
1539 * assert that the cm_scache_t mutex is held when calling
1540 buf_ForceDataVersion().
1542 * rename findNode to leafNode in bplus_Lookup
1544 * replace all OutputDebugString calls with osi_LogX calls
1546 * modify bestMatch to special case the return values for leaf nodes.
1547 If an entry is above or below the values available in the leaf node
1548 return BTLOWER or BTUPPER instead of BTERROR.
1550 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1551 if true convert to either slot 0 or Max and perform the insertion.
1552 This produces easier to read code when performing lookups.
1554 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1555 descriptive message.
1557 * the difference between osi_assert() and osi_assertx() is whether or not
1558 a description is specified as part of the call. When no message is
1559 specified afsd_notifier() is called with a NULL msgp parameter. This
1560 results in a null pointer dereference during the EventLog() call.
1562 As a result, none of the information describing the assertion is written
1565 This commit sets a default message for use when no other message is
1566 provided and it replaces all calls to osi_assert with osi_assertx and
1569 * Migrate B+ tree search key into thread local storage
1572 * Fix an uninitialized variable in lana_IsLoopback() which
1573 could result in a random determination that a loopback
1574 adapter is not a loopback adapter. This would result in
1575 a computed Netbios service name of %MACHINE%-AFS instead
1576 of AFS. This is a serious problem if the value computed
1577 by the AFS Client Service and the client process pioctl
1578 calls produce different results as it will result in a
1579 communication failure.
1581 * Fix a crash in the B+ Tree code that results when a file
1582 rename only results in a change of case.
1584 * Fix the pioctl path parsing. The previous algorithm
1585 broke the current directory of the client into two parts,
1586 the TIDpath which represented the device root and the
1587 relative path. The problem is that this is not sufficient
1588 from proper processing of symlinks and mount points.
1590 What is required is that the last component of the path
1591 must not have the symlink or mount point be evaluated
1592 whereas symlink and mount point evaluation must be performed
1593 for the rest of the relative path. Therefore, the path is
1594 now split into three parts: TIDpath, intermediate path,
1597 "fs examine" now reports the data of the symlink or mount
1598 point object instead of the target and "symlink list" and
1599 "symlink remove" now works. The "Explorer Shell Extension"
1600 menu now is displayed for symlink objects.
1602 * More informative symlink.exe error messages.
1604 * When the cmd.exe "del" command is used to remove a symlink,
1605 not only was the link object removed, but the stat cache
1606 object for the target was marked deleted. This resulted
1607 in cache corruption which would require the destruction
1608 of the cache to correct.
1610 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1611 which results in cached contents from read only volumes
1612 to be treated as valid when there is no callback registered
1613 and all servers are down or all instances are offline.
1615 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1617 * Add documented registry value, "AcceptDottedPrincipalNames"
1618 which is used to permit the acquisition of AFS tokens
1619 using Kerberos v5 principal names which include a dot in
1620 the first component.
1622 {HKLM\HKCU}\Software\OpenAFS\Client
1624 * "fs uuid" does not require "AFS Client Administrator" group
1625 membership to display the current UUID value.
1628 * Update NSIS installer to version 2.30
1630 * Update Wix installer to version 2.0.5325
1632 * 64-bit KFW support:
1633 - Network Identity Manager plug-in
1635 - 64-bit krb5 enable afscreds.exe, afs server manager,
1638 * Enable B+ tree Directory Search by default.
1640 * NIM plug-in is now compatible with FILE: ccaches
1642 * Install instloop.exe into Client/Programs/ even though
1643 it isn't required. Permit it to be used by organizations
1644 that clone Windows systems.
1646 * On a suspend, only drop callbacks for servers that
1647 are online. If the server is offline, maintain the callback
1648 until the server is once again available.
1650 * Prevent against acquisition of a directory write lock when
1651 there are no updates to be performed. This was resulting
1652 in a write-lock leak which later on would lead to a deadlock.
1654 * Add configurable file pre-fetch mode. For files that match
1655 the specified file extensions, pre-fetch the entire file into
1658 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1659 (MULTI_SZ) PrefetchExecutableExtensions
1661 * Fix the client_osi large integer support so that return values
1662 can be properly assigned to large integer variables.
1664 * Do not hold cm_scache_t mutex across cm_GetCell() calls
1665 since cm_GetCell() can block in a RPC.
1667 * Add short file name support to B+ tree searches.
1669 * Resolve a conflict between the desire to notify file servers
1670 that the client workstation is suspending and the desire to
1671 permit objects with valid callbacks when the client suspends
1672 to continue to use those objects from the cache if the client
1673 resumes without network connectivity.
1675 * Fix stack variables to ensure that all buffers storing paths
1676 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
1677 would cause a stack overflow condition.
1679 * On Vista, force afs_config.exe to run with the highest privileges
1680 available to the end user. For a user in the Administrators
1681 group running under UAC, this will cause afs_config.exe to be
1682 be run with full Administrators privileges after the user
1683 approves the privilege escalation.
1685 * More VNOVNODE handling issues. If a dirty buffer cannot be
1686 written due to VNOVNODE, clear the buffer and ignore the error.
1687 If a cm_scache_t object has been marked deleted due to a VNOVNODE
1688 error, the next time the smb_fid_t object that references it is
1689 used return either NOSUCHFILE or NOSUCHPATH and close the
1690 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
1691 result in an invalid handle error.
1693 * For Vista, set the RequestExecutionLevel of the NSIS installer
1696 * For the build environment, support the generation of manifests
1697 when using the CL1310 (VS .NET 2003) compiler/linker.
1699 * There was a bug in OutputDebugHexString() which could overflow
1700 a stack buffer resulting in unexpected process termination
1701 of afsd_service.exe if the input Security Token contained
1702 many bytes equivalent to the '%' char.
1704 * Add directory lookup statistics to "fs memdump" output.
1708 * Windows uses case-insensitive file name pattern matching
1709 but AFS is a case sensitive file system. The AFS3 directory
1710 format is block based, uses network byte order and
1711 includes a hash table for fast case sensitive lookups.
1712 This causes several problems for the Windows AFS client.
1713 (1) Traversing the directory blocks is cpu expensive
1714 (2) A hash table miss does not indicate that the desired
1715 entry does not exist.
1716 (3) Determining whether a non-ambiguous inexact match or
1717 the entry does not exist requires a linear traversal
1718 of the entire directory.
1719 These issues often result in 100% CPU utilization.
1721 These issues are addressed by building a modified B+ tree for
1722 each directory and then using the B+ tree for searches.
1724 This feature is disabled by default. To enable it set
1725 the "BPlusTrees" registry value to 0x01 in the AFS Client
1726 Service Parameters key.
1728 * Correct a null pointer dereference in the symlink recursion
1729 detection code. (Introduced in 1.5.22)
1731 * Add the Network Identity Manager AFS Provider to AMD64
1736 * A new registry value, "BlockSize", can now be used to adjust
1737 the size of the internal buffers used to store file data.
1738 The default is 4KB. When set, the values 1..1024 specify
1739 multiples of 4KB. Values > 1024 specify actual block sizes
1740 and must be multiples of 4KB. The actual block size in KB
1741 can not be larger than the chunk size which defaults to 1MB.
1743 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1746 * pts.exe and ptserver.exe built with -DSUPERGROUPS
1748 * Service Manager shutdown hints have been added. The hint
1749 specifies that it may take the AFS client service up to two
1750 minutes to shutdown.
1752 * Fix the recursion detection feature introduced in 1.5.22.
1753 The previous implementation broke the evaluation of paths
1754 using ../path or ./path. It also broke the evaluation of
1755 symlinks to absolute paths. For example, /afs/cell/path.
1757 * Fix an uninitialized variable in cm_MakeDir that could result
1758 in crash if the requested directory could not be created.
1760 * Fix a the VolumeStatusNotification log messages to save
1761 the input string in the circular log buffer so garbarge
1762 won't be output to afsd.log when the log is dumped.
1765 * Increase the SMB Maximum Path length to 260 characters
1766 from 256 as that is what CIFS supports.
1768 * On 64-bit Windows, use a unique description string for
1769 the 32-bit AFS Shell Extension.
1773 * In the AFS Admin Library, replaced all of the
1774 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
1775 allows parameter checking and proper type conversions to be
1776 performed. This is crucial when passing parameters that are
1777 64-bit wide on systems in which 'long' is a 32-bit value.
1778 Otherwise, the parameter list on the stack is interpretted
1779 incorrectly and the function not only doesn't succeed but it
1780 might corrupt memory and if lucky crash the application.
1782 The only program that we care about that uses this library
1783 is the AFS Server Manager. Now it doesn't crash.
1785 * Fixed smb directory searches for non-wildcard names. The
1786 optimized function should not have set the bulk in progress
1787 flag. This would adversely affect later directory searches
1788 that required a bulk search.
1790 * Detect infinite recursions (up to 512 stat cache objects
1791 in a chain.) If an infinite recursion is detected, return
1792 an error to stop the CIFS client from querying us forever.
1794 * Executables executed out of AFS name space can have their
1795 pages swapped out by Windows. Those pages will be reloaded
1796 when they are required. However, if the cache manager no
1797 longer has a callback and all of the file servers with
1798 the volume data are down, then the application will crash
1799 due to the inability to read the required page.
1801 We now track the down time of each server. If the callback
1802 expires after the server that issued it went down and all
1803 servers are down, we delay the expiration to permit the data
1804 in the cache to be continue to be used by Windows or an
1807 * The Windows AFS client has always had very poor performance
1808 with regards to directory management. In the case of
1809 directory updates caused by create, rename,and delete operations,
1810 the AFS cache manager will now perform local directory updates
1811 when possible in order to avoid re-reading the directory from
1812 the file server. This will be done in whenever both of the
1813 following conditions are true:
1815 - the data version on the directory increased by one as a
1816 result of the operation
1818 - all of the directory buffers necessary to make the change
1819 are present in the cache
1821 * The error message in the AFS Control Panel describing the
1822 need to be a Windows Administrator has been replaced with
1823 a message describing the need to be in the AFS Admins Group.
1825 * If the vldb server list is empty when attempting to evaluate
1826 a cell mountpoint it was possible for a crash to occur. This
1827 can only happen if there is an existing mount point and the
1828 is a failure attempting to load the server list from the
1829 CellServDB file or via AFSDB queries.
1831 * dirty buffers only have their dirty bytes stored to the file
1832 server instead of the entire chunksize
1834 * the default chunksize has been increased from 128KB to 1MB
1839 * changed the enum values for cm_serverRef_t state info to use a
1840 private name space to avoid collisions (srv_)
1842 * added a srv_deleted state for cm_serverRef_t objects. This
1843 state is set when cm_FreeServerList() is called with the
1844 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
1845 not always delete the cm_serverRef_t from the list if it is
1846 still in use by another thread. the srv_deleted state means
1847 the object's contents are no longer valid and it must be
1848 skipped. It will be deleted the next time the object is
1849 freed and the refcount hits zero.
1851 * the srv_deleted state is also used when a file server reports
1852 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
1853 as offline. This is done to prevent additional usage of the
1854 stale vldb data while waiting for the update volume request
1857 * added a state field to the cm_volume_t object (enum volstate
1858 vl_ name space) that maintains the state of the volume based
1859 upon the states of all of the cm_serverRef_t and cm_server_t
1862 * modified cm_UpdateVolume() to set the state of the cm_volume_t
1863 RW, RO, and BK to either vl_alldown or vl_online. There can't
1864 be any other states because cm_UpdateVolume() destroys any
1865 previous knowledge we might have had regarding busy or offline
1868 * modified cm_UpdateVolume() to update the volume name in the
1869 cm_volume_t to the volume base name if the previous value was
1872 * modified cm_FollowMountPoint() to check to see if the volume
1873 name is a volume ID and if so call cm_GetVolumeByID instead
1874 of cm_GetVolumeByName. This ensures that volume IDs are always
1875 looked up as numeric values. There is no longer a need to
1876 maintain a separate cm_volume_t containing the string representation
1879 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
1880 The first flag is a "CREATE" flag which is set by all existing
1881 calls. The flag is not set by calls to cm_GetVolumeByID() from
1882 the server probe code when volume status is being updated. We
1883 do not want the server probe operation to result in additional
1884 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
1885 which is set when the server probe code updates the volume status.
1886 This flag will be used to prevent the server probe operation from
1887 changing the order of the least recently used queue.
1889 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
1890 allocated for a given set of normal, readonly, and backup volumes
1891 regardless of whether or not the volume is accessed via name or
1892 ID number. The cm_volume_t namep field is always the base name
1895 * Added a new volume state, vl_unknown. This state is used as
1896 the initial state for all cm_volume_t when the cache manager starts,
1897 for each cm_volume_t at creation, and for each cm_volume_t when
1898 recycling. The cache manager does not know the state of all
1899 volumes in the world, only those that are in the cache and for
1900 which it has queried the VLDB and hosting file servers.
1902 * modified cm_GetVolumeByName() to initialize the state of a
1903 volume to vl_unknown. The actual state will be set when a
1904 cm_VolumeUpdate() call completes successfully.
1906 * changed name of scache hash table variables to avoid ambiguity
1907 when adding hash tables for volumes
1909 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
1911 * modified cm_UpdateVolume() to handle the case in which there is
1912 no RW volume but there is are RO volumes for a given base name.
1913 This is done by querying for the ".readonly" volume name if the
1914 base name does not exist in the VLDB. We never query for the
1915 .backup name because under the current usage model a .backup
1916 volume may only exist on the server that the read-write volume
1917 is located. If there is no RW volume, there can be no .backup.
1919 * Added four hash tables for cm_volume_t objects to improve the
1920 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
1921 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
1922 BKID. Three ID hash tables are necessary as long as it is
1923 desireable to maintain a single cm_volume_t containing all
1924 of the related RW, RO, and BK volume data. Having the RW and
1925 RO volume data in the same object is necessary for the
1926 implementation of cm_GetROVolumeID() which returns either the
1927 RO or RW ID depending upon the existence of RO volume instances.
1929 * Added a volume LRU queue so that volume reuse becomes fairer.
1930 This does not replace the all Volumes list which is used when
1931 it is desireable to walk a list of all the volumes whose order
1932 is not going to change out from underneath you which makes it
1933 safe to drop the cm_volumeLock.
1935 * handles volume hash table updates where volume name to
1936 volume ID number changes. The volume name remains
1937 constant in the cm_volume_t. if a vos rename is performed,
1938 the name of the volume will change and the volume IDs will be
1939 updated. Subsequent access to the old volume ID will create a
1940 new cm_volume_t with the new name.
1942 * Added a daemon thread operation to query the state of volumes
1943 listed as busy or offline. cm_CheckBusyVolumes() calls
1944 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
1945 or vl_offline. If the volume is now online, the status on the
1946 volume is updated. The default period is 600 seconds. This can
1947 be configured with the BusyVolumeCheckInterval registry value.
1949 * Added prototype for smb_IoctlPrepareRead() which was missing a
1950 return type in the function definition.
1952 * Added volume id lists to the cm_server_t. These lists are
1953 allocated in blocks of ~32 IDs. When a cm_PingServer()
1954 detects a change in server state, the state of the cm_volume_t
1957 * Added volID to the cm_serverRef_t object. volID is used
1958 to identify the volume for which the object is a referral.
1959 cm_FreeServerList() uses the volID to remove the volume
1960 from the cm_server_t.
1962 * In cm_Analyze, when VNOVOL or VMOVED are received,
1963 call cm_ForceVolumeUpdate() to force a refresh of the volume
1966 * Added cm_VolumeStatusNotification() which is used at the moment
1967 to log volume status changes to the trace log. It will also
1968 be used as the access point to the File System Filter driver
1969 notification engine.
1971 * Added an all cm_scache_t list to cm_data. This replaces the use
1972 of the stat cache LRU queue when we need to enumerate all
1973 entries. The LRU list order is not static and when using it to
1974 enumerate all entries it can result in items being missed or
1975 items being processed more than once.
1977 * Modified cm_Analyze(). Instead of reseting the busy or offline
1978 state of a volume and forcing a retry of the operation
1979 cm_Analyze will defer to the background daemon thread that will
1980 update the state once every 600 seconds.
1982 * Added the automatic generation of a Freelance ".root" read-write
1983 mountpoint that refers to the root.afs volume of the workstation
1984 cellname at the time the mountpoint is created.
1986 * re-write cm_Analyze to make better use of the known volume
1987 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
1990 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
1991 busy volumes will be reset to srv_non_busy by the function
1992 but there is no mechanism for querying the busy state other
1993 than by attempting to access the resource.
1995 * cm_Analyze will query the state of an offline volume before
1996 deciding whether or not to retry when all volume instances
1999 * the algorithm used after an IP address list change detection
2000 has been modified. we now check the state of the VLDB servers,
2001 then reset the rx connections, and finally check the state of
2002 the file servers. Previously we flushed the rx connections
2003 and then checked the servers in whatever random order they
2004 were first discovered.
2006 * fix the volume status reported by "fs examine"
2008 * various improvements to the AFS Network Identity Manager Provider.
2009 - Don't access freed objects after the window has been destroyed.
2010 - Write back the AFS cell list in the identity configuration
2011 even when its empty.
2012 - Disable all use of the provider when AFS is disabled for a
2013 particular identity.
2015 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2016 alias instead of the full cellname, double check the full cell
2017 name and see if we have a matching cell before allocating a new
2020 * Dump more data with "fs memdump". All volumes and scaches and
2023 * Reverse the order of the all cells list. This ensures that
2024 the workstation cell is always first in the list.
2026 * If we recognize the server from which a callback revoke was
2027 received, then we can ensure that we only invalidate the status
2028 information for cm_scache_t objects belonging to that cell.
2030 * When building for amd64, use APPVER 5.02 as we do not support
2031 64-bit Windows 2000 and explicitly state the machine type in
2034 * no longer treat VMWare network adapters as equivalent to a
2035 loopback adapter. For some unknown reason, publishing the "AFS"
2036 netbios name on them frequently results in a name collision
2039 * Use read locks whenever possible instead of write locks when
2040 walking the allCellsp list. This permits better use of multiple
2043 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2044 is not present on the directory optimization. It would report
2045 mount points and directory objects for which there was no
2046 status information as a zero length file. Instead we can use
2047 the vnode number of the object to determine if it should be
2048 considered to be a directory or a file. If it is odd it is a
2049 directory; even it is a file.
2051 * Give Back Callbacks to the file servers in response to suspend
2052 and shutdown power management events. Only give up callbacks
2053 to UP servers and do so with a very short RX dead timeout.
2056 * Fix a race condition in the SMB FID cleanup code that could
2057 result in a stat cache mutex never being released.
2059 * Fix the graphics files that are installed as part of the
2060 IBM Administrators Guide.
2062 * Change the behavior of pioctl path parsing to not follow
2063 symlinks or mount points during evaluation. This results in
2064 the FID being the symlink or mount point and not the object
2065 they point to. This is consistent with the behavior of the
2068 * Enable AFSDB record lookup for command line tools such as
2071 * CellServDB update 2007-07-05
2073 * Added a Volume Status Plug-in Interface that permits third parties
2074 to receive notifications when the volume state changes occur.
2076 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2078 * Prevent the afsprocmgmt.dll from being unloaded while the
2079 background thread it starts is running.
2081 * Avoid a race condition when closing smb_fid_t objects.
2084 * fix install path of IBM Quick Start Guide when installed by
2087 * fix the client so that it can write into directories with
2088 only 'li' permissions.
2090 * fix installation of 32-bit Shell Extension for 64-bit systems.
2091 Install it under its own HKCR keys.
2093 * modify the wix installer to remove and upgrade install packages
2094 with the same version number.
2096 * when flushing volume location data every two hours or in
2097 response to "fs checkvolumes", also flush the mount point
2098 links in order to force their re-evaluation.
2100 * aklog was copying an uninitialized buffer as part of the
2101 foreign cell automatic pts registration
2103 * when obtaining directory listings, check for read permission
2104 on the directory. If read permission does not exist, fake the
2105 status information for the directory entries. This improves
2106 performance, prevents the client from pounding the file server,
2107 and prevents leakage of cached status information.
2109 * "fs getcacheparms" output was incorrect.
2111 * NetIdMgr AFS Provider should only automatically add token
2112 acquisition for cell foo.com if the realm of the cell matches
2113 the realm of the identity.
2115 * Fix a null pointer dereference in cm_MergeStatus()
2119 * network identity manager provider:
2121 - Correct for the fact that pr_CreateUser tampers with the
2122 contents of the username input buffer.
2124 - Take into account that the krb4 credential data structure
2125 does not have separate realm fields for the user and service.
2126 When krb524 is used with a cross-realm ticket, the realm
2127 describes the service not the user.
2129 * Many integrated logon changes:
2131 - Fix eventlog reporting. Do not attempt to log an event
2132 if the event source registration fails. Use DebugEvent0
2133 instead of DebugEvent when there are no parameters.
2135 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2138 - Fix the reading of Domain specific configuration for
2139 LogonScript and TheseCells. Previously the dwSize value was
2140 being overwritten so that subsequent RegQueryValueEx call
2143 - Fix a memory leak in the TheseCells reading code.
2145 - Add support for Domain specific "Realm" specification. The
2146 realm is the realm to be appended to the username. When
2147 logging in as a domain or to the local machine, the specified
2148 "Domain" name is not going to be a valid realm name.
2150 - Construct a proper principal name based upon the domain
2151 specified realm for use in obtaining tokens with KFW.
2153 - If the domain specified "TheseCells" list includes the
2154 default cell, do not obtain tokens twice.
2156 * An unprotected call to the Kerberos v5 function,
2157 krb5_free_context(), can result in an invalid memory access
2158 exception if the contents of the krb5 profile, krb5.ini, cannot
2159 be parsed correctly. The error is located within the afskfw
2160 library which is used by afscreds.exe, afssvrmgr.exe, and
2161 afslogon.dll. When the error is triggered within afslogon.dll
2162 a blue screen condition can occur when the winlogon.exe system
2163 process terminates unexpectedly.
2167 * Do not stop the afs client service with an assert
2168 if the cell name cannot be determined when responding
2169 to a "fs whereis" command.
2171 * Integrated Logon (afslogon.dll). No longer attempt
2172 to preserve Kerberos 5 tickets for the logon session
2173 as that job is now supported by KFW 3.1 and above.
2175 * The Network Identity Manager plug-in will attempt to
2176 remove expired service tickets from the cache.
2180 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2182 * AFS Server Manager and AFS Account Manager data cache
2183 moved from HKCU\Software\TransarcCorporation to
2184 HKCU\Software\OpenAFS to prevent crashes due to data
2187 * Reduce the default UP server probe time from 10 minutes
2188 to 5 minutes to match UNIX. This will not solve the NAT
2189 port mapping timeout issue because most NATs default to
2190 a value less that 5 minutes and this value ensures that
2191 the afs client server will start probes 5 minutes after
2192 the last probe set ended. The probe period for any one
2193 server is therefore longer than 5 minutes. The more
2194 servers in the list the longer the time between probes.
2196 * Before marking a server as down in cm_Analyze, retry the
2197 request using a new rx connection to the server. If the
2198 cache manager is behind a NAT, it is possible that the
2199 response from the server is being sent to the original
2200 address:port from which the rx connection was established.
2201 This will appear to the cache manager as the server being
2202 down when in fact it is not. Forcing the use of a new rx
2203 connection will result in the server responding to the
2204 most recent address:port assigned to the client by the NAT.
2206 If the server is in fact down, it will simply take a little
2207 longer to notice it.
2211 * Removing a directory that contained entries was failing with
2212 the wrong error code being returned to the CIFS client.
2216 * rxdebug exponential back-off. don't take down a server that
2217 is already struggling to respond.
2219 * Network Identity Manager plug-in registers help on the Help
2220 menu if it is supported by NIM. (KFW 3.2 and above)
2222 * Plug-n-play networking did not always work after coming back
2223 from a Windows Suspend. The AFS client service failed to
2224 re-compute the LAN adapter to bind to and instead attempted
2225 to bind to all LAN adapters. This would polute the public
2226 network with the name "AFS" which could lead to conflicts
2229 * Build installers using Microsoft Hotfixes provided at
2230 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2231 Daylight Savings Time issues.
2233 * "fs checkserver" does not require admin privileges unless
2234 the probe interval is being changed.
2238 * A reference leak was discovered in the mountpoint traversal
2239 functionality which resulted in cm_volume_t objects not becoming
2240 eligible for re-use. This will trigger a panic condition when
2241 more than the maximum number of volumes is accessed.
2243 * Fixed the success test for the "fs memdump" operation. Success
2244 is now reported as such. Added cm_volume_t object information
2247 * A data corruption error has been fixed. Write requests
2248 queued for background processing were not retried upon failure.
2249 This leads to inconsistencies between the contents of the file
2250 in the local cache and the contents of the file on the file
2253 The implemented fix is to re-queue any request that has failed
2254 due to timeout, volume busy, volume offline, server busy, or
2255 server offline. Requests are not processed while all servers
2256 for a volume are inaccessible. When servers become accessible
2257 the background request will be retried.
2259 In addition, if a background write fails, do not mark the buffers
2260 as clean. Failure to preserve the dirty flag prevents the buffers
2261 from being written to the file server in the future.
2263 This fix does not address the possibility that the credentials
2264 necessary to perform a write request might expire between the
2265 time the request is queued and when the request can be processed
2266 due to server accessibility.
2268 * The Freelance volume does not support per user ACLs. The
2269 cm_HaveAccessRights function did not handle this special case.
2270 Requests for privileges greater than read|lookup would result
2271 in afsd_service.exe entering an infinite loop. ISOBuster is
2272 known to cause this behavior.
2274 * When deleting files, the SMB Server would attempt to obtain
2275 the access permissions for the parent directory instead of
2276 for the file itself. This would result in an incorrect
2277 assessment of whether or not the file can be deleted.
2279 * The Network Identity Manager AFS plug-in has had its string
2280 tables updated for consistency with the language used by NIM.
2281 All credentials are referred to as credentials instead of tokens.
2282 Drop-down combo boxes now in fact drop down.
2284 * The AFS Salvager was not properly being built on Windows.
2285 This is relevant only if the AFS servers are in use.
2287 * The default signal handlers for the AFS servers incorrectly
2288 processed SIGQUIT. Instead of stopping the server, an
2289 exception was raised.
2291 * Fix another volume refcount leak. This one was triggered by
2292 an inability to contact the vlservers in order to update the
2295 * Fix checklist control used by the drive mappings tabs in
2296 afscreds.exe and afs_control.exe.
2298 * Improve output of "fs memdump" related to buffer management.
2300 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2301 already in the process of being fetched.
2303 * Improve cache performance by computing the buffer hash table
2304 size based upon the number of buffers so that the average
2305 number of entries in a bucket is seven.
2307 * Improve cache performance of read/write data by locally
2308 updating the dataVersion of all buffers associated with the
2309 file on each StoreData or StoreStatus operation provided that
2310 the dataVersion was incremented by one and the dataVersion of
2311 the buffer matched the original dataVersion of the cached
2312 status information for the file.
2314 * When recycling cm_scache_t objects, make sure that we
2315 clear all of the previous fields. It is no longer the
2316 case that we can assume that the stat cache object is being
2317 obtained only after a successful FetchStatus call.
2318 InlineBulkStatus will allocate stat cache objects for which
2319 the no status information is available.
2321 * When a file is being opened with the OPEN_ALWAYS disposition,
2322 do not require write permissions when the file already exists
2323 unless of course the file is being opened requesting write
2326 * When a file has been opened for read-only access, if the
2327 application requests an exclusive lock, obtain a read-lock
2328 instead of a write-lock. The purpose of the exclusive lock
2329 in this situation is to prevent changes to the file and the
2330 full file read lock will provide this guarrantee.
2335 * Enforce Unix Mode Write bit to prevent file deletion of read only
2338 * The CIFS READ and LIST permissions use the same value 0x0001.
2339 When a directory is being accessed, check for PRSFS_LOOKUP
2340 instead of PRSFS_READ.
2344 * Fix computations of short names of mount points. They were
2345 broken as part of the directory search optimization. (1.5.13a)
2347 * When all buffers are in use, periodically sleep the threads
2348 waiting for a buffer to become free while the global buffer
2349 lock is not being held in order to permit those threads that
2350 can free buffers to do so.
2352 * Power Management improvements. Maintain a global flag that
2353 specifies whether or not the service is in a suspend state.
2354 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2355 lana is no longer valid. Instead, stop the listener threads
2356 and if all listener threads are stopped, reset the lana_list.
2357 Allow the cm_Daemon() thread to periodically check the state
2358 of the smb listeners. If they are all stopped and the service
2359 is not suspended, attempt to restart them. If there are no valid
2360 lanas, return to the stopped state.
2362 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2363 the user is permitted to obtain read or write locks. This function
2364 would obtain the lock and then drop it returning whether or not
2365 the lock could be obtained. If the lock was in fact required,
2366 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2367 of course being that this pattern results in three RPCs to the
2368 file server (lock, unlock, lock). This is reduced to one RPC by
2369 implementing cm_CheckNTOpenDone() which frees the allocated byte
2370 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2373 * Optimize the SMB FindFirst processing for the case in which there
2374 are no wildcards. If a cm_Lookup succeeds, generate a response
2375 from the stat cache entry we have a callback for. Its cheaper
2376 than walking each buffer of directory data associated with the
2377 directory stat cache entry when we have thousands of directory
2380 * Increase the SMB Ioctl MaxData size to support larger token sizes
2382 * Disable autorun of the AFS Server Wizard when the AFS Server is
2385 * Add AFS Server service dependencies on tcpip and PNP_TDI
2387 * Remove the AFS Server Configuration Tool shortcut
2389 * Add Freelance mode awareness to AFS Server Configuration Tool
2393 * Unlike previous versions of the OS, Vista performs a shutdown on
2394 the Microsoft Loopback adapter just like it would on a real adapter.
2395 This causes the smb_Listener threads to trigger a panic during a
2396 suspend/hibernate power event.
2398 The fix is to unbind from the network adapters in response to a
2399 suspend/hibernate power event and then rebind to the adapters
2400 when a resume power event is received. Note that the resume events
2401 are not reliably delivered so it is possible the afs service will
2402 not be accessible. However, this is the best we can do.
2404 * don't allow environment variables we set in afslogon.dll to be
2405 inherited by child processes
2407 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2410 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2411 thread walking the all buffer list periodically searching for dirty
2412 buffers to write to the file server. This had several negative
2413 results. The algorithm ate up ever increasing amounts of CPU time
2414 even when AFS is idle as the size of the cache increases. Also,
2415 buffers were written to the file server in an order based upon the
2416 original buffer allocation which has nothing to do with the order
2417 in which the buffers became dirty.
2419 The new algorithm maintains a dirty buffer list. Items are added
2420 when the buffer is originally marked dirty. A buffer is only
2421 removed from the list by the buf_IncrSyncer when the buffer is no
2422 longer dirty. If the list is empty the thread goes back to sleep
2423 immediately without additional processing requirements.
2427 * Enable renaming of directories from the Explorer Shell when
2428 Offline Folders is enabled
2430 * An error is reported in response to the SMB ReceiveCoreFlush
2431 request when an error has not occurred.
2435 * optimize dirty buffer flushes in response to suspend operations
2436 do not flush if only loopback adapter is installed
2437 do no retry dirty buffer writes
2439 * do not reset netbios when performing loopback checks. doing
2440 so will break the existing lana bind
2442 * remove a deadlock in smb3.c QPathInfo
2446 * Add NetIDMgr AFS Plug-in
2448 * "Show Tray Icon" checkbox removed from afs_config.exe
2450 * Disable DNS Registrations for the Loopback Adapter and make sure
2451 that Netbios is turned on.
2453 * A reference count leak on stat cache objects was fixed
2455 * For debugging, tread ids of writers are now tracked on locks
2457 * Corrected an incorrect lock that was being held but not
2458 freed in a rare error case
2460 * Changed trace log end of line to CR-LF so it can be read
2463 * modify the buffer synchronization thread to write all dirty
2464 buffers and sleep only when there is no work to be done
2466 * improve behavior of cache manager when large numbers of
2467 temporary files are being created by multiple clients within
2470 * do not permit dirty buffers that belong to deleted files
2471 to be written to the file server. its a waste of bandwidth
2473 * do not leak stat cache references when the file server
2476 * afscreds: do not permit an attempt to obtain tokens if the
2477 afs service could not be started. this was causing stack
2478 overflows on Vista after resuming from a suspend
2482 * better handling of dot directory
2484 * fs sq and fs mkm now behave as on UNIX
2486 * add license to installers
2488 * Return Invalid Parameter error for unknown InfoLevels in
2489 SMB Tran2 Search Directory.
2491 * Correctly use WSAStartup and WSACleanup within the RX library
2493 * Check that afsd_service.exe is running before performing
2496 * Force AFS Shell Extension to load resource DLL using
2497 afs_shl_ext.dll as a base name.
2499 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2500 OpenAFS\Client\Program
2502 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2504 * Do not panic if the maximum number of volume entries are in use
2505 and one of them can be recycled.
2509 * Fix the code that reads the backconnnectionhostnames value
2510 so that there is not an uninitialized variable
2514 * integrated logon could cause winlogin.exe to blue screen windows
2515 if it was called with a domain that does not exist on the machine.
2516 this is possible during a sysprep.
2518 * corrections to cm_buf_t handling to ensure that buffers are
2519 removed from queues before they are freed and that holds are
2520 obtained on the next buffer in a chain before releasing the
2523 * do not generate MessageBox dialogs on network errors. they
2524 cause more trouble then they are worth.
2526 * move GlobalAutoMapper into a separate thread so that it does
2527 not prevent service startup if the WNetAddConnection2 call
2530 * restrict hard dead and connection timeouts to the CIFS Session
2531 Timeout value. This is being done to try to prevent virtual
2532 circuits from being destroyed prematurely.
2534 * Fix the code that writes the backconnectionhostnames value
2535 to ensure that the data buffer is written with the correct
2538 (there were no windows specific changes between 1.5.4 and 1.5.5)
2542 * Corrections to the Integrated Logon functionality that permits
2543 Kerberos 5 TGTs to be communicated into the user session. The
2544 transfer is now more secure and succeeds when the user does not
2545 have read permission to the %SystemRoot%\TEMP directory.
2547 * NSIS installer updated to version 2.18
2549 * WiX installer updated to version 2.0.4221.0
2551 * Updates to the IFS functionality for 64-bit correctness and
2552 permit compilation. The IFS functionality works for pioctl()
2553 operations but not much else at this point.
2555 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2556 query and set functions.
2558 * Ignore the Windows CIFS client session timeout entirely. Timeout
2559 only after the AFS Hard Dead Timeout period expires.
2563 * move headers, libraries, and samples from OpenAFS\Client to
2566 * remove the auto-registration of AFS IDs for foreign cells
2567 from integrated logon. this permits afslogon.dll to avoid
2568 initializing the RX library. afslogon.dll is a network
2569 provider library and as such it is loaded into all applications
2570 which perform a CreateProcessAsUser call. Initializing the
2571 RX library spawns background threads to listen for inbound
2572 packets. When the network provider is unloaded from the
2573 process, the RX library does not cleanup the threads. If they
2574 continue to execute after the library is unloaded, the process
2577 * implement inline bulk status calls. this will significantly
2578 improve the performance of directory listings.
2580 * prevent a crash when evaluating mountpoints to volumes that do not
2583 * new command: fs uuid [-generate]
2584 displays the current uuid and allows a new one to be generated
2585 on the fly without restarting the afs client service.
2587 * fix the 64-bit installers to properly document the install
2590 * move help files to the proper location so they can be located by
2591 the system tray tool and the control panel.
2593 * "over quota" and "disk full" errors were not being reported to the
2596 * largefile support has been added. Files greater than 31-bits can no
2597 longer be written to file servers which do not support large files.
2598 When file servers support large files, files up to 64-bits can be
2601 * cm_SyncOp could return with the input buffer unlocked upon error
2602 which would result in a panic condition. This bug was introduced
2605 * vlserver failover when freelance mode did not work. This prevented
2606 the root.afs volume from being loaded if the first vlserver could
2611 * A bug in the CIFS Standard Info response packet caused directories
2612 to appear as standard files. This prevented Microsoft Vista from
2613 being able to walk the AFS directory tree.
2615 * Mount points must be reported as directories in response to
2616 Query Path Info operations.
2618 * A deadlock was detected when performing "fs flushall" if the
2619 file server reports VNOVNODE. The scp->createBufferLock is
2620 already held by the current thread and there is no mechanism
2621 to propagate the knowledge. Therefore, an alternate mechanism
2622 for clearing the cache must be developed. A new function
2623 cm_RecycleSCache(scp,flags) has been extracted from
2624 cm_GetNewSCache(). This function performs the task of recycling
2625 an cm_scache_t object. When called from cm_GetNewSCache() with
2626 no flags the expectation is that there are no associated buffers
2627 that are queued to be read or written. When called from
2628 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2629 flag, any queued buffers will be de-queued and marked as if
2630 the operations were performed so the data can be discarded.
2632 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2634 * In response to VNOVNODE, the parent is only discarded if the current
2635 cm_scache_t is not a directory.
2637 * In the Ioctl FlushFile and FlushVolume functions, there were no
2638 checks to protect against flushing the Freelance SCache entries.
2640 * In FlushFile, the wrong cm_scache_t object was being released.
2642 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2643 recycled. Choose a new entry if cm_RecycleSCache fails.
2645 * Speed up the performance of the cache manager by not
2646 holding the smb_fid_t mutex across calls to cm_SyncOp
2649 * Ensure that all smb_fid_t flag references are protected.
2651 * Remove a deadlock between smb_fid_t mutex and
2652 smb_rctLock introduced in 1.4.1.
2654 * Fix aklog AFSID lookup to use not send the realm
2655 name when the realm is the local realm for the
2656 cell. The PT Server doesn't strip the realm
2657 name when it is the local realm.
2659 * Treat "\\afs\*." as an alias for "\\afs\all"
2661 * Dynamically adjust the priority of server threads
2662 based upon the age of the CIFS request that is
2663 being processed. The priority is bumped one level
2664 for every 15 seconds of processing time.
2666 * The original openafs contribution mis-used the
2667 cm_GetCallback function. This function should only
2668 be called as a side effect of a call to
2669 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
2670 was being called results in a FetchStatus RPC being
2671 performed even when the client already has a valid
2672 callback and does so in a manner that destroys the
2673 synchronization of the threads that are calling
2674 cm_SyncOp. If nothing else this results in a
2675 significant performance penalty.
2677 * More changes to cleanup of smb_vc_t objects in
2678 order to prevent race conditions.
2680 * Reduce the amount of Power Management event logging
2683 * If the client detects that the IP address has
2684 changed, force the use of new RX connections. This
2685 enables the file server to recognize that the client
2688 * Remove a race condition and deadlock associated
2689 with cleanup of smb_vc_t objects.
2691 * Fix the ANSI filename support. One function that
2692 required it was missing the translations.
2694 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
2697 * Add a DllMain() function to afspthread.dll. Perform
2698 cleanup of thread local storage, rx queues, critical
2699 sections, event handles, etc. This was implemented
2700 to avoid crashes in afspthread.dll when applications
2701 which load/unload the dll attempt to terminate.
2703 * Lock request behavior has been altered. As of this
2704 release a locally allocated lock will be used whenever
2705 the volume is RO or the requesting user's permissions
2706 are 'rl'. This will allow end users to execute programs
2707 from RW volumes when they do not have the 'k' privilege.
2709 Note that a recent discovery indicates that the 'k' bit
2710 is only supposed to control the ability to obtain read
2711 locks. Write locks are to be controlled by the 'w' and
2714 * Code has been added to enable the cache manager to be
2715 able to obtain the AFSID of the authenticated user.
2716 This is required to be able to determine when "Insert"
2717 privileges should apply. AFS considers a file to be
2718 a new file whenever the owner's AFSID matches the
2719 AFSID of the user. The current implementation relies
2720 on the name specified during the SetToken call to be
2721 correct. This will be replaced in the future with a
2722 new RPC that allows the AFSID of the authenticated
2723 user to be obtained.
2725 * A 32-bit-tools installer for use on 64-bit systems
2726 is now available. These tools are for use with 32-bit
2727 versions of Kerberos for Windows or JPSoftware's 4NT.
2729 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
2731 * Remove the use of the "active_vcp" variable. This variable
2732 no longer serves any useful purpose and was the cause of
2733 race conditions that produced under counts of the smb_vc_t
2736 * Change default minidump output to include data segments
2738 * restore access to cm_IoctlGetVolumeStatus and
2739 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
2740 Still need to figure out why they are being called
2741 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
2743 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
2745 * Update build system to include the "sys_name" in the
2748 * optimize the number of calls to obtain the smb_user_t for a
2749 smb virtual circuit.
2751 * add logging of server addresses, status and capabilities when
2754 * prevent multiple threads from pinging the same server simultaneously.
2755 instead, threads block and then use the result of the ping that was
2756 already taking place.
2758 * Add registry values to allow the defaults for daemon checks to be altered
2759 at startup. Log values to afsd_init.log.
2761 daemonCheckDownInterval
2762 daemonCheckUpInterval
2763 daemonCheckVolInterval
2764 daemonCheckCBInterval
2765 daemonCheckLockInterval
2766 daemonCheckTokenInterval
2768 * Add EventLog messages for transitions in fs crypt state.
2770 * Byte range locking (when locks will be obtained from the
2771 is still under discussion)
2773 * 64-bit Windows operating systems now supported
2774 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
2775 This is a native 64-bit implementation of OpenAFS for Windows.
2776 There is currently no 64-bit version of KFW and therefore
2777 there is no (easy) method for obtaining AFS credentials.
2778 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
2779 to obtain AFS tokens but there is not for the feint of
2782 * A faster implementation of the fcrypt algorithms improves
2785 * The afsd_service.exe now supplies Event Log Message Text.
2786 The event log no longer generates warning text due to lack
2787 of message configuration.
2789 * When afs servers are unreachable and network stack returns
2790 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
2791 mark the attempted server as being down. This allows for
2792 very fast response to the calling application. This change
2793 has a pronounced improvement when the Windows Explorer is
2794 used on disconnected mobile computers.
2796 * Added DebugOutputString debugging of the RX library
2798 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
2800 * Fix the rx-lwp implementation to randomly generate rx call
2801 identifiers which are supposed to be unique at a given point
2802 in time. This affects some of the command line tools used
2803 for debugging but not the AFS Client Service itself which
2804 uses the rx-pthreads implementation.
2806 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
2807 Windows although they are not distributed
2809 * don't generate an exception when logging unexpected SMB packets
2811 * the data structures used by cm_IoctlGetVolumeStatus and
2812 cm_IoctlSetVolumeStatus are not compatible with the CIFS
2813 Get/Set File Information calls so don't bother executing them.
2817 * Fix data written to registry as part of BackConnectionHostnames
2820 * Fix initialization bug when allocating cm_server_t objects that
2821 would cause servers to appear to be down after each allocation
2822 or renewal. Server would appear up after the next check servers
2827 * Windows will return WSAECONNRESET to the next receive packet
2828 operation on a socket if an ICMP packet is received in response
2829 to a packet that was sent and could not be delivered because
2830 the port was unreachable. We have to treat this as
2831 WSAEWOULDBLOCK so that we can retry the select() check and
2832 not mark the connection as bad.
2834 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
2836 * The client manager maintains a least recently used list of
2837 all the ACLs. If an ACL is already the most recently used
2838 do not bother to remove it from the list and place it on the
2841 * As part of the process of adding a new server, set the
2842 server is down flag and then perform a ping. If the server
2843 is up, the flag will be reset. If the server is down, the
2844 client will not timeout unnecessarily while attempting to
2847 * If a Mountpoint String is the empty string, return PATH NOT
2850 * Modify behavior of the Network Provider to avoid querying the
2851 profile location if integrated logon is known to be disabled.
2853 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
2854 will always be preserved regardless of whether or not integrated
2855 logon was used to obtain tokens. When unset, tokens will be
2856 destroyed regardless of whether integrated logon was used to
2857 obtain the tokens unless the profile is stored in AFS.
2859 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
2860 message is being delivered even when the user is not logging
2861 off of the computer.
2863 * Fix a deadlock condition introduced in rc6.
2865 * Cell names are case-insensitive. Since the founding of OpenAFS
2866 there has been a bug that treated "cell", "Cell", and "CELL" as
2867 different cells. This would result in what appeared to be
2868 random disappearance of user tokens or a failure to authenticate
2869 even when tokens were listed by the "tokens" command.
2871 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
2873 * Additional auditing of the smb code revealed that smb ioctl
2874 operations were failing to free smb_fid_t objects
2876 * The File Identifiers assigned to smb_fid_t objects could be
2877 given the value 65535 which is 0xFFFF which is also the
2878 INVALID_HANDLE value. Both 0 and 65535 are now treated as
2881 * Audited all assignments to fields and ensured that they are
2882 protected by obtaining the relevant mutex object.
2884 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
2885 if both the current cm_scache_t object and its parent are
2886 both returning VNOVNODE.
2888 * Removed all the logic associated with the 'dead_vcp' variable
2889 which must have been added for debugging because its used
2890 was never thread safe and it truly served no purpose.
2892 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
2894 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
2895 registry settings. Default is ON and 120 seconds. If set to
2896 OFF, tokens will not be destroyed.
2898 * Fixed a number of race conditions and reference count problems
2899 associated with global queues.
2901 * Added a method to detect abandoned smb virtual circuits so their
2902 associated file handles, locks, user tokens, and smb session
2903 resources can be released.
2905 * Added a mechanism to allow user tokens to be preserved
2906 during integrated login and be freed after logoff.
2908 * Check access permissions to the file before the file is
2911 * More improvements to the handling of SMB Virtual Circuits
2912 when SMB sessions are closed.
2914 * Correct a race condition introduced in rc4 that could result
2915 in a crash of afsd_service.exe.
2917 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
2919 * Properly handle pre-mature SMB session closures.
2920 File handles, tree connections, and user sessions are now
2921 freed when the virtual circuit is closed.
2923 * smb_username_t and smb_user_t objects are now properly reference
2924 counted and freed as appropriate.
2926 * disable to use of AFS file locks. byte range locking is still
2927 implemented but all locks are managed locally. This prevents
2928 multiple processes from editing the same file on the same
2929 machine but it does prevent multiple processes on separate
2930 machines from modifying the file.
2932 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
2934 * A fix to "fs setserverprefs -vlserver". Multiple calls with
2935 the same server parameter would result in a panic condition
2936 if the server had not already been contacted by the AFS
2939 * Better warnings are written to the afsd.log file when
2940 the Windows RPC Protocols are not properly configured as
2941 is often the case on Windows 2000.
2943 * fix the procmgmt library to restore C RTL signal handlers when
2944 it is being detached from the process prior to process termination.
2946 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
2948 * close a lock leak in the rx library that could be triggered while
2951 * prevent removal of drive mappings in response to afscreds.exe -M
2953 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
2954 strings and flush the stat cache entry.
2956 * prevent server objects from being freed if user preferences are set.
2958 * Update WIX support to 2.0.2419.0
2960 * Release all locks on the server at service shutdown.
2962 * A down server should be marked up as long as it responds to the client.
2963 It doesn't matter if the server returned an error as long as the error
2964 is something other than VRESTARTING. If there is such an error, all
2965 connections on that server should be marked for replacement.
2967 * Change the check "up" servers period from once per hour to once every
2968 ten minutes. This matches the behavior of the Unix cache manager.
2970 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
2971 * Byte Range Locking support has been added. Architectural documentation
2972 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
2973 made to enforce Windows mandatory locking model using locally managed
2974 byte range locks with full file locks obtained from the file server.
2975 Read-only volumes provide the user with an implied locking privilege.
2976 Read-write volumes require that a user be granted the locking privilege
2977 in order for a lock to be obtained. As most Windows applications use
2978 Shared file access modes, they require the ability to obtain locks in
2979 order to access files.
2981 * Implement error translation functions to convert Universal Error Codes
2982 output by modern servers to local errno and winsock error values.
2983 This fixes the error message problems using pioctl() calls from fs.exe
2984 and the AFS Shell Extension.
2986 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
2987 * Allow the AFS Client Service to set the firewall configuration by
2988 retrying until after Windows determines the boot period is over.
2990 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
2993 * Detect IP address changes from within the AFS Client Service for the
2994 purpose of checking down servers sooner
2996 * If the file server or vldb server report RXKADUNKNOWNKEY, return
2997 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
2998 user that there is a Kerberos key problem. Previously the user would
2999 receive no feedback when there was something wrong.
3001 * Fix the large integer library support to properly handle values greater
3004 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3005 * Updated CellServDB
3007 * Fixed file name pattern matching. Question marks can match nothing
3008 if found at the end of a component.
3010 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3011 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3013 * The MSI installer now has a new Product ID for each build. This
3014 is to allow for automated upgrades. The Upgrade ID is remaining
3017 * Volumes that were cached and then subsequently were salvaged,
3018 moved or otherwise taken temporarily offline (as opposed to being
3019 busy) would become unreachable. A distinction is now made between
3020 a volume being offline and a server being down.
3022 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3024 * Fixed several AFS Shell Extension issues:
3025 - Checking Server Status would only work if a specific cell was
3027 - If servers were down, the Server Down dialog would write to
3028 invalid memory and cause explorer.exe to crash
3029 - Default Push Buttons were not specified for many dialogs
3030 - The extension would be unloaded by the Explorer Shell and
3031 never be reloaded under some circumstances.
3033 * Fixed the AFS System Tray menu to disappear if another Window
3034 was selected before a menu item was selected
3036 * Added volume owner and group information to "fs examine" output
3038 * Added fs and registry support for enabling/disabling/clearing
3039 rx statistics gathering.
3041 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3043 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3044 GUIDs for each platform.
3046 * Documentation directory structure and default installation rules
3049 * New OpenAFS for Windows HTML Release Notes installed
3050 as part of the documentation set.
3052 * Add thread id to rx debugging messages
3054 * When all servers are offline, return a bad network path error
3055 immediately. This is necessary to allow Explorer to do the
3056 right thing and not hang. The background daemon thread will
3057 check the down servers every 30 seconds to see if they are
3060 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3062 * The AFSCache file has been modified to store both the
3063 serial number of the volume on which the file resides
3064 and the SID of the local machine. If either change,
3065 a new UUID is generated.
3067 * add "fs flushall" command that forces all data buffers
3070 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3072 * add support for '/' instead of '\' in pioctl() calls
3074 * Apply AFS Client Admins group protect to AFS Shell Extension
3076 * Add support for \\afs\<name> to most AFS Shell Extension
3077 commands except the symlink methods.
3079 * when installing the Microsoft Loopback Adapter, enable
3080 MS Client for Networks and prevent an install failure
3081 by not calling CoInitialize twice in the same thread.
3083 * reload cell vldb values from the CellServDB every two
3084 hours in case it changes
3086 * When updating cell information from DNS, be sure to set
3089 * Add support to allow use of \\AFS\<foo> where <foo> is
3090 either a mount point or symlink. As <foo> is normally
3091 treated as a share name, we transform it into \\AFS\all\<foo>
3094 * Init 'code' to prevent false errors when integrated logon
3095 is disabled and the service is not running
3097 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3099 * Removed trace log messages that were hampering performance.
3101 * Fixed a deadlock that was being triggered by editing Word
3102 documents stored within AFS with WinWord 2003.
3104 * Bit 3 of the TraceOption registry value is now used to set the
3105 default for "fs trace" in the non-Debug builds. The new default
3106 is off for release builds and on for debug builds.
3108 * Bit 2 of the TraceOption registry value can be used to turn on
3109 real-time output of debug log entries to the Windows Debug
3110 Output monitor interface. This data can be viewed with tools
3111 such as Sysinternal's DbgView.
3113 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3114 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3115 a resource we should not be returning status codes such
3116 as Network Busy or Remote Host Down because these errors
3117 will imply that the AFS Client Service is Busy or Down
3118 and that is simply not the case. Instead we will return
3119 Bad Network Path as the path is temporarily not available.
3121 Instead of returning Sharing Paused when there is a
3122 timeout error, return Timeout. Once again, the AFS Client
3123 Service is not paused.
3125 * afscreds.exe would display an Obtain Creds dialog when
3126 the expired credentials reminder was triggered even
3127 if there was no network path to the KDC. This is prevented
3128 by adding KDC probe logic to the reminder thread.
3130 * afscreds.exe would display expired tokens no differently
3131 than unexpired ones. This would make it difficult for the
3132 user to distiguish when the tokens were expired. For the
3133 English build added a new resource string "(expired) that
3134 is displayed instead of the expiration time.
3137 * "fs wscell" when executed in freelance mode will return the
3138 name of the cell configured in the registry. The root.afs
3139 volume is not loaded from this cell, but it is used for the
3140 default for aklog and integrated login.
3142 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3143 no longer work on \\AFS\all when freelance is being used
3144 unless the user is a member of the "AFS Client Admins" Windows
3147 * some more checks performed during persistent cache validation.
3150 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3151 down servers for the current cell not all cells.
3153 * If the scache LRU list becomes corrupted, the AFS Client Service
3154 might enter an infinite loop at startup. Detect the corruption
3155 and discard the cache contents in this case.
3157 * Fixed the Explorer Shell's remove mountpoint functionality.
3159 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3160 obtain a mutex lock on a cm_scache_t object whose lock is held by
3161 a thread attempting to obtain the cm_aclLock.
3164 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3165 misinterpretted as "csail.mit.edu" when the resolver
3166 is configured to append ".mit.edu" to failed DNS queries.
3168 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3169 that can be used to force the preservation of user tokens upon logout.
3171 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3172 recommended installation categories: AFS Client, AFS Administrator,
3173 AFS Server, AFS Development Kit. Each category includes a different
3174 default subset of the OpenAFS components.
3176 The OpenAFS logo is now associated with the NSIS uninstall entry
3177 in the Add/Remove Programs control panel.
3179 * The user name associated with AFS tokens when obtained with
3180 integrated login, the afs systray tool, or aklog will always
3181 include the full Kerberos 5 user principal name regardless of
3182 whether or not the cell is local to the realm.
3184 * Modify integrated login so that it does not enter an infinite
3185 loop if the service is not set to auto start.
3187 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3190 * Added uninstall.exe to Wix installation
3192 * More modifications to algorithms used to wake sleeping threads.
3196 * Changes to the algorithms used to wake threads when they are
3197 put to sleep because of conflicts over buffer operations.
3200 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3201 the process of producing AFS tokens. The tokens are stored within
3202 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3203 New functionality has been added to temporarily copy the tickets to
3204 a file ccache during the short window between the NPLogonNotify and
3205 WinLogon Logon Event. The Logon Event handler starts a process as
3206 user, afscpcc.exe, that copies the tickets into the default ccache.
3207 The file ccache is then destroyed by the event handler.
3209 * During synchronization operations on stat cache entries, if threads
3210 are forced to sleep it was possible that the threads would not be
3211 woken under heavy load.
3213 * Enforce the delete on close flag specified by the NTCreateX()
3216 * A race condition in the rx library was fixed that resulted in
3217 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3218 was cleared when threads were still waiting to use the
3219 connection. Also, fixed a potential case in which the
3220 connection could be destroyed while threads were waiting to use it.
3222 * The DNLC freelist has been seen to become corrupted with still
3223 active entries being stored on the freelist. Changes were made
3224 to perform a better job of cleaning entries before placing them
3225 onto the freelist; marking them with the correct magic value when
3226 purging the list; and allowing the list to be purged on startup
3227 if the cache validation fails.
3229 * Windows Crash Reporting does not get triggered for afsd_service.exe
3230 because it provides its own top level exception handler. This patch
3231 places an explicit call to ReportCrash() on platforms which support
3234 If you configure Crash Reporting via AD Policy to capture crash
3235 reports within the domain, then you will receive this output.
3237 * Generate MiniDumps independent of the Windows Crash Reporting.
3239 (a) If an exception occurs,a minidump will be generated at the path
3240 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3241 can be set to any other type via use of the
3244 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3246 (b) MiniDumps can be generated at any time using the "fs minidump"
3247 command. This allows you to generate a minidump without
3248 stopping the process or even requiring the use of a debugger
3251 * an RX connection which reports bad ticket now treats it the same way
3252 the expired ticket is treated. The ticket is removed from the
3253 connection and a new anonymous connection is established.
3255 * fs memdump now requires that the caller be in the AFS Client Admins
3258 * additional debugging information is provided in cm_Analyze when the
3259 error code is ignored.
3261 * Fix a regression introduced into the pioctl() client code which
3262 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3263 MIT KFW is not installed. (1.3.82a)
3265 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3268 * Fix a race condition in cm_EndCallbackGrantingCall() that
3269 could leave 'nrevp' pointing to freed memory. In the process,
3270 optimize the processing to ensure that the applications monitoring
3271 the status of the stat cache entry are only notified at most
3274 * Conditionalize access to the Explorer Shell AFS->Mount Point
3275 ->Remove menu item. Disable the item if the selected item(s)
3276 are not mount points.
3278 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3279 Local Storage. This eliminates thread contention on the
3280 global RX free packet queue. This should improve performance
3281 on multi-CPU systems.
3283 * Fix the IP address filtering code to properly load/unload
3284 the IPHelper DLL so that Windows XP/2003 and beyond systems
3285 do not have to manually probe the registry.
3287 * CleanupACLEnt() was not being called consistently with
3288 the cm_scache_t object referred to by the back pointer
3289 mutex locked. This could in very rare conditions lead to
3290 an invalid memory access.
3292 * Added a script command to the msi installer "afs_replace"
3293 which can be used to replace one version of OpenAFS with
3294 another without requiring an uninstall and reinstall if
3295 the installed msi is no longer in its original location.
3297 * Apparently the problem with multi-domain forests with cross-
3298 realm trusts to non-Windows realms was not entirely solved.
3299 The authentication to the AFS SMB service failed because
3300 the wrong name was being used. Using ASU as an example,
3301 the authentication was being performed with the name
3302 "QAAD\user" (an account in the forest root) and not
3303 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3305 The solution was to add an additional dependency on KFW
3306 in order or to be able to easily obtain the client principal
3307 name stored in the MSLSA ccache TGT. This information is
3308 used in two locations:
3310 - the pioctl() function
3312 - a new WinLogon Event Handler for the "logon" event.
3314 The pioctl function will now be able to use the correct
3315 name when calling WNetAddConnection2() and the "logon"
3316 event handler will now be able to call WNetAddConnection2().
3317 The hope is that the "logon" event handler will be called
3318 before the profile is loaded but I have not guarrantee
3322 * Fixed a locking error in cm_TryBulkStat() which had the
3323 potential of deadlocking the system for the length of time
3324 it takes to perform a bulk status fetch operation.
3326 * Replaced time conversion code (UnixTime <-> FILETIME) to
3327 be completely arithmetic instead of relying on a bizarre
3328 algorithm involving a variety of C RTL time functions.
3329 This has the side effect that UnixTime and FILETIME which
3330 are both stored in UTC are interpretted as UTC throughout
3331 the year. Windows will apply the same localization to AFS
3332 as it does to NTFS. Applications which rely on the ability
3333 to sync files between the two file systems will no longer
3334 see the timestamps of files in AFS change an hour relative
3335 to the files stored in NTFS or Windows based backup devices.
3337 * Fixed a invalid memory access under a bizarre circumstance.
3338 Windows will allow a physical mass media device to be
3339 installed via Plug N Play to the system and will assign it
3340 the lowest available drive letter. This is true even if the
3341 drive letter is currently assigned in the user session to
3342 a network device via NET USE (or its equivalent.) When this
3343 happens, queries sent to the CIFS server will contain invalid
3344 data. This invalid data was not being caught by the AFS
3345 Client Service and was resulting in a crash when Freelance
3348 * Fixed a reference count error when registering callbacks
3349 on a stat cache object if the callback was already registered
3351 * Add a case to cm_Analyze() in order to handle the case where
3352 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3353 the server list for the volume is empty. In this case, force
3354 an update of the volume info and retry.
3356 * Insert a missing cm_EndCallbackGrantingCall() which could
3357 result in threads waiting for a callback to complete to never
3360 * In the persistent cache, there is a maximum number of volume
3361 entries. Allow volume entries to be reused if the maximum
3362 number have been allocated and their reference count is zero.
3364 * If we already have a dead virtual connection object, don't
3365 ignore the fact that additional dead vc objects must be taken
3368 * Removed a deadlock condition introduced in the summer of 2004.
3369 Do not hold mx locks around calls to RX functions including
3370 rx_NewCall(). That is what reference counts on the rx_connection
3373 * Fixed an initialization error in afslogon.dll which could
3374 result in random behavior including a failure to terminate
3375 the NPLogon function. (1.3.80b)
3377 * Fixed an error preventing the use of SMB authentication
3378 on Windows 2000. (1.3.80a)
3381 * Updated CellServDB to Public CellServDB 16 Mar 2005
3383 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3384 misinterpretted as "csail.mit.edu" when the resolver
3385 is configured to append ".mit.edu" to failed DNS queries.
3387 * Fixed another case in which the client would replace connections
3388 to the host when it was not necessary. In this case the
3389 situation would occur if cryptall was on and the connections
3390 were unauthenticated due to lack of an appropriate token.
3392 * OpenAFS for Windows has failed to work at sites which are
3393 utilizing a cross-realm trust between an MIT/Heimdal realm
3394 and a multi-domain Windows forest when the workstation being
3395 accessed is not located in the root domain. This is caused
3396 by a bug in the workstation which was triggered after the
3397 introduction of Windows 2003 Server. When the bug is triggered,
3398 the workstation attempts to authenticate users to afsd_service.exe
3399 by contacting the Domain Controller instead of using the
3400 LSA loopback authentication mechanism.
3402 One of the reasons this bug occurs is because the workstation
3403 does not have a reliable way of knowing that the service whose
3404 netbios name is "AFS" is located on the workstation. This will
3405 be fixed starting in Longhorn Beta 1 by Microsoft. The
3406 "BackConnectionHostNames" registry value will be used to
3407 indicate that the authentications to that service name should
3408 be performed using the loopback authentication mechanism.
3410 In the meantime, when Logon Caching is enabled, we can force
3411 afsd_service.exe to authenticate using the logon cache before
3412 contacting the Domain Controller. This will work with both
3413 password and smart card based logons.
3415 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3416 flag should not be cleared if a volume's server reference is
3417 marked as "offline". In the case where all of the volume's
3418 servers are either "down" or the volumes are "offline", we want
3419 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3420 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3421 should never occur in practice.
3423 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3424 5 seconds, clear the server down and volume busy flags, and
3425 then force an update from the VLDB. This allows the client to
3426 update the location of a volume if the reason for it being
3427 marked offline is because it is being moved. Calls to
3428 cm_ConnByMServers() will be retried either until success or
3429 the RDRtimeout period is reached.
3431 * Correct the Power Management code behavior in response to
3432 Standby, Suspend, and Shutdown events. Instead of flushing the
3433 buffers associated with the mounted SMB submounts, simply write
3434 all buffers which are marked dirty.
3436 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3437 provides a list of cells other than the default cell for which
3438 tokens should be obtained using the default Kerberos principal.
3439 This functionality is logon domain specific and is only available
3440 when using KFW for authentication.
3442 * Fixed FindNext Invalid Handle error caused by over agressive
3443 attempts at garbage collecting dirSearch entries when the
3444 dirSearch ID wraps from the maximum value to 1.
3446 * Add support for registry defined server preferences for VLDB and
3447 File servers. See registry.txt for details.
3449 * Increased default cache size to 96MB and default number of cache
3452 * Fixed refCount leaks related to directory lookups and pioctl calls.
3454 * Callbacks revoked during a race condition with an attempt to
3455 obtain the same callback no longer result in an inaccessible
3458 * New command line tool:
3460 afsdacl : Set or reset the DACL to allow starting or stopping
3461 the afsd service by any ordinary user.
3463 Usage : afsdacl [-set | -reset] [-show]
3464 -set : Sets the DACL
3465 -reset : Reset the DACL
3466 -show : Show current DACL (SDSF)
3468 * IP addresses are no longer queried once at startup. Instead IP
3469 addresses are obtained as needed. Loopback adapter addresses are
3470 no longer published to the server.
3472 * Pay attention to the MIT KFW registry configuration for automatic
3473 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3475 * Fix (once again) case-sensitive comparisons which was apparently
3478 * Activate support for vos listvol -format
3480 * Population of the "AFS Client Admins" group failed on non-English
3481 versions of Windows because the name of the "Administrators" group
3482 is localized. Now we lookup the name of the group by using the
3483 Administrators Group Alias Relative ID.
3485 * Multiple cell token acquisition within afscreds.exe was broken in the
3486 case of Kerberos 5 cross realm authentication. Instead of contacting
3487 the KDC associated with the cell's realm, afscreds would obtain a
3488 token from the KDC of the user's realm. This would result in
3489 "invalid kvno errors" while authenticating to the AFS servers.
3490 Unauthenticated access would work. The symptoms would vary based upon
3491 whether or not the VLDB servers had been contacted using unauthenticated
3492 connections prior to the user obtaining tokens.
3494 * The list of ACL entries was becoming corrupted. It appears as if
3495 an ACL when it expires was not being appropriately cleaned up. In
3496 fact, it was left in the list of ACLs associated with the scache
3499 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3500 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3501 and "amd64_win64" for amd 64-bit processors.
3503 * When executing executables, Windows provides a mechanisms by which
3504 the normal search for DLLs can be bypassed. If a file foo.exe is
3505 being executed and there exists a file or directory "foo.exe.local"
3506 (the contents are ignored if a file), then Windows will search for
3507 DLLs first in the "foo.exe.local" directory and second in the
3508 directory in which the "foo.exe" file is located.
3510 Previous releases of OAFW would improperly return
3511 STATUS_NOT_A_DIRECTORY instead of the expected
3512 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3513 to terminate the search for the DLL causing the execution of the
3514 application to fail.
3516 * Rework the reference counting of smb_vc_t objects. The references
3517 stored in the waiting locks were not counted. This could result in
3518 an assertion if the reference count drops to zero.
3520 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3521 when running in freelance mode.
3523 * Added support for persistent caching of file contents, stat entries,
3524 volume data, cell data, and name lookup data. When the registry
3525 value, "NonPersistentCaching", is not defined or set to 0 the file
3526 specified by "CachePath" becomes a persistent cache file. The
3527 size of the cache file is computed at run time. It has a maximum
3528 size of approximately 1.2GB. The contents of the file will be
3529 validated according to the rules specified by the "ValidateCache"
3532 * Added support for UUIDs. UUIDs are kept for as long as the
3533 cache file is intact.
3535 * Added cmdebug.exe and support cache manager debugging callback
3536 interfaces in afsd_service.exe.
3538 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3539 [-addrs] [-cache] [-help]
3540 Where: -long print all info
3541 -addrs print only host interfaces
3542 -cache print only cache configuration
3544 * Symlinks to \\AFS\[all\]... will now be treated the same as
3545 symlinks to /afs/.... However, please use /afs/... as otherwise
3546 the symlinks won't work on Unix.
3548 * Correct a problem with local CellID allocation for cells whose
3549 server list is obtained via DNS instead of a CellServDB file. If
3550 the DNS information expires the CellID assigned to the entry will
3551 be changed. This causes all of the cm_scache_t objects which refer
3552 to the old cellID number to become useless. Attempts to access
3553 files or directories with cm_scache_t objects using the old cellID
3554 will fail since the server list cannot be obtained.
3556 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3557 be triggered if the volume referenced by the cm_scache_t object is
3560 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3561 in addition to SERVICE_ACCEPT_STOP.
3563 * Move RPC shutdown until after the SMB and RX shutdown procedures
3564 complete. Block until RPC shutdown is complete.
3566 * Modify afslogon.dll (integrated logon) to wait for service if its
3567 state is START_PENDING. If the timeout period occurs, reset to
3568 the retry interval and not the sleep interval.
3570 * When renewing the server list for a cell obtained via DNS AFSDB
3571 records, the cm_cell_t entry must be removed from the list of all
3572 cells. Otherwise, the list of cells will be corrupted.
3574 * In the dcache and scache modules, use the cm_scache_t dataVersion
3575 instead of the cm_fakeDirVersion.
3577 * Synchronize fs.c with the unix version.
3579 * The variable used to determine whether a file or virtual memory
3580 mapped cache is used was not properly initialized to a default
3581 value. If the registry setting "NonPersistentCaching" was not
3582 set, the choice would be random. Properly initialized to be
3585 * The memory mapped view was never unmapped before closing the file
3586 at service shutdown. This is now properly cleaned up.
3588 * Default location of Cache file is now %TEMP%\AFSCache
3592 * A bug affecting new installations of 1.3.75/76 would result in
3593 the creating of incorrect mountpoints in the freelance root.afs
3594 volume for the default cell.
3596 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3597 of "root.cell", you have been affected by the bug. To correct the
3598 problem, execute the following commands:
3600 fs rmmount \\afs\all\<cellname>
3601 fs rmmount \\afs\all\.<cellname>
3603 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3604 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3608 * A bug has been fixed in the auto-generation of Freelance root.afs
3609 symlinks which produced random entries in the \\AFS\all directory.
3611 * Support has been added for multi-homed servers
3614 * Added a new registry value, "StoreAnsiFilenames", which can be used
3615 to force the use of ANSI character sets instead of OEM Code Pages.
3616 This feature is useful when users require the ability to create
3617 filenames with 8-bit characters and need to access the files from
3618 both Latin-1 based Unix systems as well as from Windows.
3620 Activation of this feature will prevent access to files stored with
3621 8-bit OEM characters.
3623 * Shutdown all SMB threads in a synchronized manner when stopping the
3626 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3627 by the largest contiguous block of unused memory within the 2GB process
3628 space which can be assigned to the memory mapped file. Unfortunately,
3629 when the executable digital signature verification code is activated
3630 Windows sees fit to further segment the process memory which in turn
3631 reduces the size of the maximum cache file to less then 800MB. If
3632 larger cache sizes are desired, a new registry value should be set:
3634 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3636 Setting this value will disable the runtime verification of digital
3637 signatures on afsd_service.exe and the afs dlls which it loads. It
3638 will not disable the the version number check on those same files.
3639 The signature verification is not a security messure and is only meant
3640 to enhance the ability to afsd_service.exe to detect potential
3641 destablizing mixtures of DLLs from incompatible distributions.
3643 Added code to auto-disable the signature verification check if
3644 the desired cache size is greater then 700MB.
3646 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3647 used to verify the validity of the afsd_service.exe binary
3648 as well as each of the AFS DLLs loaded by the service. Not only
3649 must the digital signature be valid but the signatures of the
3650 DLL must be signed by the same entity as the service.
3652 * Implement new functions: cm_freelanceMountPointExists and
3653 cm_freelanceSymlinkExists. Use them along with other validity
3654 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3655 ensure that name collisions do not occur and that empty strings
3656 are not valid file names.
3658 A symlink may not have a name which would resolve to a valid
3659 cell name. Doing so would prevent access to the cell.
3661 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3662 re-initialization code. The reference counts of the fake root.afs
3663 volume scache object(s) would become invalid when the mount point
3664 or symlink lists were altered.
3666 * Add registry entries to provide mappings from the afsdsbmt.ini
3667 to the new locations for applications which count on the use
3668 of the old Profile file APIs. These apps are likely to fail
3669 if the user does not have administrator privileges and the
3670 registry is locked down.
3672 * The afs_config.exe submounts dialog had two errors.
3673 First, attempts to remove entries failed because the registry
3674 key was being opened without KEY_WRITE privileges.
3675 Second, when editing a submount entry, changing the name
3676 would add a new key and leave the original one in place.
3677 Now the original submount will be removed if its name is
3680 * In recent months there have been several incidents in which
3681 users have experienced problems starting or accessing
3682 afsd_service.exe and after significant effort has been spent
3683 it has turned out that they have two versions of AFS on the
3684 machine or an inconsistent set of DLLs.
3686 Code has now been added to afsd_service.exe which will walk
3687 the list of modules loaded by afsd_service.exe and validate
3688 that the version of the AFS DLLs matches the version of the
3689 afsd_service.exe executable. If they do not match the service
3692 * When Freelance mode is enabled and there is no registry
3693 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
3694 will attempt to import the afs_freelance.ini file contents.
3695 If the file does not exist, it was creating a dummy file
3696 with a r/o and r/w entry for the default cell and then
3697 importing those values.
3699 This process has been changed. The temporary file is no
3700 longer created. Also, both the OpenAFS Client install
3701 directory as well as %WINDIR% are checked for previous
3702 afs_freelance.ini files.
3704 * Added support for VL_GetEntryByNameN(). Still need to add
3705 support for VL_GetEntryByNameU() for multi-homed support.
3707 * Fix a deadlock situation in afscreds.exe when canceling an
3708 auto-generated Obtain Tokens dialog
3711 * Fix the locking of objects during Directory Searches in the
3712 SMB/CIFS server. The failure to properly lock the reference
3713 counts was resulting in the premature freeing of smb_dirSearch_t
3714 objects while they were still in use by the SMB/CIFS client.
3715 This does not solve the "Invalid Handle" problem.
3717 * Fix Find Cell By Name pioctl call to return a valid cell
3718 name for the Freelance fake root.afs volume.
3719 "Freelance.Local.Root".
3721 * Fix the Explorer Shell Extension Symlinks->Add operation.
3722 The dialog template was missing and the link destination
3723 string was too short.
3725 * Add support for symlinks to Freelance root.afs volume
3726 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
3727 <number> = "<linkname>:<relative-path>."
3728 Use symlink.exe to create, list, or remove
3730 * Remove the fallback to the use of KFW's KRB4 library when obtaining
3731 tokens. We never obtain KRB4 tickets.
3733 * Fix AFS Client Configuration Control Panel to support new SysName
3736 * Fix a bug in afsd_service.exe which could result in the SysName
3737 not being read from the registry.
3740 * Add code to block the issuance of AFS tokens by aklog.exe or
3741 afscreds.exe when the Kerberos 5 principal name contains a dot.
3743 * Modify the IsAdmin() function to always treat the local SYSTEM
3744 account as an AFS client administrator. Affects fs.exe and
3747 * Modify the internal handling of Quota Exceeded errors
3749 * Upgrade all reference count fields in the Windows cache manager
3750 and the osi library to use unsigned long instead of signed short.
3751 A similar fix has been applied to the afs rpc (rx) library.
3753 * fix the Windows cache manager to prevent it from replacing the
3754 rx_connection object associated with the cm_conn_t object on each
3755 and every operation if "fs crypt" was set. This explains the
3756 dramatic performance difference when crypt is used vs clear.
3757 The problem: 'cryptall', a boolean flag indicating whether or not
3758 "fs crypt" is set, was being compared to the rx_connection
3759 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
3760 1 != 2 and therefore the rx_connection was always destroyed
3761 and replaced on each and every operation.
3763 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
3764 It is not safe for the cm_conn_t object to not be locked because
3765 rx_DestroyConnection might be called from another thread if:
3766 - the user's tokens have changed or expired
3767 - the crypt mode has changed
3769 This fix appears to have also taken care of the problems associated
3770 with Overlapped Writes resulting in Delayed Write errors.
3772 * fix NSIS installer's AdminGroup.exe to properly create and
3773 remove groups when given -create or -remove. The string comparison
3776 * fs sysname now accepts a list of sysname values
3778 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
3779 DWORD which when set to a non-zero value will cause error message
3780 text to be output to stderr from the pioctl() routine. Useful in
3781 debugging failures of fs.exe, tokens.exe, etc.
3783 * added a test to the power management code to only perform a
3784 flush operation if there is at least one network adapter which
3785 is not a loopback adapter.
3787 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
3788 "EnableKFW". This value will not be read if the key
3789 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
3790 value under that key does not.
3792 * provide mechanisms to force the use of krb524d for Kerberos 5
3793 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
3794 there is a new registry value "Use524" and for aklog.exe a new
3795 command line parameter "-m".
3797 * Fix the pattern matching algorithm to properly match patterns
3800 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
3801 which is used by both the original function and the new
3802 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
3803 creation of HardLinks in addition to Renaming. smb_Link()
3804 is a new function which creates HardLinks via cm_Link().
3805 cm_Link() is a new vnodeops function which creates links
3808 smb_ReceiveNTRename() does not support the File Copy and
3809 Move Cluster Information operations described in its interface.
3810 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
3812 * When opening files via symlinks, we should follow the symlinks
3813 until we reach the actual file stat cache entry. The stat cache
3814 entry of the file should then be stored in the FID instead of
3815 stat scache entry of the symlink.
3817 * return bad operation errors for all unimplemented functions
3818 even if we do not know the functions exist.
3820 * Log bad packets and unknown operation packets to the trace log
3822 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
3825 * Update list of known CIFS operations to include all those listed
3826 in CIFS-TR-1p00_FINAL.pdf.
3828 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
3829 to support the REG_EXPAND_SZ type.
3832 * A new Windows authorization group "AFS Client Admins" is now
3833 created and populated with the members of the "Administrators"
3834 group. The group is used to determine which accounts on the
3835 machine may be used to modify the AFS Client Configuration via
3836 the UI and command line tools. afs_config.exe, fs.exe,
3838 * Modify the WinLogon Logoff Event Handler to query NT4 domain
3839 controllers for the remote profile path if Active Directory
3840 services are not available.
3842 * Fix aklog.exe to not add the AFS ID to the username
3844 * PTS registration of new users to foreign cells has been added to
3847 * The cm_Daemon thread is used to perform checks for
3848 down servers, up servers, volumes, callback expirations,
3849 lock maintenance and token expiration. Due to a gaff in
3850 larger integer division the thread never performed any
3851 work. Instead the current time computation would always
3852 be less then the trigger times. This had an adverse affect
3853 on the client's ability to maintain communication with servers,
3854 keep volumes up to date, and flush user tokens and acls
3855 when they have expired. This was broken when the 1.3 branch
3856 was modified to support VC7 which no longer included
3859 * An initialization problem with the Freelance code was
3860 detected while fixing the callbackRequest. The cm_rootSCachep
3861 object is obtained during afsd_InitDaemons() but the callback
3862 information is incomplete. The callback information will not
3863 be obtained until cm_MergeStatus is called from within
3864 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
3865 test for the conditions under which the callback information
3868 * Reports have been filed indicating that callbacks were
3869 being lost. An examination of the code indicated that the
3870 cm_server_t objects were not being properly reference
3871 counted by the cm_scache_t and cm_callbackRequest_t objects.
3872 In particular, the cm_server_t objects may have been freed
3873 from beneath the cm_conn_t objects.
3875 All of the reference counting is now done via the functions:
3880 this improves the ability to track the referrals.
3882 Each cm_BeginCallbackGranting Call now allocates a reference
3883 to the cm_server_t. The cm_EndCallbackGrantingCall either
3884 frees the reference or transfers it to the cm_scache_t
3885 cbServerp field. These are then appropriately tracked
3886 through the cm_Analyze call.
3888 * Ensure that the dnlc hash table is the same size as the
3889 dir name hash table (as per original author's note).
3890 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
3891 for compatibility with 64-bit systems.
3893 * fix smb_ApplyV3DirListPatches to properly apply the hidden
3894 attribute to dotfiles when the infoLevel < 0x101 and
3895 cm_SyncOp has failed.
3897 * Fix the Freelance registry initialization code. There
3898 was a possibility that some systems could end up with
3899 garbage in the registry during a clean install.
3902 * file and directory names beginning with "." will now be given the
3903 hidden attribute when the volume access is anonymous. this matches
3904 the behavior when the volume access is via an authenticated user.
3906 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
3907 key. When a change occurs mark the root.afs data as invalid and
3908 for it to be reloaded on the next access. This allows administrators
3909 to modify the mount point list without restarting the service.
3911 The freelance client used to provide a fake modification time for
3912 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
3913 Added code to extract the last modification time of the Freelance
3914 registry key and use that instead. The time now represents the
3915 most recent mount point change.
3917 * PTS registration of new users to foreign cells has been added to
3920 * Additional Cache Control and Credential Manager options have been
3921 added to the WiX installer. See deployment guide for details.
3923 * The CachePath setting is now optionally a REG_EXPAND_SZ type
3925 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
3928 * The loopback installation code may have had a problem updating the
3929 %ETC%\HOSTS file which could have resulted in a premature failure.
3930 Work around code has been added for the case where the file cannot
3933 * The default max chunksize was increased from 15 (32K) to 17 (128K)
3934 because Windows sends 64K blocks when using overlapped writes.
3936 * The default number of server threads was increased from 4 to 25 to
3937 better handle overlapped writes.
3939 * The "AfscredsShortcutParams" registry value was not being properly
3940 loaded by afscreds.exe. Therefore, the default value was always being
3941 used instead of the value set by the installer.
3943 * Windows XP provides downgrade attack detection to prevent an attacker
3944 from being able to force the use of NTLM simply by disrupting
3945 communication with the KDC. This attack cannot exist between the
3946 Windows CIFS client and the AFS Client Service. Therefore, when a
3947 downgrade has been detected the afs pioctl library will force the
3948 establishment of a new CIFS connection using NTLM.
3950 * A locking error was discovered surrounding all references to volume
3951 server lists within the cm_cell.c source file.
3953 * The logged into Windows username was incorrect on Terminal Server
3956 * A new registry value "NonPersistentCaching" was added to the service
3957 parameters key. When set to a non-zero value, the afs cache is stored
3958 in the Windows paging file. There are two limitations to choosing
3960 1. when persistent caching is implemented it won't work with
3961 this flag set since there will be nothing to persist.
3962 2. with this flag set the initial paging allocation cannot be
3963 changed while the service is running
3965 * An initialization bug was discovered in aklog.exe which affected users
3966 who have a domain name for their afs servers which could not be mapped
3970 * afs_config.exe now validates cell names against DNS in addition
3971 to the CellServDB file.
3973 * In order to allow the freelance client to connect to a volume with ID
3974 equal to 1 on the default cell we changed the fake root.afs volume ID
3975 once again. This time we choose 0xFFFFFFFF. In addition, we change
3976 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
3977 It will now be impossible for a volume ID to match that of another
3978 cell unless the client is connected to 0xFFFFFFFD cells. That should
3979 be enough room for growth.
3981 * Fix "fs mkmount" command to work with UNC paths and when
3982 started from non-AFS drives. It is now possible to create a mount
3983 point in the freelance fake root.afs volume with the command
3985 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
3989 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
3990 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
3992 * The algorithm used to re-attempt access to the servers associated with
3993 a volume has been altered to properly address the case in which all
3994 servers have been marked down. The previous algorithm did not reset
3995 the server's down flags so the servers were never actually retried.
3996 This caused a problem with active volumes if the network connectivity
3997 was lost as could be the case with a network cable removal, wireless
3998 drop, or laptop hibernation. With the fix volume access is restored
3999 almost instantenously when network connectivity becomes available.
4001 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4002 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4003 listing of AFS submounts and freelance mount points. Support for
4004 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4005 significantly enhances the behavior of AFS volumes within the Explorer
4006 Shell. For instance, "AFS" now shows up as server in the Explorer
4007 with each submount or freelance mount point visible as a share.
4008 The right click menu in each folder now works with full functionality
4009 on a consistent basis.
4011 * The network provider can be configured to have different behavior
4012 depending on the domain that the user logs into. These settings are
4013 only relevant when using integrated login. A domain refers to an
4014 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4015 local machine (i.e. local account logins). The domain name that is
4016 used for selecting the domain would be the domain that is passed into
4017 the NPLogonNotify function of the network provider. (see registry.txt
4020 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4021 "Authentication Cell" which may be used to specify a default
4022 authentication cell for afscreds.exe which is different from
4023 the default cell for the AFS Client Service daemon.
4025 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4026 afslogon.dll moved to %WINDIR%\System32\.
4027 New registry entries added to register the dll for Winlogon events.
4029 The logoff event will now force a call to ktc_ForgetAllTokens()
4030 using the context of the user being logged off as long as the
4031 user's profile is not loaded from within AFS. If the profile
4032 was loaded from AFS we can't release the tokens since the Logoff
4033 event is triggered prior to the profile being written back to
4034 the its source location. This is now performed in an XP SP2
4037 * Windows XP SP2 Internet Connection Firewall interoperability
4040 * The %WINDIR%\afsdsbmt.ini contains four sections:
4041 Submounts, Drive Mappings, Active Maps and CSC Policies.
4042 The Submounts and CSC policies are now stored in the registry under
4043 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4044 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4045 The Drive Mappings and Active Maps are stored in the registry under
4046 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4047 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4049 There is no automatic migration of this data as it would be impossible
4050 to consistently migrate data to user profiles which may not be active
4051 when the machine is updated.
4053 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4054 fake root.afs volume. For the same reasons as for the cellservdb file,
4055 this information should not be in %WINDIR%. This information is now
4056 kept under the registry key
4057 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4059 The data from the afs_freelance.ini file will be automatically
4060 migrated to the registry on first execution of afsd_service.exe
4062 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4063 troublesome for several reasons. One, it is confusing for those who
4064 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4065 Two, this file is not a Windows Profile formatted file. Three,
4066 applications should not be reading or writing to %WINDIR%. It causes
4067 problems for Windows Terminal Server.
4069 The new location of CellServDB will be the OpenAFS Client install
4070 directory which is by default C:\Program Files\OpenAFS\Client and can
4071 be determined by querying the registry for
4072 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4074 The existing afsdcell.ini will be migrated by the NSIS installer.
4075 The Wix installer must still be updated to do the same.
4077 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4078 High Security mode; and to add Terminal Services compatibility registry
4079 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4080 configuration files in %WINDIR%.
4082 * Add support for authenticated SMB connections. This will remove
4083 the need for high security mode in most situations. Both NTLM
4084 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4085 only NTLM can be used even though Kerberos is now supported. The
4086 reason is that it is not possible to construct a service principal
4087 which is unique to each individual machine.
4089 SMB Extended Auth does not work on XP SP2 unless one of two registry
4090 modifications are made:
4092 (1) To disable the check for matching host names on loopback connections
4093 set this key. This does not require a reboot:
4095 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4096 "DisableLoopbackCheck"=dword:00000001
4098 (2) To add the AFS SMB/CIFS service name to an approved list. This
4099 does require a reboot:
4101 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4102 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4104 afsd_service.exe will automatically add the current Netbios Name
4105 to the BackConnectionHostNames list and then temporarily disable
4106 the loopback check for one cycle of startup/shutdown of the service.
4107 We assume most folks do not start/stop without a reboot so this
4108 will be adequate in most cases.
4110 * Fix security hole in afslogon.dll which allowed passwords to be
4111 sent in clear text to the KDC in a misformed principal name.
4113 * Fix cm_GetCell() to properly handle expired dns entries
4116 * If Freelance mode is active and the afs_freelance.ini
4117 file does not exist, do not create an empty file.
4118 Instead create a file containing ro and rw mountpoints
4119 to the default cell using the standard conventions.
4121 * Modify the Freelance support to handle the ability
4122 to create rw mount points in the fake root.afs volume.
4124 * Changed the RPC mechanism used for token setting from
4125 named pipes to local. Use of named pipes can be restored
4126 by setting the environment variable AFS_RPC_PROTSEQ to
4129 Named pipes were required when a Windows 9x system was
4130 using a NT system in gateway mode which is incompatible
4131 with our use of local loopback adapters.
4133 * In afscreds.exe, if a username of the form user@REALM is
4134 specified and no password is specified, do not perform a
4135 kinit operation. Only perform the aklog functionality.
4137 * Add a new registry value which allows the number of processors
4138 on which afsd_service.exe executes to be restricted. Valid
4139 values are 1..numOfProcessors
4141 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4145 * A second MSI based installer option is now available.
4147 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4148 request tickets for user/instance@REALM instead of just user@REALM
4150 * Modify the Power Management Notify routine to wait for the Hard Dead
4151 timeout period instead of a fixed 19 seconds. With the longer timeout
4152 periods Hibernation and Standby could never succeed when network
4153 connectivity is not available.
4155 * The following fs.exe commands are now restricted to Administrator:
4156 - checkservers with a non-zero timer value
4159 - sysname with a new sysname list
4168 setting the default sysname for a machine should be done via the
4169 registry and not via "fs sysname".
4171 * NSIS installer adds options to install Debugging Symbols
4172 and the Microsoft Loopback Adapter; the user is now also
4173 given the ability to select the afscreds.exe startup options.
4175 * Build system modified to generate symbols for FREE (aka RELEASE)
4176 builds as well as CHECKED (aka DEBUG) builds
4178 * Sites which have a volume ID of 0x20000001 assigned to their
4179 root.cell volumes have been experiencing problems with accessing
4180 the root.cell volume of their cell when Freelance mode has been
4181 active. This was because 0x20000001 was assigned to the fake
4182 root.afs volume created by freelance. The fake volume id is
4183 now set to 0x00000001 to prevent conflicts.
4185 * The timeout logic in the AFS Client Service has been wrong
4186 for sometime. It is based on two different assumptions.
4187 First, the SMB client timeout is a fix value as was the case
4188 with OS/2 Lan Manager. This assumption is incorrect. The
4189 SMB timeout in Windows is a dynamic value computed based upon
4190 a fixed minimum timeout to which is added time based upon the
4191 size of the request and the performance characteristics of
4192 the connection. Second, it is the responsibility of the
4193 SMB Server to enforce the timeout requirements of the client.
4194 This is untrue. The SMB Server cannot be expected to know
4195 the requirements of the client. More importantly, if the
4196 SMB server uses the SMB client timeout as a value to restrict
4197 its behavior as an RX client, the performance characteristics
4198 of the local SMB session would be used to prematurely terminate
4199 WAN connections with significantly different performance
4202 The timeout logic has therefore been modified in the following
4204 . the Lan Manager Workstation (SMB) Session Timeout is used only
4205 as a basis for configuring the Connection Dead Timeout
4206 and Hard Dead Timeout values. The Connection Dead Timeout
4207 must be at least 15 seconds longer than the SMB Timeout
4208 and the Hard Dead Timeout must be at least double the
4209 Connection Dead Timeout.
4210 . New registry entries have been added to allow the Connection
4211 Dead Timeout and Hard Dead Timeout values independent of the
4212 Lan Manager Workstation Session Timeout
4213 . The test to enforce the SMB Client Timeout has been removed.
4215 One of the side-effects of removing the enforcement of the SMB
4216 Client Timeout is that regardless of whether or not the SMB client
4217 is available to receive the response (and how would the SMB server
4218 know) the RX protocol response can be used to update the AFS
4219 Client Service state for ready access by future SMB client
4222 This should be the end of the "Server paused or restarting messages"
4224 * Add support for arbitrary UNC paths to the pioctl() support.
4225 This enables the fs commands as well as the AFS Shell Extension
4226 to work correctly when UNC paths are being used.
4228 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4229 DNS if the cell configuration cannot be determined via CellServDB
4231 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4232 are ever set more than once at a time
4234 * Fix the management of lists of cm_cell_t structures when using
4235 DNS to lookup cell information. The previous code would fail to
4236 reuse the same cellID for a cell if DNS was used more than once
4237 for a given cell name. When the ttl expired, a single cm_cell_t
4238 could be inserted into the cm_allCellsp list more than once
4239 producing a loop. In addition, the vlServerp list belonging to
4240 the cell was not freed resulting in improper refCounting of the
4243 * Add DNS support to cm_IoctlNewCell() which previous only examined
4246 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4247 to properly garbage collect cm_server_t objects
4249 * Add numVCs variable to smb.c to track the number of smb_vc_t
4250 objects created and use it to initialize the vcID field which
4251 previously was set to 0 in all objects resulting in FindByID
4254 * Fixed DNS lookups to work consistently throughout the OpenAFS
4255 product instead of just from within the afsd_service.exe
4257 * Added a runtime check to ensure that AFS Client Service SMB
4258 Server is accepting connections before attempting to mount
4261 * Read IP addresses for volume servers out of the CellServDB
4262 file if gethostbyname() on the hostname fails.
4264 * Fix getcellconfig() to populate both the Hostnames as well
4265 as IP addresses when loading cell data via DNS
4267 * Increase the Connection Dead Time to 50 from 20 seconds
4268 Increase the Hard Dead Time to 120 from 40 seconds
4269 (matches the Unix values)
4271 * Fixed an assertion validating the number of allocated NCBs
4273 * Fixed the build environment to consistently build for
4274 Windows 2000 and above. (APPVER = 5.0)
4276 * Fixed rx_debug to properly validate the receipt in incoming
4277 data with select() and recvfrom(). Do not copy data out of
4278 the socket buffer unless success is indicated.
4281 * afsd_service.exe will now display a message box to the
4282 desktop when it terminates due to an IP Address Change.
4284 * installer no longer deletes AFS Server configuration data
4287 * installer generates a warning dialog if the RPC service
4288 is not properly configured
4290 * installer compressed with lzma instead of bzip2
4292 * afsd_service.exe shutdown crash solved once and for all
4294 * reference counting of smb_vc_t data structures improved
4296 * name space collision of smb_fid_t event objects corrected
4298 * the output of "fs memdump" is now written to
4299 %WINDIR%\TEMP\afsd_alloc.log
4301 * the file TaAfsApp_1033.dll is now properly installed allowing
4302 the User Manager to start
4304 * a new algorithm is used for computing filename pattern matches
4306 * afscreds.exe now accepts user names containing instance
4309 * Fix the Directory Name Lookup Cache to be case-sensitive.
4310 This is crucial in environments in which a Windows client
4311 is accessing a directory with more than one filename that
4312 differs only by case. If the directory contains "FOO"
4313 and "Foo". You want "DEL Foo" to delete the correct one.
4314 We still have a problem in that "DEL foo" will delete a
4315 random filename. This will be addressed in a future release.
4317 * Fix afscreds.exe -M option (renewMaps) to work when High
4318 Security mode is off. Also, remember to disable the ActiveMap
4319 flag in afsdsbmt.ini when a drive mapping is removed.
4321 * Updates to NSIS installer script. AFS Server configuration
4322 data will not be destroyed on un-install or re-install.
4323 Use a better compression algorithm.
4325 * afslogon.dll now uses KFW to obtain tokens when available
4327 * afslogon.dll when given an all uppercase username will
4328 attempt to authenticate with both the uppercase name
4329 and an all lowercase variation
4331 * DST modification removed. The fix appears to make things
4332 worse after a reboot of the machine.
4334 * fs.exe: added "cscpolicy" which is used to
4335 change client side caching policy for AFS shares
4337 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4338 [-documents] [-disable] [-help]
4340 * Several uninitialized variables have been initialized
4342 * It is now possible to obtain tokens using cross realm
4343 Kerberos within afscreds.exe:
4345 user: jaltman@ATHENA.MIT.EDU
4347 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4348 will will in turn be used to obtain afs@DEMENTIA.ORG.
4349 The resulting token will be stored with the display name
4350 jaltman@ATHENA.MIT.EDU@dementia.org
4352 * aklog.exe has been added to the client
4354 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4355 [[-p | -path] pathname]
4359 -d gives debugging information.
4360 krb_realm is the kerberos realm of a cell.
4361 pathname is the name of a directory to which you wish to authenticate.
4362 -noprdb means don't try to determine AFS ID.
4363 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4364 (default is Kerberos V)
4365 No commandline arguments means authenticate to the local cell.
4368 * All of the resource files have been restructured to adhere to
4369 a set of rules IBM implemented for loading string resources.
4370 These rules had either been forgotten or were not discovered
4371 by folks working on the OpenAFS sources. The end result was
4372 memory corruption. This is primary item which was preventing
4373 the AFS Server from working.
4375 * Increased the size of the maximum ticket size stored in a token
4376 from 344 bytes to 12,000. Increased the buffers used to convey
4377 messages between the pioctl() caller and the SMB Server from
4378 1000 bytes to 12,512. The code appeared to have been writing
4379 above the top of the stack by quite a few number of bytes.
4380 (The increased ticket size is necessary for the next item.)
4382 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4383 Instead the raw Kerberos 5 ticket is used in its entirety. This
4384 is extremely important as it allows us to use pure Kerberos 5 KDCs
4385 as the source of the AFS authentication. The use of up to 12,000 byte
4386 tickets will allow tickets produced by all versions of Microsoft
4387 Active Directory to be used.
4388 - create a user account.
4389 - designate it DES only
4391 - specify its UPN to be "afs@realm"
4392 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4394 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4395 is special when using long file names. (you must use "*.*" and not "*")
4396 Instead only enforce it when performing 8dot3 searches.
4398 * Fixed the DST problem with creation times being set one hour ahead
4400 * Fixed the problem when using \\afs\cell-alias. For example,
4401 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4402 for the alias name; instead simply expand the name. One of the
4403 symptoms of this problem was a loss of acquired tokens.
4405 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4406 of strings. (Only English strings provided.)
4408 * Fixed the installer to properly replace in use files.
4410 * Fixed the build system to cleanup generated component version files
4412 * The release build compiled with MSVC 6.0 compiler to avoid the
4413 afsd_service.exe shutdown crash. This does not solve the problem
4414 but simply avoids it for the time being.
4418 * fix afslogon.dll to not corrupt memory when High Security mode
4421 * fix afsd_service.exe to not attempt to restore the stack when
4422 an exception occurs. (not safe in multi-threaded programs)
4424 * fix uninstaller to properly remove the CRT and MFC DLLs
4426 * remove a Message Box from afscreds.exe when getcellconfig()
4427 fails on a kerberos realm which is not a cell
4429 The following is a list of changes to the OpenAFS for Window client
4432 * "fs setserverprefs" will leave afsd service deadlocked
4434 * "vos listaddrs" will core dump
4436 * installer sets the appropriate keys to support Integrated Logon
4438 * installer disables the "Find Lana by Name" functionality as it
4439 was causing headaches for many users
4441 * fix the intermittent crash of the power management thread when
4442 shutting down the AFS Client Service
4444 * optimizes the obtain drive mount list functionality which is
4445 executed every time the mount tab in afscreds.exe and afs_config.exe
4446 are refreshed. (this happens a lot)
4448 * fix the service shutdown logic. add the STOP_PENDING state
4449 and do not accept additional service events after we declare
4452 The following is a list of changes to the OpenAFS for Window client
4455 * flexelint was run against the source tree and hundreds (perhaps
4456 thousands) of corrections were applied to ensure prototypes
4457 were in use; types were used consistently; variables were
4458 initialized; unused variables were removed; etc.
4460 * A wide variety of instrumentation was added including the
4461 ability to produce a stack trace from within afsd_service.exe
4464 * Dynamic configuration of the RDRtimeout value based upon the
4465 LanMan Workstation Session Timeout
4467 * The mount root no longer needs to be called "/afs". This
4468 is now set by a registry value "MountRoot" within the key
4469 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4471 * The cell list is now only read out of afsdcell.ini when the
4472 file changes instead of each time a cell is resolved.
4474 * Thread synchronization was added to cm_server.c and ktc_nt.c
4476 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4477 calloc()/free(). The Global functions were needed on Windows 3.x
4478 but have caused a variety of problems on the Win32 platforms.
4479 Avoiding them is highly recommended by several Microsoft
4480 Knowledgebase articles
4482 * Support for Symbolic Links added to the AFS Shell Extension
4484 * Added a registry value "OverlayEnabled" to determine if
4485 Shell Extension Overlays should be enabled.
4486 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4488 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4489 separate trees for checked and free builds. Build system supports
4490 a custom directory src\WINNT\extra which can be used as a grafting
4491 location of organization specific additions to the build tree.
4493 * New installer built using NSIS 2.0.
4495 * Named all kernel objects in order to allow them to be monitored
4496 with tools such as SysInternals' ProcExp.exe.
4498 * Introduced new EventLog framework for AFSD
4500 * Introduced Power Management interface to AFSD for Standby and
4501 Hibernate modes to allow cache to be flushed prior to network
4504 * Utilize Win32 DNSQuery API instead of internal routines. This
4505 allows DNS SRV queries to be sent to all current domain name
4506 servers. Not just one specified in an INI file. DNS is now
4509 * "NetbiosName" registry value may be used to specify a fixed
4510 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4511 when the loopback adapter is in use. If you need to use the
4512 old notation with a loopback adapter installed specify a registry
4515 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4517 * Refactor all modules which depend on LAN Adapter and NetbiosName
4518 determination in a new library: lanahelper.lib. This allows for
4519 consistent behavior throughout the product.
4521 * Move the afsd.log and afsd_init.log files to the directory specified
4522 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4523 for services. Added the Date to the log entries.
4525 * New registry value "RxMaxMTU" used to limit the size of the RX
4526 packets sent by the AFS Client Service to the Server. In order
4527 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4528 size must be restricted to 1264 or smaller. The latest NSIS
4529 installer sets a value of 1260 by default.
4531 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4532 packets. This is not needed in order to work across the Cisco
4533 VPN but might be needed for other network environments. This
4534 value is not set by the NSIS installer.
4536 * New registry value "HideDotFiles" is used to apply the Hidden
4537 attribute to files whose names begin with a '.'. This value
4538 is set by the NSIS installer.
4540 * New registry value "MaxMpxRequests" allows the maximum number
4541 of multiplexed sessions to be configured at run time. This
4542 value is not set by the NSIS installer. The default value is
4545 * New registry value "MaxVCPerServer" allows the maxmimum number
4546 of VCs per server to be configured at run time. This value is
4547 not set by the NSIS installer. The default value is 100.
4549 * New registry value "AllSubmount" allows the "all" submount to
4550 be disabled by setting its value to 0x00.
4552 * Allow cells names to be valid mount points
4553 \\<netbiosName>\<cellname>
4555 * Store the active state of drive mappings in order for afscreds.exe
4556 to restore them upon startup
4558 * Add exception handling to generate a Stack Trace to the afsd_init.log
4559 file if one happens to occur.
4561 * Add lots of logging to help detect the cause of invalid SMB packets
4563 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4564 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4565 renewal of expiring tokens as long as afscreds.exe is running.
4567 * New afscreds.exe command line options:
4569 -M = renew drive maps
4570 -N = ip address change detection
4573 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4574 determines whether or not MIT Kerberos for Windows should be used
4575 to obtain tokens via Kerberos 5 tickets.
4577 * New registry value "AfscredsShortcutParams" in
4578 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4579 determines the command line parameters to be specified when "fixing"
4580 the AFS Shortcut in the user's startup folder.
4582 * The "ShowTrayIcon" registry value has been moved from
4583 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4584 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4586 * The <cell name> registry values used to store the token expiration
4587 reminders have been moved from
4588 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4589 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4591 * Obtain the Logon User Name from the Explorer key when available
4593 * new text document doc\txt\winnotes\registry.txt lists all registry
4594 values used by OpenAFS (excluding the AFS Server)
4596 * BUG: rx_securityClass objects were not properly reference
4597 counted and were never freed.
4599 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4600 would be generated. The existence of a server does not imply
4601 that it is not down. If all of the servers for a cell are down
4602 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4605 * BUG: the directory name lookup cache failed to free the entries
4606 in the cache when the name cache entries cycled. The entries
4607 in the cache would become dereferenced without being freed.
4609 * BUG: fs setserverprefs could be executed without Administrator
4612 * BUG: the number of allocated NCB objects (100) exceeded the number
4613 which could actually be waited upon by the kernel (64). Any objects
4614 which were utilized above the limit could never have event completions
4617 * BUG: smb_username_t objects were not being reference counted and
4618 were not properly freed.
4620 * BUG: smb_tid_t objects could under unusual circumstances be freed
4621 before they were no longer referenced.
4623 * BUG: smb_fid_t object pointer were frequently used even when
4624 their value could be NULL. They were not properly released and
4625 therefore they were never freed.
4627 * BUG: smb_packet_t data structures were not completely initialized
4630 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4631 Denied" because that causes the Explorer Shell to try again until
4632 access is obtained. Instead return "Remote Resources" which allows
4633 the shell to move on and treat the error as transient.
4635 * BUG: when initializing the NCBreturns structure, separate Event objects
4636 were created for each NCB although a single Event object was supposed
4637 to be shared by all.
4639 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4642 * BUG: smb_tran2Packet_t objects were not being properly referenced
4645 * BUG: directory path creation did not handle the case of multiple
4646 directories requiring creation in one attempt
4648 * BUG: SMB requests which required an Extended Response were ignored.
4649 This prevented some files from being written to AFS volumes.
4651 * BUG: character strings were being freed even after they were
4652 inserted into in use data structures
4654 * BUG: inconsistent usernames were used when High Security mode was
4655 enabled. (there is still much to do in this area)
4657 * BUG: pioctl() calls which require out of band RPC operations were
4658 susceptible to race conditions when performed by multiple processes
4660 * BUG: memory allocation and deallocation crossed instances of the
4661 C Runtime Library producing memory leakage and corruption in
4662 afscreds and the client configurator.