3 * Restore use of DNS AFSDB and SRV records by kaserver clients.
7 * Avoid a potential Freelance deadlock during initial execution
8 of afsd_service.exe if the old ini file data has to be
11 * Three rx library corrections. (1) Idle data connection
12 processing could timeout if the send window filled and
13 took longer than the idle data timeout period for the
14 transmit window to re-open. (2) The transmit queue
15 could be emptied prematurely. A required check for the
16 queue being in use was forgotten. (3) The function that
17 is supposed to implement a wait for the transmit queue
18 to cease being busy failed to wait.
21 * Restore use of DNS AFSDB and SRV records which were
22 unintentionally disabled in 1.5.69
25 * Add a context menu to the NetIdMgr AFS Provider
28 * Prevent an empty directory Btree from being created and
29 marked as valid if cm_BPlusDirBuildTree fails.
31 * cm_BPlusEnumAlloc should not fail if the enumeration
32 contains zero entries.
34 * In cm_BPlusXXX functions, return ENOMEM if malloc() fails.
36 * Do not leak the cm_scache_t dirlock if cm_BPlusDirBuildTree
39 * fs examine should report owner and group ids as signed values.
41 * Protect buffers in smb_WriteData from simultaneous writes.
43 * Query network interface MTU values from the registry because
44 the IP Helper GetAdapterAddresses() fails to report a manually
47 * Improve performance by not dropping and reacquiring the rx_call
48 lock in rx_WriteProc* and rx_ReadProc*.
50 * When configuring the rx library, the network interface MTU
51 must be applied to both the maximum send size as well as the
52 maximum receive size as is done when rx_SetMaxMTU is called
55 There was no 1.5.67 release
58 * Convert afsauthent.dll to use the xdr_alloc and xdr_free
59 routines from afsrpc.dll in place of malloc and free.
60 This permits the two libraries to be built with different
61 C runtime libraries without crashing.
63 * Set the DOS Readonly attribute on a file/directory
64 whenever the unix mode combined with the mask 0200
65 is true. Previously there was a discrepency between
66 the mask used for testing for readonly behavior and
67 that used for setting the attribute.
69 * Disable AFSVolSync based .readonly "whole-volume callback"
70 support because the all file servers prior to 1.5.67
71 (and perhaps 1.4.12) do not properly assign a value to
72 the AFSVolSync structure in bulk status RPC responses.
74 * Improve the error output from aklog to output the value
75 from krb5 error_message() if the afs_com_err output
76 indicates an unknown value.
78 * Hold a lock on the cm_scache_t object when the smb
79 server is assigning a new clientModTime value.
81 * Permit custom build version numbers to be set
82 without modifying the src/config/NTMakefile.<platform>
85 * Add support for binary code signing with cross-signed
86 certificates and arbitrary versions of signtool.exe.
88 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and
89 do not permit them to be exposed to the smb redirector.
91 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT
92 to avoid confusion within the smb redirector.
94 * Fix the byte order assigned to port numbers associated
95 with AFSDB record lookups. They must be network byte
96 order not host byte order.
98 * Add dynamic server ranking based on RPC round trip
102 * Official support for Windows 7 and Server 2008 R2
104 * Improved Rx RTT computation following the lessons
107 * Prevent a file server bug (FetchData returning an
108 invalid length instead of zero) from causing an
109 "unexpected network error" when writing to files.
111 * Promote DNS SRV records as superior to DNS AFSDB
112 records. Support arbitrary port numbers for vldb
115 * Add "fs listacl -cmd" support.
117 * Add AFSVolSync based .readonly "whole-volume callback"
118 support. With this functionality, multiple objects from
119 a .readonly volume can have their status validated by
120 issuing a single RXAFS_FetchStatus RPC.
122 * Remove drive mapping functionality and service start/stop
125 * Remove drive mapping functionality from afs_config.exe.
127 * Adjust SMB error return codes to avoid returning
128 STATUS_TIMEOUT which results in the SMB redirector
131 * Network Identity Manager OpenAFS Provider now provides its
132 own "AFS lock" notification icon to report the status of
133 "have tokens, have no tokens, service not started, service
134 started but inaccessible". Hovering over the icon lists the
135 cells for which tokens exist (if any) and the OpenAFS version
136 number. Double-clicking executes the Network Identity Manager
139 * When merging the current status for an AFS object, ensure
140 that the volume status for that object is set to 'vl_online'.
142 * Permit cm_Analyze to request a retry even when the cm_req_t is
143 flagged CM_REQ_NORETRY if the RPC failed due to RX_CALL_DEAD.
144 The retry will force the use of a new connection.
146 * Add digital signatures to resource only dlls as required by
149 * Prevent pioctl calls from retrying indefinitely when a sharing
150 violation error occurs.
152 * Add {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab"
153 to restore access to drive mapping functionality in afscreds.exe
156 * Permit Rx minimum peer timeout, maximum receive window size and
157 maximum send window size to be configured via the registry.
159 * The MSI installer now sets the ARPINSTALLLOCATION as required by
162 * DNS lookups set the port numbers in network byte order within
165 * Replace use of the 16-bit compatibility API WinExec with
166 ShellExecuteEx in afs_config.exe and afscreds.exe. WinExec
167 is incompatible with Windows 7 / Server 2008 R2 UAC
171 * Short circuit background volume checks if a shutdown
172 or suspend request is received.
174 * Avoid contacting the file server in order to allocate
175 an empty buffer beyond the current length of the file.
177 * When a request for streams on a directory or mount
178 point object is received, do not offer a default stream.
179 This was affecting the saving of roaming profiles.
182 * Fix Windows 2000 compatibility. Do not call rand_s()
183 even if it is compatible with the C run time library
186 * Fix a data consistency error between the output of
187 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
188 workstation group name "AFS", and report server name as
189 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
191 * Enable executables to be run from \\AFS on Windows 7.
192 Return "Name not found" instead of "File not found" when
193 a directory or file name cannot be found.
195 * Prevent cache manager from marking file server "down" when
196 the data returned in response to either RXAFS_FetchData64
197 or RXAFS_StoreData64 is invalid.
199 * Add pioctl data validation to the AFS Explorer Shell extension.
202 * After 1.5.62 afslogon.dll would no longer crash but it
203 also would not load "domain" specific configurations.
206 * An uninitialized variable was present in the symlink
207 recursion detection loop.
209 * Reverse addition of unique per file GUID during
210 Create/Open responses. Inclusion of the GUID prevents
211 Cygwin applications (and others) from accessing //afs.
213 * Treat "filename::$DATA" as a synonym for "filename".
214 This addresses the recent report of roaming profiles
215 failing to save back to the file server.
217 * When given a choice between an error from rx_Write()
218 or RXAFS_EndStoreData(), use the rx_Write() error.
219 This ensures that over quota and access denied errors
222 * If an error occurs during RXAFS_StoreData operations,
223 apply that error to all of the buffers in the BIOD.
224 This ensures that failures will not be retried if they
225 will only fail again.
227 * Do not perform a DNS lookup for a cellname if the
228 cellname prefix is _._AFS_IOCTL_._. This will avoid
229 DNS lookups when the query is for
230 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
232 * Fix Freelance root.afs volume object callback processing.
233 Object synchronization and status merging bugs.
235 * Fix Explorer Shell "invalid parameter" bug introduced
238 * Log RXKAD errors to the trace log.
240 * Prevent simultaneous pioctl calls from multiple processes
241 from stepping on each other's data.
243 * Prevent simultaneous pioctl calls from crashing the
246 * Add data validation to pioctl output processing in
247 ktc_GetToken and ktc_ListTokens.
249 * Add data validation to pioctl output processing in
250 fs.exe and symlink.exe.
253 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
254 No longer will browsing \\AFS produce truncated share names.
256 * Registry specified server preferences did not result in
257 the CM_SERVERFLAG_PREF_SET flag being applied to the
260 * Avoid unnecessary DNS lookups of share names as cells
261 when it is known that the name cannot be a cell name.
262 Any name that does not contain a dot is skipped.
264 * When processing Dfs Referral requests, do not return an
265 error if the server is down or busy; if the volume is
266 offline; the cell vldb cannot be reached; etc.
267 These paths are still \\AFS paths and so the client
268 should not be sent to search elsewhere for resolution.
270 * CRITICAL: Prevent the cm_Daemon thread from terminating when
271 the machine enters the suspend state. IF the cm_Daemon thread
272 dies, there is nothing to execute down server checks.
274 * Prevent use of smb_StartedLock before initialization which
275 can result in a panic in the lock package.
277 * Improve error checking in afslogon.dll in order to prevent
278 NULL pointer dereferences within GetLogonDomainOptions()
279 if the Lsa operations fail.
281 * Add support to the build system to automatically update
282 a Microsoft Symbol Store.
284 * Do not access the cm_conn_t after executing the cm_Analyze
285 loops. Doing so can result in an access to memory that has
288 * CRITICAL: If a StoreData request offset is not aligned on the
289 buffer module blockSize, (offset % blockSize) bytes of dirty
290 data will not be written to the file server.
292 * CRITICAL: If a StoreData64 request is sent to a file server
293 that does not support large files (>2GB) and the client is
294 unaware that the server has no large file support, the first
295 chunksize worth of data will not be written to the file server.
296 This bug was introduced in 1.5.3. It affects all IBM AFS file
297 servers and all OpenAFS file servers older than 1.4.0.
300 * If a file server becomes inaccessible while the cache manager has
301 dirty buffers to write, the afsd_service buf_IncrSync thread can
302 attempt to use 100% of the cpu.
304 * Fix the locking that protects the cm_cell_t lists
305 (all, name, id, free).
307 * Remove cm_cell_t objects from the name and id hash tables before
308 placing them on the free list.
310 * Fix "fs newcell" which was broken in 1.5.60.
312 * Do not attempt to synchronize dirty buffers if the associated
313 volume is known to be unavailable.
315 * Log to the Windows Event Log dirty buffers that cannot be written
316 during the shutdown process.
318 * Modify behavior of a Freelance mountpoint target that does not
319 specify a cell. Instead of assuming the target volume is in the
320 Freelance.Local cell, use the workstation "Cell" specified in the
321 registry. A mountpoint target of "#root.cell." will now mean the
322 root.cell volume in the workstation cell for the current session.
323 If the workstation cell changes from "athena.mit.edu" to
324 "andrew.cmu.edu", the referenced volume will also change without
325 requiring that the mount point targets be altered.
327 * Avoid false infinite loop errors when validating the cm_cell_t
328 free list during startup.
330 * Fix two more locations where xdr_free() should be used instead of
331 free() to permit a checked built of afsd_service.exe to work with
332 release builds of afsrpc.dll.
334 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
335 to permit the server Uuid to be used to lookup the server object
336 and from that determine the cell. This permits callbacks that are
337 received from alternate addresses to be processed with a known server
338 object. Previously a request from an unknown server would clear all
339 callbacks from all cells.
341 * Fix a bug that prevented optimal performance when using a non-zero
342 value for 'daemonCheckVolCBInterval'. As a reminder, when
343 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
344 volume callbacks are automatically renewed 90 minutes before their
347 * Fix automatic ranking of vldb servers whose values are obtained from
350 * Add logging for RX CALL DEAD errors which are generated if the remote
353 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
354 or the call is to a vldb server.
356 * Add registry based cell search functionality to NetIdMgr,
357 afs_config.exe, and klog.exe.
359 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
360 lookups on the host names in the CellServDB instead of using the
361 specified IP addresses. This provides aklog, pts, vos, etc. the same
362 CellServDB behavior that the Windows Cache Manager uses.
364 * Completely a behind the scenes change. The snprintf() routines have
365 been completely rewritten.
367 * When updating the stat cache entry callback of a .readonly object
368 from the volume group object, update the file server reference to
369 ensure it matches the most update to date callback.
371 * Add proper support for processing callbacks from multi-homed file
372 servers. Instead of comparing servers by cm_server_t pointer,
373 compare them by UUID when the UUID is known.
375 * During a shutdown short circuit the offline volume check daemon
378 * Properly track link counts during cm_Link and cm_Unlink operations.
380 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
381 preference to an error code reported by rx_EndCall.
383 * Add "PerFileAccessCheck" registry value to permit testing against
384 experimental file servers that include per-file acl support. This
385 value is intentionally undocumented. It is not to be used by
386 production environment deployments.
388 * Log a warning to trace log if a volume group is marked as a DFS
389 File Set. The Windows OpenAFS client does not include DFS support.
391 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
392 provider from obtaining tokens when referrals are in play.
394 * Add "fs chown" and "fs chgrp" commands to permit the owner and
395 group of objects stored in AFS to be set from Windows.
397 * Avoid performing background daemon operations when the machine is
398 going into suspend mode.
400 * Perform offline volume checks in most recently used order.
402 * Prevent crash when a data version for a cache object goes backwards.
404 * fs setquota, fs setcachesize, vos setfields, and vos create now
405 accept human readable orders of magnitude. (K, M, G)
407 * fs listquota fixed to permit large quota sizes to be displayed.
409 * rx packet resend and data packets sent counts were incorrect.
411 * Multi-thread safe library versions are now being generated and used.
412 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
414 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
415 values are now available on XP through Windows 7. Add functionality
416 to autodetect if such support is present on the machine. If so,
417 configure it if necessary and dynamically adjust the AFS Rx timeout
420 * Correct another race condition in the Rx library that could result
421 in an unexpected panic while freeing the Rx call iovq.
424 * A fix to the pioctl library to support drive substitution
425 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
427 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
428 that corrects a deadlock in the smb redirector and also adds new
429 functionality that permits the AFS SMB server to be given a longer
430 timeout than is normally the case. New functionality has been added
431 to configure these additional LanmanWorkstation\Parameter values.
432 (This functionality has been backported to XP SP3 and is scheduled
433 to be released on June 5th.)
435 * The BackConnectionHostNames registry value configuration was broken
436 when dynamic re-establishment of Netbios Name registrations was added.
439 * Hidden vos commands are revealed.
441 * If the "DisableLoopbackCheck" registry value is set, do not unset
442 it during the same service session.
444 * Reorganize code that prevents multiple Store operations to the same
447 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
448 permit broken symlinks to be treated as being in AFS.
450 * Fix vos commands that output 64-bit integer values.
452 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
453 Permits building cygwin applications against OpenAFS libraries.
455 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
456 as Windows HTML Help (.CHM) files.
458 * NSIS installer does a much better job of cleaning up files left
459 over from previous installs.
461 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
462 or "fs flushall" and on-going directory operations that can result in
463 afsd_service.exe crashing.
465 * Add support for DNS SRV records in place of AFSDB records.Â
466 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
467 Weight is currently ignored.
469 * Add a method of specifying Client CellServDB information within the
470 registry that can be used to either override the CellServDB file or
471 force the use of DNS lookups for a given cell.
473 The registry schema is as follows:
475 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
476 "LinkedCell" REG_SZ "[cellname]"
477 "Description" REG_SZ "[comment]"
478 "ForceDNS" DWORD {0,1}
480 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
481 "HostName" REG_SZ "[hostname]" Default: [servername]
482 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
483 "IPv6Address" REG_SZ "[address]" <future>
484 "Comment" REG_SZ "[comment]"
485 "Rank" DWORD "0..65535" Default: 0
486 "Clone" DWORD "{0,1}" <future: server only>
487 "vlserver" DWORD "7003" <future>
488 "ptserver" DWORD ... <future>
490 ForceDNS is implied non-zero if there are no [servername]
491 keys under the [cellname] key. Otherwise, ForceDNS is zero.
492 If [servername] keys are specified and none of them evaluate
493 to a valid server configuration, the return code is success.
494 This prevents failover to the CellServDB file or DNS.
496 Only one of "HostName" or "IPv4Address" is required. "HostName"
497 is optional if [servername] is the host name.
499 * Extend registry based CellServDB functionality to
500 afsconf_GetCellInfo() interface used by aklog and many
501 other command line utilities.
503 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
504 as it is now used by both client and server components.
507 * PriorityClass of afsd_service.exe process raised to
508 "High" to match the priority of the system services
509 that are waiting for afsd_service.exe to respond.
510 Configurable using [TransarcAFSDaemon\Parameters]
511 DWORD "PriorityClass" value. (use values from Win32
514 * Change error reporting of CM_ERROR_BADSHARENAME to
515 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
516 The error we were sending would cause the smb
517 redirector to drop the connection.
519 * Change the error reporting of CM_ERROR_ALLOFFLINE
520 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
521 Previous we would return Bad Network Path which
522 would cause the smb redirector to cancel all file
523 handles on the same "share".
525 * A debug version of afsd_service.exe can now be used
526 with a release version of afsrpc.dll (or vice versa).
527 afsrpc.dll now implements and exports xdr_free().
529 * Removed all unprotected OutputDebugString() calls as
530 those calls can block for up to 10 seconds if there
531 is contention for the globally shared memory or if
532 the debugger fails to read data in a timely manner.
534 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
535 supported for builds. The production releases still
539 * Discovered that XP/2003 supports a variant on recvmsg
540 called WSARecvMsg. The input parameters are mostly
541 compatible with the Posix recvmsg. This will save
542 multiple memcpy() calls per received packet.
544 * Discovered that Vista/2008 supports a variant on
545 sendmsg called WSASendMsg. The input parameters
546 are mostly compatible with the Posix sendmsg. This
547 will save multiple memcpy() calls per sent packet.
549 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
550 and utilizing rx_ReadProc32 from within the cache manager
551 as part of the FetchData operations. rx_ReadProc32 and
552 rx_WriteProc32 are optimized versions of rx_ReadProc and
553 rx_WriteProc for use when reading 32-bit integer values.
555 * The Network Identity Manager AFS provider now reports
556 progress events. These provide a better user experience
557 with Network Identity Manager 2.0.
559 * [RT 124293] A race condition exists which permits the
560 scp field of the an smb_fid_t object to become invalid
561 after a request on the smb_fid_t has begun. Remove the
562 race by protecting all access to the scp field with the
563 smb_fid_t mx mutex and obtaining a local reference on
564 the cm_scache_t object for the length of the request.
566 * [RT 124276] If the vldb is out of sync with the contents
567 of the file servers, afsd_service will retry too many
568 times when a file server reports a volume as not being
569 present. Now if the list reported by the vldb is the
570 same as the previously seen list, then the retry is
573 * [RT 124276] Read-only volume failover was broken in
574 1.5.53 whenever accessing a volume results in VNOVOL or
577 * [RT 124276] Prior to 1.3.70 the volume server reference
578 list was not reference counted and would be prematurely
579 freed while in use. When reference counting was added
580 in 1.3.70 a bug was introduced that could result in
581 service reference list corruption.
583 * Add Windows Application Event Log warning messages for
584 "Client SMB MPX value too large" and
585 "Client SMB Buffer Size too small".
587 * Renaming of files across directory boundaries would result
588 in an invalid handle error when attempting to access the
589 files after the move.
591 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
592 when a smb file descriptor cannot be found. The whole point
593 of using a Path Info function is because an smb file
594 descriptor wasn't allocated.
596 * More edge cases in which dynamic addition of Freelance
597 root.afs entries would get the wrong FID or where the
598 root.afs directory would not be refreshed.
600 * Buffer overflow could occur if the workstation cell name was
601 longer than 64 characters. Crashes could occur in
602 afscreds.exe, afslogon.dll, and afsd_service.exe.
604 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
607 * Improve consistency of the bad data version value used for
610 * Log file server uuid values as part of the cm_server object
611 when available. Dump the cm_server object list in response
614 * Optimize the performance of resetting access control lists
615 when tokens are set or removed.
617 * Fix vos.exe so that it can work with very large volume
620 * Remove symlink recursion tests and increase max symlink count
624 * Undo the "UAC manifest fix" applied to afs_config.exe.
626 * Fixed some additional race conditions in the Rx library
628 * Ensure that Freelance allocation of vnodes follow
629 the AFS convention of odd vnodes are directories
630 and everything else is an even vnode.
632 * Add Freelance logic to mount point and symlink
633 evaluation functions.
635 * Enhance smb_ParseASCIIBlock() so that it can handle
636 all of the STRING formats defined by the CIFS Technical
639 * Validate the output of smb_ParseASCIIBlock() in all
640 callers. Return CM_ERROR_BADSMB if the STRING field
641 cannot be parsed. CM_ERROR_BADSMB will cause the
642 contents of the packet to be logged.
644 * If multiple SMB Raw Write operations were taking
645 place at the same time, there could be data corruption
646 because unique event objects were not generated for
647 each Netbios receive operation.
651 * Fix a race in rxi_WriteProc which could result
652 in a freed or in use packet being passed to
653 rxi_PrepareSendPacket if the call is currently
654 in a transmit wait state. Passing a freed packet
655 will result in a panic. Passing an in use packet
656 might result in a call failure or perhaps data
659 * Kerberos referrals broke the NIM AFS provider
660 logic for automated configuration of the workstation
661 cell with a new identity. Now if a referrals
662 response is returned, assume the cell belongs to the
665 * The interface to buf_CleanAsync() and
666 buf_CleanAsyncLocked() has been changed to permit
667 errors to be propagated to the caller.
669 * During an enumeration of a directory, if the btree
670 has not yet been generated, return an error.
672 * When mapping CM_ERROR values to NTSTATUS values
673 do not treat 0 as an error.
675 * Avoid a cm_scache_t object reference count underflow
676 when renaming via the SMB interface.
678 * smb_IsMask() can be called with a non-8.3 input buffer.
679 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
680 input buffers and smb_IsMask() for non-8.3 input buffers.
682 * Add cm_buf_t error state to the "fs memdump" output.
684 * pioctl drive mapping checks need to take into account
685 Global Drive Mappings and SUBST assignments.
687 * Rename operations should not be permitted if the
688 destination name collides with an existing name that
689 differs only by case UNLESS the collision is the object
690 that is being renamed.
692 * Increase the maximum number of background daemon threads
693 to 64. Leave the default at 4.
695 * Set the FILE_UNICODE_ON_DISK flag when reporting the
696 file system characteristics.
698 * Report the maximum component name length as 255 even
699 for Unicode as part of the file system characteristics.
701 * Add RX debugging information to "fs memdump" output
704 * Refcount the cm_server_t objects returned by
707 * Fix the UAC manifest applied to afs_config.exe
709 * rxdebug now indicates if any RX_CONN flag is set
710 not just some of them
712 * Change the mapping of CM_ERROR_WOULDBLOCK from
713 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
717 * Includes #GCO Public CellServDB 07 Nov 2008
719 * Prevent separate cell entries from being created for
720 "foobar.org" and "foobar.org." which can happen if
721 an AFSDB record is found instead of a CellServDB
724 * Fix a panic when checking the SMB virtual circuit list.
725 If a VC is destroyed while the list is being checked
726 it is possible for the next object to become invalid.
728 * Increased logging to afsd_init.log of netbios name
731 * Fix a DNLC panic when a UTF-16 string that cannot be
732 converted to a utf-8 string is provided as a filename.
734 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
735 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
736 location to refer directly to the new location in
737 %programfiles%\openafs\client\program
739 * Be more aggressive about responding to Netbios
740 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
741 the LANA within the listener thread without performing a
744 * Use the WNet Enumeration of all connected drives as a filter
745 in the pioctl GetIoctlHandle operation. This should remove
746 the last situations in which there was a delay when opening
747 an Explorer Shell properties menu with the AFS Shell
750 * Fix the parsing of CellServDB in Windows to handle linked
751 cells as is done on Unix.
753 * Add support for linked cells to the cache manager,
754 netidmgr afs provider, and aklog. When two cells are
755 linked, a volume lookup that fails in one cell is looked
756 up in the other. This can be used for construction of
757 test cells that override specific volumes; cell renaming;
760 * Increase the length of cell and realm names that can
761 be entered in the NIM AFS provider.
763 * Actively prevent a 32-bit afsd_service.exe from
764 executing in the WOW64 environment.
766 * Protect against invalid IP address values in CellServDB
771 * Fix an Rx panic caused by an attempt to double free
772 an rx_packet when an error occurs while writing the
775 * Fix an exception cause by a race condition when
776 processing multiple file change notifications registered
777 by the Explorer Shell or other applications.
779 * Modify the Rx debug interface to use var_args and
780 permit Rx debugging to be used on 64-bit Windows
781 which previously had it disabled.
785 * Automated lock hierarchy enforcement added to the
786 osi base lock package. This permits detection of lock
787 violations that might but often does not result in
788 deadlocks. Enabled by default in the checked builds.
789 May be activated in release builds via the registry
790 value "LockOrderValidation" in the TransarcAFSDaemon
793 When a violation is detection an assertion is thrown
794 generating a dump file and a message indicating which
795 locks were obtained out of order is logged to the
796 Windows Application Event Log.
798 * Attempt to ensure that a dump file will be generated
799 whenever the service panics.
801 * Removed unused lock cm_bufGetMutex
803 * Many out of order lock acquisitions were detected
804 using the lock order verifier and were corrected.
806 * A race condition within the function path cm_GetSCache() ->
807 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
808 cm_scache_t object to be issued simultaneously to two
809 threads. This would eventually result in a panic due
810 to the resulting undercount.
812 * Fix interpretation of the empty string as the ioctl path
813 to mean the current directory. "fs lsm", "symlink list",
814 etc. now return a "not a ..." error instead of "does not
817 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
818 to indicate that the input string is a nul terminated string.
819 Assign ti when input strings are nul terminated instead of
820 length counted. (from 1.5.50)
822 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
823 does not match the observed behavor. The 'nameLength' is specified
824 in bytes, not in characters. Fix the implementation to match
825 the observed behavior. (1.5.50)
827 * The cm_HaveAccessRights() attempt at deadlock avoidance by
828 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
829 does not avoid the deadlock. Avoid the deadlock by enforcing the
830 lock order of the lowest vnode first. Then remove the infinite
831 loop avoidance in cm_SyncOp() that was returning an unwarranted
834 * File Server lock synchronization was not propertly enforced in
835 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
837 * Another change to the short name generator. Plus and equal are
838 not valid characters in a short name according to the
839 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
840 with left and right brace in order to minimize the impact.
842 * Fix a race condition that can result in a heap overwrite error if
843 the number of known file servers or vldb servers increases during
844 the preparation phase of a cm_CheckServers() operation.
846 * If the SMB Authentication would fail, a LSA allocated memory
847 buffer would be leaked.
849 * Rewrite the cm_cell_t allocation code to permit cell vldb server
850 lookups to fail without leaking cm_cell_t objects.
852 * Add "fs setcrypt auth" to permit the use of rxkad authentication
853 only mode. Not documented.
855 * Initialize Rx mutex and lock objects once and only once as part of
856 the DLL load process.
858 * When reporting the Volume Name "AFS" do not include the terminating
859 NUL in the reported length.
861 * Refresh Volume Location information upon receipt of a VNOVOL error
862 from the file server and avoid a race condition that prevented the
863 updated volume location information from being destroyed immediately
864 after it was acquired.
866 * In response to VNOVOL and VMOVED, remove the volume id from the server's
867 volume list so that the volume's up and down status is not affected
868 when the machine is shutdown.
870 * Many fixes to the Rx stack related to packet allocation and deallocation;
871 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
872 by overwriting the rx_call currentPacket pointer when it was already
873 referring to a packet.
875 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
876 in the TransarcAFSDeamon/Parameters key.
878 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
879 constructed directory B+ trees for flushed directory objects.
881 * When cached pages for directory objects have been locally modified, do
882 not permit the locally modified pages to be mixed with pages subsequently
883 obtained from the file server. Doing so can result in directory
886 * If Ubix errors are received in response to VL_xxx operations record
887 the error on the server instance and permit the operation request to
888 be retried in case another server is able to answer the query. We have
889 seen at least one instance of a cell in which a VL server returned
890 UNOQUORUM (possible database truncation) when the other servers were
891 responding successfully.
894 * Optimize the assignment of write errors to dirty
895 buffers by applying the first error to all buffers
896 belonging to the same object instead of trying each
899 * Add resource version info to xstat_fs_test and xstat_cm_test
901 * Fix a delete vs delete[] usage problem in the WiX custom
904 * When creating a mount point, validate the volume name to
907 * When processing a mount point, pay attention to VL_BADNAME
910 * Remove a race condition from cm_volume_t recycling that could
911 have resulted in two threads making use of the same object at
914 * Fix a reference count leak on smb_vc_t objects introduced in
915 1.5.50. Permit the objects to be freed once again.
917 * Fix the output of "fs memdump" for smb_vc_t objects that was
920 * Add routines for debugging smb refcount errors.
922 * When shutting down the cache manager, do not permit AFS
923 callback requests to be processed. Doing so can result in
924 invalid memory accesses.
926 * Do not recycle a deleted cm_scache_t object unless the
927 refCount is 0. Doing so can result in a deadlock and
928 invalid memory access.
930 * Minimize the number of RPCs sent to the file server for
931 directory bulk status requests. Only request status for
932 objects that we are actually reporting on and only if we
933 do not already have status for the object and then make
934 sure that we return as much as we can at once via the SMB
935 interface but not so much that it exceeds the RDR timeout
937 * Enable bos restricted operations
939 * Fix the creation of submounts used by the afscreds and afs_config
942 * Avoid repeated retries when attempting to obtain access rights.
943 If the access rights were obtained successfully but do not
944 satisfy the request, fail the request.
946 * Fix error code reported when attempting delete a file on a
947 readonly volume or one that is marked with the readonly DOS
950 * Fix a short name truncation bug.
952 * Fix a heap corruption error when reading the CellServDB file
955 * Fix directory listings when the directory has "lookup" permission
956 only and the directory has never been previously accessed.
958 * Fix several tracking issues for cm_scache_t rw locks.
960 * Ensure that all buffers used to store cell names are the
963 * Restore Windows 2000 support (broken in 1.5.50).
964 LOCALE_INVARIANT was introduced in XP.
966 * Add RxUdpBufSize registry value. Default is 256K.
968 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
970 * Do not include trailing NULs when returning file names in
971 a directory search response.
973 * Prevent RX from crashing the afsd_service by attempting
974 to allocate to many buffers at once when memory is fragmented.
976 * Avoid a deadlock when freeing cm_server_t objects
978 * Optimize search cell file. Do not search to end of file
979 if the cell has already been found.
981 * Remove explorer shell extension delay on local disks.
983 * Automatically pre-allocate 64 Rx packets per thread
985 * Add smb lock requests and stat cache lock allocations to
986 the output from fs memdump.
989 * Fix Large File Support (broken in 1.5.50)
991 * Fix "fs setacl" (broken in 1.5.50)
993 * Fix pioctl path evaluation of drive letter
994 mapped paths. (broken in 1.5.50)
996 * Fix installation of VC8 Runtime library in the EXE
997 Retail installer. (broken in 1.5.50)
999 * Avoid a potential deadlock during Volume Location
1000 updates if all the file servers are unreachable at
1001 the time the VL update succeeds.
1003 * Ensure that rx connection object is references
1006 * Fix propagation of over quota errors during cm_FSync()
1009 Since 1.5.39 [1.5.50 released 16 July 2008]
1010 * There were no releases between 1.5.39 and 1.5.50.
1012 * FIRST RELEASE WITH UNICODE SUPPORT
1014 * Pioctl interfaces to the cache manager refactored
1015 to provide layering between the SMB specific code
1016 and the general purpose ioctl operation.
1018 * Garbage collect dead SMB virtual circuits as soon
1019 as they are no longer being referenced. This avoids
1020 problems with outstanding locks not being dropped
1021 when the virtual circuit becomes invalid.
1023 * Remove the IBM Administration Reference documentation
1024 and replace it with the OpenAFS Command Reference
1027 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
1028 functions each time a connection is about to be used.
1029 Do not hold a lock on the rx connection object while
1030 it is being selected. This avoids a race between
1031 threads attempting to set the timeout values and
1032 removes a bottleneck that was hampering performance.
1034 * Ensure that the smb directory attribute is set
1035 for all directory objects.
1037 * Add vs2008 support to the NSIS installer scripts
1039 * Replace the VC Runtime EXE installer with the MSI
1040 installer in the NSIS installer scripts
1042 * Properly delete the VC7.1 C Runtime libraries.
1044 Since 1.5.36 [1.5.39 released 23 June 2008]
1045 * There were no 1.5.37 or 1.5.38 releases for Windows
1047 * Use Visual Studio version of vsprintf() instead of
1048 home grown version that resulted in linking conflicts
1050 * Prevent buserver, ptserver and fileserver from crashing
1051 due to non-portable strftime() format strings.
1053 * Prevent server crashes caused by rx_InitMorePackets()
1054 being called prior to allocation of all of the required
1055 synchronization objects.
1057 * Fixed a memory leak in the internal B+ tree directory.
1058 For objects with non-8.3 names, the 'longname' memory
1059 allocation would be leaked during object deletion or
1062 * Fixed a memory leak in the background daemon cm_CheckServer()
1063 routine. The memory allocated to store the server list
1066 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
1068 * Correct an abstraction layer violation. cm_scache_t objects
1069 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
1070 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
1071 performed in cm_ReleaseSCache() when the reference count hits zero.
1073 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
1075 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
1078 * Ensure that scache object 'nextp' pointers are set to NULL
1079 when the objects are removed from the hash table linked list.
1081 * If a dirty buffer cannot be written to the file server because
1082 of an access denied error, mark the error in the cm_buf_t
1083 structure so that it can be dealt with instead of retrying
1087 * Update CellServDB file to GCO Public 23 Apr 2008
1089 * Fix a cm_buf_t reference count leak when attempts to write
1090 dirty buffers to the file server from within cm_IncrSyncer()
1093 * Prevent udebug from crashing.
1095 * Activate RX Idle Timeouts. If the file server is busy for more
1096 then 30 seconds, attempt to failover to another server without
1097 marking the busy server down.
1099 * Another VNOVNODE issue fixed. When writing a dirty buffer
1100 to the file server, if VNOVNODE is received, mark all buffers
1101 as invalid without further attempts to contact the file server.
1103 * Improved performance on high latency links.
1106 * Fix a bug preventing the re-initialization of the Freelance
1107 root.cell contents when a change is made.
1109 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
1110 or mountpoint that does not exist.
1112 * aklog, afscreds, integrated logon, and the NIM AFS credential
1113 provider will all attempt to first acquire an afs service ticket
1114 using the service principal: afs/<cell>@<CLIENT_REALM>. This
1115 permits the proper detection of the cell realm when multiple
1116 realms are local realms for the cell.
1118 * Fix fs commands that re-use file identifiers instead of forcing
1119 a path evaluation each time. This was broken in 1.5.33 when
1120 hashes were added to the cm_fid_t structure.
1122 * Fix symlink and mountpoint removal and then recreation when
1123 applied to the fake Freelance root.afs volume.
1125 * Fix SMB change notifications for the Freelance root.afs volume
1126 when creating or removing symlinks and mount points.
1128 * Force DNS AFSDB cell name lookups to lower case if the DNS
1129 response is not all lower case.
1131 * 32-bit installers are now built with Visual Studio 2005
1133 * All vos.exe commands now support the "-noresolve" option which
1134 forces the display of IP addresses instead of DNS hostnames.
1136 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
1137 permits operations that do not require up to date volume
1138 location data to continue while waiting for the RPC to complete.
1139 Multiple threads waiting for a volume location update on the
1140 same volume set no longer trigger multiple calls VL_xxx RPCs.
1142 * Volume server references could be marked offline and never
1143 reset to the not_busy state.
1145 * Further optimizations to cm_Analyze() retry logic for
1146 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
1149 * Optimize the DNLC by applying Interlocked incrementing for
1150 statistics gathering and enabling greater use of read
1151 locks instead of write locks
1153 * Further optimize osi_Log operations when logging is disabled
1154 by removing the need for the function call overhead.
1156 * Add a new lock operation, Convert Read to Write, and apply
1157 it throughout the cache manager. This function provides a
1158 fast transition from a read lock to a write lock when the
1159 caller is the only reader.
1161 * Fix a timeout problem when opening files or request locks
1162 when the file in question is already write-locked by another
1165 * Fix a deadlock that can occur if "fs flushall" is executed
1166 while applications are writing to AFS.
1168 * The cache manager would re-read data from the file server
1169 after file truncation when obtaining buffers to write to the
1170 same file until the first StoreData RPC completes.
1172 * Fixes error handling in cm_NTCheckDelete()
1174 * Replaces cm_scache_t mutex with a read-write lock permitting
1175 additional parallel access.
1177 * Prevent an cm_scache_t reference undercount when evaluating
1178 symlinks containing @sys when none of the active sysnames
1179 match any of the available targets.
1181 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1184 * Set RxMaxMTU registry value to 0 which is the equivalent of
1185 it not being set at all.
1187 * Reduce contention between BkgDaemon threads
1189 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1192 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1193 objects to use the faster Interlocked operations which avoid
1194 the need to hold write locks.
1196 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1197 non-fatal error that forces a retry against another service
1200 * Fix a crash in smb_WriteData that would occur if an application
1201 opened the file as read-only and then sent a write request anyway.
1204 * Respond to VL_NOENT errors by removing cm_volume_t object and
1205 setting it to be the first object to be recycled. This is not
1206 only a more efficient use of resources but it also prevents
1207 repeated attempts to query the VLDB server for the volume from
1208 the CheckOfflineVolumes routine.
1210 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1213 * A first cut at a cache manager statistics monitor that can be used
1214 to determine the necessary cache parameters to support the
1215 application working set.
1217 Off by default, the performance package can be activated by setting
1218 "daemonPerformanceTuningInterval" in the service Parameters key.
1219 As with the other daemon interval values the unit is in seconds.
1221 At service start and each succeeding interval the cache manager
1222 will write statistics to %TEMP%\afsd_performance.txt showing the
1223 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1224 The FID statistics keep track of all FIDs seen by the cache manager
1225 during the service session whether or not they are backed by any
1226 live object in the cache.
1228 These statistics are not stored in the cache file.
1230 * Fix windows event logging broken in 1.5.33.
1232 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1233 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1234 as well as when computing the server preference list.
1236 * When merging status after a StoreData operation prevent data
1237 buffers from being discarded that we would prefer to keep.
1238 (Bug introduced in 1.5.33.)
1240 * Modify installers the set RxMaxMTU to 0.
1242 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1243 dialog, the Background Daemons and Service Threads fields were
1246 * Increase default number of Background Daemons to 4.
1248 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1249 cm_volume_t objects to be recycled.
1251 * Windows Error Reporting has shown a number of previously unseen
1252 stack corruption errors in the Explorer Shell extension. The
1253 minidump appeared to indicate the problem was in ParseAcl().
1254 Reviewing the function it appears that numerous things could go
1255 wrong if invalid input was provided. Added error checking and
1256 data validation. Hopefully this will address the problem.
1258 * Windows Error Reporting has shown crashes in aklog as a result
1259 of krb5_cc_get_default() failures not being handled. They are
1263 * The Rx library used a 32-bit type for sockets which was
1264 truncating the socket value on 64-bit Windows. This
1265 was not an issue in general but became an issue when
1266 Rx "hot thread" support was enabled.
1268 * Enable "Rx hot thread" support which ensures that while
1269 a received message is being processed that a new thread
1270 is allocated to receive the next message to arrive.
1272 * Increased parallel access to smb_rctLock protected
1273 resources. Reduce the number of times that the
1274 lock is dropped and re-obtained.
1276 * Increased parallel access to cm_scache_t and cm_buf_t
1277 resources by replacing the previous global write lock
1278 protected reference counting mechanism with one built
1279 upon the use of InterlockedIncrement and InterlockedDecrement.
1281 * Implement a hash for cm_fid_t to reduce comparison costs
1282 and inline the cm_FidCmp() function.
1284 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1285 hash which has a better distribution
1287 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1288 the most recent read request instead of the next chunkSize.
1289 cm_chunkSize can be dozens or hundreds of buffers. As a result
1290 too much time is spent performing the evaluation.
1292 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1293 lock is to prevent the creation of new buffers while a truncation
1294 is being performed. All references to bufCreateLock have been
1295 removed except in two places:
1296 i. a write-lock surrounding the function that calls buf_Truncate()
1297 ii. a read-lock within buf_GetNewLocked() that actually allocates
1300 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1301 CM_CONFIG_DATA_VERSION number which value be used to force the
1302 replacement of the cache file contents when incompatible changes
1303 are made between releases.
1305 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1306 cm_SyncOpCheckContinue
1308 * Avoid calls to multi_Rx if nconns == 0
1310 * Modify smb_WriteData to perform background writes based upon the
1311 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1312 boundaries. This will slow down writes from the SMB interface but
1313 will avoid the risk of the CIFS client disconnecting from the AFS
1314 client SMB server. The default value for 'smb_AsyncStoreSize'
1317 * SMB AsyncStore functionality is on by default but can be
1318 disabled using the registry value:
1320 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1322 smb_AsyncStoreSize can be adjusted with
1324 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1326 The value must be a multiple of the buffer block size
1327 and cannot be larger than the chunk size.
1329 * Do not leak a read lock on the bufCreateLock
1331 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1334 * In MergeStatus, only remove clean buffers from the hash table.
1336 * Avoid the need to update the data version number on each buffer
1337 associated with a scache when MergeStatus is called after a
1338 StoreData by maintaining a range of valid data versions as part
1339 of the cm_scache_t object.
1341 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1342 from cm_HaveAccess... We can safely avoid the call so do so.
1344 * Add "ForceAnonVLDB" service parameters registry value. Default
1345 is 0. Set to 1 to force the use of anonymous connections to
1346 the VLDB servers. This is useful when using krb5 tokens and the
1347 VLDB servers have not been updated.
1350 * Do not attempt to search for a cell name that matches one
1351 of the special share names "svrsvc", "wkssvc", "ipc$"
1353 * Issue permission denied warnings from afscreds if attempts
1354 to modify the startup shortcut or registry values fail.
1355 This happens under Vista UAC.
1357 * Do not return BAD_NETWORK_ERROR in response to attempts
1358 to create directories or files in the Freelance root.afs
1359 volume. Instead return ACCESS_DENIED.
1361 BAD_NETWORK_ERROR was being sent because the Freelance cell
1362 has no vldb servers. Short circuit the error by testing for
1363 freelance in CreateFile and MakeDir
1365 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1366 to be opened and closed without actually existing. For now
1367 we treat them as equivalent to ioctls but this should
1370 * Fix RAP processing to not reject names that exist in the
1371 root.afs volume for GetInfo queries.
1373 * In afscreds, use GetUserNameEx to obtain the username instead
1374 of searching the registry. Try UserPrincipalName first and
1375 fallback to SamCompatible.
1377 * Stop linking to rpcns4.dll which is no longer supported on
1378 Vista and 2008. TaAfsAdmSvr is going to have to be
1381 * When processing dfs referral requests, if the sharename is
1382 a partial match and therefore will not be added automatically
1383 to the share list, make sure that we return an error.
1385 * Avoid a crash when logging VNOVOL errors returned from
1386 cm_CheckOfflineVolume.
1388 * When trace logging is disabled, do not permit osi_LogSaveString
1389 to perform the string copying. By disabling the string
1390 copies, throughput of the service more than doubles.
1392 * when checking offline volumes, update the cell vlserver info
1393 if required and don't perform a check if there are no servers
1396 * Add volumes in the vl_unknown state to the list of objects to be
1397 checked as offline or retried when an allbusy error is received
1400 * Add support for configuring rx_ExtraPackets
1402 * All calls to buf_Get() must be protected by a read lock on the
1403 cm_scache_t bufCreateLock
1405 * When a MergeStatus results in the invalidation of buffers, do
1406 not remove buffers from the hash tables that have active references.
1408 * Upon receiving a VNOVNODE error from a file server, be sure to
1409 invalidate any file locks. This prevents on-going attempts to
1410 extend the locks and will cause subsequent attempts to access
1411 the file to receive a invalid handle error.
1413 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1414 to directory entry names.
1416 * No longer discard tokens when RXKADBADTICKET error is received.
1417 Instead return STATUS_NO_KERB_KEY error.
1420 * Add "cmdebug <host> -cellservdb" option and client service
1421 implementation permits querying the vldb server for each
1422 cell known to the cache manager. The output is in a form
1423 that permits it to be used as a CellServDB file.
1425 * Add Symlinks->Show option to the AFS Explorer Shell Context
1428 * Prevent buffer overrun in the cmd parser used by command
1429 line tools such as fs, vos, pts, etc.
1431 * The path ioctl operations have several issues:
1433 (1) the specified path for "fs examine, whereis, whichcell,
1434 flush" and so always has the follow symlinks and mount
1435 points semantics. This makes it impossible to determine
1436 what the FID of a symlink or mount point is.
1438 (2) "fs examine" out is not the result of a single pioctl
1439 operation but is actually the combined output of half
1440 a dozen operations. Path evaluation is an expensive
1441 operation. It would be faster if the caller could
1442 evaluate the FID first and then perform all of the
1443 rest of the operations by specifying the FID instead
1446 (3) fs output reports all objects as files. By adding a
1447 GetFileType pioctl more informative output can be provided
1448 that indicates what type of object the path evaluates to.
1450 (4) the Windows fs command includes a number of commands that
1451 do nothing but exist only because the Unix cache manager
1454 A new extendible data structure cm_ioctl_query_opts_t has been
1455 added which can be optionally specified with pioctls that do not
1456 already require input data. The first two fields of this structure
1457 are 'literal' and 'fid'. The literal field is used to indicate
1458 whether the last component of the path should be evaluated following
1459 symlinks and mount points.
1461 The fid field permits a fid to be specified.
1463 * A new GetFileType pioctl has been added. The type of objects are
1466 * A new "-literal" option is available for "fs examine, flush, whereis,
1469 * Unimplemented fs commands have been removed.
1471 * There are circumstances where a volume object is being accessed and
1472 the volume is marked indicating that the volume location information
1473 is out of date but where it is also pointless and perhaps dangerous
1474 to block waiting for the rpc to complete. One example is when
1475 processing the cmdebug requests. If we know that we are not going
1476 to use the volume object to contact a server, then we can now set
1477 the CM_GETVOL_FLAG_NO_RESET flag.
1479 * when moving up in the directory tree we search the recorded fid list
1480 to find a matching fid that we have already crossed. we must also
1481 reset the fid count based upon what we discover.
1483 * Cell names are published as share names. As a result they are searched
1484 for as part of the Dfs Referral evaluation. If share "foo" can not be
1485 reached, the CIFS client will fallback to searching for "fo" as well.
1487 Since the freelance client automatically adds a symlink for prefixes
1488 of the cell name, this results in both "foo" and "fo" being added to
1489 the freelance root.afs volume. This patch prevents that negative side
1492 * when tracking Kerberos credential cache names be sure to include
1493 the cache type prefix.
1495 * Microsoft has assigned OpenAFS a network type value
1497 * The algorithms that were used to produce a Netbios name were broken.
1498 The name that was produced when the hostname was too long was in fact
1499 longer than the maximum permitted netbios name.
1501 Also, the "NetbiosName" value was not used as a suffix when the loopback
1502 adapter was not installed. The hardcoded string "AFS" was used instead.
1504 * avoid another deadlock during server probes initiated by the ipaddr
1505 change daemon thread.
1507 * prevent the afs client service from crashing when fs newcell is
1510 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1511 to \Program Files\OpenAFS\Client\Program in order to get them out
1512 of the %windir% tree
1514 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1515 end STATUS_OUT_OF_QUOTA can be returned to the application.
1517 * Modify the search order for determining the location of CellServDB
1518 and other client configuration files.
1520 1. AFSCONF environment variable
1522 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1523 4. use Program Files\OpenAFS\Client
1525 * More Vista Power Management tweaks
1527 * Close a socket handle leak in the kauth routines
1529 * Don't hold a cm_volume_t mutex across RPCs
1531 * Convert downTimes to use clock time instead of relative times.
1532 This permits correct comparisons with expiration times.
1534 * Update the Wix installer to properly install the prerequisites
1535 for translate_et.exe
1537 * Do not translate pioctl paths as they are not converted by the
1540 * Implement multi_rx probes for checking whether servers are up
1541 or down. Turns probes into a constant time activity instead
1542 of one dependent upon the number of known servers.
1544 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1545 update of the volume location data and retry. VNOVOL means
1546 that the volume could not be attached. It doesn't mean that
1547 the volume is not on this server.
1549 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1550 not attempt to validate the target path. There is no reliable
1553 * Improve performance of \\afs\<cell><mount-type><volume>\
1554 access by eliminating extraneous attempts to search the
1555 CellServDB file and DNS AFSDB records. Share names
1556 containing '%' or '#' are not valid cell names.
1558 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1559 to unlock a range that is not currently locked.
1561 * Enable the 32-bit tools to read the 64-bit registry keys
1562 when executing in the WOW64 environment.
1566 * Fix buffer refcount leak introduced in 1.5.29
1568 * Prevent the NIM AFS Provider from crashing when configured to use
1569 Kerberos v4 for token acquisition in situations where no Kerberos v4
1570 support is available. For example, 64-bit KFW or sites that delete
1571 the Kerberos v4 DLL.
1573 * In afsd_init.log, support logging cache sizes greater than 2GB.
1575 * The volume status output of "fs examine" was being determined based
1576 upon the wrong error value. Instead of using the pioctl() return
1577 code, the errno value must be used.
1579 * Prevent invalid pts auto-registration attempts by aklog and the
1580 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1582 * Add synchronization protection to all of the SMB Listener State
1583 variables. This prevents race conditions when the addition or
1584 loss of a network adapter takes place.
1586 * On Vista, there is a race condition between the restart of the
1587 afs client service and the network adapters. Ensure that the
1588 afs client service can handle switching between loopback and
1589 non-loopback modes of operation.
1591 * Fix FollowBackupPath to work with mount points that are not
1592 explicitly read/write.
1594 * Fix a deadlock that has been observed on Vista while resuming
1595 after at least four hours of sleep.
1597 * Add new HKLM registry configuration options that can be used
1598 to provide configuration hints to the NIM AFS Provider when
1599 creating new identities. See release notes for details on
1600 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1603 * Speed up write buffer operations by holding a global read-lock
1604 instead of a write-lock when updating the buffer data version
1605 number. The data version is protect by the associated stat
1608 * Speed up read buffer operations by removing out of date buffers
1609 from the buffer hash tables.
1611 * Protect against a null pointer dereference in afslogon.dll
1612 if QueryAdHomePathFromSid() fails.
1614 * Remove a race condition while accessing private freelance data
1617 * Update the buffer data versions when Freelance directory data
1620 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1623 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1624 instead of returning zero length.
1626 * Return ERROR_RETRY when a server sends VBUSY.
1628 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1629 a volstat plug-in is in use.
1631 * Enhance the volstat interface to support Dfs path mappings.
1633 * Fix "symlink list" to do the right thing for Dfs links
1635 * Fix callback expiration that was broken in 1.5.27.
1637 * Fix a crash at shutdown caused by the lack of synchronization
1638 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1640 * Hold the cm_server_t mutex while walking or modifying the
1641 cm_server_vols_t list in order to avoid a race condition.
1643 * Ensure that the cm_server_t refCount is safely modified.
1645 * The pthread library implementations of pthread_cond_wait
1646 and pthread_cond_timedwait were failing to return with
1647 the mutex held when error conditions occurred. This could
1648 result in race conditions wherever they are used. One case
1649 in which problems were seen was the Rx event_handler().
1651 * Prevent cm_GetBuffer() from locking all buffers associated with
1652 a stat cache object before deciding if the current buffer is up
1655 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1656 object for each offset that we will attempt to reserve.
1658 * Reduce the default number of background daemon threads used to
1659 prefetch or store to one in order to prevent two or more threads
1660 from stepping on each other.
1662 * Reduce the default chunk size to 256KB in order to prevent read
1663 timeouts on low bandwidth connections.
1665 * Fix a race condition that could result in threads becoming stranded
1666 when synchronizing multiple requests on the same stat cache object.
1669 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1670 has been disabled. It has been disabled because it turns out
1671 that the RPC is improperly implemented in all file servers
1672 previous to 1.4.6. The RPC is executed without the proper
1673 locks being held resulting in data corruption and eventual file
1674 server failure. The functionality will be re-enabled in a
1677 * a couple of .readonly Volume Callback optimizations were
1679 - Apply most recent volume callback to all stat cache objects
1680 in the volume. This avoids premature callback expirations
1681 and unnecessary FetchStatus calls.
1682 - Add option to permit attempts to renew .readonly Volume
1683 Callbacks on a periodic basis to avoid invalidation of stat
1684 cache data. [Requires daemonCheckVolCBInterval be set to
1685 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1687 * Adds support for 64-bit data version values. Previous releases
1688 only supported 32-bit values. Larger values sent by the file server
1689 would result in rollover.
1691 * aklog and asetkey no longer generate an exception if Kerberos for
1692 Windows is not installed.
1694 * If the LSA authentication functions report a temporary out of memory
1695 condition, fail the current authentication and do not cause the AFS
1696 client service to terminate unexpectedly.
1698 * When the AFS client service is shutting down, prevent the daemon check
1699 thread from executing additional checks which might attempt to access
1700 discarded locks or unmapped memory.
1702 * Correct a double free condition in the NIM AFS credentials provider.
1704 * Add "FollowBackupPath" service parameter registry value.
1705 This provides equivalent functionality to the UNIX afsd -backuptree
1706 option. When set to a non-zero value, normal mount points originating
1707 in a .backup volume will prefer a .backup volume.
1710 * Fix NSIS uninstall from the .MSI installer.
1712 * Add Stack Frame logging for AMD64
1714 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1715 server returns EWOULDBLOCK, retry the request every two
1716 seconds for up to the RDRtimeout.
1718 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1719 is not assigned the same value as EIO.
1721 * Update CellServDB to GCO Public 25 Oct 2007
1723 * Add additional validation and error handling code after
1724 each call to getSlot(). If an invalid slot is returned,
1725 return NONODE. If the invalid slot is returned when
1726 extracting a data node, invalidate the tree.
1728 * Modify compareKeys() to always perform a case-insensitive
1729 comparison and only perform a case sensistive comparison
1730 if the case-insensitive one matches. This ensures the
1731 ordering is consistently reported.
1733 * Add lock assertions to ensure that all calls are being
1734 performed with the correct locks being held. There have been
1735 some crash reports that provide stack data that does not appear
1736 to be possible unless there is a race. However, there are no
1737 obvious locations where the race is taking place and the test
1738 suite indicates that all of the correct locks are being held.
1739 We shall see what happens in the field.
1741 * For consistency replace all calls to findKey in which the range i
1742 s (1,numentries) with calls to getSlot().
1744 * Optimize the depth search loop by testing the slot value in the
1745 for statement instead of forcing the loop to be broken later.
1747 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1748 There are a number of locations where locks are obtained, dropped,
1749 and reobtained. This reorganization attempts to accomplish several
1752 (1) be optimistic for the most common case so it will be fast
1754 (2) add consistency checks after each location where locks are
1755 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1756 and the bplus tree is out of date, retry until we get to a
1757 consistent state that we can use.
1759 (3) Ensure that all operations take place with the correct locks.
1761 * One of the issues that has become a serious problem since the
1762 addition of the local directory updates is that although cm_SyncOp
1763 synchronizes operations, it does not preserve the order of requests.
1764 This has always been a problem in that it has been possible for a
1765 request to fail to complete due to its worker thread's bad luck.
1766 When a request takes longer than the Windows SMB Redirector's timeout,
1767 the SMB Redirector tears down the SMB virtual circuit.
1769 When using the local directory updates it is really important that
1770 the directory update operations complete in the order that they were
1771 sent to the file server. If they don't, then the local directory state
1772 and the file server state will not match and the local directory state
1773 must be discarded which in turn forces a new read of the entire
1774 directory contents over the network.
1776 This patch adds a new cm_scache_waiter_t object that is used to store
1777 the current thread, buffer, and syncop flags within a waiters queue
1778 on each cm_scache_t object. If a thread is forced to sleep in
1779 cm_SyncOp, upon waking it will check to see if there are any other
1780 threads waiting that are attempting to perform a similar task ahead
1781 of it in the queue. If yes, the thread goes back to sleep. If not,
1782 it goes ahead and enters the cm_SyncOp conflict resolution block.
1784 This patch has the additional side effect of reducing the number of
1785 competing threads that must obtain the cm_scache_t mutex and process
1786 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1787 utilization of the service and the clock time associated with processing
1788 requests will be reduced.
1790 * assert that the cm_scache_t mutex is held when calling
1791 buf_ForceDataVersion().
1793 * rename findNode to leafNode in bplus_Lookup
1795 * replace all OutputDebugString calls with osi_LogX calls
1797 * modify bestMatch to special case the return values for leaf nodes.
1798 If an entry is above or below the values available in the leaf node
1799 return BTLOWER or BTUPPER instead of BTERROR.
1801 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1802 if true convert to either slot 0 or Max and perform the insertion.
1803 This produces easier to read code when performing lookups.
1805 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1806 descriptive message.
1808 * the difference between osi_assert() and osi_assertx() is whether or not
1809 a description is specified as part of the call. When no message is
1810 specified afsd_notifier() is called with a NULL msgp parameter. This
1811 results in a null pointer dereference during the EventLog() call.
1813 As a result, none of the information describing the assertion is written
1816 This commit sets a default message for use when no other message is
1817 provided and it replaces all calls to osi_assert with osi_assertx and
1820 * Migrate B+ tree search key into thread local storage
1823 * Fix an uninitialized variable in lana_IsLoopback() which
1824 could result in a random determination that a loopback
1825 adapter is not a loopback adapter. This would result in
1826 a computed Netbios service name of %MACHINE%-AFS instead
1827 of AFS. This is a serious problem if the value computed
1828 by the AFS Client Service and the client process pioctl
1829 calls produce different results as it will result in a
1830 communication failure.
1832 * Fix a crash in the B+ Tree code that results when a file
1833 rename only results in a change of case.
1835 * Fix the pioctl path parsing. The previous algorithm
1836 broke the current directory of the client into two parts,
1837 the TIDpath which represented the device root and the
1838 relative path. The problem is that this is not sufficient
1839 from proper processing of symlinks and mount points.
1841 What is required is that the last component of the path
1842 must not have the symlink or mount point be evaluated
1843 whereas symlink and mount point evaluation must be performed
1844 for the rest of the relative path. Therefore, the path is
1845 now split into three parts: TIDpath, intermediate path,
1848 "fs examine" now reports the data of the symlink or mount
1849 point object instead of the target and "symlink list" and
1850 "symlink remove" now works. The "Explorer Shell Extension"
1851 menu now is displayed for symlink objects.
1853 * More informative symlink.exe error messages.
1855 * When the cmd.exe "del" command is used to remove a symlink,
1856 not only was the link object removed, but the stat cache
1857 object for the target was marked deleted. This resulted
1858 in cache corruption which would require the destruction
1859 of the cache to correct.
1861 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1862 which results in cached contents from read only volumes
1863 to be treated as valid when there is no callback registered
1864 and all servers are down or all instances are offline.
1866 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1868 * Add documented registry value, "AcceptDottedPrincipalNames"
1869 which is used to permit the acquisition of AFS tokens
1870 using Kerberos v5 principal names which include a dot in
1871 the first component.
1873 {HKLM\HKCU}\Software\OpenAFS\Client
1875 * "fs uuid" does not require "AFS Client Administrator" group
1876 membership to display the current UUID value.
1879 * Update NSIS installer to version 2.30
1881 * Update Wix installer to version 2.0.5325
1883 * 64-bit KFW support:
1884 - Network Identity Manager plug-in
1886 - 64-bit krb5 enable afscreds.exe, afs server manager,
1889 * Enable B+ tree Directory Search by default.
1891 * NIM plug-in is now compatible with FILE: ccaches
1893 * Install instloop.exe into Client/Programs/ even though
1894 it isn't required. Permit it to be used by organizations
1895 that clone Windows systems.
1897 * On a suspend, only drop callbacks for servers that
1898 are online. If the server is offline, maintain the callback
1899 until the server is once again available.
1901 * Prevent against acquisition of a directory write lock when
1902 there are no updates to be performed. This was resulting
1903 in a write-lock leak which later on would lead to a deadlock.
1905 * Add configurable file pre-fetch mode. For files that match
1906 the specified file extensions, pre-fetch the entire file into
1909 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1910 (MULTI_SZ) PrefetchExecutableExtensions
1912 * Fix the client_osi large integer support so that return values
1913 can be properly assigned to large integer variables.
1915 * Do not hold cm_scache_t mutex across cm_GetCell() calls
1916 since cm_GetCell() can block in a RPC.
1918 * Add short file name support to B+ tree searches.
1920 * Resolve a conflict between the desire to notify file servers
1921 that the client workstation is suspending and the desire to
1922 permit objects with valid callbacks when the client suspends
1923 to continue to use those objects from the cache if the client
1924 resumes without network connectivity.
1926 * Fix stack variables to ensure that all buffers storing paths
1927 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
1928 would cause a stack overflow condition.
1930 * On Vista, force afs_config.exe to run with the highest privileges
1931 available to the end user. For a user in the Administrators
1932 group running under UAC, this will cause afs_config.exe to be
1933 be run with full Administrators privileges after the user
1934 approves the privilege escalation.
1936 * More VNOVNODE handling issues. If a dirty buffer cannot be
1937 written due to VNOVNODE, clear the buffer and ignore the error.
1938 If a cm_scache_t object has been marked deleted due to a VNOVNODE
1939 error, the next time the smb_fid_t object that references it is
1940 used return either NOSUCHFILE or NOSUCHPATH and close the
1941 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
1942 result in an invalid handle error.
1944 * For Vista, set the RequestExecutionLevel of the NSIS installer
1947 * For the build environment, support the generation of manifests
1948 when using the CL1310 (VS .NET 2003) compiler/linker.
1950 * There was a bug in OutputDebugHexString() which could overflow
1951 a stack buffer resulting in unexpected process termination
1952 of afsd_service.exe if the input Security Token contained
1953 many bytes equivalent to the '%' char.
1955 * Add directory lookup statistics to "fs memdump" output.
1959 * Windows uses case-insensitive file name pattern matching
1960 but AFS is a case sensitive file system. The AFS3 directory
1961 format is block based, uses network byte order and
1962 includes a hash table for fast case sensitive lookups.
1963 This causes several problems for the Windows AFS client.
1964 (1) Traversing the directory blocks is cpu expensive
1965 (2) A hash table miss does not indicate that the desired
1966 entry does not exist.
1967 (3) Determining whether a non-ambiguous inexact match or
1968 the entry does not exist requires a linear traversal
1969 of the entire directory.
1970 These issues often result in 100% CPU utilization.
1972 These issues are addressed by building a modified B+ tree for
1973 each directory and then using the B+ tree for searches.
1975 This feature is disabled by default. To enable it set
1976 the "BPlusTrees" registry value to 0x01 in the AFS Client
1977 Service Parameters key.
1979 * Correct a null pointer dereference in the symlink recursion
1980 detection code. (Introduced in 1.5.22)
1982 * Add the Network Identity Manager AFS Provider to AMD64
1987 * A new registry value, "BlockSize", can now be used to adjust
1988 the size of the internal buffers used to store file data.
1989 The default is 4KB. When set, the values 1..1024 specify
1990 multiples of 4KB. Values > 1024 specify actual block sizes
1991 and must be multiples of 4KB. The actual block size in KB
1992 can not be larger than the chunk size which defaults to 1MB.
1994 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1997 * pts.exe and ptserver.exe built with -DSUPERGROUPS
1999 * Service Manager shutdown hints have been added. The hint
2000 specifies that it may take the AFS client service up to two
2001 minutes to shutdown.
2003 * Fix the recursion detection feature introduced in 1.5.22.
2004 The previous implementation broke the evaluation of paths
2005 using ../path or ./path. It also broke the evaluation of
2006 symlinks to absolute paths. For example, /afs/cell/path.
2008 * Fix an uninitialized variable in cm_MakeDir that could result
2009 in crash if the requested directory could not be created.
2011 * Fix a the VolumeStatusNotification log messages to save
2012 the input string in the circular log buffer so garbarge
2013 won't be output to afsd.log when the log is dumped.
2016 * Increase the SMB Maximum Path length to 260 characters
2017 from 256 as that is what CIFS supports.
2019 * On 64-bit Windows, use a unique description string for
2020 the 32-bit AFS Shell Extension.
2024 * In the AFS Admin Library, replaced all of the
2025 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
2026 allows parameter checking and proper type conversions to be
2027 performed. This is crucial when passing parameters that are
2028 64-bit wide on systems in which 'long' is a 32-bit value.
2029 Otherwise, the parameter list on the stack is interpretted
2030 incorrectly and the function not only doesn't succeed but it
2031 might corrupt memory and if lucky crash the application.
2033 The only program that we care about that uses this library
2034 is the AFS Server Manager. Now it doesn't crash.
2036 * Fixed smb directory searches for non-wildcard names. The
2037 optimized function should not have set the bulk in progress
2038 flag. This would adversely affect later directory searches
2039 that required a bulk search.
2041 * Detect infinite recursions (up to 512 stat cache objects
2042 in a chain.) If an infinite recursion is detected, return
2043 an error to stop the CIFS client from querying us forever.
2045 * Executables executed out of AFS name space can have their
2046 pages swapped out by Windows. Those pages will be reloaded
2047 when they are required. However, if the cache manager no
2048 longer has a callback and all of the file servers with
2049 the volume data are down, then the application will crash
2050 due to the inability to read the required page.
2052 We now track the down time of each server. If the callback
2053 expires after the server that issued it went down and all
2054 servers are down, we delay the expiration to permit the data
2055 in the cache to be continue to be used by Windows or an
2058 * The Windows AFS client has always had very poor performance
2059 with regards to directory management. In the case of
2060 directory updates caused by create, rename,and delete operations,
2061 the AFS cache manager will now perform local directory updates
2062 when possible in order to avoid re-reading the directory from
2063 the file server. This will be done in whenever both of the
2064 following conditions are true:
2066 - the data version on the directory increased by one as a
2067 result of the operation
2069 - all of the directory buffers necessary to make the change
2070 are present in the cache
2072 * The error message in the AFS Control Panel describing the
2073 need to be a Windows Administrator has been replaced with
2074 a message describing the need to be in the AFS Admins Group.
2076 * If the vldb server list is empty when attempting to evaluate
2077 a cell mountpoint it was possible for a crash to occur. This
2078 can only happen if there is an existing mount point and the
2079 is a failure attempting to load the server list from the
2080 CellServDB file or via AFSDB queries.
2082 * dirty buffers only have their dirty bytes stored to the file
2083 server instead of the entire chunksize
2085 * the default chunksize has been increased from 128KB to 1MB
2090 * changed the enum values for cm_serverRef_t state info to use a
2091 private name space to avoid collisions (srv_)
2093 * added a srv_deleted state for cm_serverRef_t objects. This
2094 state is set when cm_FreeServerList() is called with the
2095 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
2096 not always delete the cm_serverRef_t from the list if it is
2097 still in use by another thread. the srv_deleted state means
2098 the object's contents are no longer valid and it must be
2099 skipped. It will be deleted the next time the object is
2100 freed and the refcount hits zero.
2102 * the srv_deleted state is also used when a file server reports
2103 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
2104 as offline. This is done to prevent additional usage of the
2105 stale vldb data while waiting for the update volume request
2108 * added a state field to the cm_volume_t object (enum volstate
2109 vl_ name space) that maintains the state of the volume based
2110 upon the states of all of the cm_serverRef_t and cm_server_t
2113 * modified cm_UpdateVolume() to set the state of the cm_volume_t
2114 RW, RO, and BK to either vl_alldown or vl_online. There can't
2115 be any other states because cm_UpdateVolume() destroys any
2116 previous knowledge we might have had regarding busy or offline
2119 * modified cm_UpdateVolume() to update the volume name in the
2120 cm_volume_t to the volume base name if the previous value was
2123 * modified cm_FollowMountPoint() to check to see if the volume
2124 name is a volume ID and if so call cm_GetVolumeByID instead
2125 of cm_GetVolumeByName. This ensures that volume IDs are always
2126 looked up as numeric values. There is no longer a need to
2127 maintain a separate cm_volume_t containing the string representation
2130 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
2131 The first flag is a "CREATE" flag which is set by all existing
2132 calls. The flag is not set by calls to cm_GetVolumeByID() from
2133 the server probe code when volume status is being updated. We
2134 do not want the server probe operation to result in additional
2135 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
2136 which is set when the server probe code updates the volume status.
2137 This flag will be used to prevent the server probe operation from
2138 changing the order of the least recently used queue.
2140 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
2141 allocated for a given set of normal, readonly, and backup volumes
2142 regardless of whether or not the volume is accessed via name or
2143 ID number. The cm_volume_t namep field is always the base name
2146 * Added a new volume state, vl_unknown. This state is used as
2147 the initial state for all cm_volume_t when the cache manager starts,
2148 for each cm_volume_t at creation, and for each cm_volume_t when
2149 recycling. The cache manager does not know the state of all
2150 volumes in the world, only those that are in the cache and for
2151 which it has queried the VLDB and hosting file servers.
2153 * modified cm_GetVolumeByName() to initialize the state of a
2154 volume to vl_unknown. The actual state will be set when a
2155 cm_VolumeUpdate() call completes successfully.
2157 * changed name of scache hash table variables to avoid ambiguity
2158 when adding hash tables for volumes
2160 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
2162 * modified cm_UpdateVolume() to handle the case in which there is
2163 no RW volume but there is are RO volumes for a given base name.
2164 This is done by querying for the ".readonly" volume name if the
2165 base name does not exist in the VLDB. We never query for the
2166 .backup name because under the current usage model a .backup
2167 volume may only exist on the server that the read-write volume
2168 is located. If there is no RW volume, there can be no .backup.
2170 * Added four hash tables for cm_volume_t objects to improve the
2171 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2172 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2173 BKID. Three ID hash tables are necessary as long as it is
2174 desireable to maintain a single cm_volume_t containing all
2175 of the related RW, RO, and BK volume data. Having the RW and
2176 RO volume data in the same object is necessary for the
2177 implementation of cm_GetROVolumeID() which returns either the
2178 RO or RW ID depending upon the existence of RO volume instances.
2180 * Added a volume LRU queue so that volume reuse becomes fairer.
2181 This does not replace the all Volumes list which is used when
2182 it is desireable to walk a list of all the volumes whose order
2183 is not going to change out from underneath you which makes it
2184 safe to drop the cm_volumeLock.
2186 * handles volume hash table updates where volume name to
2187 volume ID number changes. The volume name remains
2188 constant in the cm_volume_t. if a vos rename is performed,
2189 the name of the volume will change and the volume IDs will be
2190 updated. Subsequent access to the old volume ID will create a
2191 new cm_volume_t with the new name.
2193 * Added a daemon thread operation to query the state of volumes
2194 listed as busy or offline. cm_CheckBusyVolumes() calls
2195 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2196 or vl_offline. If the volume is now online, the status on the
2197 volume is updated. The default period is 600 seconds. This can
2198 be configured with the BusyVolumeCheckInterval registry value.
2200 * Added prototype for smb_IoctlPrepareRead() which was missing a
2201 return type in the function definition.
2203 * Added volume id lists to the cm_server_t. These lists are
2204 allocated in blocks of ~32 IDs. When a cm_PingServer()
2205 detects a change in server state, the state of the cm_volume_t
2208 * Added volID to the cm_serverRef_t object. volID is used
2209 to identify the volume for which the object is a referral.
2210 cm_FreeServerList() uses the volID to remove the volume
2211 from the cm_server_t.
2213 * In cm_Analyze, when VNOVOL or VMOVED are received,
2214 call cm_ForceVolumeUpdate() to force a refresh of the volume
2217 * Added cm_VolumeStatusNotification() which is used at the moment
2218 to log volume status changes to the trace log. It will also
2219 be used as the access point to the File System Filter driver
2220 notification engine.
2222 * Added an all cm_scache_t list to cm_data. This replaces the use
2223 of the stat cache LRU queue when we need to enumerate all
2224 entries. The LRU list order is not static and when using it to
2225 enumerate all entries it can result in items being missed or
2226 items being processed more than once.
2228 * Modified cm_Analyze(). Instead of reseting the busy or offline
2229 state of a volume and forcing a retry of the operation
2230 cm_Analyze will defer to the background daemon thread that will
2231 update the state once every 600 seconds.
2233 * Added the automatic generation of a Freelance ".root" read-write
2234 mountpoint that refers to the root.afs volume of the workstation
2235 cellname at the time the mountpoint is created.
2237 * re-write cm_Analyze to make better use of the known volume
2238 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2241 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2242 busy volumes will be reset to srv_non_busy by the function
2243 but there is no mechanism for querying the busy state other
2244 than by attempting to access the resource.
2246 * cm_Analyze will query the state of an offline volume before
2247 deciding whether or not to retry when all volume instances
2250 * the algorithm used after an IP address list change detection
2251 has been modified. we now check the state of the VLDB servers,
2252 then reset the rx connections, and finally check the state of
2253 the file servers. Previously we flushed the rx connections
2254 and then checked the servers in whatever random order they
2255 were first discovered.
2257 * fix the volume status reported by "fs examine"
2259 * various improvements to the AFS Network Identity Manager Provider.
2260 - Don't access freed objects after the window has been destroyed.
2261 - Write back the AFS cell list in the identity configuration
2262 even when its empty.
2263 - Disable all use of the provider when AFS is disabled for a
2264 particular identity.
2266 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2267 alias instead of the full cellname, double check the full cell
2268 name and see if we have a matching cell before allocating a new
2271 * Dump more data with "fs memdump". All volumes and scaches and
2274 * Reverse the order of the all cells list. This ensures that
2275 the workstation cell is always first in the list.
2277 * If we recognize the server from which a callback revoke was
2278 received, then we can ensure that we only invalidate the status
2279 information for cm_scache_t objects belonging to that cell.
2281 * When building for amd64, use APPVER 5.02 as we do not support
2282 64-bit Windows 2000 and explicitly state the machine type in
2285 * no longer treat VMWare network adapters as equivalent to a
2286 loopback adapter. For some unknown reason, publishing the "AFS"
2287 netbios name on them frequently results in a name collision
2290 * Use read locks whenever possible instead of write locks when
2291 walking the allCellsp list. This permits better use of multiple
2294 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2295 is not present on the directory optimization. It would report
2296 mount points and directory objects for which there was no
2297 status information as a zero length file. Instead we can use
2298 the vnode number of the object to determine if it should be
2299 considered to be a directory or a file. If it is odd it is a
2300 directory; even it is a file.
2302 * Give Back Callbacks to the file servers in response to suspend
2303 and shutdown power management events. Only give up callbacks
2304 to UP servers and do so with a very short RX dead timeout.
2307 * Fix a race condition in the SMB FID cleanup code that could
2308 result in a stat cache mutex never being released.
2310 * Fix the graphics files that are installed as part of the
2311 IBM Administrators Guide.
2313 * Change the behavior of pioctl path parsing to not follow
2314 symlinks or mount points during evaluation. This results in
2315 the FID being the symlink or mount point and not the object
2316 they point to. This is consistent with the behavior of the
2319 * Enable AFSDB record lookup for command line tools such as
2322 * CellServDB update 2007-07-05
2324 * Added a Volume Status Plug-in Interface that permits third parties
2325 to receive notifications when the volume state changes occur.
2327 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2329 * Prevent the afsprocmgmt.dll from being unloaded while the
2330 background thread it starts is running.
2332 * Avoid a race condition when closing smb_fid_t objects.
2335 * fix install path of IBM Quick Start Guide when installed by
2338 * fix the client so that it can write into directories with
2339 only 'li' permissions.
2341 * fix installation of 32-bit Shell Extension for 64-bit systems.
2342 Install it under its own HKCR keys.
2344 * modify the wix installer to remove and upgrade install packages
2345 with the same version number.
2347 * when flushing volume location data every two hours or in
2348 response to "fs checkvolumes", also flush the mount point
2349 links in order to force their re-evaluation.
2351 * aklog was copying an uninitialized buffer as part of the
2352 foreign cell automatic pts registration
2354 * when obtaining directory listings, check for read permission
2355 on the directory. If read permission does not exist, fake the
2356 status information for the directory entries. This improves
2357 performance, prevents the client from pounding the file server,
2358 and prevents leakage of cached status information.
2360 * "fs getcacheparms" output was incorrect.
2362 * NetIdMgr AFS Provider should only automatically add token
2363 acquisition for cell foo.com if the realm of the cell matches
2364 the realm of the identity.
2366 * Fix a null pointer dereference in cm_MergeStatus()
2370 * network identity manager provider:
2372 - Correct for the fact that pr_CreateUser tampers with the
2373 contents of the username input buffer.
2375 - Take into account that the krb4 credential data structure
2376 does not have separate realm fields for the user and service.
2377 When krb524 is used with a cross-realm ticket, the realm
2378 describes the service not the user.
2380 * Many integrated logon changes:
2382 - Fix eventlog reporting. Do not attempt to log an event
2383 if the event source registration fails. Use DebugEvent0
2384 instead of DebugEvent when there are no parameters.
2386 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2389 - Fix the reading of Domain specific configuration for
2390 LogonScript and TheseCells. Previously the dwSize value was
2391 being overwritten so that subsequent RegQueryValueEx call
2394 - Fix a memory leak in the TheseCells reading code.
2396 - Add support for Domain specific "Realm" specification. The
2397 realm is the realm to be appended to the username. When
2398 logging in as a domain or to the local machine, the specified
2399 "Domain" name is not going to be a valid realm name.
2401 - Construct a proper principal name based upon the domain
2402 specified realm for use in obtaining tokens with KFW.
2404 - If the domain specified "TheseCells" list includes the
2405 default cell, do not obtain tokens twice.
2407 * An unprotected call to the Kerberos v5 function,
2408 krb5_free_context(), can result in an invalid memory access
2409 exception if the contents of the krb5 profile, krb5.ini, cannot
2410 be parsed correctly. The error is located within the afskfw
2411 library which is used by afscreds.exe, afssvrmgr.exe, and
2412 afslogon.dll. When the error is triggered within afslogon.dll
2413 a blue screen condition can occur when the winlogon.exe system
2414 process terminates unexpectedly.
2418 * Do not stop the afs client service with an assert
2419 if the cell name cannot be determined when responding
2420 to a "fs whereis" command.
2422 * Integrated Logon (afslogon.dll). No longer attempt
2423 to preserve Kerberos 5 tickets for the logon session
2424 as that job is now supported by KFW 3.1 and above.
2426 * The Network Identity Manager plug-in will attempt to
2427 remove expired service tickets from the cache.
2431 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2433 * AFS Server Manager and AFS Account Manager data cache
2434 moved from HKCU\Software\TransarcCorporation to
2435 HKCU\Software\OpenAFS to prevent crashes due to data
2438 * Reduce the default UP server probe time from 10 minutes
2439 to 5 minutes to match UNIX. This will not solve the NAT
2440 port mapping timeout issue because most NATs default to
2441 a value less that 5 minutes and this value ensures that
2442 the afs client server will start probes 5 minutes after
2443 the last probe set ended. The probe period for any one
2444 server is therefore longer than 5 minutes. The more
2445 servers in the list the longer the time between probes.
2447 * Before marking a server as down in cm_Analyze, retry the
2448 request using a new rx connection to the server. If the
2449 cache manager is behind a NAT, it is possible that the
2450 response from the server is being sent to the original
2451 address:port from which the rx connection was established.
2452 This will appear to the cache manager as the server being
2453 down when in fact it is not. Forcing the use of a new rx
2454 connection will result in the server responding to the
2455 most recent address:port assigned to the client by the NAT.
2457 If the server is in fact down, it will simply take a little
2458 longer to notice it.
2462 * Removing a directory that contained entries was failing with
2463 the wrong error code being returned to the CIFS client.
2467 * rxdebug exponential back-off. don't take down a server that
2468 is already struggling to respond.
2470 * Network Identity Manager plug-in registers help on the Help
2471 menu if it is supported by NIM. (KFW 3.2 and above)
2473 * Plug-n-play networking did not always work after coming back
2474 from a Windows Suspend. The AFS client service failed to
2475 re-compute the LAN adapter to bind to and instead attempted
2476 to bind to all LAN adapters. This would polute the public
2477 network with the name "AFS" which could lead to conflicts
2480 * Build installers using Microsoft Hotfixes provided at
2481 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2482 Daylight Savings Time issues.
2484 * "fs checkserver" does not require admin privileges unless
2485 the probe interval is being changed.
2489 * A reference leak was discovered in the mountpoint traversal
2490 functionality which resulted in cm_volume_t objects not becoming
2491 eligible for re-use. This will trigger a panic condition when
2492 more than the maximum number of volumes is accessed.
2494 * Fixed the success test for the "fs memdump" operation. Success
2495 is now reported as such. Added cm_volume_t object information
2498 * A data corruption error has been fixed. Write requests
2499 queued for background processing were not retried upon failure.
2500 This leads to inconsistencies between the contents of the file
2501 in the local cache and the contents of the file on the file
2504 The implemented fix is to re-queue any request that has failed
2505 due to timeout, volume busy, volume offline, server busy, or
2506 server offline. Requests are not processed while all servers
2507 for a volume are inaccessible. When servers become accessible
2508 the background request will be retried.
2510 In addition, if a background write fails, do not mark the buffers
2511 as clean. Failure to preserve the dirty flag prevents the buffers
2512 from being written to the file server in the future.
2514 This fix does not address the possibility that the credentials
2515 necessary to perform a write request might expire between the
2516 time the request is queued and when the request can be processed
2517 due to server accessibility.
2519 * The Freelance volume does not support per user ACLs. The
2520 cm_HaveAccessRights function did not handle this special case.
2521 Requests for privileges greater than read|lookup would result
2522 in afsd_service.exe entering an infinite loop. ISOBuster is
2523 known to cause this behavior.
2525 * When deleting files, the SMB Server would attempt to obtain
2526 the access permissions for the parent directory instead of
2527 for the file itself. This would result in an incorrect
2528 assessment of whether or not the file can be deleted.
2530 * The Network Identity Manager AFS plug-in has had its string
2531 tables updated for consistency with the language used by NIM.
2532 All credentials are referred to as credentials instead of tokens.
2533 Drop-down combo boxes now in fact drop down.
2535 * The AFS Salvager was not properly being built on Windows.
2536 This is relevant only if the AFS servers are in use.
2538 * The default signal handlers for the AFS servers incorrectly
2539 processed SIGQUIT. Instead of stopping the server, an
2540 exception was raised.
2542 * Fix another volume refcount leak. This one was triggered by
2543 an inability to contact the vlservers in order to update the
2546 * Fix checklist control used by the drive mappings tabs in
2547 afscreds.exe and afs_control.exe.
2549 * Improve output of "fs memdump" related to buffer management.
2551 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2552 already in the process of being fetched.
2554 * Improve cache performance by computing the buffer hash table
2555 size based upon the number of buffers so that the average
2556 number of entries in a bucket is seven.
2558 * Improve cache performance of read/write data by locally
2559 updating the dataVersion of all buffers associated with the
2560 file on each StoreData or StoreStatus operation provided that
2561 the dataVersion was incremented by one and the dataVersion of
2562 the buffer matched the original dataVersion of the cached
2563 status information for the file.
2565 * When recycling cm_scache_t objects, make sure that we
2566 clear all of the previous fields. It is no longer the
2567 case that we can assume that the stat cache object is being
2568 obtained only after a successful FetchStatus call.
2569 InlineBulkStatus will allocate stat cache objects for which
2570 the no status information is available.
2572 * When a file is being opened with the OPEN_ALWAYS disposition,
2573 do not require write permissions when the file already exists
2574 unless of course the file is being opened requesting write
2577 * When a file has been opened for read-only access, if the
2578 application requests an exclusive lock, obtain a read-lock
2579 instead of a write-lock. The purpose of the exclusive lock
2580 in this situation is to prevent changes to the file and the
2581 full file read lock will provide this guarrantee.
2586 * Enforce Unix Mode Write bit to prevent file deletion of read only
2589 * The CIFS READ and LIST permissions use the same value 0x0001.
2590 When a directory is being accessed, check for PRSFS_LOOKUP
2591 instead of PRSFS_READ.
2595 * Fix computations of short names of mount points. They were
2596 broken as part of the directory search optimization. (1.5.13a)
2598 * When all buffers are in use, periodically sleep the threads
2599 waiting for a buffer to become free while the global buffer
2600 lock is not being held in order to permit those threads that
2601 can free buffers to do so.
2603 * Power Management improvements. Maintain a global flag that
2604 specifies whether or not the service is in a suspend state.
2605 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2606 lana is no longer valid. Instead, stop the listener threads
2607 and if all listener threads are stopped, reset the lana_list.
2608 Allow the cm_Daemon() thread to periodically check the state
2609 of the smb listeners. If they are all stopped and the service
2610 is not suspended, attempt to restart them. If there are no valid
2611 lanas, return to the stopped state.
2613 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2614 the user is permitted to obtain read or write locks. This function
2615 would obtain the lock and then drop it returning whether or not
2616 the lock could be obtained. If the lock was in fact required,
2617 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2618 of course being that this pattern results in three RPCs to the
2619 file server (lock, unlock, lock). This is reduced to one RPC by
2620 implementing cm_CheckNTOpenDone() which frees the allocated byte
2621 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2624 * Optimize the SMB FindFirst processing for the case in which there
2625 are no wildcards. If a cm_Lookup succeeds, generate a response
2626 from the stat cache entry we have a callback for. Its cheaper
2627 than walking each buffer of directory data associated with the
2628 directory stat cache entry when we have thousands of directory
2631 * Increase the SMB Ioctl MaxData size to support larger token sizes
2633 * Disable autorun of the AFS Server Wizard when the AFS Server is
2636 * Add AFS Server service dependencies on tcpip and PNP_TDI
2638 * Remove the AFS Server Configuration Tool shortcut
2640 * Add Freelance mode awareness to AFS Server Configuration Tool
2644 * Unlike previous versions of the OS, Vista performs a shutdown on
2645 the Microsoft Loopback adapter just like it would on a real adapter.
2646 This causes the smb_Listener threads to trigger a panic during a
2647 suspend/hibernate power event.
2649 The fix is to unbind from the network adapters in response to a
2650 suspend/hibernate power event and then rebind to the adapters
2651 when a resume power event is received. Note that the resume events
2652 are not reliably delivered so it is possible the afs service will
2653 not be accessible. However, this is the best we can do.
2655 * don't allow environment variables we set in afslogon.dll to be
2656 inherited by child processes
2658 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2661 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2662 thread walking the all buffer list periodically searching for dirty
2663 buffers to write to the file server. This had several negative
2664 results. The algorithm ate up ever increasing amounts of CPU time
2665 even when AFS is idle as the size of the cache increases. Also,
2666 buffers were written to the file server in an order based upon the
2667 original buffer allocation which has nothing to do with the order
2668 in which the buffers became dirty.
2670 The new algorithm maintains a dirty buffer list. Items are added
2671 when the buffer is originally marked dirty. A buffer is only
2672 removed from the list by the buf_IncrSyncer when the buffer is no
2673 longer dirty. If the list is empty the thread goes back to sleep
2674 immediately without additional processing requirements.
2678 * Enable renaming of directories from the Explorer Shell when
2679 Offline Folders is enabled
2681 * An error is reported in response to the SMB ReceiveCoreFlush
2682 request when an error has not occurred.
2686 * optimize dirty buffer flushes in response to suspend operations
2687 do not flush if only loopback adapter is installed
2688 do no retry dirty buffer writes
2690 * do not reset netbios when performing loopback checks. doing
2691 so will break the existing lana bind
2693 * remove a deadlock in smb3.c QPathInfo
2697 * Add NetIDMgr AFS Plug-in
2699 * "Show Tray Icon" checkbox removed from afs_config.exe
2701 * Disable DNS Registrations for the Loopback Adapter and make sure
2702 that Netbios is turned on.
2704 * A reference count leak on stat cache objects was fixed
2706 * For debugging, tread ids of writers are now tracked on locks
2708 * Corrected an incorrect lock that was being held but not
2709 freed in a rare error case
2711 * Changed trace log end of line to CR-LF so it can be read
2714 * modify the buffer synchronization thread to write all dirty
2715 buffers and sleep only when there is no work to be done
2717 * improve behavior of cache manager when large numbers of
2718 temporary files are being created by multiple clients within
2721 * do not permit dirty buffers that belong to deleted files
2722 to be written to the file server. its a waste of bandwidth
2724 * do not leak stat cache references when the file server
2727 * afscreds: do not permit an attempt to obtain tokens if the
2728 afs service could not be started. this was causing stack
2729 overflows on Vista after resuming from a suspend
2733 * better handling of dot directory
2735 * fs sq and fs mkm now behave as on UNIX
2737 * add license to installers
2739 * Return Invalid Parameter error for unknown InfoLevels in
2740 SMB Tran2 Search Directory.
2742 * Correctly use WSAStartup and WSACleanup within the RX library
2744 * Check that afsd_service.exe is running before performing
2747 * Force AFS Shell Extension to load resource DLL using
2748 afs_shl_ext.dll as a base name.
2750 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2751 OpenAFS\Client\Program
2753 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2755 * Do not panic if the maximum number of volume entries are in use
2756 and one of them can be recycled.
2760 * Fix the code that reads the backconnnectionhostnames value
2761 so that there is not an uninitialized variable
2765 * integrated logon could cause winlogin.exe to blue screen windows
2766 if it was called with a domain that does not exist on the machine.
2767 this is possible during a sysprep.
2769 * corrections to cm_buf_t handling to ensure that buffers are
2770 removed from queues before they are freed and that holds are
2771 obtained on the next buffer in a chain before releasing the
2774 * do not generate MessageBox dialogs on network errors. they
2775 cause more trouble then they are worth.
2777 * move GlobalAutoMapper into a separate thread so that it does
2778 not prevent service startup if the WNetAddConnection2 call
2781 * restrict hard dead and connection timeouts to the CIFS Session
2782 Timeout value. This is being done to try to prevent virtual
2783 circuits from being destroyed prematurely.
2785 * Fix the code that writes the backconnectionhostnames value
2786 to ensure that the data buffer is written with the correct
2789 (there were no windows specific changes between 1.5.4 and 1.5.5)
2793 * Corrections to the Integrated Logon functionality that permits
2794 Kerberos 5 TGTs to be communicated into the user session. The
2795 transfer is now more secure and succeeds when the user does not
2796 have read permission to the %SystemRoot%\TEMP directory.
2798 * NSIS installer updated to version 2.18
2800 * WiX installer updated to version 2.0.4221.0
2802 * Updates to the IFS functionality for 64-bit correctness and
2803 permit compilation. The IFS functionality works for pioctl()
2804 operations but not much else at this point.
2806 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2807 query and set functions.
2809 * Ignore the Windows CIFS client session timeout entirely. Timeout
2810 only after the AFS Hard Dead Timeout period expires.
2814 * move headers, libraries, and samples from OpenAFS\Client to
2817 * remove the auto-registration of AFS IDs for foreign cells
2818 from integrated logon. this permits afslogon.dll to avoid
2819 initializing the RX library. afslogon.dll is a network
2820 provider library and as such it is loaded into all applications
2821 which perform a CreateProcessAsUser call. Initializing the
2822 RX library spawns background threads to listen for inbound
2823 packets. When the network provider is unloaded from the
2824 process, the RX library does not cleanup the threads. If they
2825 continue to execute after the library is unloaded, the process
2828 * implement inline bulk status calls. this will significantly
2829 improve the performance of directory listings.
2831 * prevent a crash when evaluating mountpoints to volumes that do not
2834 * new command: fs uuid [-generate]
2835 displays the current uuid and allows a new one to be generated
2836 on the fly without restarting the afs client service.
2838 * fix the 64-bit installers to properly document the install
2841 * move help files to the proper location so they can be located by
2842 the system tray tool and the control panel.
2844 * "over quota" and "disk full" errors were not being reported to the
2847 * largefile support has been added. Files greater than 31-bits can no
2848 longer be written to file servers which do not support large files.
2849 When file servers support large files, files up to 64-bits can be
2852 * cm_SyncOp could return with the input buffer unlocked upon error
2853 which would result in a panic condition. This bug was introduced
2856 * vlserver failover when freelance mode did not work. This prevented
2857 the root.afs volume from being loaded if the first vlserver could
2862 * A bug in the CIFS Standard Info response packet caused directories
2863 to appear as standard files. This prevented Microsoft Vista from
2864 being able to walk the AFS directory tree.
2866 * Mount points must be reported as directories in response to
2867 Query Path Info operations.
2869 * A deadlock was detected when performing "fs flushall" if the
2870 file server reports VNOVNODE. The scp->createBufferLock is
2871 already held by the current thread and there is no mechanism
2872 to propagate the knowledge. Therefore, an alternate mechanism
2873 for clearing the cache must be developed. A new function
2874 cm_RecycleSCache(scp,flags) has been extracted from
2875 cm_GetNewSCache(). This function performs the task of recycling
2876 an cm_scache_t object. When called from cm_GetNewSCache() with
2877 no flags the expectation is that there are no associated buffers
2878 that are queued to be read or written. When called from
2879 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2880 flag, any queued buffers will be de-queued and marked as if
2881 the operations were performed so the data can be discarded.
2883 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2885 * In response to VNOVNODE, the parent is only discarded if the current
2886 cm_scache_t is not a directory.
2888 * In the Ioctl FlushFile and FlushVolume functions, there were no
2889 checks to protect against flushing the Freelance SCache entries.
2891 * In FlushFile, the wrong cm_scache_t object was being released.
2893 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2894 recycled. Choose a new entry if cm_RecycleSCache fails.
2896 * Speed up the performance of the cache manager by not
2897 holding the smb_fid_t mutex across calls to cm_SyncOp
2900 * Ensure that all smb_fid_t flag references are protected.
2902 * Remove a deadlock between smb_fid_t mutex and
2903 smb_rctLock introduced in 1.4.1.
2905 * Fix aklog AFSID lookup to use not send the realm
2906 name when the realm is the local realm for the
2907 cell. The PT Server doesn't strip the realm
2908 name when it is the local realm.
2910 * Treat "\\afs\*." as an alias for "\\afs\all"
2912 * Dynamically adjust the priority of server threads
2913 based upon the age of the CIFS request that is
2914 being processed. The priority is bumped one level
2915 for every 15 seconds of processing time.
2917 * The original openafs contribution mis-used the
2918 cm_GetCallback function. This function should only
2919 be called as a side effect of a call to
2920 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
2921 was being called results in a FetchStatus RPC being
2922 performed even when the client already has a valid
2923 callback and does so in a manner that destroys the
2924 synchronization of the threads that are calling
2925 cm_SyncOp. If nothing else this results in a
2926 significant performance penalty.
2928 * More changes to cleanup of smb_vc_t objects in
2929 order to prevent race conditions.
2931 * Reduce the amount of Power Management event logging
2934 * If the client detects that the IP address has
2935 changed, force the use of new RX connections. This
2936 enables the file server to recognize that the client
2939 * Remove a race condition and deadlock associated
2940 with cleanup of smb_vc_t objects.
2942 * Fix the ANSI filename support. One function that
2943 required it was missing the translations.
2945 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
2948 * Add a DllMain() function to afspthread.dll. Perform
2949 cleanup of thread local storage, rx queues, critical
2950 sections, event handles, etc. This was implemented
2951 to avoid crashes in afspthread.dll when applications
2952 which load/unload the dll attempt to terminate.
2954 * Lock request behavior has been altered. As of this
2955 release a locally allocated lock will be used whenever
2956 the volume is RO or the requesting user's permissions
2957 are 'rl'. This will allow end users to execute programs
2958 from RW volumes when they do not have the 'k' privilege.
2960 Note that a recent discovery indicates that the 'k' bit
2961 is only supposed to control the ability to obtain read
2962 locks. Write locks are to be controlled by the 'w' and
2965 * Code has been added to enable the cache manager to be
2966 able to obtain the AFSID of the authenticated user.
2967 This is required to be able to determine when "Insert"
2968 privileges should apply. AFS considers a file to be
2969 a new file whenever the owner's AFSID matches the
2970 AFSID of the user. The current implementation relies
2971 on the name specified during the SetToken call to be
2972 correct. This will be replaced in the future with a
2973 new RPC that allows the AFSID of the authenticated
2974 user to be obtained.
2976 * A 32-bit-tools installer for use on 64-bit systems
2977 is now available. These tools are for use with 32-bit
2978 versions of Kerberos for Windows or JPSoftware's 4NT.
2980 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
2982 * Remove the use of the "active_vcp" variable. This variable
2983 no longer serves any useful purpose and was the cause of
2984 race conditions that produced under counts of the smb_vc_t
2987 * Change default minidump output to include data segments
2989 * restore access to cm_IoctlGetVolumeStatus and
2990 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
2991 Still need to figure out why they are being called
2992 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
2994 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
2996 * Update build system to include the "sys_name" in the
2999 * optimize the number of calls to obtain the smb_user_t for a
3000 smb virtual circuit.
3002 * add logging of server addresses, status and capabilities when
3005 * prevent multiple threads from pinging the same server simultaneously.
3006 instead, threads block and then use the result of the ping that was
3007 already taking place.
3009 * Add registry values to allow the defaults for daemon checks to be altered
3010 at startup. Log values to afsd_init.log.
3012 daemonCheckDownInterval
3013 daemonCheckUpInterval
3014 daemonCheckVolInterval
3015 daemonCheckCBInterval
3016 daemonCheckLockInterval
3017 daemonCheckTokenInterval
3019 * Add EventLog messages for transitions in fs crypt state.
3021 * Byte range locking (when locks will be obtained from the
3022 is still under discussion)
3024 * 64-bit Windows operating systems now supported
3025 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
3026 This is a native 64-bit implementation of OpenAFS for Windows.
3027 There is currently no 64-bit version of KFW and therefore
3028 there is no (easy) method for obtaining AFS credentials.
3029 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
3030 to obtain AFS tokens but there is not for the feint of
3033 * A faster implementation of the fcrypt algorithms improves
3036 * The afsd_service.exe now supplies Event Log Message Text.
3037 The event log no longer generates warning text due to lack
3038 of message configuration.
3040 * When afs servers are unreachable and network stack returns
3041 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
3042 mark the attempted server as being down. This allows for
3043 very fast response to the calling application. This change
3044 has a pronounced improvement when the Windows Explorer is
3045 used on disconnected mobile computers.
3047 * Added DebugOutputString debugging of the RX library
3049 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
3051 * Fix the rx-lwp implementation to randomly generate rx call
3052 identifiers which are supposed to be unique at a given point
3053 in time. This affects some of the command line tools used
3054 for debugging but not the AFS Client Service itself which
3055 uses the rx-pthreads implementation.
3057 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
3058 Windows although they are not distributed
3060 * don't generate an exception when logging unexpected SMB packets
3062 * the data structures used by cm_IoctlGetVolumeStatus and
3063 cm_IoctlSetVolumeStatus are not compatible with the CIFS
3064 Get/Set File Information calls so don't bother executing them.
3068 * Fix data written to registry as part of BackConnectionHostnames
3071 * Fix initialization bug when allocating cm_server_t objects that
3072 would cause servers to appear to be down after each allocation
3073 or renewal. Server would appear up after the next check servers
3078 * Windows will return WSAECONNRESET to the next receive packet
3079 operation on a socket if an ICMP packet is received in response
3080 to a packet that was sent and could not be delivered because
3081 the port was unreachable. We have to treat this as
3082 WSAEWOULDBLOCK so that we can retry the select() check and
3083 not mark the connection as bad.
3085 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
3087 * The client manager maintains a least recently used list of
3088 all the ACLs. If an ACL is already the most recently used
3089 do not bother to remove it from the list and place it on the
3092 * As part of the process of adding a new server, set the
3093 server is down flag and then perform a ping. If the server
3094 is up, the flag will be reset. If the server is down, the
3095 client will not timeout unnecessarily while attempting to
3098 * If a Mountpoint String is the empty string, return PATH NOT
3101 * Modify behavior of the Network Provider to avoid querying the
3102 profile location if integrated logon is known to be disabled.
3104 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
3105 will always be preserved regardless of whether or not integrated
3106 logon was used to obtain tokens. When unset, tokens will be
3107 destroyed regardless of whether integrated logon was used to
3108 obtain the tokens unless the profile is stored in AFS.
3110 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
3111 message is being delivered even when the user is not logging
3112 off of the computer.
3114 * Fix a deadlock condition introduced in rc6.
3116 * Cell names are case-insensitive. Since the founding of OpenAFS
3117 there has been a bug that treated "cell", "Cell", and "CELL" as
3118 different cells. This would result in what appeared to be
3119 random disappearance of user tokens or a failure to authenticate
3120 even when tokens were listed by the "tokens" command.
3122 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
3124 * Additional auditing of the smb code revealed that smb ioctl
3125 operations were failing to free smb_fid_t objects
3127 * The File Identifiers assigned to smb_fid_t objects could be
3128 given the value 65535 which is 0xFFFF which is also the
3129 INVALID_HANDLE value. Both 0 and 65535 are now treated as
3132 * Audited all assignments to fields and ensured that they are
3133 protected by obtaining the relevant mutex object.
3135 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
3136 if both the current cm_scache_t object and its parent are
3137 both returning VNOVNODE.
3139 * Removed all the logic associated with the 'dead_vcp' variable
3140 which must have been added for debugging because its used
3141 was never thread safe and it truly served no purpose.
3143 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
3145 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
3146 registry settings. Default is ON and 120 seconds. If set to
3147 OFF, tokens will not be destroyed.
3149 * Fixed a number of race conditions and reference count problems
3150 associated with global queues.
3152 * Added a method to detect abandoned smb virtual circuits so their
3153 associated file handles, locks, user tokens, and smb session
3154 resources can be released.
3156 * Added a mechanism to allow user tokens to be preserved
3157 during integrated login and be freed after logoff.
3159 * Check access permissions to the file before the file is
3162 * More improvements to the handling of SMB Virtual Circuits
3163 when SMB sessions are closed.
3165 * Correct a race condition introduced in rc4 that could result
3166 in a crash of afsd_service.exe.
3168 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3170 * Properly handle pre-mature SMB session closures.
3171 File handles, tree connections, and user sessions are now
3172 freed when the virtual circuit is closed.
3174 * smb_username_t and smb_user_t objects are now properly reference
3175 counted and freed as appropriate.
3177 * disable to use of AFS file locks. byte range locking is still
3178 implemented but all locks are managed locally. This prevents
3179 multiple processes from editing the same file on the same
3180 machine but it does prevent multiple processes on separate
3181 machines from modifying the file.
3183 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3185 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3186 the same server parameter would result in a panic condition
3187 if the server had not already been contacted by the AFS
3190 * Better warnings are written to the afsd.log file when
3191 the Windows RPC Protocols are not properly configured as
3192 is often the case on Windows 2000.
3194 * fix the procmgmt library to restore C RTL signal handlers when
3195 it is being detached from the process prior to process termination.
3197 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3199 * close a lock leak in the rx library that could be triggered while
3202 * prevent removal of drive mappings in response to afscreds.exe -M
3204 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3205 strings and flush the stat cache entry.
3207 * prevent server objects from being freed if user preferences are set.
3209 * Update WIX support to 2.0.2419.0
3211 * Release all locks on the server at service shutdown.
3213 * A down server should be marked up as long as it responds to the client.
3214 It doesn't matter if the server returned an error as long as the error
3215 is something other than VRESTARTING. If there is such an error, all
3216 connections on that server should be marked for replacement.
3218 * Change the check "up" servers period from once per hour to once every
3219 ten minutes. This matches the behavior of the Unix cache manager.
3221 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3222 * Byte Range Locking support has been added. Architectural documentation
3223 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3224 made to enforce Windows mandatory locking model using locally managed
3225 byte range locks with full file locks obtained from the file server.
3226 Read-only volumes provide the user with an implied locking privilege.
3227 Read-write volumes require that a user be granted the locking privilege
3228 in order for a lock to be obtained. As most Windows applications use
3229 Shared file access modes, they require the ability to obtain locks in
3230 order to access files.
3232 * Implement error translation functions to convert Universal Error Codes
3233 output by modern servers to local errno and winsock error values.
3234 This fixes the error message problems using pioctl() calls from fs.exe
3235 and the AFS Shell Extension.
3237 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3238 * Allow the AFS Client Service to set the firewall configuration by
3239 retrying until after Windows determines the boot period is over.
3241 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3244 * Detect IP address changes from within the AFS Client Service for the
3245 purpose of checking down servers sooner
3247 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3248 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3249 user that there is a Kerberos key problem. Previously the user would
3250 receive no feedback when there was something wrong.
3252 * Fix the large integer library support to properly handle values greater
3255 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3256 * Updated CellServDB
3258 * Fixed file name pattern matching. Question marks can match nothing
3259 if found at the end of a component.
3261 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3262 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3264 * The MSI installer now has a new Product ID for each build. This
3265 is to allow for automated upgrades. The Upgrade ID is remaining
3268 * Volumes that were cached and then subsequently were salvaged,
3269 moved or otherwise taken temporarily offline (as opposed to being
3270 busy) would become unreachable. A distinction is now made between
3271 a volume being offline and a server being down.
3273 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3275 * Fixed several AFS Shell Extension issues:
3276 - Checking Server Status would only work if a specific cell was
3278 - If servers were down, the Server Down dialog would write to
3279 invalid memory and cause explorer.exe to crash
3280 - Default Push Buttons were not specified for many dialogs
3281 - The extension would be unloaded by the Explorer Shell and
3282 never be reloaded under some circumstances.
3284 * Fixed the AFS System Tray menu to disappear if another Window
3285 was selected before a menu item was selected
3287 * Added volume owner and group information to "fs examine" output
3289 * Added fs and registry support for enabling/disabling/clearing
3290 rx statistics gathering.
3292 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3294 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3295 GUIDs for each platform.
3297 * Documentation directory structure and default installation rules
3300 * New OpenAFS for Windows HTML Release Notes installed
3301 as part of the documentation set.
3303 * Add thread id to rx debugging messages
3305 * When all servers are offline, return a bad network path error
3306 immediately. This is necessary to allow Explorer to do the
3307 right thing and not hang. The background daemon thread will
3308 check the down servers every 30 seconds to see if they are
3311 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3313 * The AFSCache file has been modified to store both the
3314 serial number of the volume on which the file resides
3315 and the SID of the local machine. If either change,
3316 a new UUID is generated.
3318 * add "fs flushall" command that forces all data buffers
3321 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3323 * add support for '/' instead of '\' in pioctl() calls
3325 * Apply AFS Client Admins group protect to AFS Shell Extension
3327 * Add support for \\afs\<name> to most AFS Shell Extension
3328 commands except the symlink methods.
3330 * when installing the Microsoft Loopback Adapter, enable
3331 MS Client for Networks and prevent an install failure
3332 by not calling CoInitialize twice in the same thread.
3334 * reload cell vldb values from the CellServDB every two
3335 hours in case it changes
3337 * When updating cell information from DNS, be sure to set
3340 * Add support to allow use of \\AFS\<foo> where <foo> is
3341 either a mount point or symlink. As <foo> is normally
3342 treated as a share name, we transform it into \\AFS\all\<foo>
3345 * Init 'code' to prevent false errors when integrated logon
3346 is disabled and the service is not running
3348 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3350 * Removed trace log messages that were hampering performance.
3352 * Fixed a deadlock that was being triggered by editing Word
3353 documents stored within AFS with WinWord 2003.
3355 * Bit 3 of the TraceOption registry value is now used to set the
3356 default for "fs trace" in the non-Debug builds. The new default
3357 is off for release builds and on for debug builds.
3359 * Bit 2 of the TraceOption registry value can be used to turn on
3360 real-time output of debug log entries to the Windows Debug
3361 Output monitor interface. This data can be viewed with tools
3362 such as Sysinternal's DbgView.
3364 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3365 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3366 a resource we should not be returning status codes such
3367 as Network Busy or Remote Host Down because these errors
3368 will imply that the AFS Client Service is Busy or Down
3369 and that is simply not the case. Instead we will return
3370 Bad Network Path as the path is temporarily not available.
3372 Instead of returning Sharing Paused when there is a
3373 timeout error, return Timeout. Once again, the AFS Client
3374 Service is not paused.
3376 * afscreds.exe would display an Obtain Creds dialog when
3377 the expired credentials reminder was triggered even
3378 if there was no network path to the KDC. This is prevented
3379 by adding KDC probe logic to the reminder thread.
3381 * afscreds.exe would display expired tokens no differently
3382 than unexpired ones. This would make it difficult for the
3383 user to distiguish when the tokens were expired. For the
3384 English build added a new resource string "(expired) that
3385 is displayed instead of the expiration time.
3388 * "fs wscell" when executed in freelance mode will return the
3389 name of the cell configured in the registry. The root.afs
3390 volume is not loaded from this cell, but it is used for the
3391 default for aklog and integrated login.
3393 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3394 no longer work on \\AFS\all when freelance is being used
3395 unless the user is a member of the "AFS Client Admins" Windows
3398 * some more checks performed during persistent cache validation.
3401 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3402 down servers for the current cell not all cells.
3404 * If the scache LRU list becomes corrupted, the AFS Client Service
3405 might enter an infinite loop at startup. Detect the corruption
3406 and discard the cache contents in this case.
3408 * Fixed the Explorer Shell's remove mountpoint functionality.
3410 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3411 obtain a mutex lock on a cm_scache_t object whose lock is held by
3412 a thread attempting to obtain the cm_aclLock.
3415 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3416 misinterpretted as "csail.mit.edu" when the resolver
3417 is configured to append ".mit.edu" to failed DNS queries.
3419 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3420 that can be used to force the preservation of user tokens upon logout.
3422 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3423 recommended installation categories: AFS Client, AFS Administrator,
3424 AFS Server, AFS Development Kit. Each category includes a different
3425 default subset of the OpenAFS components.
3427 The OpenAFS logo is now associated with the NSIS uninstall entry
3428 in the Add/Remove Programs control panel.
3430 * The user name associated with AFS tokens when obtained with
3431 integrated login, the afs systray tool, or aklog will always
3432 include the full Kerberos 5 user principal name regardless of
3433 whether or not the cell is local to the realm.
3435 * Modify integrated login so that it does not enter an infinite
3436 loop if the service is not set to auto start.
3438 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3441 * Added uninstall.exe to Wix installation
3443 * More modifications to algorithms used to wake sleeping threads.
3447 * Changes to the algorithms used to wake threads when they are
3448 put to sleep because of conflicts over buffer operations.
3451 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3452 the process of producing AFS tokens. The tokens are stored within
3453 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3454 New functionality has been added to temporarily copy the tickets to
3455 a file ccache during the short window between the NPLogonNotify and
3456 WinLogon Logon Event. The Logon Event handler starts a process as
3457 user, afscpcc.exe, that copies the tickets into the default ccache.
3458 The file ccache is then destroyed by the event handler.
3460 * During synchronization operations on stat cache entries, if threads
3461 are forced to sleep it was possible that the threads would not be
3462 woken under heavy load.
3464 * Enforce the delete on close flag specified by the NTCreateX()
3467 * A race condition in the rx library was fixed that resulted in
3468 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3469 was cleared when threads were still waiting to use the
3470 connection. Also, fixed a potential case in which the
3471 connection could be destroyed while threads were waiting to use it.
3473 * The DNLC freelist has been seen to become corrupted with still
3474 active entries being stored on the freelist. Changes were made
3475 to perform a better job of cleaning entries before placing them
3476 onto the freelist; marking them with the correct magic value when
3477 purging the list; and allowing the list to be purged on startup
3478 if the cache validation fails.
3480 * Windows Crash Reporting does not get triggered for afsd_service.exe
3481 because it provides its own top level exception handler. This patch
3482 places an explicit call to ReportCrash() on platforms which support
3485 If you configure Crash Reporting via AD Policy to capture crash
3486 reports within the domain, then you will receive this output.
3488 * Generate MiniDumps independent of the Windows Crash Reporting.
3490 (a) If an exception occurs,a minidump will be generated at the path
3491 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3492 can be set to any other type via use of the
3495 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3497 (b) MiniDumps can be generated at any time using the "fs minidump"
3498 command. This allows you to generate a minidump without
3499 stopping the process or even requiring the use of a debugger
3502 * an RX connection which reports bad ticket now treats it the same way
3503 the expired ticket is treated. The ticket is removed from the
3504 connection and a new anonymous connection is established.
3506 * fs memdump now requires that the caller be in the AFS Client Admins
3509 * additional debugging information is provided in cm_Analyze when the
3510 error code is ignored.
3512 * Fix a regression introduced into the pioctl() client code which
3513 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3514 MIT KFW is not installed. (1.3.82a)
3516 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3519 * Fix a race condition in cm_EndCallbackGrantingCall() that
3520 could leave 'nrevp' pointing to freed memory. In the process,
3521 optimize the processing to ensure that the applications monitoring
3522 the status of the stat cache entry are only notified at most
3525 * Conditionalize access to the Explorer Shell AFS->Mount Point
3526 ->Remove menu item. Disable the item if the selected item(s)
3527 are not mount points.
3529 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3530 Local Storage. This eliminates thread contention on the
3531 global RX free packet queue. This should improve performance
3532 on multi-CPU systems.
3534 * Fix the IP address filtering code to properly load/unload
3535 the IPHelper DLL so that Windows XP/2003 and beyond systems
3536 do not have to manually probe the registry.
3538 * CleanupACLEnt() was not being called consistently with
3539 the cm_scache_t object referred to by the back pointer
3540 mutex locked. This could in very rare conditions lead to
3541 an invalid memory access.
3543 * Added a script command to the msi installer "afs_replace"
3544 which can be used to replace one version of OpenAFS with
3545 another without requiring an uninstall and reinstall if
3546 the installed msi is no longer in its original location.
3548 * Apparently the problem with multi-domain forests with cross-
3549 realm trusts to non-Windows realms was not entirely solved.
3550 The authentication to the AFS SMB service failed because
3551 the wrong name was being used. Using ASU as an example,
3552 the authentication was being performed with the name
3553 "QAAD\user" (an account in the forest root) and not
3554 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3556 The solution was to add an additional dependency on KFW
3557 in order or to be able to easily obtain the client principal
3558 name stored in the MSLSA ccache TGT. This information is
3559 used in two locations:
3561 - the pioctl() function
3563 - a new WinLogon Event Handler for the "logon" event.
3565 The pioctl function will now be able to use the correct
3566 name when calling WNetAddConnection2() and the "logon"
3567 event handler will now be able to call WNetAddConnection2().
3568 The hope is that the "logon" event handler will be called
3569 before the profile is loaded but I have not guarrantee
3573 * Fixed a locking error in cm_TryBulkStat() which had the
3574 potential of deadlocking the system for the length of time
3575 it takes to perform a bulk status fetch operation.
3577 * Replaced time conversion code (UnixTime <-> FILETIME) to
3578 be completely arithmetic instead of relying on a bizarre
3579 algorithm involving a variety of C RTL time functions.
3580 This has the side effect that UnixTime and FILETIME which
3581 are both stored in UTC are interpretted as UTC throughout
3582 the year. Windows will apply the same localization to AFS
3583 as it does to NTFS. Applications which rely on the ability
3584 to sync files between the two file systems will no longer
3585 see the timestamps of files in AFS change an hour relative
3586 to the files stored in NTFS or Windows based backup devices.
3588 * Fixed a invalid memory access under a bizarre circumstance.
3589 Windows will allow a physical mass media device to be
3590 installed via Plug N Play to the system and will assign it
3591 the lowest available drive letter. This is true even if the
3592 drive letter is currently assigned in the user session to
3593 a network device via NET USE (or its equivalent.) When this
3594 happens, queries sent to the CIFS server will contain invalid
3595 data. This invalid data was not being caught by the AFS
3596 Client Service and was resulting in a crash when Freelance
3599 * Fixed a reference count error when registering callbacks
3600 on a stat cache object if the callback was already registered
3602 * Add a case to cm_Analyze() in order to handle the case where
3603 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3604 the server list for the volume is empty. In this case, force
3605 an update of the volume info and retry.
3607 * Insert a missing cm_EndCallbackGrantingCall() which could
3608 result in threads waiting for a callback to complete to never
3611 * In the persistent cache, there is a maximum number of volume
3612 entries. Allow volume entries to be reused if the maximum
3613 number have been allocated and their reference count is zero.
3615 * If we already have a dead virtual connection object, don't
3616 ignore the fact that additional dead vc objects must be taken
3619 * Removed a deadlock condition introduced in the summer of 2004.
3620 Do not hold mx locks around calls to RX functions including
3621 rx_NewCall(). That is what reference counts on the rx_connection
3624 * Fixed an initialization error in afslogon.dll which could
3625 result in random behavior including a failure to terminate
3626 the NPLogon function. (1.3.80b)
3628 * Fixed an error preventing the use of SMB authentication
3629 on Windows 2000. (1.3.80a)
3632 * Updated CellServDB to Public CellServDB 16 Mar 2005
3634 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3635 misinterpretted as "csail.mit.edu" when the resolver
3636 is configured to append ".mit.edu" to failed DNS queries.
3638 * Fixed another case in which the client would replace connections
3639 to the host when it was not necessary. In this case the
3640 situation would occur if cryptall was on and the connections
3641 were unauthenticated due to lack of an appropriate token.
3643 * OpenAFS for Windows has failed to work at sites which are
3644 utilizing a cross-realm trust between an MIT/Heimdal realm
3645 and a multi-domain Windows forest when the workstation being
3646 accessed is not located in the root domain. This is caused
3647 by a bug in the workstation which was triggered after the
3648 introduction of Windows 2003 Server. When the bug is triggered,
3649 the workstation attempts to authenticate users to afsd_service.exe
3650 by contacting the Domain Controller instead of using the
3651 LSA loopback authentication mechanism.
3653 One of the reasons this bug occurs is because the workstation
3654 does not have a reliable way of knowing that the service whose
3655 netbios name is "AFS" is located on the workstation. This will
3656 be fixed starting in Longhorn Beta 1 by Microsoft. The
3657 "BackConnectionHostNames" registry value will be used to
3658 indicate that the authentications to that service name should
3659 be performed using the loopback authentication mechanism.
3661 In the meantime, when Logon Caching is enabled, we can force
3662 afsd_service.exe to authenticate using the logon cache before
3663 contacting the Domain Controller. This will work with both
3664 password and smart card based logons.
3666 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3667 flag should not be cleared if a volume's server reference is
3668 marked as "offline". In the case where all of the volume's
3669 servers are either "down" or the volumes are "offline", we want
3670 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3671 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3672 should never occur in practice.
3674 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3675 5 seconds, clear the server down and volume busy flags, and
3676 then force an update from the VLDB. This allows the client to
3677 update the location of a volume if the reason for it being
3678 marked offline is because it is being moved. Calls to
3679 cm_ConnByMServers() will be retried either until success or
3680 the RDRtimeout period is reached.
3682 * Correct the Power Management code behavior in response to
3683 Standby, Suspend, and Shutdown events. Instead of flushing the
3684 buffers associated with the mounted SMB submounts, simply write
3685 all buffers which are marked dirty.
3687 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3688 provides a list of cells other than the default cell for which
3689 tokens should be obtained using the default Kerberos principal.
3690 This functionality is logon domain specific and is only available
3691 when using KFW for authentication.
3693 * Fixed FindNext Invalid Handle error caused by over agressive
3694 attempts at garbage collecting dirSearch entries when the
3695 dirSearch ID wraps from the maximum value to 1.
3697 * Add support for registry defined server preferences for VLDB and
3698 File servers. See registry.txt for details.
3700 * Increased default cache size to 96MB and default number of cache
3703 * Fixed refCount leaks related to directory lookups and pioctl calls.
3705 * Callbacks revoked during a race condition with an attempt to
3706 obtain the same callback no longer result in an inaccessible
3709 * New command line tool:
3711 afsdacl : Set or reset the DACL to allow starting or stopping
3712 the afsd service by any ordinary user.
3714 Usage : afsdacl [-set | -reset] [-show]
3715 -set : Sets the DACL
3716 -reset : Reset the DACL
3717 -show : Show current DACL (SDSF)
3719 * IP addresses are no longer queried once at startup. Instead IP
3720 addresses are obtained as needed. Loopback adapter addresses are
3721 no longer published to the server.
3723 * Pay attention to the MIT KFW registry configuration for automatic
3724 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3726 * Fix (once again) case-sensitive comparisons which was apparently
3729 * Activate support for vos listvol -format
3731 * Population of the "AFS Client Admins" group failed on non-English
3732 versions of Windows because the name of the "Administrators" group
3733 is localized. Now we lookup the name of the group by using the
3734 Administrators Group Alias Relative ID.
3736 * Multiple cell token acquisition within afscreds.exe was broken in the
3737 case of Kerberos 5 cross realm authentication. Instead of contacting
3738 the KDC associated with the cell's realm, afscreds would obtain a
3739 token from the KDC of the user's realm. This would result in
3740 "invalid kvno errors" while authenticating to the AFS servers.
3741 Unauthenticated access would work. The symptoms would vary based upon
3742 whether or not the VLDB servers had been contacted using unauthenticated
3743 connections prior to the user obtaining tokens.
3745 * The list of ACL entries was becoming corrupted. It appears as if
3746 an ACL when it expires was not being appropriately cleaned up. In
3747 fact, it was left in the list of ACLs associated with the scache
3750 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3751 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3752 and "amd64_win64" for amd 64-bit processors.
3754 * When executing executables, Windows provides a mechanisms by which
3755 the normal search for DLLs can be bypassed. If a file foo.exe is
3756 being executed and there exists a file or directory "foo.exe.local"
3757 (the contents are ignored if a file), then Windows will search for
3758 DLLs first in the "foo.exe.local" directory and second in the
3759 directory in which the "foo.exe" file is located.
3761 Previous releases of OAFW would improperly return
3762 STATUS_NOT_A_DIRECTORY instead of the expected
3763 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3764 to terminate the search for the DLL causing the execution of the
3765 application to fail.
3767 * Rework the reference counting of smb_vc_t objects. The references
3768 stored in the waiting locks were not counted. This could result in
3769 an assertion if the reference count drops to zero.
3771 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3772 when running in freelance mode.
3774 * Added support for persistent caching of file contents, stat entries,
3775 volume data, cell data, and name lookup data. When the registry
3776 value, "NonPersistentCaching", is not defined or set to 0 the file
3777 specified by "CachePath" becomes a persistent cache file. The
3778 size of the cache file is computed at run time. It has a maximum
3779 size of approximately 1.2GB. The contents of the file will be
3780 validated according to the rules specified by the "ValidateCache"
3783 * Added support for UUIDs. UUIDs are kept for as long as the
3784 cache file is intact.
3786 * Added cmdebug.exe and support cache manager debugging callback
3787 interfaces in afsd_service.exe.
3789 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3790 [-addrs] [-cache] [-help]
3791 Where: -long print all info
3792 -addrs print only host interfaces
3793 -cache print only cache configuration
3795 * Symlinks to \\AFS\[all\]... will now be treated the same as
3796 symlinks to /afs/.... However, please use /afs/... as otherwise
3797 the symlinks won't work on Unix.
3799 * Correct a problem with local CellID allocation for cells whose
3800 server list is obtained via DNS instead of a CellServDB file. If
3801 the DNS information expires the CellID assigned to the entry will
3802 be changed. This causes all of the cm_scache_t objects which refer
3803 to the old cellID number to become useless. Attempts to access
3804 files or directories with cm_scache_t objects using the old cellID
3805 will fail since the server list cannot be obtained.
3807 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3808 be triggered if the volume referenced by the cm_scache_t object is
3811 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3812 in addition to SERVICE_ACCEPT_STOP.
3814 * Move RPC shutdown until after the SMB and RX shutdown procedures
3815 complete. Block until RPC shutdown is complete.
3817 * Modify afslogon.dll (integrated logon) to wait for service if its
3818 state is START_PENDING. If the timeout period occurs, reset to
3819 the retry interval and not the sleep interval.
3821 * When renewing the server list for a cell obtained via DNS AFSDB
3822 records, the cm_cell_t entry must be removed from the list of all
3823 cells. Otherwise, the list of cells will be corrupted.
3825 * In the dcache and scache modules, use the cm_scache_t dataVersion
3826 instead of the cm_fakeDirVersion.
3828 * Synchronize fs.c with the unix version.
3830 * The variable used to determine whether a file or virtual memory
3831 mapped cache is used was not properly initialized to a default
3832 value. If the registry setting "NonPersistentCaching" was not
3833 set, the choice would be random. Properly initialized to be
3836 * The memory mapped view was never unmapped before closing the file
3837 at service shutdown. This is now properly cleaned up.
3839 * Default location of Cache file is now %TEMP%\AFSCache
3843 * A bug affecting new installations of 1.3.75/76 would result in
3844 the creating of incorrect mountpoints in the freelance root.afs
3845 volume for the default cell.
3847 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3848 of "root.cell", you have been affected by the bug. To correct the
3849 problem, execute the following commands:
3851 fs rmmount \\afs\all\<cellname>
3852 fs rmmount \\afs\all\.<cellname>
3854 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3855 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3859 * A bug has been fixed in the auto-generation of Freelance root.afs
3860 symlinks which produced random entries in the \\AFS\all directory.
3862 * Support has been added for multi-homed servers
3865 * Added a new registry value, "StoreAnsiFilenames", which can be used
3866 to force the use of ANSI character sets instead of OEM Code Pages.
3867 This feature is useful when users require the ability to create
3868 filenames with 8-bit characters and need to access the files from
3869 both Latin-1 based Unix systems as well as from Windows.
3871 Activation of this feature will prevent access to files stored with
3872 8-bit OEM characters.
3874 * Shutdown all SMB threads in a synchronized manner when stopping the
3877 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3878 by the largest contiguous block of unused memory within the 2GB process
3879 space which can be assigned to the memory mapped file. Unfortunately,
3880 when the executable digital signature verification code is activated
3881 Windows sees fit to further segment the process memory which in turn
3882 reduces the size of the maximum cache file to less then 800MB. If
3883 larger cache sizes are desired, a new registry value should be set:
3885 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3887 Setting this value will disable the runtime verification of digital
3888 signatures on afsd_service.exe and the afs dlls which it loads. It
3889 will not disable the the version number check on those same files.
3890 The signature verification is not a security messure and is only meant
3891 to enhance the ability to afsd_service.exe to detect potential
3892 destablizing mixtures of DLLs from incompatible distributions.
3894 Added code to auto-disable the signature verification check if
3895 the desired cache size is greater then 700MB.
3897 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3898 used to verify the validity of the afsd_service.exe binary
3899 as well as each of the AFS DLLs loaded by the service. Not only
3900 must the digital signature be valid but the signatures of the
3901 DLL must be signed by the same entity as the service.
3903 * Implement new functions: cm_freelanceMountPointExists and
3904 cm_freelanceSymlinkExists. Use them along with other validity
3905 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3906 ensure that name collisions do not occur and that empty strings
3907 are not valid file names.
3909 A symlink may not have a name which would resolve to a valid
3910 cell name. Doing so would prevent access to the cell.
3912 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3913 re-initialization code. The reference counts of the fake root.afs
3914 volume scache object(s) would become invalid when the mount point
3915 or symlink lists were altered.
3917 * Add registry entries to provide mappings from the afsdsbmt.ini
3918 to the new locations for applications which count on the use
3919 of the old Profile file APIs. These apps are likely to fail
3920 if the user does not have administrator privileges and the
3921 registry is locked down.
3923 * The afs_config.exe submounts dialog had two errors.
3924 First, attempts to remove entries failed because the registry
3925 key was being opened without KEY_WRITE privileges.
3926 Second, when editing a submount entry, changing the name
3927 would add a new key and leave the original one in place.
3928 Now the original submount will be removed if its name is
3931 * In recent months there have been several incidents in which
3932 users have experienced problems starting or accessing
3933 afsd_service.exe and after significant effort has been spent
3934 it has turned out that they have two versions of AFS on the
3935 machine or an inconsistent set of DLLs.
3937 Code has now been added to afsd_service.exe which will walk
3938 the list of modules loaded by afsd_service.exe and validate
3939 that the version of the AFS DLLs matches the version of the
3940 afsd_service.exe executable. If they do not match the service
3943 * When Freelance mode is enabled and there is no registry
3944 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
3945 will attempt to import the afs_freelance.ini file contents.
3946 If the file does not exist, it was creating a dummy file
3947 with a r/o and r/w entry for the default cell and then
3948 importing those values.
3950 This process has been changed. The temporary file is no
3951 longer created. Also, both the OpenAFS Client install
3952 directory as well as %WINDIR% are checked for previous
3953 afs_freelance.ini files.
3955 * Added support for VL_GetEntryByNameN(). Still need to add
3956 support for VL_GetEntryByNameU() for multi-homed support.
3958 * Fix a deadlock situation in afscreds.exe when canceling an
3959 auto-generated Obtain Tokens dialog
3962 * Fix the locking of objects during Directory Searches in the
3963 SMB/CIFS server. The failure to properly lock the reference
3964 counts was resulting in the premature freeing of smb_dirSearch_t
3965 objects while they were still in use by the SMB/CIFS client.
3966 This does not solve the "Invalid Handle" problem.
3968 * Fix Find Cell By Name pioctl call to return a valid cell
3969 name for the Freelance fake root.afs volume.
3970 "Freelance.Local.Root".
3972 * Fix the Explorer Shell Extension Symlinks->Add operation.
3973 The dialog template was missing and the link destination
3974 string was too short.
3976 * Add support for symlinks to Freelance root.afs volume
3977 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
3978 <number> = "<linkname>:<relative-path>."
3979 Use symlink.exe to create, list, or remove
3981 * Remove the fallback to the use of KFW's KRB4 library when obtaining
3982 tokens. We never obtain KRB4 tickets.
3984 * Fix AFS Client Configuration Control Panel to support new SysName
3987 * Fix a bug in afsd_service.exe which could result in the SysName
3988 not being read from the registry.
3991 * Add code to block the issuance of AFS tokens by aklog.exe or
3992 afscreds.exe when the Kerberos 5 principal name contains a dot.
3994 * Modify the IsAdmin() function to always treat the local SYSTEM
3995 account as an AFS client administrator. Affects fs.exe and
3998 * Modify the internal handling of Quota Exceeded errors
4000 * Upgrade all reference count fields in the Windows cache manager
4001 and the osi library to use unsigned long instead of signed short.
4002 A similar fix has been applied to the afs rpc (rx) library.
4004 * fix the Windows cache manager to prevent it from replacing the
4005 rx_connection object associated with the cm_conn_t object on each
4006 and every operation if "fs crypt" was set. This explains the
4007 dramatic performance difference when crypt is used vs clear.
4008 The problem: 'cryptall', a boolean flag indicating whether or not
4009 "fs crypt" is set, was being compared to the rx_connection
4010 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
4011 1 != 2 and therefore the rx_connection was always destroyed
4012 and replaced on each and every operation.
4014 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
4015 It is not safe for the cm_conn_t object to not be locked because
4016 rx_DestroyConnection might be called from another thread if:
4017 - the user's tokens have changed or expired
4018 - the crypt mode has changed
4020 This fix appears to have also taken care of the problems associated
4021 with Overlapped Writes resulting in Delayed Write errors.
4023 * fix NSIS installer's AdminGroup.exe to properly create and
4024 remove groups when given -create or -remove. The string comparison
4027 * fs sysname now accepts a list of sysname values
4029 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
4030 DWORD which when set to a non-zero value will cause error message
4031 text to be output to stderr from the pioctl() routine. Useful in
4032 debugging failures of fs.exe, tokens.exe, etc.
4034 * added a test to the power management code to only perform a
4035 flush operation if there is at least one network adapter which
4036 is not a loopback adapter.
4038 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
4039 "EnableKFW". This value will not be read if the key
4040 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
4041 value under that key does not.
4043 * provide mechanisms to force the use of krb524d for Kerberos 5
4044 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
4045 there is a new registry value "Use524" and for aklog.exe a new
4046 command line parameter "-m".
4048 * Fix the pattern matching algorithm to properly match patterns
4051 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
4052 which is used by both the original function and the new
4053 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
4054 creation of HardLinks in addition to Renaming. smb_Link()
4055 is a new function which creates HardLinks via cm_Link().
4056 cm_Link() is a new vnodeops function which creates links
4059 smb_ReceiveNTRename() does not support the File Copy and
4060 Move Cluster Information operations described in its interface.
4061 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
4063 * When opening files via symlinks, we should follow the symlinks
4064 until we reach the actual file stat cache entry. The stat cache
4065 entry of the file should then be stored in the FID instead of
4066 stat scache entry of the symlink.
4068 * return bad operation errors for all unimplemented functions
4069 even if we do not know the functions exist.
4071 * Log bad packets and unknown operation packets to the trace log
4073 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
4076 * Update list of known CIFS operations to include all those listed
4077 in CIFS-TR-1p00_FINAL.pdf.
4079 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
4080 to support the REG_EXPAND_SZ type.
4083 * A new Windows authorization group "AFS Client Admins" is now
4084 created and populated with the members of the "Administrators"
4085 group. The group is used to determine which accounts on the
4086 machine may be used to modify the AFS Client Configuration via
4087 the UI and command line tools. afs_config.exe, fs.exe,
4089 * Modify the WinLogon Logoff Event Handler to query NT4 domain
4090 controllers for the remote profile path if Active Directory
4091 services are not available.
4093 * Fix aklog.exe to not add the AFS ID to the username
4095 * PTS registration of new users to foreign cells has been added to
4098 * The cm_Daemon thread is used to perform checks for
4099 down servers, up servers, volumes, callback expirations,
4100 lock maintenance and token expiration. Due to a gaff in
4101 larger integer division the thread never performed any
4102 work. Instead the current time computation would always
4103 be less then the trigger times. This had an adverse affect
4104 on the client's ability to maintain communication with servers,
4105 keep volumes up to date, and flush user tokens and acls
4106 when they have expired. This was broken when the 1.3 branch
4107 was modified to support VC7 which no longer included
4110 * An initialization problem with the Freelance code was
4111 detected while fixing the callbackRequest. The cm_rootSCachep
4112 object is obtained during afsd_InitDaemons() but the callback
4113 information is incomplete. The callback information will not
4114 be obtained until cm_MergeStatus is called from within
4115 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
4116 test for the conditions under which the callback information
4119 * Reports have been filed indicating that callbacks were
4120 being lost. An examination of the code indicated that the
4121 cm_server_t objects were not being properly reference
4122 counted by the cm_scache_t and cm_callbackRequest_t objects.
4123 In particular, the cm_server_t objects may have been freed
4124 from beneath the cm_conn_t objects.
4126 All of the reference counting is now done via the functions:
4131 this improves the ability to track the referrals.
4133 Each cm_BeginCallbackGranting Call now allocates a reference
4134 to the cm_server_t. The cm_EndCallbackGrantingCall either
4135 frees the reference or transfers it to the cm_scache_t
4136 cbServerp field. These are then appropriately tracked
4137 through the cm_Analyze call.
4139 * Ensure that the dnlc hash table is the same size as the
4140 dir name hash table (as per original author's note).
4141 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
4142 for compatibility with 64-bit systems.
4144 * fix smb_ApplyV3DirListPatches to properly apply the hidden
4145 attribute to dotfiles when the infoLevel < 0x101 and
4146 cm_SyncOp has failed.
4148 * Fix the Freelance registry initialization code. There
4149 was a possibility that some systems could end up with
4150 garbage in the registry during a clean install.
4153 * file and directory names beginning with "." will now be given the
4154 hidden attribute when the volume access is anonymous. this matches
4155 the behavior when the volume access is via an authenticated user.
4157 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
4158 key. When a change occurs mark the root.afs data as invalid and
4159 for it to be reloaded on the next access. This allows administrators
4160 to modify the mount point list without restarting the service.
4162 The freelance client used to provide a fake modification time for
4163 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
4164 Added code to extract the last modification time of the Freelance
4165 registry key and use that instead. The time now represents the
4166 most recent mount point change.
4168 * PTS registration of new users to foreign cells has been added to
4171 * Additional Cache Control and Credential Manager options have been
4172 added to the WiX installer. See deployment guide for details.
4174 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4176 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4179 * The loopback installation code may have had a problem updating the
4180 %ETC%\HOSTS file which could have resulted in a premature failure.
4181 Work around code has been added for the case where the file cannot
4184 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4185 because Windows sends 64K blocks when using overlapped writes.
4187 * The default number of server threads was increased from 4 to 25 to
4188 better handle overlapped writes.
4190 * The "AfscredsShortcutParams" registry value was not being properly
4191 loaded by afscreds.exe. Therefore, the default value was always being
4192 used instead of the value set by the installer.
4194 * Windows XP provides downgrade attack detection to prevent an attacker
4195 from being able to force the use of NTLM simply by disrupting
4196 communication with the KDC. This attack cannot exist between the
4197 Windows CIFS client and the AFS Client Service. Therefore, when a
4198 downgrade has been detected the afs pioctl library will force the
4199 establishment of a new CIFS connection using NTLM.
4201 * A locking error was discovered surrounding all references to volume
4202 server lists within the cm_cell.c source file.
4204 * The logged into Windows username was incorrect on Terminal Server
4207 * A new registry value "NonPersistentCaching" was added to the service
4208 parameters key. When set to a non-zero value, the afs cache is stored
4209 in the Windows paging file. There are two limitations to choosing
4211 1. when persistent caching is implemented it won't work with
4212 this flag set since there will be nothing to persist.
4213 2. with this flag set the initial paging allocation cannot be
4214 changed while the service is running
4216 * An initialization bug was discovered in aklog.exe which affected users
4217 who have a domain name for their afs servers which could not be mapped
4221 * afs_config.exe now validates cell names against DNS in addition
4222 to the CellServDB file.
4224 * In order to allow the freelance client to connect to a volume with ID
4225 equal to 1 on the default cell we changed the fake root.afs volume ID
4226 once again. This time we choose 0xFFFFFFFF. In addition, we change
4227 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4228 It will now be impossible for a volume ID to match that of another
4229 cell unless the client is connected to 0xFFFFFFFD cells. That should
4230 be enough room for growth.
4232 * Fix "fs mkmount" command to work with UNC paths and when
4233 started from non-AFS drives. It is now possible to create a mount
4234 point in the freelance fake root.afs volume with the command
4236 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4240 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4241 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4243 * The algorithm used to re-attempt access to the servers associated with
4244 a volume has been altered to properly address the case in which all
4245 servers have been marked down. The previous algorithm did not reset
4246 the server's down flags so the servers were never actually retried.
4247 This caused a problem with active volumes if the network connectivity
4248 was lost as could be the case with a network cable removal, wireless
4249 drop, or laptop hibernation. With the fix volume access is restored
4250 almost instantenously when network connectivity becomes available.
4252 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4253 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4254 listing of AFS submounts and freelance mount points. Support for
4255 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4256 significantly enhances the behavior of AFS volumes within the Explorer
4257 Shell. For instance, "AFS" now shows up as server in the Explorer
4258 with each submount or freelance mount point visible as a share.
4259 The right click menu in each folder now works with full functionality
4260 on a consistent basis.
4262 * The network provider can be configured to have different behavior
4263 depending on the domain that the user logs into. These settings are
4264 only relevant when using integrated login. A domain refers to an
4265 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4266 local machine (i.e. local account logins). The domain name that is
4267 used for selecting the domain would be the domain that is passed into
4268 the NPLogonNotify function of the network provider. (see registry.txt
4271 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4272 "Authentication Cell" which may be used to specify a default
4273 authentication cell for afscreds.exe which is different from
4274 the default cell for the AFS Client Service daemon.
4276 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4277 afslogon.dll moved to %WINDIR%\System32\.
4278 New registry entries added to register the dll for Winlogon events.
4280 The logoff event will now force a call to ktc_ForgetAllTokens()
4281 using the context of the user being logged off as long as the
4282 user's profile is not loaded from within AFS. If the profile
4283 was loaded from AFS we can't release the tokens since the Logoff
4284 event is triggered prior to the profile being written back to
4285 the its source location. This is now performed in an XP SP2
4288 * Windows XP SP2 Internet Connection Firewall interoperability
4291 * The %WINDIR%\afsdsbmt.ini contains four sections:
4292 Submounts, Drive Mappings, Active Maps and CSC Policies.
4293 The Submounts and CSC policies are now stored in the registry under
4294 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4295 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4296 The Drive Mappings and Active Maps are stored in the registry under
4297 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4298 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4300 There is no automatic migration of this data as it would be impossible
4301 to consistently migrate data to user profiles which may not be active
4302 when the machine is updated.
4304 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4305 fake root.afs volume. For the same reasons as for the cellservdb file,
4306 this information should not be in %WINDIR%. This information is now
4307 kept under the registry key
4308 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4310 The data from the afs_freelance.ini file will be automatically
4311 migrated to the registry on first execution of afsd_service.exe
4313 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4314 troublesome for several reasons. One, it is confusing for those who
4315 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4316 Two, this file is not a Windows Profile formatted file. Three,
4317 applications should not be reading or writing to %WINDIR%. It causes
4318 problems for Windows Terminal Server.
4320 The new location of CellServDB will be the OpenAFS Client install
4321 directory which is by default C:\Program Files\OpenAFS\Client and can
4322 be determined by querying the registry for
4323 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4325 The existing afsdcell.ini will be migrated by the NSIS installer.
4326 The Wix installer must still be updated to do the same.
4328 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4329 High Security mode; and to add Terminal Services compatibility registry
4330 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4331 configuration files in %WINDIR%.
4333 * Add support for authenticated SMB connections. This will remove
4334 the need for high security mode in most situations. Both NTLM
4335 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4336 only NTLM can be used even though Kerberos is now supported. The
4337 reason is that it is not possible to construct a service principal
4338 which is unique to each individual machine.
4340 SMB Extended Auth does not work on XP SP2 unless one of two registry
4341 modifications are made:
4343 (1) To disable the check for matching host names on loopback connections
4344 set this key. This does not require a reboot:
4346 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4347 "DisableLoopbackCheck"=dword:00000001
4349 (2) To add the AFS SMB/CIFS service name to an approved list. This
4350 does require a reboot:
4352 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4353 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4355 afsd_service.exe will automatically add the current Netbios Name
4356 to the BackConnectionHostNames list and then temporarily disable
4357 the loopback check for one cycle of startup/shutdown of the service.
4358 We assume most folks do not start/stop without a reboot so this
4359 will be adequate in most cases.
4361 * Fix security hole in afslogon.dll which allowed passwords to be
4362 sent in clear text to the KDC in a misformed principal name.
4364 * Fix cm_GetCell() to properly handle expired dns entries
4367 * If Freelance mode is active and the afs_freelance.ini
4368 file does not exist, do not create an empty file.
4369 Instead create a file containing ro and rw mountpoints
4370 to the default cell using the standard conventions.
4372 * Modify the Freelance support to handle the ability
4373 to create rw mount points in the fake root.afs volume.
4375 * Changed the RPC mechanism used for token setting from
4376 named pipes to local. Use of named pipes can be restored
4377 by setting the environment variable AFS_RPC_PROTSEQ to
4380 Named pipes were required when a Windows 9x system was
4381 using a NT system in gateway mode which is incompatible
4382 with our use of local loopback adapters.
4384 * In afscreds.exe, if a username of the form user@REALM is
4385 specified and no password is specified, do not perform a
4386 kinit operation. Only perform the aklog functionality.
4388 * Add a new registry value which allows the number of processors
4389 on which afsd_service.exe executes to be restricted. Valid
4390 values are 1..numOfProcessors
4392 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4396 * A second MSI based installer option is now available.
4398 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4399 request tickets for user/instance@REALM instead of just user@REALM
4401 * Modify the Power Management Notify routine to wait for the Hard Dead
4402 timeout period instead of a fixed 19 seconds. With the longer timeout
4403 periods Hibernation and Standby could never succeed when network
4404 connectivity is not available.
4406 * The following fs.exe commands are now restricted to Administrator:
4407 - checkservers with a non-zero timer value
4410 - sysname with a new sysname list
4419 setting the default sysname for a machine should be done via the
4420 registry and not via "fs sysname".
4422 * NSIS installer adds options to install Debugging Symbols
4423 and the Microsoft Loopback Adapter; the user is now also
4424 given the ability to select the afscreds.exe startup options.
4426 * Build system modified to generate symbols for FREE (aka RELEASE)
4427 builds as well as CHECKED (aka DEBUG) builds
4429 * Sites which have a volume ID of 0x20000001 assigned to their
4430 root.cell volumes have been experiencing problems with accessing
4431 the root.cell volume of their cell when Freelance mode has been
4432 active. This was because 0x20000001 was assigned to the fake
4433 root.afs volume created by freelance. The fake volume id is
4434 now set to 0x00000001 to prevent conflicts.
4436 * The timeout logic in the AFS Client Service has been wrong
4437 for sometime. It is based on two different assumptions.
4438 First, the SMB client timeout is a fix value as was the case
4439 with OS/2 Lan Manager. This assumption is incorrect. The
4440 SMB timeout in Windows is a dynamic value computed based upon
4441 a fixed minimum timeout to which is added time based upon the
4442 size of the request and the performance characteristics of
4443 the connection. Second, it is the responsibility of the
4444 SMB Server to enforce the timeout requirements of the client.
4445 This is untrue. The SMB Server cannot be expected to know
4446 the requirements of the client. More importantly, if the
4447 SMB server uses the SMB client timeout as a value to restrict
4448 its behavior as an RX client, the performance characteristics
4449 of the local SMB session would be used to prematurely terminate
4450 WAN connections with significantly different performance
4453 The timeout logic has therefore been modified in the following
4455 . the Lan Manager Workstation (SMB) Session Timeout is used only
4456 as a basis for configuring the Connection Dead Timeout
4457 and Hard Dead Timeout values. The Connection Dead Timeout
4458 must be at least 15 seconds longer than the SMB Timeout
4459 and the Hard Dead Timeout must be at least double the
4460 Connection Dead Timeout.
4461 . New registry entries have been added to allow the Connection
4462 Dead Timeout and Hard Dead Timeout values independent of the
4463 Lan Manager Workstation Session Timeout
4464 . The test to enforce the SMB Client Timeout has been removed.
4466 One of the side-effects of removing the enforcement of the SMB
4467 Client Timeout is that regardless of whether or not the SMB client
4468 is available to receive the response (and how would the SMB server
4469 know) the RX protocol response can be used to update the AFS
4470 Client Service state for ready access by future SMB client
4473 This should be the end of the "Server paused or restarting messages"
4475 * Add support for arbitrary UNC paths to the pioctl() support.
4476 This enables the fs commands as well as the AFS Shell Extension
4477 to work correctly when UNC paths are being used.
4479 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4480 DNS if the cell configuration cannot be determined via CellServDB
4482 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4483 are ever set more than once at a time
4485 * Fix the management of lists of cm_cell_t structures when using
4486 DNS to lookup cell information. The previous code would fail to
4487 reuse the same cellID for a cell if DNS was used more than once
4488 for a given cell name. When the ttl expired, a single cm_cell_t
4489 could be inserted into the cm_allCellsp list more than once
4490 producing a loop. In addition, the vlServerp list belonging to
4491 the cell was not freed resulting in improper refCounting of the
4494 * Add DNS support to cm_IoctlNewCell() which previous only examined
4497 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4498 to properly garbage collect cm_server_t objects
4500 * Add numVCs variable to smb.c to track the number of smb_vc_t
4501 objects created and use it to initialize the vcID field which
4502 previously was set to 0 in all objects resulting in FindByID
4505 * Fixed DNS lookups to work consistently throughout the OpenAFS
4506 product instead of just from within the afsd_service.exe
4508 * Added a runtime check to ensure that AFS Client Service SMB
4509 Server is accepting connections before attempting to mount
4512 * Read IP addresses for volume servers out of the CellServDB
4513 file if gethostbyname() on the hostname fails.
4515 * Fix getcellconfig() to populate both the Hostnames as well
4516 as IP addresses when loading cell data via DNS
4518 * Increase the Connection Dead Time to 50 from 20 seconds
4519 Increase the Hard Dead Time to 120 from 40 seconds
4520 (matches the Unix values)
4522 * Fixed an assertion validating the number of allocated NCBs
4524 * Fixed the build environment to consistently build for
4525 Windows 2000 and above. (APPVER = 5.0)
4527 * Fixed rx_debug to properly validate the receipt in incoming
4528 data with select() and recvfrom(). Do not copy data out of
4529 the socket buffer unless success is indicated.
4532 * afsd_service.exe will now display a message box to the
4533 desktop when it terminates due to an IP Address Change.
4535 * installer no longer deletes AFS Server configuration data
4538 * installer generates a warning dialog if the RPC service
4539 is not properly configured
4541 * installer compressed with lzma instead of bzip2
4543 * afsd_service.exe shutdown crash solved once and for all
4545 * reference counting of smb_vc_t data structures improved
4547 * name space collision of smb_fid_t event objects corrected
4549 * the output of "fs memdump" is now written to
4550 %WINDIR%\TEMP\afsd_alloc.log
4552 * the file TaAfsApp_1033.dll is now properly installed allowing
4553 the User Manager to start
4555 * a new algorithm is used for computing filename pattern matches
4557 * afscreds.exe now accepts user names containing instance
4560 * Fix the Directory Name Lookup Cache to be case-sensitive.
4561 This is crucial in environments in which a Windows client
4562 is accessing a directory with more than one filename that
4563 differs only by case. If the directory contains "FOO"
4564 and "Foo". You want "DEL Foo" to delete the correct one.
4565 We still have a problem in that "DEL foo" will delete a
4566 random filename. This will be addressed in a future release.
4568 * Fix afscreds.exe -M option (renewMaps) to work when High
4569 Security mode is off. Also, remember to disable the ActiveMap
4570 flag in afsdsbmt.ini when a drive mapping is removed.
4572 * Updates to NSIS installer script. AFS Server configuration
4573 data will not be destroyed on un-install or re-install.
4574 Use a better compression algorithm.
4576 * afslogon.dll now uses KFW to obtain tokens when available
4578 * afslogon.dll when given an all uppercase username will
4579 attempt to authenticate with both the uppercase name
4580 and an all lowercase variation
4582 * DST modification removed. The fix appears to make things
4583 worse after a reboot of the machine.
4585 * fs.exe: added "cscpolicy" which is used to
4586 change client side caching policy for AFS shares
4588 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4589 [-documents] [-disable] [-help]
4591 * Several uninitialized variables have been initialized
4593 * It is now possible to obtain tokens using cross realm
4594 Kerberos within afscreds.exe:
4596 user: jaltman@ATHENA.MIT.EDU
4598 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4599 will will in turn be used to obtain afs@DEMENTIA.ORG.
4600 The resulting token will be stored with the display name
4601 jaltman@ATHENA.MIT.EDU@dementia.org
4603 * aklog.exe has been added to the client
4605 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4606 [[-p | -path] pathname]
4610 -d gives debugging information.
4611 krb_realm is the kerberos realm of a cell.
4612 pathname is the name of a directory to which you wish to authenticate.
4613 -noprdb means don't try to determine AFS ID.
4614 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4615 (default is Kerberos V)
4616 No commandline arguments means authenticate to the local cell.
4619 * All of the resource files have been restructured to adhere to
4620 a set of rules IBM implemented for loading string resources.
4621 These rules had either been forgotten or were not discovered
4622 by folks working on the OpenAFS sources. The end result was
4623 memory corruption. This is primary item which was preventing
4624 the AFS Server from working.
4626 * Increased the size of the maximum ticket size stored in a token
4627 from 344 bytes to 12,000. Increased the buffers used to convey
4628 messages between the pioctl() caller and the SMB Server from
4629 1000 bytes to 12,512. The code appeared to have been writing
4630 above the top of the stack by quite a few number of bytes.
4631 (The increased ticket size is necessary for the next item.)
4633 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4634 Instead the raw Kerberos 5 ticket is used in its entirety. This
4635 is extremely important as it allows us to use pure Kerberos 5 KDCs
4636 as the source of the AFS authentication. The use of up to 12,000 byte
4637 tickets will allow tickets produced by all versions of Microsoft
4638 Active Directory to be used.
4639 - create a user account.
4640 - designate it DES only
4642 - specify its UPN to be "afs@realm"
4643 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4645 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4646 is special when using long file names. (you must use "*.*" and not "*")
4647 Instead only enforce it when performing 8dot3 searches.
4649 * Fixed the DST problem with creation times being set one hour ahead
4651 * Fixed the problem when using \\afs\cell-alias. For example,
4652 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4653 for the alias name; instead simply expand the name. One of the
4654 symptoms of this problem was a loss of acquired tokens.
4656 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4657 of strings. (Only English strings provided.)
4659 * Fixed the installer to properly replace in use files.
4661 * Fixed the build system to cleanup generated component version files
4663 * The release build compiled with MSVC 6.0 compiler to avoid the
4664 afsd_service.exe shutdown crash. This does not solve the problem
4665 but simply avoids it for the time being.
4669 * fix afslogon.dll to not corrupt memory when High Security mode
4672 * fix afsd_service.exe to not attempt to restore the stack when
4673 an exception occurs. (not safe in multi-threaded programs)
4675 * fix uninstaller to properly remove the CRT and MFC DLLs
4677 * remove a Message Box from afscreds.exe when getcellconfig()
4678 fails on a kerberos realm which is not a cell
4680 The following is a list of changes to the OpenAFS for Window client
4683 * "fs setserverprefs" will leave afsd service deadlocked
4685 * "vos listaddrs" will core dump
4687 * installer sets the appropriate keys to support Integrated Logon
4689 * installer disables the "Find Lana by Name" functionality as it
4690 was causing headaches for many users
4692 * fix the intermittent crash of the power management thread when
4693 shutting down the AFS Client Service
4695 * optimizes the obtain drive mount list functionality which is
4696 executed every time the mount tab in afscreds.exe and afs_config.exe
4697 are refreshed. (this happens a lot)
4699 * fix the service shutdown logic. add the STOP_PENDING state
4700 and do not accept additional service events after we declare
4703 The following is a list of changes to the OpenAFS for Window client
4706 * flexelint was run against the source tree and hundreds (perhaps
4707 thousands) of corrections were applied to ensure prototypes
4708 were in use; types were used consistently; variables were
4709 initialized; unused variables were removed; etc.
4711 * A wide variety of instrumentation was added including the
4712 ability to produce a stack trace from within afsd_service.exe
4715 * Dynamic configuration of the RDRtimeout value based upon the
4716 LanMan Workstation Session Timeout
4718 * The mount root no longer needs to be called "/afs". This
4719 is now set by a registry value "MountRoot" within the key
4720 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4722 * The cell list is now only read out of afsdcell.ini when the
4723 file changes instead of each time a cell is resolved.
4725 * Thread synchronization was added to cm_server.c and ktc_nt.c
4727 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4728 calloc()/free(). The Global functions were needed on Windows 3.x
4729 but have caused a variety of problems on the Win32 platforms.
4730 Avoiding them is highly recommended by several Microsoft
4731 Knowledgebase articles
4733 * Support for Symbolic Links added to the AFS Shell Extension
4735 * Added a registry value "OverlayEnabled" to determine if
4736 Shell Extension Overlays should be enabled.
4737 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4739 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4740 separate trees for checked and free builds. Build system supports
4741 a custom directory src\WINNT\extra which can be used as a grafting
4742 location of organization specific additions to the build tree.
4744 * New installer built using NSIS 2.0.
4746 * Named all kernel objects in order to allow them to be monitored
4747 with tools such as SysInternals' ProcExp.exe.
4749 * Introduced new EventLog framework for AFSD
4751 * Introduced Power Management interface to AFSD for Standby and
4752 Hibernate modes to allow cache to be flushed prior to network
4755 * Utilize Win32 DNSQuery API instead of internal routines. This
4756 allows DNS SRV queries to be sent to all current domain name
4757 servers. Not just one specified in an INI file. DNS is now
4760 * "NetbiosName" registry value may be used to specify a fixed
4761 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4762 when the loopback adapter is in use. If you need to use the
4763 old notation with a loopback adapter installed specify a registry
4766 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4768 * Refactor all modules which depend on LAN Adapter and NetbiosName
4769 determination in a new library: lanahelper.lib. This allows for
4770 consistent behavior throughout the product.
4772 * Move the afsd.log and afsd_init.log files to the directory specified
4773 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4774 for services. Added the Date to the log entries.
4776 * New registry value "RxMaxMTU" used to limit the size of the RX
4777 packets sent by the AFS Client Service to the Server. In order
4778 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4779 size must be restricted to 1264 or smaller. The latest NSIS
4780 installer sets a value of 1260 by default.
4782 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4783 packets. This is not needed in order to work across the Cisco
4784 VPN but might be needed for other network environments. This
4785 value is not set by the NSIS installer.
4787 * New registry value "HideDotFiles" is used to apply the Hidden
4788 attribute to files whose names begin with a '.'. This value
4789 is set by the NSIS installer.
4791 * New registry value "MaxMpxRequests" allows the maximum number
4792 of multiplexed sessions to be configured at run time. This
4793 value is not set by the NSIS installer. The default value is
4796 * New registry value "MaxVCPerServer" allows the maxmimum number
4797 of VCs per server to be configured at run time. This value is
4798 not set by the NSIS installer. The default value is 100.
4800 * New registry value "AllSubmount" allows the "all" submount to
4801 be disabled by setting its value to 0x00.
4803 * Allow cells names to be valid mount points
4804 \\<netbiosName>\<cellname>
4806 * Store the active state of drive mappings in order for afscreds.exe
4807 to restore them upon startup
4809 * Add exception handling to generate a Stack Trace to the afsd_init.log
4810 file if one happens to occur.
4812 * Add lots of logging to help detect the cause of invalid SMB packets
4814 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4815 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4816 renewal of expiring tokens as long as afscreds.exe is running.
4818 * New afscreds.exe command line options:
4820 -M = renew drive maps
4821 -N = ip address change detection
4824 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4825 determines whether or not MIT Kerberos for Windows should be used
4826 to obtain tokens via Kerberos 5 tickets.
4828 * New registry value "AfscredsShortcutParams" in
4829 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4830 determines the command line parameters to be specified when "fixing"
4831 the AFS Shortcut in the user's startup folder.
4833 * The "ShowTrayIcon" registry value has been moved from
4834 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4835 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4837 * The <cell name> registry values used to store the token expiration
4838 reminders have been moved from
4839 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4840 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4842 * Obtain the Logon User Name from the Explorer key when available
4844 * new text document doc\txt\winnotes\registry.txt lists all registry
4845 values used by OpenAFS (excluding the AFS Server)
4847 * BUG: rx_securityClass objects were not properly reference
4848 counted and were never freed.
4850 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4851 would be generated. The existence of a server does not imply
4852 that it is not down. If all of the servers for a cell are down
4853 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4856 * BUG: the directory name lookup cache failed to free the entries
4857 in the cache when the name cache entries cycled. The entries
4858 in the cache would become dereferenced without being freed.
4860 * BUG: fs setserverprefs could be executed without Administrator
4863 * BUG: the number of allocated NCB objects (100) exceeded the number
4864 which could actually be waited upon by the kernel (64). Any objects
4865 which were utilized above the limit could never have event completions
4868 * BUG: smb_username_t objects were not being reference counted and
4869 were not properly freed.
4871 * BUG: smb_tid_t objects could under unusual circumstances be freed
4872 before they were no longer referenced.
4874 * BUG: smb_fid_t object pointer were frequently used even when
4875 their value could be NULL. They were not properly released and
4876 therefore they were never freed.
4878 * BUG: smb_packet_t data structures were not completely initialized
4881 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4882 Denied" because that causes the Explorer Shell to try again until
4883 access is obtained. Instead return "Remote Resources" which allows
4884 the shell to move on and treat the error as transient.
4886 * BUG: when initializing the NCBreturns structure, separate Event objects
4887 were created for each NCB although a single Event object was supposed
4888 to be shared by all.
4890 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4893 * BUG: smb_tran2Packet_t objects were not being properly referenced
4896 * BUG: directory path creation did not handle the case of multiple
4897 directories requiring creation in one attempt
4899 * BUG: SMB requests which required an Extended Response were ignored.
4900 This prevented some files from being written to AFS volumes.
4902 * BUG: character strings were being freed even after they were
4903 inserted into in use data structures
4905 * BUG: inconsistent usernames were used when High Security mode was
4906 enabled. (there is still much to do in this area)
4908 * BUG: pioctl() calls which require out of band RPC operations were
4909 susceptible to race conditions when performed by multiple processes
4911 * BUG: memory allocation and deallocation crossed instances of the
4912 C Runtime Library producing memory leakage and corruption in
4913 afscreds and the client configurator.