2 * Convert afsauthent.dll to use the xdr_alloc and xdr_free
3 routines from afsrpc.dll in place of malloc and free.
4 This permits the two libraries to be built with different
5 C runtime libraries without crashing.
7 * Set the DOS Readonly attribute on a file/directory
8 whenever the unix mode combined with the mask 0200
9 is true. Previously there was a discrepency between
10 the mask used for testing for readonly behavior and
11 that used for setting the attribute.
13 * Disable AFSVolSync based .readonly "whole-volume callback"
14 support because the all file servers prior to 1.5.67
15 (and perhaps 1.4.12) do not properly assign a value to
16 the AFSVolSync structure in bulk status RPC responses.
18 * Improve the error output from aklog to output the value
19 from krb5 error_message() if the afs_com_err output
20 indicates an unknown value.
22 * Hold a lock on the cm_scache_t object when the smb
23 server is assigning a new clientModTime value.
25 * Permit custom build version numbers to be set
26 without modifying the src/config/NTMakefile.<platform>
29 * Add support for binary code signing with cross-signed
30 certificates and arbitrary versions of signtool.exe.
32 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and
33 do not permit them to be exposed to the smb redirector.
35 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT
36 to avoid confusion within the smb redirector.
38 * Fix the byte order assigned to port numbers associated
39 with AFSDB record lookups. They must be network byte
40 order not host byte order.
42 * Add dynamic server ranking based on RPC round trip
46 * Official support for Windows 7 and Server 2008 R2
48 * Improved Rx RTT computation following the lessons
51 * Prevent a file server bug (FetchData returning an
52 invalid length instead of zero) from causing an
53 "unexpected network error" when writing to files.
55 * Promote DNS SRV records as superior to DNS AFSDB
56 records. Support arbitrary port numbers for vldb
59 * Add "fs listacl -cmd" support.
61 * Add AFSVolSync based .readonly "whole-volume callback"
62 support. With this functionality, multiple objects from
63 a .readonly volume can have their status validated by
64 issuing a single RXAFS_FetchStatus RPC.
66 * Remove drive mapping functionality and service start/stop
69 * Remove drive mapping functionality from afs_config.exe.
71 * Adjust SMB error return codes to avoid returning
72 STATUS_TIMEOUT which results in the SMB redirector
75 * Network Identity Manager OpenAFS Provider now provides its
76 own "AFS lock" notification icon to report the status of
77 "have tokens, have no tokens, service not started, service
78 started but inaccessible". Hovering over the icon lists the
79 cells for which tokens exist (if any) and the OpenAFS version
80 number. Double-clicking executes the Network Identity Manager
83 * When merging the current status for an AFS object, ensure
84 that the volume status for that object is set to 'vl_online'.
86 * Permit cm_Analyze to request a retry even when the cm_req_t is
87 flagged CM_REQ_NORETRY if the RPC failed due to RX_CALL_DEAD.
88 The retry will force the use of a new connection.
90 * Add digital signatures to resource only dlls as required by
93 * Prevent pioctl calls from retrying indefinitely when a sharing
94 violation error occurs.
96 * Add {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab"
97 to restore access to drive mapping functionality in afscreds.exe
100 * Permit Rx minimum peer timeout, maximum receive window size and
101 maximum send window size to be configured via the registry.
103 * The MSI installer now sets the ARPINSTALLLOCATION as required by
106 * DNS lookups set the port numbers in network byte order within
109 * Replace use of the 16-bit compatibility API WinExec with
110 ShellExecuteEx in afs_config.exe and afscreds.exe. WinExec
111 is incompatible with Windows 7 / Server 2008 R2 UAC
115 * Short circuit background volume checks if a shutdown
116 or suspend request is received.
118 * Avoid contacting the file server in order to allocate
119 an empty buffer beyond the current length of the file.
121 * When a request for streams on a directory or mount
122 point object is received, do not offer a default stream.
123 This was affecting the saving of roaming profiles.
126 * Fix Windows 2000 compatibility. Do not call rand_s()
127 even if it is compatible with the C run time library
130 * Fix a data consistency error between the output of
131 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
132 workstation group name "AFS", and report server name as
133 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
135 * Enable executables to be run from \\AFS on Windows 7.
136 Return "Name not found" instead of "File not found" when
137 a directory or file name cannot be found.
139 * Prevent cache manager from marking file server "down" when
140 the data returned in response to either RXAFS_FetchData64
141 or RXAFS_StoreData64 is invalid.
143 * Add pioctl data validation to the AFS Explorer Shell extension.
146 * After 1.5.62 afslogon.dll would no longer crash but it
147 also would not load "domain" specific configurations.
150 * An uninitialized variable was present in the symlink
151 recursion detection loop.
153 * Reverse addition of unique per file GUID during
154 Create/Open responses. Inclusion of the GUID prevents
155 Cygwin applications (and others) from accessing //afs.
157 * Treat "filename::$DATA" as a synonym for "filename".
158 This addresses the recent report of roaming profiles
159 failing to save back to the file server.
161 * When given a choice between an error from rx_Write()
162 or RXAFS_EndStoreData(), use the rx_Write() error.
163 This ensures that over quota and access denied errors
166 * If an error occurs during RXAFS_StoreData operations,
167 apply that error to all of the buffers in the BIOD.
168 This ensures that failures will not be retried if they
169 will only fail again.
171 * Do not perform a DNS lookup for a cellname if the
172 cellname prefix is _._AFS_IOCTL_._. This will avoid
173 DNS lookups when the query is for
174 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
176 * Fix Freelance root.afs volume object callback processing.
177 Object synchronization and status merging bugs.
179 * Fix Explorer Shell "invalid parameter" bug introduced
182 * Log RXKAD errors to the trace log.
184 * Prevent simultaneous pioctl calls from multiple processes
185 from stepping on each other's data.
187 * Prevent simultaneous pioctl calls from crashing the
190 * Add data validation to pioctl output processing in
191 ktc_GetToken and ktc_ListTokens.
193 * Add data validation to pioctl output processing in
194 fs.exe and symlink.exe.
197 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
198 No longer will browsing \\AFS produce truncated share names.
200 * Registry specified server preferences did not result in
201 the CM_SERVERFLAG_PREF_SET flag being applied to the
204 * Avoid unnecessary DNS lookups of share names as cells
205 when it is known that the name cannot be a cell name.
206 Any name that does not contain a dot is skipped.
208 * When processing Dfs Referral requests, do not return an
209 error if the server is down or busy; if the volume is
210 offline; the cell vldb cannot be reached; etc.
211 These paths are still \\AFS paths and so the client
212 should not be sent to search elsewhere for resolution.
214 * CRITICAL: Prevent the cm_Daemon thread from terminating when
215 the machine enters the suspend state. IF the cm_Daemon thread
216 dies, there is nothing to execute down server checks.
218 * Prevent use of smb_StartedLock before initialization which
219 can result in a panic in the lock package.
221 * Improve error checking in afslogon.dll in order to prevent
222 NULL pointer dereferences within GetLogonDomainOptions()
223 if the Lsa operations fail.
225 * Add support to the build system to automatically update
226 a Microsoft Symbol Store.
228 * Do not access the cm_conn_t after executing the cm_Analyze
229 loops. Doing so can result in an access to memory that has
232 * CRITICAL: If a StoreData request offset is not aligned on the
233 buffer module blockSize, (offset % blockSize) bytes of dirty
234 data will not be written to the file server.
236 * CRITICAL: If a StoreData64 request is sent to a file server
237 that does not support large files (>2GB) and the client is
238 unaware that the server has no large file support, the first
239 chunksize worth of data will not be written to the file server.
240 This bug was introduced in 1.5.3. It affects all IBM AFS file
241 servers and all OpenAFS file servers older than 1.4.0.
244 * If a file server becomes inaccessible while the cache manager has
245 dirty buffers to write, the afsd_service buf_IncrSync thread can
246 attempt to use 100% of the cpu.
248 * Fix the locking that protects the cm_cell_t lists
249 (all, name, id, free).
251 * Remove cm_cell_t objects from the name and id hash tables before
252 placing them on the free list.
254 * Fix "fs newcell" which was broken in 1.5.60.
256 * Do not attempt to synchronize dirty buffers if the associated
257 volume is known to be unavailable.
259 * Log to the Windows Event Log dirty buffers that cannot be written
260 during the shutdown process.
262 * Modify behavior of a Freelance mountpoint target that does not
263 specify a cell. Instead of assuming the target volume is in the
264 Freelance.Local cell, use the workstation "Cell" specified in the
265 registry. A mountpoint target of "#root.cell." will now mean the
266 root.cell volume in the workstation cell for the current session.
267 If the workstation cell changes from "athena.mit.edu" to
268 "andrew.cmu.edu", the referenced volume will also change without
269 requiring that the mount point targets be altered.
271 * Avoid false infinite loop errors when validating the cm_cell_t
272 free list during startup.
274 * Fix two more locations where xdr_free() should be used instead of
275 free() to permit a checked built of afsd_service.exe to work with
276 release builds of afsrpc.dll.
278 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
279 to permit the server Uuid to be used to lookup the server object
280 and from that determine the cell. This permits callbacks that are
281 received from alternate addresses to be processed with a known server
282 object. Previously a request from an unknown server would clear all
283 callbacks from all cells.
285 * Fix a bug that prevented optimal performance when using a non-zero
286 value for 'daemonCheckVolCBInterval'. As a reminder, when
287 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
288 volume callbacks are automatically renewed 90 minutes before their
291 * Fix automatic ranking of vldb servers whose values are obtained from
294 * Add logging for RX CALL DEAD errors which are generated if the remote
297 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
298 or the call is to a vldb server.
300 * Add registry based cell search functionality to NetIdMgr,
301 afs_config.exe, and klog.exe.
303 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
304 lookups on the host names in the CellServDB instead of using the
305 specified IP addresses. This provides aklog, pts, vos, etc. the same
306 CellServDB behavior that the Windows Cache Manager uses.
308 * Completely a behind the scenes change. The snprintf() routines have
309 been completely rewritten.
311 * When updating the stat cache entry callback of a .readonly object
312 from the volume group object, update the file server reference to
313 ensure it matches the most update to date callback.
315 * Add proper support for processing callbacks from multi-homed file
316 servers. Instead of comparing servers by cm_server_t pointer,
317 compare them by UUID when the UUID is known.
319 * During a shutdown short circuit the offline volume check daemon
322 * Properly track link counts during cm_Link and cm_Unlink operations.
324 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
325 preference to an error code reported by rx_EndCall.
327 * Add "PerFileAccessCheck" registry value to permit testing against
328 experimental file servers that include per-file acl support. This
329 value is intentionally undocumented. It is not to be used by
330 production environment deployments.
332 * Log a warning to trace log if a volume group is marked as a DFS
333 File Set. The Windows OpenAFS client does not include DFS support.
335 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
336 provider from obtaining tokens when referrals are in play.
338 * Add "fs chown" and "fs chgrp" commands to permit the owner and
339 group of objects stored in AFS to be set from Windows.
341 * Avoid performing background daemon operations when the machine is
342 going into suspend mode.
344 * Perform offline volume checks in most recently used order.
346 * Prevent crash when a data version for a cache object goes backwards.
348 * fs setquota, fs setcachesize, vos setfields, and vos create now
349 accept human readable orders of magnitude. (K, M, G)
351 * fs listquota fixed to permit large quota sizes to be displayed.
353 * rx packet resend and data packets sent counts were incorrect.
355 * Multi-thread safe library versions are now being generated and used.
356 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
358 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
359 values are now available on XP through Windows 7. Add functionality
360 to autodetect if such support is present on the machine. If so,
361 configure it if necessary and dynamically adjust the AFS Rx timeout
364 * Correct another race condition in the Rx library that could result
365 in an unexpected panic while freeing the Rx call iovq.
368 * A fix to the pioctl library to support drive substitution
369 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
371 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
372 that corrects a deadlock in the smb redirector and also adds new
373 functionality that permits the AFS SMB server to be given a longer
374 timeout than is normally the case. New functionality has been added
375 to configure these additional LanmanWorkstation\Parameter values.
376 (This functionality has been backported to XP SP3 and is scheduled
377 to be released on June 5th.)
379 * The BackConnectionHostNames registry value configuration was broken
380 when dynamic re-establishment of Netbios Name registrations was added.
383 * Hidden vos commands are revealed.
385 * If the "DisableLoopbackCheck" registry value is set, do not unset
386 it during the same service session.
388 * Reorganize code that prevents multiple Store operations to the same
391 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
392 permit broken symlinks to be treated as being in AFS.
394 * Fix vos commands that output 64-bit integer values.
396 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
397 Permits building cygwin applications against OpenAFS libraries.
399 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
400 as Windows HTML Help (.CHM) files.
402 * NSIS installer does a much better job of cleaning up files left
403 over from previous installs.
405 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
406 or "fs flushall" and on-going directory operations that can result in
407 afsd_service.exe crashing.
409 * Add support for DNS SRV records in place of AFSDB records.Â
410 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
411 Weight is currently ignored.
413 * Add a method of specifying Client CellServDB information within the
414 registry that can be used to either override the CellServDB file or
415 force the use of DNS lookups for a given cell.
417 The registry schema is as follows:
419 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
420 "LinkedCell" REG_SZ "[cellname]"
421 "Description" REG_SZ "[comment]"
422 "ForceDNS" DWORD {0,1}
424 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
425 "HostName" REG_SZ "[hostname]" Default: [servername]
426 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
427 "IPv6Address" REG_SZ "[address]" <future>
428 "Comment" REG_SZ "[comment]"
429 "Rank" DWORD "0..65535" Default: 0
430 "Clone" DWORD "{0,1}" <future: server only>
431 "vlserver" DWORD "7003" <future>
432 "ptserver" DWORD ... <future>
434 ForceDNS is implied non-zero if there are no [servername]
435 keys under the [cellname] key. Otherwise, ForceDNS is zero.
436 If [servername] keys are specified and none of them evaluate
437 to a valid server configuration, the return code is success.
438 This prevents failover to the CellServDB file or DNS.
440 Only one of "HostName" or "IPv4Address" is required. "HostName"
441 is optional if [servername] is the host name.
443 * Extend registry based CellServDB functionality to
444 afsconf_GetCellInfo() interface used by aklog and many
445 other command line utilities.
447 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
448 as it is now used by both client and server components.
451 * PriorityClass of afsd_service.exe process raised to
452 "High" to match the priority of the system services
453 that are waiting for afsd_service.exe to respond.
454 Configurable using [TransarcAFSDaemon\Parameters]
455 DWORD "PriorityClass" value. (use values from Win32
458 * Change error reporting of CM_ERROR_BADSHARENAME to
459 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
460 The error we were sending would cause the smb
461 redirector to drop the connection.
463 * Change the error reporting of CM_ERROR_ALLOFFLINE
464 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
465 Previous we would return Bad Network Path which
466 would cause the smb redirector to cancel all file
467 handles on the same "share".
469 * A debug version of afsd_service.exe can now be used
470 with a release version of afsrpc.dll (or vice versa).
471 afsrpc.dll now implements and exports xdr_free().
473 * Removed all unprotected OutputDebugString() calls as
474 those calls can block for up to 10 seconds if there
475 is contention for the globally shared memory or if
476 the debugger fails to read data in a timely manner.
478 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
479 supported for builds. The production releases still
483 * Discovered that XP/2003 supports a variant on recvmsg
484 called WSARecvMsg. The input parameters are mostly
485 compatible with the Posix recvmsg. This will save
486 multiple memcpy() calls per received packet.
488 * Discovered that Vista/2008 supports a variant on
489 sendmsg called WSASendMsg. The input parameters
490 are mostly compatible with the Posix sendmsg. This
491 will save multiple memcpy() calls per sent packet.
493 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
494 and utilizing rx_ReadProc32 from within the cache manager
495 as part of the FetchData operations. rx_ReadProc32 and
496 rx_WriteProc32 are optimized versions of rx_ReadProc and
497 rx_WriteProc for use when reading 32-bit integer values.
499 * The Network Identity Manager AFS provider now reports
500 progress events. These provide a better user experience
501 with Network Identity Manager 2.0.
503 * [RT 124293] A race condition exists which permits the
504 scp field of the an smb_fid_t object to become invalid
505 after a request on the smb_fid_t has begun. Remove the
506 race by protecting all access to the scp field with the
507 smb_fid_t mx mutex and obtaining a local reference on
508 the cm_scache_t object for the length of the request.
510 * [RT 124276] If the vldb is out of sync with the contents
511 of the file servers, afsd_service will retry too many
512 times when a file server reports a volume as not being
513 present. Now if the list reported by the vldb is the
514 same as the previously seen list, then the retry is
517 * [RT 124276] Read-only volume failover was broken in
518 1.5.53 whenever accessing a volume results in VNOVOL or
521 * [RT 124276] Prior to 1.3.70 the volume server reference
522 list was not reference counted and would be prematurely
523 freed while in use. When reference counting was added
524 in 1.3.70 a bug was introduced that could result in
525 service reference list corruption.
527 * Add Windows Application Event Log warning messages for
528 "Client SMB MPX value too large" and
529 "Client SMB Buffer Size too small".
531 * Renaming of files across directory boundaries would result
532 in an invalid handle error when attempting to access the
533 files after the move.
535 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
536 when a smb file descriptor cannot be found. The whole point
537 of using a Path Info function is because an smb file
538 descriptor wasn't allocated.
540 * More edge cases in which dynamic addition of Freelance
541 root.afs entries would get the wrong FID or where the
542 root.afs directory would not be refreshed.
544 * Buffer overflow could occur if the workstation cell name was
545 longer than 64 characters. Crashes could occur in
546 afscreds.exe, afslogon.dll, and afsd_service.exe.
548 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
551 * Improve consistency of the bad data version value used for
554 * Log file server uuid values as part of the cm_server object
555 when available. Dump the cm_server object list in response
558 * Optimize the performance of resetting access control lists
559 when tokens are set or removed.
561 * Fix vos.exe so that it can work with very large volume
564 * Remove symlink recursion tests and increase max symlink count
568 * Undo the "UAC manifest fix" applied to afs_config.exe.
570 * Fixed some additional race conditions in the Rx library
572 * Ensure that Freelance allocation of vnodes follow
573 the AFS convention of odd vnodes are directories
574 and everything else is an even vnode.
576 * Add Freelance logic to mount point and symlink
577 evaluation functions.
579 * Enhance smb_ParseASCIIBlock() so that it can handle
580 all of the STRING formats defined by the CIFS Technical
583 * Validate the output of smb_ParseASCIIBlock() in all
584 callers. Return CM_ERROR_BADSMB if the STRING field
585 cannot be parsed. CM_ERROR_BADSMB will cause the
586 contents of the packet to be logged.
588 * If multiple SMB Raw Write operations were taking
589 place at the same time, there could be data corruption
590 because unique event objects were not generated for
591 each Netbios receive operation.
595 * Fix a race in rxi_WriteProc which could result
596 in a freed or in use packet being passed to
597 rxi_PrepareSendPacket if the call is currently
598 in a transmit wait state. Passing a freed packet
599 will result in a panic. Passing an in use packet
600 might result in a call failure or perhaps data
603 * Kerberos referrals broke the NIM AFS provider
604 logic for automated configuration of the workstation
605 cell with a new identity. Now if a referrals
606 response is returned, assume the cell belongs to the
609 * The interface to buf_CleanAsync() and
610 buf_CleanAsyncLocked() has been changed to permit
611 errors to be propagated to the caller.
613 * During an enumeration of a directory, if the btree
614 has not yet been generated, return an error.
616 * When mapping CM_ERROR values to NTSTATUS values
617 do not treat 0 as an error.
619 * Avoid a cm_scache_t object reference count underflow
620 when renaming via the SMB interface.
622 * smb_IsMask() can be called with a non-8.3 input buffer.
623 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
624 input buffers and smb_IsMask() for non-8.3 input buffers.
626 * Add cm_buf_t error state to the "fs memdump" output.
628 * pioctl drive mapping checks need to take into account
629 Global Drive Mappings and SUBST assignments.
631 * Rename operations should not be permitted if the
632 destination name collides with an existing name that
633 differs only by case UNLESS the collision is the object
634 that is being renamed.
636 * Increase the maximum number of background daemon threads
637 to 64. Leave the default at 4.
639 * Set the FILE_UNICODE_ON_DISK flag when reporting the
640 file system characteristics.
642 * Report the maximum component name length as 255 even
643 for Unicode as part of the file system characteristics.
645 * Add RX debugging information to "fs memdump" output
648 * Refcount the cm_server_t objects returned by
651 * Fix the UAC manifest applied to afs_config.exe
653 * rxdebug now indicates if any RX_CONN flag is set
654 not just some of them
656 * Change the mapping of CM_ERROR_WOULDBLOCK from
657 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
661 * Includes #GCO Public CellServDB 07 Nov 2008
663 * Prevent separate cell entries from being created for
664 "foobar.org" and "foobar.org." which can happen if
665 an AFSDB record is found instead of a CellServDB
668 * Fix a panic when checking the SMB virtual circuit list.
669 If a VC is destroyed while the list is being checked
670 it is possible for the next object to become invalid.
672 * Increased logging to afsd_init.log of netbios name
675 * Fix a DNLC panic when a UTF-16 string that cannot be
676 converted to a utf-8 string is provided as a filename.
678 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
679 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
680 location to refer directly to the new location in
681 %programfiles%\openafs\client\program
683 * Be more aggressive about responding to Netbios
684 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
685 the LANA within the listener thread without performing a
688 * Use the WNet Enumeration of all connected drives as a filter
689 in the pioctl GetIoctlHandle operation. This should remove
690 the last situations in which there was a delay when opening
691 an Explorer Shell properties menu with the AFS Shell
694 * Fix the parsing of CellServDB in Windows to handle linked
695 cells as is done on Unix.
697 * Add support for linked cells to the cache manager,
698 netidmgr afs provider, and aklog. When two cells are
699 linked, a volume lookup that fails in one cell is looked
700 up in the other. This can be used for construction of
701 test cells that override specific volumes; cell renaming;
704 * Increase the length of cell and realm names that can
705 be entered in the NIM AFS provider.
707 * Actively prevent a 32-bit afsd_service.exe from
708 executing in the WOW64 environment.
710 * Protect against invalid IP address values in CellServDB
715 * Fix an Rx panic caused by an attempt to double free
716 an rx_packet when an error occurs while writing the
719 * Fix an exception cause by a race condition when
720 processing multiple file change notifications registered
721 by the Explorer Shell or other applications.
723 * Modify the Rx debug interface to use var_args and
724 permit Rx debugging to be used on 64-bit Windows
725 which previously had it disabled.
729 * Automated lock hierarchy enforcement added to the
730 osi base lock package. This permits detection of lock
731 violations that might but often does not result in
732 deadlocks. Enabled by default in the checked builds.
733 May be activated in release builds via the registry
734 value "LockOrderValidation" in the TransarcAFSDaemon
737 When a violation is detection an assertion is thrown
738 generating a dump file and a message indicating which
739 locks were obtained out of order is logged to the
740 Windows Application Event Log.
742 * Attempt to ensure that a dump file will be generated
743 whenever the service panics.
745 * Removed unused lock cm_bufGetMutex
747 * Many out of order lock acquisitions were detected
748 using the lock order verifier and were corrected.
750 * A race condition within the function path cm_GetSCache() ->
751 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
752 cm_scache_t object to be issued simultaneously to two
753 threads. This would eventually result in a panic due
754 to the resulting undercount.
756 * Fix interpretation of the empty string as the ioctl path
757 to mean the current directory. "fs lsm", "symlink list",
758 etc. now return a "not a ..." error instead of "does not
761 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
762 to indicate that the input string is a nul terminated string.
763 Assign ti when input strings are nul terminated instead of
764 length counted. (from 1.5.50)
766 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
767 does not match the observed behavor. The 'nameLength' is specified
768 in bytes, not in characters. Fix the implementation to match
769 the observed behavior. (1.5.50)
771 * The cm_HaveAccessRights() attempt at deadlock avoidance by
772 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
773 does not avoid the deadlock. Avoid the deadlock by enforcing the
774 lock order of the lowest vnode first. Then remove the infinite
775 loop avoidance in cm_SyncOp() that was returning an unwarranted
778 * File Server lock synchronization was not propertly enforced in
779 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
781 * Another change to the short name generator. Plus and equal are
782 not valid characters in a short name according to the
783 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
784 with left and right brace in order to minimize the impact.
786 * Fix a race condition that can result in a heap overwrite error if
787 the number of known file servers or vldb servers increases during
788 the preparation phase of a cm_CheckServers() operation.
790 * If the SMB Authentication would fail, a LSA allocated memory
791 buffer would be leaked.
793 * Rewrite the cm_cell_t allocation code to permit cell vldb server
794 lookups to fail without leaking cm_cell_t objects.
796 * Add "fs setcrypt auth" to permit the use of rxkad authentication
797 only mode. Not documented.
799 * Initialize Rx mutex and lock objects once and only once as part of
800 the DLL load process.
802 * When reporting the Volume Name "AFS" do not include the terminating
803 NUL in the reported length.
805 * Refresh Volume Location information upon receipt of a VNOVOL error
806 from the file server and avoid a race condition that prevented the
807 updated volume location information from being destroyed immediately
808 after it was acquired.
810 * In response to VNOVOL and VMOVED, remove the volume id from the server's
811 volume list so that the volume's up and down status is not affected
812 when the machine is shutdown.
814 * Many fixes to the Rx stack related to packet allocation and deallocation;
815 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
816 by overwriting the rx_call currentPacket pointer when it was already
817 referring to a packet.
819 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
820 in the TransarcAFSDeamon/Parameters key.
822 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
823 constructed directory B+ trees for flushed directory objects.
825 * When cached pages for directory objects have been locally modified, do
826 not permit the locally modified pages to be mixed with pages subsequently
827 obtained from the file server. Doing so can result in directory
830 * If Ubix errors are received in response to VL_xxx operations record
831 the error on the server instance and permit the operation request to
832 be retried in case another server is able to answer the query. We have
833 seen at least one instance of a cell in which a VL server returned
834 UNOQUORUM (possible database truncation) when the other servers were
835 responding successfully.
838 * Optimize the assignment of write errors to dirty
839 buffers by applying the first error to all buffers
840 belonging to the same object instead of trying each
843 * Add resource version info to xstat_fs_test and xstat_cm_test
845 * Fix a delete vs delete[] usage problem in the WiX custom
848 * When creating a mount point, validate the volume name to
851 * When processing a mount point, pay attention to VL_BADNAME
854 * Remove a race condition from cm_volume_t recycling that could
855 have resulted in two threads making use of the same object at
858 * Fix a reference count leak on smb_vc_t objects introduced in
859 1.5.50. Permit the objects to be freed once again.
861 * Fix the output of "fs memdump" for smb_vc_t objects that was
864 * Add routines for debugging smb refcount errors.
866 * When shutting down the cache manager, do not permit AFS
867 callback requests to be processed. Doing so can result in
868 invalid memory accesses.
870 * Do not recycle a deleted cm_scache_t object unless the
871 refCount is 0. Doing so can result in a deadlock and
872 invalid memory access.
874 * Minimize the number of RPCs sent to the file server for
875 directory bulk status requests. Only request status for
876 objects that we are actually reporting on and only if we
877 do not already have status for the object and then make
878 sure that we return as much as we can at once via the SMB
879 interface but not so much that it exceeds the RDR timeout
881 * Enable bos restricted operations
883 * Fix the creation of submounts used by the afscreds and afs_config
886 * Avoid repeated retries when attempting to obtain access rights.
887 If the access rights were obtained successfully but do not
888 satisfy the request, fail the request.
890 * Fix error code reported when attempting delete a file on a
891 readonly volume or one that is marked with the readonly DOS
894 * Fix a short name truncation bug.
896 * Fix a heap corruption error when reading the CellServDB file
899 * Fix directory listings when the directory has "lookup" permission
900 only and the directory has never been previously accessed.
902 * Fix several tracking issues for cm_scache_t rw locks.
904 * Ensure that all buffers used to store cell names are the
907 * Restore Windows 2000 support (broken in 1.5.50).
908 LOCALE_INVARIANT was introduced in XP.
910 * Add RxUdpBufSize registry value. Default is 256K.
912 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
914 * Do not include trailing NULs when returning file names in
915 a directory search response.
917 * Prevent RX from crashing the afsd_service by attempting
918 to allocate to many buffers at once when memory is fragmented.
920 * Avoid a deadlock when freeing cm_server_t objects
922 * Optimize search cell file. Do not search to end of file
923 if the cell has already been found.
925 * Remove explorer shell extension delay on local disks.
927 * Automatically pre-allocate 64 Rx packets per thread
929 * Add smb lock requests and stat cache lock allocations to
930 the output from fs memdump.
933 * Fix Large File Support (broken in 1.5.50)
935 * Fix "fs setacl" (broken in 1.5.50)
937 * Fix pioctl path evaluation of drive letter
938 mapped paths. (broken in 1.5.50)
940 * Fix installation of VC8 Runtime library in the EXE
941 Retail installer. (broken in 1.5.50)
943 * Avoid a potential deadlock during Volume Location
944 updates if all the file servers are unreachable at
945 the time the VL update succeeds.
947 * Ensure that rx connection object is references
950 * Fix propagation of over quota errors during cm_FSync()
953 Since 1.5.39 [1.5.50 released 16 July 2008]
954 * There were no releases between 1.5.39 and 1.5.50.
956 * FIRST RELEASE WITH UNICODE SUPPORT
958 * Pioctl interfaces to the cache manager refactored
959 to provide layering between the SMB specific code
960 and the general purpose ioctl operation.
962 * Garbage collect dead SMB virtual circuits as soon
963 as they are no longer being referenced. This avoids
964 problems with outstanding locks not being dropped
965 when the virtual circuit becomes invalid.
967 * Remove the IBM Administration Reference documentation
968 and replace it with the OpenAFS Command Reference
971 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
972 functions each time a connection is about to be used.
973 Do not hold a lock on the rx connection object while
974 it is being selected. This avoids a race between
975 threads attempting to set the timeout values and
976 removes a bottleneck that was hampering performance.
978 * Ensure that the smb directory attribute is set
979 for all directory objects.
981 * Add vs2008 support to the NSIS installer scripts
983 * Replace the VC Runtime EXE installer with the MSI
984 installer in the NSIS installer scripts
986 * Properly delete the VC7.1 C Runtime libraries.
988 Since 1.5.36 [1.5.39 released 23 June 2008]
989 * There were no 1.5.37 or 1.5.38 releases for Windows
991 * Use Visual Studio version of vsprintf() instead of
992 home grown version that resulted in linking conflicts
994 * Prevent buserver, ptserver and fileserver from crashing
995 due to non-portable strftime() format strings.
997 * Prevent server crashes caused by rx_InitMorePackets()
998 being called prior to allocation of all of the required
999 synchronization objects.
1001 * Fixed a memory leak in the internal B+ tree directory.
1002 For objects with non-8.3 names, the 'longname' memory
1003 allocation would be leaked during object deletion or
1006 * Fixed a memory leak in the background daemon cm_CheckServer()
1007 routine. The memory allocated to store the server list
1010 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
1012 * Correct an abstraction layer violation. cm_scache_t objects
1013 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
1014 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
1015 performed in cm_ReleaseSCache() when the reference count hits zero.
1017 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
1019 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
1022 * Ensure that scache object 'nextp' pointers are set to NULL
1023 when the objects are removed from the hash table linked list.
1025 * If a dirty buffer cannot be written to the file server because
1026 of an access denied error, mark the error in the cm_buf_t
1027 structure so that it can be dealt with instead of retrying
1031 * Update CellServDB file to GCO Public 23 Apr 2008
1033 * Fix a cm_buf_t reference count leak when attempts to write
1034 dirty buffers to the file server from within cm_IncrSyncer()
1037 * Prevent udebug from crashing.
1039 * Activate RX Idle Timeouts. If the file server is busy for more
1040 then 30 seconds, attempt to failover to another server without
1041 marking the busy server down.
1043 * Another VNOVNODE issue fixed. When writing a dirty buffer
1044 to the file server, if VNOVNODE is received, mark all buffers
1045 as invalid without further attempts to contact the file server.
1047 * Improved performance on high latency links.
1050 * Fix a bug preventing the re-initialization of the Freelance
1051 root.cell contents when a change is made.
1053 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
1054 or mountpoint that does not exist.
1056 * aklog, afscreds, integrated logon, and the NIM AFS credential
1057 provider will all attempt to first acquire an afs service ticket
1058 using the service principal: afs/<cell>@<CLIENT_REALM>. This
1059 permits the proper detection of the cell realm when multiple
1060 realms are local realms for the cell.
1062 * Fix fs commands that re-use file identifiers instead of forcing
1063 a path evaluation each time. This was broken in 1.5.33 when
1064 hashes were added to the cm_fid_t structure.
1066 * Fix symlink and mountpoint removal and then recreation when
1067 applied to the fake Freelance root.afs volume.
1069 * Fix SMB change notifications for the Freelance root.afs volume
1070 when creating or removing symlinks and mount points.
1072 * Force DNS AFSDB cell name lookups to lower case if the DNS
1073 response is not all lower case.
1075 * 32-bit installers are now built with Visual Studio 2005
1077 * All vos.exe commands now support the "-noresolve" option which
1078 forces the display of IP addresses instead of DNS hostnames.
1080 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
1081 permits operations that do not require up to date volume
1082 location data to continue while waiting for the RPC to complete.
1083 Multiple threads waiting for a volume location update on the
1084 same volume set no longer trigger multiple calls VL_xxx RPCs.
1086 * Volume server references could be marked offline and never
1087 reset to the not_busy state.
1089 * Further optimizations to cm_Analyze() retry logic for
1090 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
1093 * Optimize the DNLC by applying Interlocked incrementing for
1094 statistics gathering and enabling greater use of read
1095 locks instead of write locks
1097 * Further optimize osi_Log operations when logging is disabled
1098 by removing the need for the function call overhead.
1100 * Add a new lock operation, Convert Read to Write, and apply
1101 it throughout the cache manager. This function provides a
1102 fast transition from a read lock to a write lock when the
1103 caller is the only reader.
1105 * Fix a timeout problem when opening files or request locks
1106 when the file in question is already write-locked by another
1109 * Fix a deadlock that can occur if "fs flushall" is executed
1110 while applications are writing to AFS.
1112 * The cache manager would re-read data from the file server
1113 after file truncation when obtaining buffers to write to the
1114 same file until the first StoreData RPC completes.
1116 * Fixes error handling in cm_NTCheckDelete()
1118 * Replaces cm_scache_t mutex with a read-write lock permitting
1119 additional parallel access.
1121 * Prevent an cm_scache_t reference undercount when evaluating
1122 symlinks containing @sys when none of the active sysnames
1123 match any of the available targets.
1125 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1128 * Set RxMaxMTU registry value to 0 which is the equivalent of
1129 it not being set at all.
1131 * Reduce contention between BkgDaemon threads
1133 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1136 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1137 objects to use the faster Interlocked operations which avoid
1138 the need to hold write locks.
1140 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1141 non-fatal error that forces a retry against another service
1144 * Fix a crash in smb_WriteData that would occur if an application
1145 opened the file as read-only and then sent a write request anyway.
1148 * Respond to VL_NOENT errors by removing cm_volume_t object and
1149 setting it to be the first object to be recycled. This is not
1150 only a more efficient use of resources but it also prevents
1151 repeated attempts to query the VLDB server for the volume from
1152 the CheckOfflineVolumes routine.
1154 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1157 * A first cut at a cache manager statistics monitor that can be used
1158 to determine the necessary cache parameters to support the
1159 application working set.
1161 Off by default, the performance package can be activated by setting
1162 "daemonPerformanceTuningInterval" in the service Parameters key.
1163 As with the other daemon interval values the unit is in seconds.
1165 At service start and each succeeding interval the cache manager
1166 will write statistics to %TEMP%\afsd_performance.txt showing the
1167 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1168 The FID statistics keep track of all FIDs seen by the cache manager
1169 during the service session whether or not they are backed by any
1170 live object in the cache.
1172 These statistics are not stored in the cache file.
1174 * Fix windows event logging broken in 1.5.33.
1176 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1177 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1178 as well as when computing the server preference list.
1180 * When merging status after a StoreData operation prevent data
1181 buffers from being discarded that we would prefer to keep.
1182 (Bug introduced in 1.5.33.)
1184 * Modify installers the set RxMaxMTU to 0.
1186 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1187 dialog, the Background Daemons and Service Threads fields were
1190 * Increase default number of Background Daemons to 4.
1192 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1193 cm_volume_t objects to be recycled.
1195 * Windows Error Reporting has shown a number of previously unseen
1196 stack corruption errors in the Explorer Shell extension. The
1197 minidump appeared to indicate the problem was in ParseAcl().
1198 Reviewing the function it appears that numerous things could go
1199 wrong if invalid input was provided. Added error checking and
1200 data validation. Hopefully this will address the problem.
1202 * Windows Error Reporting has shown crashes in aklog as a result
1203 of krb5_cc_get_default() failures not being handled. They are
1207 * The Rx library used a 32-bit type for sockets which was
1208 truncating the socket value on 64-bit Windows. This
1209 was not an issue in general but became an issue when
1210 Rx "hot thread" support was enabled.
1212 * Enable "Rx hot thread" support which ensures that while
1213 a received message is being processed that a new thread
1214 is allocated to receive the next message to arrive.
1216 * Increased parallel access to smb_rctLock protected
1217 resources. Reduce the number of times that the
1218 lock is dropped and re-obtained.
1220 * Increased parallel access to cm_scache_t and cm_buf_t
1221 resources by replacing the previous global write lock
1222 protected reference counting mechanism with one built
1223 upon the use of InterlockedIncrement and InterlockedDecrement.
1225 * Implement a hash for cm_fid_t to reduce comparison costs
1226 and inline the cm_FidCmp() function.
1228 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1229 hash which has a better distribution
1231 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1232 the most recent read request instead of the next chunkSize.
1233 cm_chunkSize can be dozens or hundreds of buffers. As a result
1234 too much time is spent performing the evaluation.
1236 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1237 lock is to prevent the creation of new buffers while a truncation
1238 is being performed. All references to bufCreateLock have been
1239 removed except in two places:
1240 i. a write-lock surrounding the function that calls buf_Truncate()
1241 ii. a read-lock within buf_GetNewLocked() that actually allocates
1244 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1245 CM_CONFIG_DATA_VERSION number which value be used to force the
1246 replacement of the cache file contents when incompatible changes
1247 are made between releases.
1249 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1250 cm_SyncOpCheckContinue
1252 * Avoid calls to multi_Rx if nconns == 0
1254 * Modify smb_WriteData to perform background writes based upon the
1255 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1256 boundaries. This will slow down writes from the SMB interface but
1257 will avoid the risk of the CIFS client disconnecting from the AFS
1258 client SMB server. The default value for 'smb_AsyncStoreSize'
1261 * SMB AsyncStore functionality is on by default but can be
1262 disabled using the registry value:
1264 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1266 smb_AsyncStoreSize can be adjusted with
1268 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1270 The value must be a multiple of the buffer block size
1271 and cannot be larger than the chunk size.
1273 * Do not leak a read lock on the bufCreateLock
1275 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1278 * In MergeStatus, only remove clean buffers from the hash table.
1280 * Avoid the need to update the data version number on each buffer
1281 associated with a scache when MergeStatus is called after a
1282 StoreData by maintaining a range of valid data versions as part
1283 of the cm_scache_t object.
1285 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1286 from cm_HaveAccess... We can safely avoid the call so do so.
1288 * Add "ForceAnonVLDB" service parameters registry value. Default
1289 is 0. Set to 1 to force the use of anonymous connections to
1290 the VLDB servers. This is useful when using krb5 tokens and the
1291 VLDB servers have not been updated.
1294 * Do not attempt to search for a cell name that matches one
1295 of the special share names "svrsvc", "wkssvc", "ipc$"
1297 * Issue permission denied warnings from afscreds if attempts
1298 to modify the startup shortcut or registry values fail.
1299 This happens under Vista UAC.
1301 * Do not return BAD_NETWORK_ERROR in response to attempts
1302 to create directories or files in the Freelance root.afs
1303 volume. Instead return ACCESS_DENIED.
1305 BAD_NETWORK_ERROR was being sent because the Freelance cell
1306 has no vldb servers. Short circuit the error by testing for
1307 freelance in CreateFile and MakeDir
1309 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1310 to be opened and closed without actually existing. For now
1311 we treat them as equivalent to ioctls but this should
1314 * Fix RAP processing to not reject names that exist in the
1315 root.afs volume for GetInfo queries.
1317 * In afscreds, use GetUserNameEx to obtain the username instead
1318 of searching the registry. Try UserPrincipalName first and
1319 fallback to SamCompatible.
1321 * Stop linking to rpcns4.dll which is no longer supported on
1322 Vista and 2008. TaAfsAdmSvr is going to have to be
1325 * When processing dfs referral requests, if the sharename is
1326 a partial match and therefore will not be added automatically
1327 to the share list, make sure that we return an error.
1329 * Avoid a crash when logging VNOVOL errors returned from
1330 cm_CheckOfflineVolume.
1332 * When trace logging is disabled, do not permit osi_LogSaveString
1333 to perform the string copying. By disabling the string
1334 copies, throughput of the service more than doubles.
1336 * when checking offline volumes, update the cell vlserver info
1337 if required and don't perform a check if there are no servers
1340 * Add volumes in the vl_unknown state to the list of objects to be
1341 checked as offline or retried when an allbusy error is received
1344 * Add support for configuring rx_ExtraPackets
1346 * All calls to buf_Get() must be protected by a read lock on the
1347 cm_scache_t bufCreateLock
1349 * When a MergeStatus results in the invalidation of buffers, do
1350 not remove buffers from the hash tables that have active references.
1352 * Upon receiving a VNOVNODE error from a file server, be sure to
1353 invalidate any file locks. This prevents on-going attempts to
1354 extend the locks and will cause subsequent attempts to access
1355 the file to receive a invalid handle error.
1357 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1358 to directory entry names.
1360 * No longer discard tokens when RXKADBADTICKET error is received.
1361 Instead return STATUS_NO_KERB_KEY error.
1364 * Add "cmdebug <host> -cellservdb" option and client service
1365 implementation permits querying the vldb server for each
1366 cell known to the cache manager. The output is in a form
1367 that permits it to be used as a CellServDB file.
1369 * Add Symlinks->Show option to the AFS Explorer Shell Context
1372 * Prevent buffer overrun in the cmd parser used by command
1373 line tools such as fs, vos, pts, etc.
1375 * The path ioctl operations have several issues:
1377 (1) the specified path for "fs examine, whereis, whichcell,
1378 flush" and so always has the follow symlinks and mount
1379 points semantics. This makes it impossible to determine
1380 what the FID of a symlink or mount point is.
1382 (2) "fs examine" out is not the result of a single pioctl
1383 operation but is actually the combined output of half
1384 a dozen operations. Path evaluation is an expensive
1385 operation. It would be faster if the caller could
1386 evaluate the FID first and then perform all of the
1387 rest of the operations by specifying the FID instead
1390 (3) fs output reports all objects as files. By adding a
1391 GetFileType pioctl more informative output can be provided
1392 that indicates what type of object the path evaluates to.
1394 (4) the Windows fs command includes a number of commands that
1395 do nothing but exist only because the Unix cache manager
1398 A new extendible data structure cm_ioctl_query_opts_t has been
1399 added which can be optionally specified with pioctls that do not
1400 already require input data. The first two fields of this structure
1401 are 'literal' and 'fid'. The literal field is used to indicate
1402 whether the last component of the path should be evaluated following
1403 symlinks and mount points.
1405 The fid field permits a fid to be specified.
1407 * A new GetFileType pioctl has been added. The type of objects are
1410 * A new "-literal" option is available for "fs examine, flush, whereis,
1413 * Unimplemented fs commands have been removed.
1415 * There are circumstances where a volume object is being accessed and
1416 the volume is marked indicating that the volume location information
1417 is out of date but where it is also pointless and perhaps dangerous
1418 to block waiting for the rpc to complete. One example is when
1419 processing the cmdebug requests. If we know that we are not going
1420 to use the volume object to contact a server, then we can now set
1421 the CM_GETVOL_FLAG_NO_RESET flag.
1423 * when moving up in the directory tree we search the recorded fid list
1424 to find a matching fid that we have already crossed. we must also
1425 reset the fid count based upon what we discover.
1427 * Cell names are published as share names. As a result they are searched
1428 for as part of the Dfs Referral evaluation. If share "foo" can not be
1429 reached, the CIFS client will fallback to searching for "fo" as well.
1431 Since the freelance client automatically adds a symlink for prefixes
1432 of the cell name, this results in both "foo" and "fo" being added to
1433 the freelance root.afs volume. This patch prevents that negative side
1436 * when tracking Kerberos credential cache names be sure to include
1437 the cache type prefix.
1439 * Microsoft has assigned OpenAFS a network type value
1441 * The algorithms that were used to produce a Netbios name were broken.
1442 The name that was produced when the hostname was too long was in fact
1443 longer than the maximum permitted netbios name.
1445 Also, the "NetbiosName" value was not used as a suffix when the loopback
1446 adapter was not installed. The hardcoded string "AFS" was used instead.
1448 * avoid another deadlock during server probes initiated by the ipaddr
1449 change daemon thread.
1451 * prevent the afs client service from crashing when fs newcell is
1454 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1455 to \Program Files\OpenAFS\Client\Program in order to get them out
1456 of the %windir% tree
1458 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1459 end STATUS_OUT_OF_QUOTA can be returned to the application.
1461 * Modify the search order for determining the location of CellServDB
1462 and other client configuration files.
1464 1. AFSCONF environment variable
1466 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1467 4. use Program Files\OpenAFS\Client
1469 * More Vista Power Management tweaks
1471 * Close a socket handle leak in the kauth routines
1473 * Don't hold a cm_volume_t mutex across RPCs
1475 * Convert downTimes to use clock time instead of relative times.
1476 This permits correct comparisons with expiration times.
1478 * Update the Wix installer to properly install the prerequisites
1479 for translate_et.exe
1481 * Do not translate pioctl paths as they are not converted by the
1484 * Implement multi_rx probes for checking whether servers are up
1485 or down. Turns probes into a constant time activity instead
1486 of one dependent upon the number of known servers.
1488 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1489 update of the volume location data and retry. VNOVOL means
1490 that the volume could not be attached. It doesn't mean that
1491 the volume is not on this server.
1493 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1494 not attempt to validate the target path. There is no reliable
1497 * Improve performance of \\afs\<cell><mount-type><volume>\
1498 access by eliminating extraneous attempts to search the
1499 CellServDB file and DNS AFSDB records. Share names
1500 containing '%' or '#' are not valid cell names.
1502 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1503 to unlock a range that is not currently locked.
1505 * Enable the 32-bit tools to read the 64-bit registry keys
1506 when executing in the WOW64 environment.
1510 * Fix buffer refcount leak introduced in 1.5.29
1512 * Prevent the NIM AFS Provider from crashing when configured to use
1513 Kerberos v4 for token acquisition in situations where no Kerberos v4
1514 support is available. For example, 64-bit KFW or sites that delete
1515 the Kerberos v4 DLL.
1517 * In afsd_init.log, support logging cache sizes greater than 2GB.
1519 * The volume status output of "fs examine" was being determined based
1520 upon the wrong error value. Instead of using the pioctl() return
1521 code, the errno value must be used.
1523 * Prevent invalid pts auto-registration attempts by aklog and the
1524 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1526 * Add synchronization protection to all of the SMB Listener State
1527 variables. This prevents race conditions when the addition or
1528 loss of a network adapter takes place.
1530 * On Vista, there is a race condition between the restart of the
1531 afs client service and the network adapters. Ensure that the
1532 afs client service can handle switching between loopback and
1533 non-loopback modes of operation.
1535 * Fix FollowBackupPath to work with mount points that are not
1536 explicitly read/write.
1538 * Fix a deadlock that has been observed on Vista while resuming
1539 after at least four hours of sleep.
1541 * Add new HKLM registry configuration options that can be used
1542 to provide configuration hints to the NIM AFS Provider when
1543 creating new identities. See release notes for details on
1544 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1547 * Speed up write buffer operations by holding a global read-lock
1548 instead of a write-lock when updating the buffer data version
1549 number. The data version is protect by the associated stat
1552 * Speed up read buffer operations by removing out of date buffers
1553 from the buffer hash tables.
1555 * Protect against a null pointer dereference in afslogon.dll
1556 if QueryAdHomePathFromSid() fails.
1558 * Remove a race condition while accessing private freelance data
1561 * Update the buffer data versions when Freelance directory data
1564 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1567 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1568 instead of returning zero length.
1570 * Return ERROR_RETRY when a server sends VBUSY.
1572 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1573 a volstat plug-in is in use.
1575 * Enhance the volstat interface to support Dfs path mappings.
1577 * Fix "symlink list" to do the right thing for Dfs links
1579 * Fix callback expiration that was broken in 1.5.27.
1581 * Fix a crash at shutdown caused by the lack of synchronization
1582 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1584 * Hold the cm_server_t mutex while walking or modifying the
1585 cm_server_vols_t list in order to avoid a race condition.
1587 * Ensure that the cm_server_t refCount is safely modified.
1589 * The pthread library implementations of pthread_cond_wait
1590 and pthread_cond_timedwait were failing to return with
1591 the mutex held when error conditions occurred. This could
1592 result in race conditions wherever they are used. One case
1593 in which problems were seen was the Rx event_handler().
1595 * Prevent cm_GetBuffer() from locking all buffers associated with
1596 a stat cache object before deciding if the current buffer is up
1599 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1600 object for each offset that we will attempt to reserve.
1602 * Reduce the default number of background daemon threads used to
1603 prefetch or store to one in order to prevent two or more threads
1604 from stepping on each other.
1606 * Reduce the default chunk size to 256KB in order to prevent read
1607 timeouts on low bandwidth connections.
1609 * Fix a race condition that could result in threads becoming stranded
1610 when synchronizing multiple requests on the same stat cache object.
1613 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1614 has been disabled. It has been disabled because it turns out
1615 that the RPC is improperly implemented in all file servers
1616 previous to 1.4.6. The RPC is executed without the proper
1617 locks being held resulting in data corruption and eventual file
1618 server failure. The functionality will be re-enabled in a
1621 * a couple of .readonly Volume Callback optimizations were
1623 - Apply most recent volume callback to all stat cache objects
1624 in the volume. This avoids premature callback expirations
1625 and unnecessary FetchStatus calls.
1626 - Add option to permit attempts to renew .readonly Volume
1627 Callbacks on a periodic basis to avoid invalidation of stat
1628 cache data. [Requires daemonCheckVolCBInterval be set to
1629 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1631 * Adds support for 64-bit data version values. Previous releases
1632 only supported 32-bit values. Larger values sent by the file server
1633 would result in rollover.
1635 * aklog and asetkey no longer generate an exception if Kerberos for
1636 Windows is not installed.
1638 * If the LSA authentication functions report a temporary out of memory
1639 condition, fail the current authentication and do not cause the AFS
1640 client service to terminate unexpectedly.
1642 * When the AFS client service is shutting down, prevent the daemon check
1643 thread from executing additional checks which might attempt to access
1644 discarded locks or unmapped memory.
1646 * Correct a double free condition in the NIM AFS credentials provider.
1648 * Add "FollowBackupPath" service parameter registry value.
1649 This provides equivalent functionality to the UNIX afsd -backuptree
1650 option. When set to a non-zero value, normal mount points originating
1651 in a .backup volume will prefer a .backup volume.
1654 * Fix NSIS uninstall from the .MSI installer.
1656 * Add Stack Frame logging for AMD64
1658 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1659 server returns EWOULDBLOCK, retry the request every two
1660 seconds for up to the RDRtimeout.
1662 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1663 is not assigned the same value as EIO.
1665 * Update CellServDB to GCO Public 25 Oct 2007
1667 * Add additional validation and error handling code after
1668 each call to getSlot(). If an invalid slot is returned,
1669 return NONODE. If the invalid slot is returned when
1670 extracting a data node, invalidate the tree.
1672 * Modify compareKeys() to always perform a case-insensitive
1673 comparison and only perform a case sensistive comparison
1674 if the case-insensitive one matches. This ensures the
1675 ordering is consistently reported.
1677 * Add lock assertions to ensure that all calls are being
1678 performed with the correct locks being held. There have been
1679 some crash reports that provide stack data that does not appear
1680 to be possible unless there is a race. However, there are no
1681 obvious locations where the race is taking place and the test
1682 suite indicates that all of the correct locks are being held.
1683 We shall see what happens in the field.
1685 * For consistency replace all calls to findKey in which the range i
1686 s (1,numentries) with calls to getSlot().
1688 * Optimize the depth search loop by testing the slot value in the
1689 for statement instead of forcing the loop to be broken later.
1691 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1692 There are a number of locations where locks are obtained, dropped,
1693 and reobtained. This reorganization attempts to accomplish several
1696 (1) be optimistic for the most common case so it will be fast
1698 (2) add consistency checks after each location where locks are
1699 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1700 and the bplus tree is out of date, retry until we get to a
1701 consistent state that we can use.
1703 (3) Ensure that all operations take place with the correct locks.
1705 * One of the issues that has become a serious problem since the
1706 addition of the local directory updates is that although cm_SyncOp
1707 synchronizes operations, it does not preserve the order of requests.
1708 This has always been a problem in that it has been possible for a
1709 request to fail to complete due to its worker thread's bad luck.
1710 When a request takes longer than the Windows SMB Redirector's timeout,
1711 the SMB Redirector tears down the SMB virtual circuit.
1713 When using the local directory updates it is really important that
1714 the directory update operations complete in the order that they were
1715 sent to the file server. If they don't, then the local directory state
1716 and the file server state will not match and the local directory state
1717 must be discarded which in turn forces a new read of the entire
1718 directory contents over the network.
1720 This patch adds a new cm_scache_waiter_t object that is used to store
1721 the current thread, buffer, and syncop flags within a waiters queue
1722 on each cm_scache_t object. If a thread is forced to sleep in
1723 cm_SyncOp, upon waking it will check to see if there are any other
1724 threads waiting that are attempting to perform a similar task ahead
1725 of it in the queue. If yes, the thread goes back to sleep. If not,
1726 it goes ahead and enters the cm_SyncOp conflict resolution block.
1728 This patch has the additional side effect of reducing the number of
1729 competing threads that must obtain the cm_scache_t mutex and process
1730 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1731 utilization of the service and the clock time associated with processing
1732 requests will be reduced.
1734 * assert that the cm_scache_t mutex is held when calling
1735 buf_ForceDataVersion().
1737 * rename findNode to leafNode in bplus_Lookup
1739 * replace all OutputDebugString calls with osi_LogX calls
1741 * modify bestMatch to special case the return values for leaf nodes.
1742 If an entry is above or below the values available in the leaf node
1743 return BTLOWER or BTUPPER instead of BTERROR.
1745 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1746 if true convert to either slot 0 or Max and perform the insertion.
1747 This produces easier to read code when performing lookups.
1749 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1750 descriptive message.
1752 * the difference between osi_assert() and osi_assertx() is whether or not
1753 a description is specified as part of the call. When no message is
1754 specified afsd_notifier() is called with a NULL msgp parameter. This
1755 results in a null pointer dereference during the EventLog() call.
1757 As a result, none of the information describing the assertion is written
1760 This commit sets a default message for use when no other message is
1761 provided and it replaces all calls to osi_assert with osi_assertx and
1764 * Migrate B+ tree search key into thread local storage
1767 * Fix an uninitialized variable in lana_IsLoopback() which
1768 could result in a random determination that a loopback
1769 adapter is not a loopback adapter. This would result in
1770 a computed Netbios service name of %MACHINE%-AFS instead
1771 of AFS. This is a serious problem if the value computed
1772 by the AFS Client Service and the client process pioctl
1773 calls produce different results as it will result in a
1774 communication failure.
1776 * Fix a crash in the B+ Tree code that results when a file
1777 rename only results in a change of case.
1779 * Fix the pioctl path parsing. The previous algorithm
1780 broke the current directory of the client into two parts,
1781 the TIDpath which represented the device root and the
1782 relative path. The problem is that this is not sufficient
1783 from proper processing of symlinks and mount points.
1785 What is required is that the last component of the path
1786 must not have the symlink or mount point be evaluated
1787 whereas symlink and mount point evaluation must be performed
1788 for the rest of the relative path. Therefore, the path is
1789 now split into three parts: TIDpath, intermediate path,
1792 "fs examine" now reports the data of the symlink or mount
1793 point object instead of the target and "symlink list" and
1794 "symlink remove" now works. The "Explorer Shell Extension"
1795 menu now is displayed for symlink objects.
1797 * More informative symlink.exe error messages.
1799 * When the cmd.exe "del" command is used to remove a symlink,
1800 not only was the link object removed, but the stat cache
1801 object for the target was marked deleted. This resulted
1802 in cache corruption which would require the destruction
1803 of the cache to correct.
1805 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1806 which results in cached contents from read only volumes
1807 to be treated as valid when there is no callback registered
1808 and all servers are down or all instances are offline.
1810 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1812 * Add documented registry value, "AcceptDottedPrincipalNames"
1813 which is used to permit the acquisition of AFS tokens
1814 using Kerberos v5 principal names which include a dot in
1815 the first component.
1817 {HKLM\HKCU}\Software\OpenAFS\Client
1819 * "fs uuid" does not require "AFS Client Administrator" group
1820 membership to display the current UUID value.
1823 * Update NSIS installer to version 2.30
1825 * Update Wix installer to version 2.0.5325
1827 * 64-bit KFW support:
1828 - Network Identity Manager plug-in
1830 - 64-bit krb5 enable afscreds.exe, afs server manager,
1833 * Enable B+ tree Directory Search by default.
1835 * NIM plug-in is now compatible with FILE: ccaches
1837 * Install instloop.exe into Client/Programs/ even though
1838 it isn't required. Permit it to be used by organizations
1839 that clone Windows systems.
1841 * On a suspend, only drop callbacks for servers that
1842 are online. If the server is offline, maintain the callback
1843 until the server is once again available.
1845 * Prevent against acquisition of a directory write lock when
1846 there are no updates to be performed. This was resulting
1847 in a write-lock leak which later on would lead to a deadlock.
1849 * Add configurable file pre-fetch mode. For files that match
1850 the specified file extensions, pre-fetch the entire file into
1853 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1854 (MULTI_SZ) PrefetchExecutableExtensions
1856 * Fix the client_osi large integer support so that return values
1857 can be properly assigned to large integer variables.
1859 * Do not hold cm_scache_t mutex across cm_GetCell() calls
1860 since cm_GetCell() can block in a RPC.
1862 * Add short file name support to B+ tree searches.
1864 * Resolve a conflict between the desire to notify file servers
1865 that the client workstation is suspending and the desire to
1866 permit objects with valid callbacks when the client suspends
1867 to continue to use those objects from the cache if the client
1868 resumes without network connectivity.
1870 * Fix stack variables to ensure that all buffers storing paths
1871 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
1872 would cause a stack overflow condition.
1874 * On Vista, force afs_config.exe to run with the highest privileges
1875 available to the end user. For a user in the Administrators
1876 group running under UAC, this will cause afs_config.exe to be
1877 be run with full Administrators privileges after the user
1878 approves the privilege escalation.
1880 * More VNOVNODE handling issues. If a dirty buffer cannot be
1881 written due to VNOVNODE, clear the buffer and ignore the error.
1882 If a cm_scache_t object has been marked deleted due to a VNOVNODE
1883 error, the next time the smb_fid_t object that references it is
1884 used return either NOSUCHFILE or NOSUCHPATH and close the
1885 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
1886 result in an invalid handle error.
1888 * For Vista, set the RequestExecutionLevel of the NSIS installer
1891 * For the build environment, support the generation of manifests
1892 when using the CL1310 (VS .NET 2003) compiler/linker.
1894 * There was a bug in OutputDebugHexString() which could overflow
1895 a stack buffer resulting in unexpected process termination
1896 of afsd_service.exe if the input Security Token contained
1897 many bytes equivalent to the '%' char.
1899 * Add directory lookup statistics to "fs memdump" output.
1903 * Windows uses case-insensitive file name pattern matching
1904 but AFS is a case sensitive file system. The AFS3 directory
1905 format is block based, uses network byte order and
1906 includes a hash table for fast case sensitive lookups.
1907 This causes several problems for the Windows AFS client.
1908 (1) Traversing the directory blocks is cpu expensive
1909 (2) A hash table miss does not indicate that the desired
1910 entry does not exist.
1911 (3) Determining whether a non-ambiguous inexact match or
1912 the entry does not exist requires a linear traversal
1913 of the entire directory.
1914 These issues often result in 100% CPU utilization.
1916 These issues are addressed by building a modified B+ tree for
1917 each directory and then using the B+ tree for searches.
1919 This feature is disabled by default. To enable it set
1920 the "BPlusTrees" registry value to 0x01 in the AFS Client
1921 Service Parameters key.
1923 * Correct a null pointer dereference in the symlink recursion
1924 detection code. (Introduced in 1.5.22)
1926 * Add the Network Identity Manager AFS Provider to AMD64
1931 * A new registry value, "BlockSize", can now be used to adjust
1932 the size of the internal buffers used to store file data.
1933 The default is 4KB. When set, the values 1..1024 specify
1934 multiples of 4KB. Values > 1024 specify actual block sizes
1935 and must be multiples of 4KB. The actual block size in KB
1936 can not be larger than the chunk size which defaults to 1MB.
1938 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1941 * pts.exe and ptserver.exe built with -DSUPERGROUPS
1943 * Service Manager shutdown hints have been added. The hint
1944 specifies that it may take the AFS client service up to two
1945 minutes to shutdown.
1947 * Fix the recursion detection feature introduced in 1.5.22.
1948 The previous implementation broke the evaluation of paths
1949 using ../path or ./path. It also broke the evaluation of
1950 symlinks to absolute paths. For example, /afs/cell/path.
1952 * Fix an uninitialized variable in cm_MakeDir that could result
1953 in crash if the requested directory could not be created.
1955 * Fix a the VolumeStatusNotification log messages to save
1956 the input string in the circular log buffer so garbarge
1957 won't be output to afsd.log when the log is dumped.
1960 * Increase the SMB Maximum Path length to 260 characters
1961 from 256 as that is what CIFS supports.
1963 * On 64-bit Windows, use a unique description string for
1964 the 32-bit AFS Shell Extension.
1968 * In the AFS Admin Library, replaced all of the
1969 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
1970 allows parameter checking and proper type conversions to be
1971 performed. This is crucial when passing parameters that are
1972 64-bit wide on systems in which 'long' is a 32-bit value.
1973 Otherwise, the parameter list on the stack is interpretted
1974 incorrectly and the function not only doesn't succeed but it
1975 might corrupt memory and if lucky crash the application.
1977 The only program that we care about that uses this library
1978 is the AFS Server Manager. Now it doesn't crash.
1980 * Fixed smb directory searches for non-wildcard names. The
1981 optimized function should not have set the bulk in progress
1982 flag. This would adversely affect later directory searches
1983 that required a bulk search.
1985 * Detect infinite recursions (up to 512 stat cache objects
1986 in a chain.) If an infinite recursion is detected, return
1987 an error to stop the CIFS client from querying us forever.
1989 * Executables executed out of AFS name space can have their
1990 pages swapped out by Windows. Those pages will be reloaded
1991 when they are required. However, if the cache manager no
1992 longer has a callback and all of the file servers with
1993 the volume data are down, then the application will crash
1994 due to the inability to read the required page.
1996 We now track the down time of each server. If the callback
1997 expires after the server that issued it went down and all
1998 servers are down, we delay the expiration to permit the data
1999 in the cache to be continue to be used by Windows or an
2002 * The Windows AFS client has always had very poor performance
2003 with regards to directory management. In the case of
2004 directory updates caused by create, rename,and delete operations,
2005 the AFS cache manager will now perform local directory updates
2006 when possible in order to avoid re-reading the directory from
2007 the file server. This will be done in whenever both of the
2008 following conditions are true:
2010 - the data version on the directory increased by one as a
2011 result of the operation
2013 - all of the directory buffers necessary to make the change
2014 are present in the cache
2016 * The error message in the AFS Control Panel describing the
2017 need to be a Windows Administrator has been replaced with
2018 a message describing the need to be in the AFS Admins Group.
2020 * If the vldb server list is empty when attempting to evaluate
2021 a cell mountpoint it was possible for a crash to occur. This
2022 can only happen if there is an existing mount point and the
2023 is a failure attempting to load the server list from the
2024 CellServDB file or via AFSDB queries.
2026 * dirty buffers only have their dirty bytes stored to the file
2027 server instead of the entire chunksize
2029 * the default chunksize has been increased from 128KB to 1MB
2034 * changed the enum values for cm_serverRef_t state info to use a
2035 private name space to avoid collisions (srv_)
2037 * added a srv_deleted state for cm_serverRef_t objects. This
2038 state is set when cm_FreeServerList() is called with the
2039 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
2040 not always delete the cm_serverRef_t from the list if it is
2041 still in use by another thread. the srv_deleted state means
2042 the object's contents are no longer valid and it must be
2043 skipped. It will be deleted the next time the object is
2044 freed and the refcount hits zero.
2046 * the srv_deleted state is also used when a file server reports
2047 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
2048 as offline. This is done to prevent additional usage of the
2049 stale vldb data while waiting for the update volume request
2052 * added a state field to the cm_volume_t object (enum volstate
2053 vl_ name space) that maintains the state of the volume based
2054 upon the states of all of the cm_serverRef_t and cm_server_t
2057 * modified cm_UpdateVolume() to set the state of the cm_volume_t
2058 RW, RO, and BK to either vl_alldown or vl_online. There can't
2059 be any other states because cm_UpdateVolume() destroys any
2060 previous knowledge we might have had regarding busy or offline
2063 * modified cm_UpdateVolume() to update the volume name in the
2064 cm_volume_t to the volume base name if the previous value was
2067 * modified cm_FollowMountPoint() to check to see if the volume
2068 name is a volume ID and if so call cm_GetVolumeByID instead
2069 of cm_GetVolumeByName. This ensures that volume IDs are always
2070 looked up as numeric values. There is no longer a need to
2071 maintain a separate cm_volume_t containing the string representation
2074 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
2075 The first flag is a "CREATE" flag which is set by all existing
2076 calls. The flag is not set by calls to cm_GetVolumeByID() from
2077 the server probe code when volume status is being updated. We
2078 do not want the server probe operation to result in additional
2079 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
2080 which is set when the server probe code updates the volume status.
2081 This flag will be used to prevent the server probe operation from
2082 changing the order of the least recently used queue.
2084 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
2085 allocated for a given set of normal, readonly, and backup volumes
2086 regardless of whether or not the volume is accessed via name or
2087 ID number. The cm_volume_t namep field is always the base name
2090 * Added a new volume state, vl_unknown. This state is used as
2091 the initial state for all cm_volume_t when the cache manager starts,
2092 for each cm_volume_t at creation, and for each cm_volume_t when
2093 recycling. The cache manager does not know the state of all
2094 volumes in the world, only those that are in the cache and for
2095 which it has queried the VLDB and hosting file servers.
2097 * modified cm_GetVolumeByName() to initialize the state of a
2098 volume to vl_unknown. The actual state will be set when a
2099 cm_VolumeUpdate() call completes successfully.
2101 * changed name of scache hash table variables to avoid ambiguity
2102 when adding hash tables for volumes
2104 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
2106 * modified cm_UpdateVolume() to handle the case in which there is
2107 no RW volume but there is are RO volumes for a given base name.
2108 This is done by querying for the ".readonly" volume name if the
2109 base name does not exist in the VLDB. We never query for the
2110 .backup name because under the current usage model a .backup
2111 volume may only exist on the server that the read-write volume
2112 is located. If there is no RW volume, there can be no .backup.
2114 * Added four hash tables for cm_volume_t objects to improve the
2115 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2116 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2117 BKID. Three ID hash tables are necessary as long as it is
2118 desireable to maintain a single cm_volume_t containing all
2119 of the related RW, RO, and BK volume data. Having the RW and
2120 RO volume data in the same object is necessary for the
2121 implementation of cm_GetROVolumeID() which returns either the
2122 RO or RW ID depending upon the existence of RO volume instances.
2124 * Added a volume LRU queue so that volume reuse becomes fairer.
2125 This does not replace the all Volumes list which is used when
2126 it is desireable to walk a list of all the volumes whose order
2127 is not going to change out from underneath you which makes it
2128 safe to drop the cm_volumeLock.
2130 * handles volume hash table updates where volume name to
2131 volume ID number changes. The volume name remains
2132 constant in the cm_volume_t. if a vos rename is performed,
2133 the name of the volume will change and the volume IDs will be
2134 updated. Subsequent access to the old volume ID will create a
2135 new cm_volume_t with the new name.
2137 * Added a daemon thread operation to query the state of volumes
2138 listed as busy or offline. cm_CheckBusyVolumes() calls
2139 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2140 or vl_offline. If the volume is now online, the status on the
2141 volume is updated. The default period is 600 seconds. This can
2142 be configured with the BusyVolumeCheckInterval registry value.
2144 * Added prototype for smb_IoctlPrepareRead() which was missing a
2145 return type in the function definition.
2147 * Added volume id lists to the cm_server_t. These lists are
2148 allocated in blocks of ~32 IDs. When a cm_PingServer()
2149 detects a change in server state, the state of the cm_volume_t
2152 * Added volID to the cm_serverRef_t object. volID is used
2153 to identify the volume for which the object is a referral.
2154 cm_FreeServerList() uses the volID to remove the volume
2155 from the cm_server_t.
2157 * In cm_Analyze, when VNOVOL or VMOVED are received,
2158 call cm_ForceVolumeUpdate() to force a refresh of the volume
2161 * Added cm_VolumeStatusNotification() which is used at the moment
2162 to log volume status changes to the trace log. It will also
2163 be used as the access point to the File System Filter driver
2164 notification engine.
2166 * Added an all cm_scache_t list to cm_data. This replaces the use
2167 of the stat cache LRU queue when we need to enumerate all
2168 entries. The LRU list order is not static and when using it to
2169 enumerate all entries it can result in items being missed or
2170 items being processed more than once.
2172 * Modified cm_Analyze(). Instead of reseting the busy or offline
2173 state of a volume and forcing a retry of the operation
2174 cm_Analyze will defer to the background daemon thread that will
2175 update the state once every 600 seconds.
2177 * Added the automatic generation of a Freelance ".root" read-write
2178 mountpoint that refers to the root.afs volume of the workstation
2179 cellname at the time the mountpoint is created.
2181 * re-write cm_Analyze to make better use of the known volume
2182 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2185 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2186 busy volumes will be reset to srv_non_busy by the function
2187 but there is no mechanism for querying the busy state other
2188 than by attempting to access the resource.
2190 * cm_Analyze will query the state of an offline volume before
2191 deciding whether or not to retry when all volume instances
2194 * the algorithm used after an IP address list change detection
2195 has been modified. we now check the state of the VLDB servers,
2196 then reset the rx connections, and finally check the state of
2197 the file servers. Previously we flushed the rx connections
2198 and then checked the servers in whatever random order they
2199 were first discovered.
2201 * fix the volume status reported by "fs examine"
2203 * various improvements to the AFS Network Identity Manager Provider.
2204 - Don't access freed objects after the window has been destroyed.
2205 - Write back the AFS cell list in the identity configuration
2206 even when its empty.
2207 - Disable all use of the provider when AFS is disabled for a
2208 particular identity.
2210 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2211 alias instead of the full cellname, double check the full cell
2212 name and see if we have a matching cell before allocating a new
2215 * Dump more data with "fs memdump". All volumes and scaches and
2218 * Reverse the order of the all cells list. This ensures that
2219 the workstation cell is always first in the list.
2221 * If we recognize the server from which a callback revoke was
2222 received, then we can ensure that we only invalidate the status
2223 information for cm_scache_t objects belonging to that cell.
2225 * When building for amd64, use APPVER 5.02 as we do not support
2226 64-bit Windows 2000 and explicitly state the machine type in
2229 * no longer treat VMWare network adapters as equivalent to a
2230 loopback adapter. For some unknown reason, publishing the "AFS"
2231 netbios name on them frequently results in a name collision
2234 * Use read locks whenever possible instead of write locks when
2235 walking the allCellsp list. This permits better use of multiple
2238 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2239 is not present on the directory optimization. It would report
2240 mount points and directory objects for which there was no
2241 status information as a zero length file. Instead we can use
2242 the vnode number of the object to determine if it should be
2243 considered to be a directory or a file. If it is odd it is a
2244 directory; even it is a file.
2246 * Give Back Callbacks to the file servers in response to suspend
2247 and shutdown power management events. Only give up callbacks
2248 to UP servers and do so with a very short RX dead timeout.
2251 * Fix a race condition in the SMB FID cleanup code that could
2252 result in a stat cache mutex never being released.
2254 * Fix the graphics files that are installed as part of the
2255 IBM Administrators Guide.
2257 * Change the behavior of pioctl path parsing to not follow
2258 symlinks or mount points during evaluation. This results in
2259 the FID being the symlink or mount point and not the object
2260 they point to. This is consistent with the behavior of the
2263 * Enable AFSDB record lookup for command line tools such as
2266 * CellServDB update 2007-07-05
2268 * Added a Volume Status Plug-in Interface that permits third parties
2269 to receive notifications when the volume state changes occur.
2271 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2273 * Prevent the afsprocmgmt.dll from being unloaded while the
2274 background thread it starts is running.
2276 * Avoid a race condition when closing smb_fid_t objects.
2279 * fix install path of IBM Quick Start Guide when installed by
2282 * fix the client so that it can write into directories with
2283 only 'li' permissions.
2285 * fix installation of 32-bit Shell Extension for 64-bit systems.
2286 Install it under its own HKCR keys.
2288 * modify the wix installer to remove and upgrade install packages
2289 with the same version number.
2291 * when flushing volume location data every two hours or in
2292 response to "fs checkvolumes", also flush the mount point
2293 links in order to force their re-evaluation.
2295 * aklog was copying an uninitialized buffer as part of the
2296 foreign cell automatic pts registration
2298 * when obtaining directory listings, check for read permission
2299 on the directory. If read permission does not exist, fake the
2300 status information for the directory entries. This improves
2301 performance, prevents the client from pounding the file server,
2302 and prevents leakage of cached status information.
2304 * "fs getcacheparms" output was incorrect.
2306 * NetIdMgr AFS Provider should only automatically add token
2307 acquisition for cell foo.com if the realm of the cell matches
2308 the realm of the identity.
2310 * Fix a null pointer dereference in cm_MergeStatus()
2314 * network identity manager provider:
2316 - Correct for the fact that pr_CreateUser tampers with the
2317 contents of the username input buffer.
2319 - Take into account that the krb4 credential data structure
2320 does not have separate realm fields for the user and service.
2321 When krb524 is used with a cross-realm ticket, the realm
2322 describes the service not the user.
2324 * Many integrated logon changes:
2326 - Fix eventlog reporting. Do not attempt to log an event
2327 if the event source registration fails. Use DebugEvent0
2328 instead of DebugEvent when there are no parameters.
2330 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2333 - Fix the reading of Domain specific configuration for
2334 LogonScript and TheseCells. Previously the dwSize value was
2335 being overwritten so that subsequent RegQueryValueEx call
2338 - Fix a memory leak in the TheseCells reading code.
2340 - Add support for Domain specific "Realm" specification. The
2341 realm is the realm to be appended to the username. When
2342 logging in as a domain or to the local machine, the specified
2343 "Domain" name is not going to be a valid realm name.
2345 - Construct a proper principal name based upon the domain
2346 specified realm for use in obtaining tokens with KFW.
2348 - If the domain specified "TheseCells" list includes the
2349 default cell, do not obtain tokens twice.
2351 * An unprotected call to the Kerberos v5 function,
2352 krb5_free_context(), can result in an invalid memory access
2353 exception if the contents of the krb5 profile, krb5.ini, cannot
2354 be parsed correctly. The error is located within the afskfw
2355 library which is used by afscreds.exe, afssvrmgr.exe, and
2356 afslogon.dll. When the error is triggered within afslogon.dll
2357 a blue screen condition can occur when the winlogon.exe system
2358 process terminates unexpectedly.
2362 * Do not stop the afs client service with an assert
2363 if the cell name cannot be determined when responding
2364 to a "fs whereis" command.
2366 * Integrated Logon (afslogon.dll). No longer attempt
2367 to preserve Kerberos 5 tickets for the logon session
2368 as that job is now supported by KFW 3.1 and above.
2370 * The Network Identity Manager plug-in will attempt to
2371 remove expired service tickets from the cache.
2375 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2377 * AFS Server Manager and AFS Account Manager data cache
2378 moved from HKCU\Software\TransarcCorporation to
2379 HKCU\Software\OpenAFS to prevent crashes due to data
2382 * Reduce the default UP server probe time from 10 minutes
2383 to 5 minutes to match UNIX. This will not solve the NAT
2384 port mapping timeout issue because most NATs default to
2385 a value less that 5 minutes and this value ensures that
2386 the afs client server will start probes 5 minutes after
2387 the last probe set ended. The probe period for any one
2388 server is therefore longer than 5 minutes. The more
2389 servers in the list the longer the time between probes.
2391 * Before marking a server as down in cm_Analyze, retry the
2392 request using a new rx connection to the server. If the
2393 cache manager is behind a NAT, it is possible that the
2394 response from the server is being sent to the original
2395 address:port from which the rx connection was established.
2396 This will appear to the cache manager as the server being
2397 down when in fact it is not. Forcing the use of a new rx
2398 connection will result in the server responding to the
2399 most recent address:port assigned to the client by the NAT.
2401 If the server is in fact down, it will simply take a little
2402 longer to notice it.
2406 * Removing a directory that contained entries was failing with
2407 the wrong error code being returned to the CIFS client.
2411 * rxdebug exponential back-off. don't take down a server that
2412 is already struggling to respond.
2414 * Network Identity Manager plug-in registers help on the Help
2415 menu if it is supported by NIM. (KFW 3.2 and above)
2417 * Plug-n-play networking did not always work after coming back
2418 from a Windows Suspend. The AFS client service failed to
2419 re-compute the LAN adapter to bind to and instead attempted
2420 to bind to all LAN adapters. This would polute the public
2421 network with the name "AFS" which could lead to conflicts
2424 * Build installers using Microsoft Hotfixes provided at
2425 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2426 Daylight Savings Time issues.
2428 * "fs checkserver" does not require admin privileges unless
2429 the probe interval is being changed.
2433 * A reference leak was discovered in the mountpoint traversal
2434 functionality which resulted in cm_volume_t objects not becoming
2435 eligible for re-use. This will trigger a panic condition when
2436 more than the maximum number of volumes is accessed.
2438 * Fixed the success test for the "fs memdump" operation. Success
2439 is now reported as such. Added cm_volume_t object information
2442 * A data corruption error has been fixed. Write requests
2443 queued for background processing were not retried upon failure.
2444 This leads to inconsistencies between the contents of the file
2445 in the local cache and the contents of the file on the file
2448 The implemented fix is to re-queue any request that has failed
2449 due to timeout, volume busy, volume offline, server busy, or
2450 server offline. Requests are not processed while all servers
2451 for a volume are inaccessible. When servers become accessible
2452 the background request will be retried.
2454 In addition, if a background write fails, do not mark the buffers
2455 as clean. Failure to preserve the dirty flag prevents the buffers
2456 from being written to the file server in the future.
2458 This fix does not address the possibility that the credentials
2459 necessary to perform a write request might expire between the
2460 time the request is queued and when the request can be processed
2461 due to server accessibility.
2463 * The Freelance volume does not support per user ACLs. The
2464 cm_HaveAccessRights function did not handle this special case.
2465 Requests for privileges greater than read|lookup would result
2466 in afsd_service.exe entering an infinite loop. ISOBuster is
2467 known to cause this behavior.
2469 * When deleting files, the SMB Server would attempt to obtain
2470 the access permissions for the parent directory instead of
2471 for the file itself. This would result in an incorrect
2472 assessment of whether or not the file can be deleted.
2474 * The Network Identity Manager AFS plug-in has had its string
2475 tables updated for consistency with the language used by NIM.
2476 All credentials are referred to as credentials instead of tokens.
2477 Drop-down combo boxes now in fact drop down.
2479 * The AFS Salvager was not properly being built on Windows.
2480 This is relevant only if the AFS servers are in use.
2482 * The default signal handlers for the AFS servers incorrectly
2483 processed SIGQUIT. Instead of stopping the server, an
2484 exception was raised.
2486 * Fix another volume refcount leak. This one was triggered by
2487 an inability to contact the vlservers in order to update the
2490 * Fix checklist control used by the drive mappings tabs in
2491 afscreds.exe and afs_control.exe.
2493 * Improve output of "fs memdump" related to buffer management.
2495 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2496 already in the process of being fetched.
2498 * Improve cache performance by computing the buffer hash table
2499 size based upon the number of buffers so that the average
2500 number of entries in a bucket is seven.
2502 * Improve cache performance of read/write data by locally
2503 updating the dataVersion of all buffers associated with the
2504 file on each StoreData or StoreStatus operation provided that
2505 the dataVersion was incremented by one and the dataVersion of
2506 the buffer matched the original dataVersion of the cached
2507 status information for the file.
2509 * When recycling cm_scache_t objects, make sure that we
2510 clear all of the previous fields. It is no longer the
2511 case that we can assume that the stat cache object is being
2512 obtained only after a successful FetchStatus call.
2513 InlineBulkStatus will allocate stat cache objects for which
2514 the no status information is available.
2516 * When a file is being opened with the OPEN_ALWAYS disposition,
2517 do not require write permissions when the file already exists
2518 unless of course the file is being opened requesting write
2521 * When a file has been opened for read-only access, if the
2522 application requests an exclusive lock, obtain a read-lock
2523 instead of a write-lock. The purpose of the exclusive lock
2524 in this situation is to prevent changes to the file and the
2525 full file read lock will provide this guarrantee.
2530 * Enforce Unix Mode Write bit to prevent file deletion of read only
2533 * The CIFS READ and LIST permissions use the same value 0x0001.
2534 When a directory is being accessed, check for PRSFS_LOOKUP
2535 instead of PRSFS_READ.
2539 * Fix computations of short names of mount points. They were
2540 broken as part of the directory search optimization. (1.5.13a)
2542 * When all buffers are in use, periodically sleep the threads
2543 waiting for a buffer to become free while the global buffer
2544 lock is not being held in order to permit those threads that
2545 can free buffers to do so.
2547 * Power Management improvements. Maintain a global flag that
2548 specifies whether or not the service is in a suspend state.
2549 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2550 lana is no longer valid. Instead, stop the listener threads
2551 and if all listener threads are stopped, reset the lana_list.
2552 Allow the cm_Daemon() thread to periodically check the state
2553 of the smb listeners. If they are all stopped and the service
2554 is not suspended, attempt to restart them. If there are no valid
2555 lanas, return to the stopped state.
2557 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2558 the user is permitted to obtain read or write locks. This function
2559 would obtain the lock and then drop it returning whether or not
2560 the lock could be obtained. If the lock was in fact required,
2561 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2562 of course being that this pattern results in three RPCs to the
2563 file server (lock, unlock, lock). This is reduced to one RPC by
2564 implementing cm_CheckNTOpenDone() which frees the allocated byte
2565 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2568 * Optimize the SMB FindFirst processing for the case in which there
2569 are no wildcards. If a cm_Lookup succeeds, generate a response
2570 from the stat cache entry we have a callback for. Its cheaper
2571 than walking each buffer of directory data associated with the
2572 directory stat cache entry when we have thousands of directory
2575 * Increase the SMB Ioctl MaxData size to support larger token sizes
2577 * Disable autorun of the AFS Server Wizard when the AFS Server is
2580 * Add AFS Server service dependencies on tcpip and PNP_TDI
2582 * Remove the AFS Server Configuration Tool shortcut
2584 * Add Freelance mode awareness to AFS Server Configuration Tool
2588 * Unlike previous versions of the OS, Vista performs a shutdown on
2589 the Microsoft Loopback adapter just like it would on a real adapter.
2590 This causes the smb_Listener threads to trigger a panic during a
2591 suspend/hibernate power event.
2593 The fix is to unbind from the network adapters in response to a
2594 suspend/hibernate power event and then rebind to the adapters
2595 when a resume power event is received. Note that the resume events
2596 are not reliably delivered so it is possible the afs service will
2597 not be accessible. However, this is the best we can do.
2599 * don't allow environment variables we set in afslogon.dll to be
2600 inherited by child processes
2602 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2605 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2606 thread walking the all buffer list periodically searching for dirty
2607 buffers to write to the file server. This had several negative
2608 results. The algorithm ate up ever increasing amounts of CPU time
2609 even when AFS is idle as the size of the cache increases. Also,
2610 buffers were written to the file server in an order based upon the
2611 original buffer allocation which has nothing to do with the order
2612 in which the buffers became dirty.
2614 The new algorithm maintains a dirty buffer list. Items are added
2615 when the buffer is originally marked dirty. A buffer is only
2616 removed from the list by the buf_IncrSyncer when the buffer is no
2617 longer dirty. If the list is empty the thread goes back to sleep
2618 immediately without additional processing requirements.
2622 * Enable renaming of directories from the Explorer Shell when
2623 Offline Folders is enabled
2625 * An error is reported in response to the SMB ReceiveCoreFlush
2626 request when an error has not occurred.
2630 * optimize dirty buffer flushes in response to suspend operations
2631 do not flush if only loopback adapter is installed
2632 do no retry dirty buffer writes
2634 * do not reset netbios when performing loopback checks. doing
2635 so will break the existing lana bind
2637 * remove a deadlock in smb3.c QPathInfo
2641 * Add NetIDMgr AFS Plug-in
2643 * "Show Tray Icon" checkbox removed from afs_config.exe
2645 * Disable DNS Registrations for the Loopback Adapter and make sure
2646 that Netbios is turned on.
2648 * A reference count leak on stat cache objects was fixed
2650 * For debugging, tread ids of writers are now tracked on locks
2652 * Corrected an incorrect lock that was being held but not
2653 freed in a rare error case
2655 * Changed trace log end of line to CR-LF so it can be read
2658 * modify the buffer synchronization thread to write all dirty
2659 buffers and sleep only when there is no work to be done
2661 * improve behavior of cache manager when large numbers of
2662 temporary files are being created by multiple clients within
2665 * do not permit dirty buffers that belong to deleted files
2666 to be written to the file server. its a waste of bandwidth
2668 * do not leak stat cache references when the file server
2671 * afscreds: do not permit an attempt to obtain tokens if the
2672 afs service could not be started. this was causing stack
2673 overflows on Vista after resuming from a suspend
2677 * better handling of dot directory
2679 * fs sq and fs mkm now behave as on UNIX
2681 * add license to installers
2683 * Return Invalid Parameter error for unknown InfoLevels in
2684 SMB Tran2 Search Directory.
2686 * Correctly use WSAStartup and WSACleanup within the RX library
2688 * Check that afsd_service.exe is running before performing
2691 * Force AFS Shell Extension to load resource DLL using
2692 afs_shl_ext.dll as a base name.
2694 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2695 OpenAFS\Client\Program
2697 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2699 * Do not panic if the maximum number of volume entries are in use
2700 and one of them can be recycled.
2704 * Fix the code that reads the backconnnectionhostnames value
2705 so that there is not an uninitialized variable
2709 * integrated logon could cause winlogin.exe to blue screen windows
2710 if it was called with a domain that does not exist on the machine.
2711 this is possible during a sysprep.
2713 * corrections to cm_buf_t handling to ensure that buffers are
2714 removed from queues before they are freed and that holds are
2715 obtained on the next buffer in a chain before releasing the
2718 * do not generate MessageBox dialogs on network errors. they
2719 cause more trouble then they are worth.
2721 * move GlobalAutoMapper into a separate thread so that it does
2722 not prevent service startup if the WNetAddConnection2 call
2725 * restrict hard dead and connection timeouts to the CIFS Session
2726 Timeout value. This is being done to try to prevent virtual
2727 circuits from being destroyed prematurely.
2729 * Fix the code that writes the backconnectionhostnames value
2730 to ensure that the data buffer is written with the correct
2733 (there were no windows specific changes between 1.5.4 and 1.5.5)
2737 * Corrections to the Integrated Logon functionality that permits
2738 Kerberos 5 TGTs to be communicated into the user session. The
2739 transfer is now more secure and succeeds when the user does not
2740 have read permission to the %SystemRoot%\TEMP directory.
2742 * NSIS installer updated to version 2.18
2744 * WiX installer updated to version 2.0.4221.0
2746 * Updates to the IFS functionality for 64-bit correctness and
2747 permit compilation. The IFS functionality works for pioctl()
2748 operations but not much else at this point.
2750 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2751 query and set functions.
2753 * Ignore the Windows CIFS client session timeout entirely. Timeout
2754 only after the AFS Hard Dead Timeout period expires.
2758 * move headers, libraries, and samples from OpenAFS\Client to
2761 * remove the auto-registration of AFS IDs for foreign cells
2762 from integrated logon. this permits afslogon.dll to avoid
2763 initializing the RX library. afslogon.dll is a network
2764 provider library and as such it is loaded into all applications
2765 which perform a CreateProcessAsUser call. Initializing the
2766 RX library spawns background threads to listen for inbound
2767 packets. When the network provider is unloaded from the
2768 process, the RX library does not cleanup the threads. If they
2769 continue to execute after the library is unloaded, the process
2772 * implement inline bulk status calls. this will significantly
2773 improve the performance of directory listings.
2775 * prevent a crash when evaluating mountpoints to volumes that do not
2778 * new command: fs uuid [-generate]
2779 displays the current uuid and allows a new one to be generated
2780 on the fly without restarting the afs client service.
2782 * fix the 64-bit installers to properly document the install
2785 * move help files to the proper location so they can be located by
2786 the system tray tool and the control panel.
2788 * "over quota" and "disk full" errors were not being reported to the
2791 * largefile support has been added. Files greater than 31-bits can no
2792 longer be written to file servers which do not support large files.
2793 When file servers support large files, files up to 64-bits can be
2796 * cm_SyncOp could return with the input buffer unlocked upon error
2797 which would result in a panic condition. This bug was introduced
2800 * vlserver failover when freelance mode did not work. This prevented
2801 the root.afs volume from being loaded if the first vlserver could
2806 * A bug in the CIFS Standard Info response packet caused directories
2807 to appear as standard files. This prevented Microsoft Vista from
2808 being able to walk the AFS directory tree.
2810 * Mount points must be reported as directories in response to
2811 Query Path Info operations.
2813 * A deadlock was detected when performing "fs flushall" if the
2814 file server reports VNOVNODE. The scp->createBufferLock is
2815 already held by the current thread and there is no mechanism
2816 to propagate the knowledge. Therefore, an alternate mechanism
2817 for clearing the cache must be developed. A new function
2818 cm_RecycleSCache(scp,flags) has been extracted from
2819 cm_GetNewSCache(). This function performs the task of recycling
2820 an cm_scache_t object. When called from cm_GetNewSCache() with
2821 no flags the expectation is that there are no associated buffers
2822 that are queued to be read or written. When called from
2823 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2824 flag, any queued buffers will be de-queued and marked as if
2825 the operations were performed so the data can be discarded.
2827 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2829 * In response to VNOVNODE, the parent is only discarded if the current
2830 cm_scache_t is not a directory.
2832 * In the Ioctl FlushFile and FlushVolume functions, there were no
2833 checks to protect against flushing the Freelance SCache entries.
2835 * In FlushFile, the wrong cm_scache_t object was being released.
2837 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2838 recycled. Choose a new entry if cm_RecycleSCache fails.
2840 * Speed up the performance of the cache manager by not
2841 holding the smb_fid_t mutex across calls to cm_SyncOp
2844 * Ensure that all smb_fid_t flag references are protected.
2846 * Remove a deadlock between smb_fid_t mutex and
2847 smb_rctLock introduced in 1.4.1.
2849 * Fix aklog AFSID lookup to use not send the realm
2850 name when the realm is the local realm for the
2851 cell. The PT Server doesn't strip the realm
2852 name when it is the local realm.
2854 * Treat "\\afs\*." as an alias for "\\afs\all"
2856 * Dynamically adjust the priority of server threads
2857 based upon the age of the CIFS request that is
2858 being processed. The priority is bumped one level
2859 for every 15 seconds of processing time.
2861 * The original openafs contribution mis-used the
2862 cm_GetCallback function. This function should only
2863 be called as a side effect of a call to
2864 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
2865 was being called results in a FetchStatus RPC being
2866 performed even when the client already has a valid
2867 callback and does so in a manner that destroys the
2868 synchronization of the threads that are calling
2869 cm_SyncOp. If nothing else this results in a
2870 significant performance penalty.
2872 * More changes to cleanup of smb_vc_t objects in
2873 order to prevent race conditions.
2875 * Reduce the amount of Power Management event logging
2878 * If the client detects that the IP address has
2879 changed, force the use of new RX connections. This
2880 enables the file server to recognize that the client
2883 * Remove a race condition and deadlock associated
2884 with cleanup of smb_vc_t objects.
2886 * Fix the ANSI filename support. One function that
2887 required it was missing the translations.
2889 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
2892 * Add a DllMain() function to afspthread.dll. Perform
2893 cleanup of thread local storage, rx queues, critical
2894 sections, event handles, etc. This was implemented
2895 to avoid crashes in afspthread.dll when applications
2896 which load/unload the dll attempt to terminate.
2898 * Lock request behavior has been altered. As of this
2899 release a locally allocated lock will be used whenever
2900 the volume is RO or the requesting user's permissions
2901 are 'rl'. This will allow end users to execute programs
2902 from RW volumes when they do not have the 'k' privilege.
2904 Note that a recent discovery indicates that the 'k' bit
2905 is only supposed to control the ability to obtain read
2906 locks. Write locks are to be controlled by the 'w' and
2909 * Code has been added to enable the cache manager to be
2910 able to obtain the AFSID of the authenticated user.
2911 This is required to be able to determine when "Insert"
2912 privileges should apply. AFS considers a file to be
2913 a new file whenever the owner's AFSID matches the
2914 AFSID of the user. The current implementation relies
2915 on the name specified during the SetToken call to be
2916 correct. This will be replaced in the future with a
2917 new RPC that allows the AFSID of the authenticated
2918 user to be obtained.
2920 * A 32-bit-tools installer for use on 64-bit systems
2921 is now available. These tools are for use with 32-bit
2922 versions of Kerberos for Windows or JPSoftware's 4NT.
2924 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
2926 * Remove the use of the "active_vcp" variable. This variable
2927 no longer serves any useful purpose and was the cause of
2928 race conditions that produced under counts of the smb_vc_t
2931 * Change default minidump output to include data segments
2933 * restore access to cm_IoctlGetVolumeStatus and
2934 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
2935 Still need to figure out why they are being called
2936 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
2938 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
2940 * Update build system to include the "sys_name" in the
2943 * optimize the number of calls to obtain the smb_user_t for a
2944 smb virtual circuit.
2946 * add logging of server addresses, status and capabilities when
2949 * prevent multiple threads from pinging the same server simultaneously.
2950 instead, threads block and then use the result of the ping that was
2951 already taking place.
2953 * Add registry values to allow the defaults for daemon checks to be altered
2954 at startup. Log values to afsd_init.log.
2956 daemonCheckDownInterval
2957 daemonCheckUpInterval
2958 daemonCheckVolInterval
2959 daemonCheckCBInterval
2960 daemonCheckLockInterval
2961 daemonCheckTokenInterval
2963 * Add EventLog messages for transitions in fs crypt state.
2965 * Byte range locking (when locks will be obtained from the
2966 is still under discussion)
2968 * 64-bit Windows operating systems now supported
2969 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
2970 This is a native 64-bit implementation of OpenAFS for Windows.
2971 There is currently no 64-bit version of KFW and therefore
2972 there is no (easy) method for obtaining AFS credentials.
2973 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
2974 to obtain AFS tokens but there is not for the feint of
2977 * A faster implementation of the fcrypt algorithms improves
2980 * The afsd_service.exe now supplies Event Log Message Text.
2981 The event log no longer generates warning text due to lack
2982 of message configuration.
2984 * When afs servers are unreachable and network stack returns
2985 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
2986 mark the attempted server as being down. This allows for
2987 very fast response to the calling application. This change
2988 has a pronounced improvement when the Windows Explorer is
2989 used on disconnected mobile computers.
2991 * Added DebugOutputString debugging of the RX library
2993 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
2995 * Fix the rx-lwp implementation to randomly generate rx call
2996 identifiers which are supposed to be unique at a given point
2997 in time. This affects some of the command line tools used
2998 for debugging but not the AFS Client Service itself which
2999 uses the rx-pthreads implementation.
3001 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
3002 Windows although they are not distributed
3004 * don't generate an exception when logging unexpected SMB packets
3006 * the data structures used by cm_IoctlGetVolumeStatus and
3007 cm_IoctlSetVolumeStatus are not compatible with the CIFS
3008 Get/Set File Information calls so don't bother executing them.
3012 * Fix data written to registry as part of BackConnectionHostnames
3015 * Fix initialization bug when allocating cm_server_t objects that
3016 would cause servers to appear to be down after each allocation
3017 or renewal. Server would appear up after the next check servers
3022 * Windows will return WSAECONNRESET to the next receive packet
3023 operation on a socket if an ICMP packet is received in response
3024 to a packet that was sent and could not be delivered because
3025 the port was unreachable. We have to treat this as
3026 WSAEWOULDBLOCK so that we can retry the select() check and
3027 not mark the connection as bad.
3029 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
3031 * The client manager maintains a least recently used list of
3032 all the ACLs. If an ACL is already the most recently used
3033 do not bother to remove it from the list and place it on the
3036 * As part of the process of adding a new server, set the
3037 server is down flag and then perform a ping. If the server
3038 is up, the flag will be reset. If the server is down, the
3039 client will not timeout unnecessarily while attempting to
3042 * If a Mountpoint String is the empty string, return PATH NOT
3045 * Modify behavior of the Network Provider to avoid querying the
3046 profile location if integrated logon is known to be disabled.
3048 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
3049 will always be preserved regardless of whether or not integrated
3050 logon was used to obtain tokens. When unset, tokens will be
3051 destroyed regardless of whether integrated logon was used to
3052 obtain the tokens unless the profile is stored in AFS.
3054 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
3055 message is being delivered even when the user is not logging
3056 off of the computer.
3058 * Fix a deadlock condition introduced in rc6.
3060 * Cell names are case-insensitive. Since the founding of OpenAFS
3061 there has been a bug that treated "cell", "Cell", and "CELL" as
3062 different cells. This would result in what appeared to be
3063 random disappearance of user tokens or a failure to authenticate
3064 even when tokens were listed by the "tokens" command.
3066 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
3068 * Additional auditing of the smb code revealed that smb ioctl
3069 operations were failing to free smb_fid_t objects
3071 * The File Identifiers assigned to smb_fid_t objects could be
3072 given the value 65535 which is 0xFFFF which is also the
3073 INVALID_HANDLE value. Both 0 and 65535 are now treated as
3076 * Audited all assignments to fields and ensured that they are
3077 protected by obtaining the relevant mutex object.
3079 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
3080 if both the current cm_scache_t object and its parent are
3081 both returning VNOVNODE.
3083 * Removed all the logic associated with the 'dead_vcp' variable
3084 which must have been added for debugging because its used
3085 was never thread safe and it truly served no purpose.
3087 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
3089 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
3090 registry settings. Default is ON and 120 seconds. If set to
3091 OFF, tokens will not be destroyed.
3093 * Fixed a number of race conditions and reference count problems
3094 associated with global queues.
3096 * Added a method to detect abandoned smb virtual circuits so their
3097 associated file handles, locks, user tokens, and smb session
3098 resources can be released.
3100 * Added a mechanism to allow user tokens to be preserved
3101 during integrated login and be freed after logoff.
3103 * Check access permissions to the file before the file is
3106 * More improvements to the handling of SMB Virtual Circuits
3107 when SMB sessions are closed.
3109 * Correct a race condition introduced in rc4 that could result
3110 in a crash of afsd_service.exe.
3112 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3114 * Properly handle pre-mature SMB session closures.
3115 File handles, tree connections, and user sessions are now
3116 freed when the virtual circuit is closed.
3118 * smb_username_t and smb_user_t objects are now properly reference
3119 counted and freed as appropriate.
3121 * disable to use of AFS file locks. byte range locking is still
3122 implemented but all locks are managed locally. This prevents
3123 multiple processes from editing the same file on the same
3124 machine but it does prevent multiple processes on separate
3125 machines from modifying the file.
3127 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3129 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3130 the same server parameter would result in a panic condition
3131 if the server had not already been contacted by the AFS
3134 * Better warnings are written to the afsd.log file when
3135 the Windows RPC Protocols are not properly configured as
3136 is often the case on Windows 2000.
3138 * fix the procmgmt library to restore C RTL signal handlers when
3139 it is being detached from the process prior to process termination.
3141 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3143 * close a lock leak in the rx library that could be triggered while
3146 * prevent removal of drive mappings in response to afscreds.exe -M
3148 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3149 strings and flush the stat cache entry.
3151 * prevent server objects from being freed if user preferences are set.
3153 * Update WIX support to 2.0.2419.0
3155 * Release all locks on the server at service shutdown.
3157 * A down server should be marked up as long as it responds to the client.
3158 It doesn't matter if the server returned an error as long as the error
3159 is something other than VRESTARTING. If there is such an error, all
3160 connections on that server should be marked for replacement.
3162 * Change the check "up" servers period from once per hour to once every
3163 ten minutes. This matches the behavior of the Unix cache manager.
3165 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3166 * Byte Range Locking support has been added. Architectural documentation
3167 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3168 made to enforce Windows mandatory locking model using locally managed
3169 byte range locks with full file locks obtained from the file server.
3170 Read-only volumes provide the user with an implied locking privilege.
3171 Read-write volumes require that a user be granted the locking privilege
3172 in order for a lock to be obtained. As most Windows applications use
3173 Shared file access modes, they require the ability to obtain locks in
3174 order to access files.
3176 * Implement error translation functions to convert Universal Error Codes
3177 output by modern servers to local errno and winsock error values.
3178 This fixes the error message problems using pioctl() calls from fs.exe
3179 and the AFS Shell Extension.
3181 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3182 * Allow the AFS Client Service to set the firewall configuration by
3183 retrying until after Windows determines the boot period is over.
3185 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3188 * Detect IP address changes from within the AFS Client Service for the
3189 purpose of checking down servers sooner
3191 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3192 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3193 user that there is a Kerberos key problem. Previously the user would
3194 receive no feedback when there was something wrong.
3196 * Fix the large integer library support to properly handle values greater
3199 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3200 * Updated CellServDB
3202 * Fixed file name pattern matching. Question marks can match nothing
3203 if found at the end of a component.
3205 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3206 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3208 * The MSI installer now has a new Product ID for each build. This
3209 is to allow for automated upgrades. The Upgrade ID is remaining
3212 * Volumes that were cached and then subsequently were salvaged,
3213 moved or otherwise taken temporarily offline (as opposed to being
3214 busy) would become unreachable. A distinction is now made between
3215 a volume being offline and a server being down.
3217 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3219 * Fixed several AFS Shell Extension issues:
3220 - Checking Server Status would only work if a specific cell was
3222 - If servers were down, the Server Down dialog would write to
3223 invalid memory and cause explorer.exe to crash
3224 - Default Push Buttons were not specified for many dialogs
3225 - The extension would be unloaded by the Explorer Shell and
3226 never be reloaded under some circumstances.
3228 * Fixed the AFS System Tray menu to disappear if another Window
3229 was selected before a menu item was selected
3231 * Added volume owner and group information to "fs examine" output
3233 * Added fs and registry support for enabling/disabling/clearing
3234 rx statistics gathering.
3236 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3238 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3239 GUIDs for each platform.
3241 * Documentation directory structure and default installation rules
3244 * New OpenAFS for Windows HTML Release Notes installed
3245 as part of the documentation set.
3247 * Add thread id to rx debugging messages
3249 * When all servers are offline, return a bad network path error
3250 immediately. This is necessary to allow Explorer to do the
3251 right thing and not hang. The background daemon thread will
3252 check the down servers every 30 seconds to see if they are
3255 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3257 * The AFSCache file has been modified to store both the
3258 serial number of the volume on which the file resides
3259 and the SID of the local machine. If either change,
3260 a new UUID is generated.
3262 * add "fs flushall" command that forces all data buffers
3265 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3267 * add support for '/' instead of '\' in pioctl() calls
3269 * Apply AFS Client Admins group protect to AFS Shell Extension
3271 * Add support for \\afs\<name> to most AFS Shell Extension
3272 commands except the symlink methods.
3274 * when installing the Microsoft Loopback Adapter, enable
3275 MS Client for Networks and prevent an install failure
3276 by not calling CoInitialize twice in the same thread.
3278 * reload cell vldb values from the CellServDB every two
3279 hours in case it changes
3281 * When updating cell information from DNS, be sure to set
3284 * Add support to allow use of \\AFS\<foo> where <foo> is
3285 either a mount point or symlink. As <foo> is normally
3286 treated as a share name, we transform it into \\AFS\all\<foo>
3289 * Init 'code' to prevent false errors when integrated logon
3290 is disabled and the service is not running
3292 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3294 * Removed trace log messages that were hampering performance.
3296 * Fixed a deadlock that was being triggered by editing Word
3297 documents stored within AFS with WinWord 2003.
3299 * Bit 3 of the TraceOption registry value is now used to set the
3300 default for "fs trace" in the non-Debug builds. The new default
3301 is off for release builds and on for debug builds.
3303 * Bit 2 of the TraceOption registry value can be used to turn on
3304 real-time output of debug log entries to the Windows Debug
3305 Output monitor interface. This data can be viewed with tools
3306 such as Sysinternal's DbgView.
3308 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3309 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3310 a resource we should not be returning status codes such
3311 as Network Busy or Remote Host Down because these errors
3312 will imply that the AFS Client Service is Busy or Down
3313 and that is simply not the case. Instead we will return
3314 Bad Network Path as the path is temporarily not available.
3316 Instead of returning Sharing Paused when there is a
3317 timeout error, return Timeout. Once again, the AFS Client
3318 Service is not paused.
3320 * afscreds.exe would display an Obtain Creds dialog when
3321 the expired credentials reminder was triggered even
3322 if there was no network path to the KDC. This is prevented
3323 by adding KDC probe logic to the reminder thread.
3325 * afscreds.exe would display expired tokens no differently
3326 than unexpired ones. This would make it difficult for the
3327 user to distiguish when the tokens were expired. For the
3328 English build added a new resource string "(expired) that
3329 is displayed instead of the expiration time.
3332 * "fs wscell" when executed in freelance mode will return the
3333 name of the cell configured in the registry. The root.afs
3334 volume is not loaded from this cell, but it is used for the
3335 default for aklog and integrated login.
3337 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3338 no longer work on \\AFS\all when freelance is being used
3339 unless the user is a member of the "AFS Client Admins" Windows
3342 * some more checks performed during persistent cache validation.
3345 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3346 down servers for the current cell not all cells.
3348 * If the scache LRU list becomes corrupted, the AFS Client Service
3349 might enter an infinite loop at startup. Detect the corruption
3350 and discard the cache contents in this case.
3352 * Fixed the Explorer Shell's remove mountpoint functionality.
3354 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3355 obtain a mutex lock on a cm_scache_t object whose lock is held by
3356 a thread attempting to obtain the cm_aclLock.
3359 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3360 misinterpretted as "csail.mit.edu" when the resolver
3361 is configured to append ".mit.edu" to failed DNS queries.
3363 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3364 that can be used to force the preservation of user tokens upon logout.
3366 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3367 recommended installation categories: AFS Client, AFS Administrator,
3368 AFS Server, AFS Development Kit. Each category includes a different
3369 default subset of the OpenAFS components.
3371 The OpenAFS logo is now associated with the NSIS uninstall entry
3372 in the Add/Remove Programs control panel.
3374 * The user name associated with AFS tokens when obtained with
3375 integrated login, the afs systray tool, or aklog will always
3376 include the full Kerberos 5 user principal name regardless of
3377 whether or not the cell is local to the realm.
3379 * Modify integrated login so that it does not enter an infinite
3380 loop if the service is not set to auto start.
3382 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3385 * Added uninstall.exe to Wix installation
3387 * More modifications to algorithms used to wake sleeping threads.
3391 * Changes to the algorithms used to wake threads when they are
3392 put to sleep because of conflicts over buffer operations.
3395 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3396 the process of producing AFS tokens. The tokens are stored within
3397 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3398 New functionality has been added to temporarily copy the tickets to
3399 a file ccache during the short window between the NPLogonNotify and
3400 WinLogon Logon Event. The Logon Event handler starts a process as
3401 user, afscpcc.exe, that copies the tickets into the default ccache.
3402 The file ccache is then destroyed by the event handler.
3404 * During synchronization operations on stat cache entries, if threads
3405 are forced to sleep it was possible that the threads would not be
3406 woken under heavy load.
3408 * Enforce the delete on close flag specified by the NTCreateX()
3411 * A race condition in the rx library was fixed that resulted in
3412 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3413 was cleared when threads were still waiting to use the
3414 connection. Also, fixed a potential case in which the
3415 connection could be destroyed while threads were waiting to use it.
3417 * The DNLC freelist has been seen to become corrupted with still
3418 active entries being stored on the freelist. Changes were made
3419 to perform a better job of cleaning entries before placing them
3420 onto the freelist; marking them with the correct magic value when
3421 purging the list; and allowing the list to be purged on startup
3422 if the cache validation fails.
3424 * Windows Crash Reporting does not get triggered for afsd_service.exe
3425 because it provides its own top level exception handler. This patch
3426 places an explicit call to ReportCrash() on platforms which support
3429 If you configure Crash Reporting via AD Policy to capture crash
3430 reports within the domain, then you will receive this output.
3432 * Generate MiniDumps independent of the Windows Crash Reporting.
3434 (a) If an exception occurs,a minidump will be generated at the path
3435 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3436 can be set to any other type via use of the
3439 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3441 (b) MiniDumps can be generated at any time using the "fs minidump"
3442 command. This allows you to generate a minidump without
3443 stopping the process or even requiring the use of a debugger
3446 * an RX connection which reports bad ticket now treats it the same way
3447 the expired ticket is treated. The ticket is removed from the
3448 connection and a new anonymous connection is established.
3450 * fs memdump now requires that the caller be in the AFS Client Admins
3453 * additional debugging information is provided in cm_Analyze when the
3454 error code is ignored.
3456 * Fix a regression introduced into the pioctl() client code which
3457 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3458 MIT KFW is not installed. (1.3.82a)
3460 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3463 * Fix a race condition in cm_EndCallbackGrantingCall() that
3464 could leave 'nrevp' pointing to freed memory. In the process,
3465 optimize the processing to ensure that the applications monitoring
3466 the status of the stat cache entry are only notified at most
3469 * Conditionalize access to the Explorer Shell AFS->Mount Point
3470 ->Remove menu item. Disable the item if the selected item(s)
3471 are not mount points.
3473 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3474 Local Storage. This eliminates thread contention on the
3475 global RX free packet queue. This should improve performance
3476 on multi-CPU systems.
3478 * Fix the IP address filtering code to properly load/unload
3479 the IPHelper DLL so that Windows XP/2003 and beyond systems
3480 do not have to manually probe the registry.
3482 * CleanupACLEnt() was not being called consistently with
3483 the cm_scache_t object referred to by the back pointer
3484 mutex locked. This could in very rare conditions lead to
3485 an invalid memory access.
3487 * Added a script command to the msi installer "afs_replace"
3488 which can be used to replace one version of OpenAFS with
3489 another without requiring an uninstall and reinstall if
3490 the installed msi is no longer in its original location.
3492 * Apparently the problem with multi-domain forests with cross-
3493 realm trusts to non-Windows realms was not entirely solved.
3494 The authentication to the AFS SMB service failed because
3495 the wrong name was being used. Using ASU as an example,
3496 the authentication was being performed with the name
3497 "QAAD\user" (an account in the forest root) and not
3498 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3500 The solution was to add an additional dependency on KFW
3501 in order or to be able to easily obtain the client principal
3502 name stored in the MSLSA ccache TGT. This information is
3503 used in two locations:
3505 - the pioctl() function
3507 - a new WinLogon Event Handler for the "logon" event.
3509 The pioctl function will now be able to use the correct
3510 name when calling WNetAddConnection2() and the "logon"
3511 event handler will now be able to call WNetAddConnection2().
3512 The hope is that the "logon" event handler will be called
3513 before the profile is loaded but I have not guarrantee
3517 * Fixed a locking error in cm_TryBulkStat() which had the
3518 potential of deadlocking the system for the length of time
3519 it takes to perform a bulk status fetch operation.
3521 * Replaced time conversion code (UnixTime <-> FILETIME) to
3522 be completely arithmetic instead of relying on a bizarre
3523 algorithm involving a variety of C RTL time functions.
3524 This has the side effect that UnixTime and FILETIME which
3525 are both stored in UTC are interpretted as UTC throughout
3526 the year. Windows will apply the same localization to AFS
3527 as it does to NTFS. Applications which rely on the ability
3528 to sync files between the two file systems will no longer
3529 see the timestamps of files in AFS change an hour relative
3530 to the files stored in NTFS or Windows based backup devices.
3532 * Fixed a invalid memory access under a bizarre circumstance.
3533 Windows will allow a physical mass media device to be
3534 installed via Plug N Play to the system and will assign it
3535 the lowest available drive letter. This is true even if the
3536 drive letter is currently assigned in the user session to
3537 a network device via NET USE (or its equivalent.) When this
3538 happens, queries sent to the CIFS server will contain invalid
3539 data. This invalid data was not being caught by the AFS
3540 Client Service and was resulting in a crash when Freelance
3543 * Fixed a reference count error when registering callbacks
3544 on a stat cache object if the callback was already registered
3546 * Add a case to cm_Analyze() in order to handle the case where
3547 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3548 the server list for the volume is empty. In this case, force
3549 an update of the volume info and retry.
3551 * Insert a missing cm_EndCallbackGrantingCall() which could
3552 result in threads waiting for a callback to complete to never
3555 * In the persistent cache, there is a maximum number of volume
3556 entries. Allow volume entries to be reused if the maximum
3557 number have been allocated and their reference count is zero.
3559 * If we already have a dead virtual connection object, don't
3560 ignore the fact that additional dead vc objects must be taken
3563 * Removed a deadlock condition introduced in the summer of 2004.
3564 Do not hold mx locks around calls to RX functions including
3565 rx_NewCall(). That is what reference counts on the rx_connection
3568 * Fixed an initialization error in afslogon.dll which could
3569 result in random behavior including a failure to terminate
3570 the NPLogon function. (1.3.80b)
3572 * Fixed an error preventing the use of SMB authentication
3573 on Windows 2000. (1.3.80a)
3576 * Updated CellServDB to Public CellServDB 16 Mar 2005
3578 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3579 misinterpretted as "csail.mit.edu" when the resolver
3580 is configured to append ".mit.edu" to failed DNS queries.
3582 * Fixed another case in which the client would replace connections
3583 to the host when it was not necessary. In this case the
3584 situation would occur if cryptall was on and the connections
3585 were unauthenticated due to lack of an appropriate token.
3587 * OpenAFS for Windows has failed to work at sites which are
3588 utilizing a cross-realm trust between an MIT/Heimdal realm
3589 and a multi-domain Windows forest when the workstation being
3590 accessed is not located in the root domain. This is caused
3591 by a bug in the workstation which was triggered after the
3592 introduction of Windows 2003 Server. When the bug is triggered,
3593 the workstation attempts to authenticate users to afsd_service.exe
3594 by contacting the Domain Controller instead of using the
3595 LSA loopback authentication mechanism.
3597 One of the reasons this bug occurs is because the workstation
3598 does not have a reliable way of knowing that the service whose
3599 netbios name is "AFS" is located on the workstation. This will
3600 be fixed starting in Longhorn Beta 1 by Microsoft. The
3601 "BackConnectionHostNames" registry value will be used to
3602 indicate that the authentications to that service name should
3603 be performed using the loopback authentication mechanism.
3605 In the meantime, when Logon Caching is enabled, we can force
3606 afsd_service.exe to authenticate using the logon cache before
3607 contacting the Domain Controller. This will work with both
3608 password and smart card based logons.
3610 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3611 flag should not be cleared if a volume's server reference is
3612 marked as "offline". In the case where all of the volume's
3613 servers are either "down" or the volumes are "offline", we want
3614 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3615 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3616 should never occur in practice.
3618 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3619 5 seconds, clear the server down and volume busy flags, and
3620 then force an update from the VLDB. This allows the client to
3621 update the location of a volume if the reason for it being
3622 marked offline is because it is being moved. Calls to
3623 cm_ConnByMServers() will be retried either until success or
3624 the RDRtimeout period is reached.
3626 * Correct the Power Management code behavior in response to
3627 Standby, Suspend, and Shutdown events. Instead of flushing the
3628 buffers associated with the mounted SMB submounts, simply write
3629 all buffers which are marked dirty.
3631 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3632 provides a list of cells other than the default cell for which
3633 tokens should be obtained using the default Kerberos principal.
3634 This functionality is logon domain specific and is only available
3635 when using KFW for authentication.
3637 * Fixed FindNext Invalid Handle error caused by over agressive
3638 attempts at garbage collecting dirSearch entries when the
3639 dirSearch ID wraps from the maximum value to 1.
3641 * Add support for registry defined server preferences for VLDB and
3642 File servers. See registry.txt for details.
3644 * Increased default cache size to 96MB and default number of cache
3647 * Fixed refCount leaks related to directory lookups and pioctl calls.
3649 * Callbacks revoked during a race condition with an attempt to
3650 obtain the same callback no longer result in an inaccessible
3653 * New command line tool:
3655 afsdacl : Set or reset the DACL to allow starting or stopping
3656 the afsd service by any ordinary user.
3658 Usage : afsdacl [-set | -reset] [-show]
3659 -set : Sets the DACL
3660 -reset : Reset the DACL
3661 -show : Show current DACL (SDSF)
3663 * IP addresses are no longer queried once at startup. Instead IP
3664 addresses are obtained as needed. Loopback adapter addresses are
3665 no longer published to the server.
3667 * Pay attention to the MIT KFW registry configuration for automatic
3668 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3670 * Fix (once again) case-sensitive comparisons which was apparently
3673 * Activate support for vos listvol -format
3675 * Population of the "AFS Client Admins" group failed on non-English
3676 versions of Windows because the name of the "Administrators" group
3677 is localized. Now we lookup the name of the group by using the
3678 Administrators Group Alias Relative ID.
3680 * Multiple cell token acquisition within afscreds.exe was broken in the
3681 case of Kerberos 5 cross realm authentication. Instead of contacting
3682 the KDC associated with the cell's realm, afscreds would obtain a
3683 token from the KDC of the user's realm. This would result in
3684 "invalid kvno errors" while authenticating to the AFS servers.
3685 Unauthenticated access would work. The symptoms would vary based upon
3686 whether or not the VLDB servers had been contacted using unauthenticated
3687 connections prior to the user obtaining tokens.
3689 * The list of ACL entries was becoming corrupted. It appears as if
3690 an ACL when it expires was not being appropriately cleaned up. In
3691 fact, it was left in the list of ACLs associated with the scache
3694 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3695 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3696 and "amd64_win64" for amd 64-bit processors.
3698 * When executing executables, Windows provides a mechanisms by which
3699 the normal search for DLLs can be bypassed. If a file foo.exe is
3700 being executed and there exists a file or directory "foo.exe.local"
3701 (the contents are ignored if a file), then Windows will search for
3702 DLLs first in the "foo.exe.local" directory and second in the
3703 directory in which the "foo.exe" file is located.
3705 Previous releases of OAFW would improperly return
3706 STATUS_NOT_A_DIRECTORY instead of the expected
3707 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3708 to terminate the search for the DLL causing the execution of the
3709 application to fail.
3711 * Rework the reference counting of smb_vc_t objects. The references
3712 stored in the waiting locks were not counted. This could result in
3713 an assertion if the reference count drops to zero.
3715 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3716 when running in freelance mode.
3718 * Added support for persistent caching of file contents, stat entries,
3719 volume data, cell data, and name lookup data. When the registry
3720 value, "NonPersistentCaching", is not defined or set to 0 the file
3721 specified by "CachePath" becomes a persistent cache file. The
3722 size of the cache file is computed at run time. It has a maximum
3723 size of approximately 1.2GB. The contents of the file will be
3724 validated according to the rules specified by the "ValidateCache"
3727 * Added support for UUIDs. UUIDs are kept for as long as the
3728 cache file is intact.
3730 * Added cmdebug.exe and support cache manager debugging callback
3731 interfaces in afsd_service.exe.
3733 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3734 [-addrs] [-cache] [-help]
3735 Where: -long print all info
3736 -addrs print only host interfaces
3737 -cache print only cache configuration
3739 * Symlinks to \\AFS\[all\]... will now be treated the same as
3740 symlinks to /afs/.... However, please use /afs/... as otherwise
3741 the symlinks won't work on Unix.
3743 * Correct a problem with local CellID allocation for cells whose
3744 server list is obtained via DNS instead of a CellServDB file. If
3745 the DNS information expires the CellID assigned to the entry will
3746 be changed. This causes all of the cm_scache_t objects which refer
3747 to the old cellID number to become useless. Attempts to access
3748 files or directories with cm_scache_t objects using the old cellID
3749 will fail since the server list cannot be obtained.
3751 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3752 be triggered if the volume referenced by the cm_scache_t object is
3755 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3756 in addition to SERVICE_ACCEPT_STOP.
3758 * Move RPC shutdown until after the SMB and RX shutdown procedures
3759 complete. Block until RPC shutdown is complete.
3761 * Modify afslogon.dll (integrated logon) to wait for service if its
3762 state is START_PENDING. If the timeout period occurs, reset to
3763 the retry interval and not the sleep interval.
3765 * When renewing the server list for a cell obtained via DNS AFSDB
3766 records, the cm_cell_t entry must be removed from the list of all
3767 cells. Otherwise, the list of cells will be corrupted.
3769 * In the dcache and scache modules, use the cm_scache_t dataVersion
3770 instead of the cm_fakeDirVersion.
3772 * Synchronize fs.c with the unix version.
3774 * The variable used to determine whether a file or virtual memory
3775 mapped cache is used was not properly initialized to a default
3776 value. If the registry setting "NonPersistentCaching" was not
3777 set, the choice would be random. Properly initialized to be
3780 * The memory mapped view was never unmapped before closing the file
3781 at service shutdown. This is now properly cleaned up.
3783 * Default location of Cache file is now %TEMP%\AFSCache
3787 * A bug affecting new installations of 1.3.75/76 would result in
3788 the creating of incorrect mountpoints in the freelance root.afs
3789 volume for the default cell.
3791 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3792 of "root.cell", you have been affected by the bug. To correct the
3793 problem, execute the following commands:
3795 fs rmmount \\afs\all\<cellname>
3796 fs rmmount \\afs\all\.<cellname>
3798 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3799 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3803 * A bug has been fixed in the auto-generation of Freelance root.afs
3804 symlinks which produced random entries in the \\AFS\all directory.
3806 * Support has been added for multi-homed servers
3809 * Added a new registry value, "StoreAnsiFilenames", which can be used
3810 to force the use of ANSI character sets instead of OEM Code Pages.
3811 This feature is useful when users require the ability to create
3812 filenames with 8-bit characters and need to access the files from
3813 both Latin-1 based Unix systems as well as from Windows.
3815 Activation of this feature will prevent access to files stored with
3816 8-bit OEM characters.
3818 * Shutdown all SMB threads in a synchronized manner when stopping the
3821 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3822 by the largest contiguous block of unused memory within the 2GB process
3823 space which can be assigned to the memory mapped file. Unfortunately,
3824 when the executable digital signature verification code is activated
3825 Windows sees fit to further segment the process memory which in turn
3826 reduces the size of the maximum cache file to less then 800MB. If
3827 larger cache sizes are desired, a new registry value should be set:
3829 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3831 Setting this value will disable the runtime verification of digital
3832 signatures on afsd_service.exe and the afs dlls which it loads. It
3833 will not disable the the version number check on those same files.
3834 The signature verification is not a security messure and is only meant
3835 to enhance the ability to afsd_service.exe to detect potential
3836 destablizing mixtures of DLLs from incompatible distributions.
3838 Added code to auto-disable the signature verification check if
3839 the desired cache size is greater then 700MB.
3841 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3842 used to verify the validity of the afsd_service.exe binary
3843 as well as each of the AFS DLLs loaded by the service. Not only
3844 must the digital signature be valid but the signatures of the
3845 DLL must be signed by the same entity as the service.
3847 * Implement new functions: cm_freelanceMountPointExists and
3848 cm_freelanceSymlinkExists. Use them along with other validity
3849 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3850 ensure that name collisions do not occur and that empty strings
3851 are not valid file names.
3853 A symlink may not have a name which would resolve to a valid
3854 cell name. Doing so would prevent access to the cell.
3856 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3857 re-initialization code. The reference counts of the fake root.afs
3858 volume scache object(s) would become invalid when the mount point
3859 or symlink lists were altered.
3861 * Add registry entries to provide mappings from the afsdsbmt.ini
3862 to the new locations for applications which count on the use
3863 of the old Profile file APIs. These apps are likely to fail
3864 if the user does not have administrator privileges and the
3865 registry is locked down.
3867 * The afs_config.exe submounts dialog had two errors.
3868 First, attempts to remove entries failed because the registry
3869 key was being opened without KEY_WRITE privileges.
3870 Second, when editing a submount entry, changing the name
3871 would add a new key and leave the original one in place.
3872 Now the original submount will be removed if its name is
3875 * In recent months there have been several incidents in which
3876 users have experienced problems starting or accessing
3877 afsd_service.exe and after significant effort has been spent
3878 it has turned out that they have two versions of AFS on the
3879 machine or an inconsistent set of DLLs.
3881 Code has now been added to afsd_service.exe which will walk
3882 the list of modules loaded by afsd_service.exe and validate
3883 that the version of the AFS DLLs matches the version of the
3884 afsd_service.exe executable. If they do not match the service
3887 * When Freelance mode is enabled and there is no registry
3888 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
3889 will attempt to import the afs_freelance.ini file contents.
3890 If the file does not exist, it was creating a dummy file
3891 with a r/o and r/w entry for the default cell and then
3892 importing those values.
3894 This process has been changed. The temporary file is no
3895 longer created. Also, both the OpenAFS Client install
3896 directory as well as %WINDIR% are checked for previous
3897 afs_freelance.ini files.
3899 * Added support for VL_GetEntryByNameN(). Still need to add
3900 support for VL_GetEntryByNameU() for multi-homed support.
3902 * Fix a deadlock situation in afscreds.exe when canceling an
3903 auto-generated Obtain Tokens dialog
3906 * Fix the locking of objects during Directory Searches in the
3907 SMB/CIFS server. The failure to properly lock the reference
3908 counts was resulting in the premature freeing of smb_dirSearch_t
3909 objects while they were still in use by the SMB/CIFS client.
3910 This does not solve the "Invalid Handle" problem.
3912 * Fix Find Cell By Name pioctl call to return a valid cell
3913 name for the Freelance fake root.afs volume.
3914 "Freelance.Local.Root".
3916 * Fix the Explorer Shell Extension Symlinks->Add operation.
3917 The dialog template was missing and the link destination
3918 string was too short.
3920 * Add support for symlinks to Freelance root.afs volume
3921 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
3922 <number> = "<linkname>:<relative-path>."
3923 Use symlink.exe to create, list, or remove
3925 * Remove the fallback to the use of KFW's KRB4 library when obtaining
3926 tokens. We never obtain KRB4 tickets.
3928 * Fix AFS Client Configuration Control Panel to support new SysName
3931 * Fix a bug in afsd_service.exe which could result in the SysName
3932 not being read from the registry.
3935 * Add code to block the issuance of AFS tokens by aklog.exe or
3936 afscreds.exe when the Kerberos 5 principal name contains a dot.
3938 * Modify the IsAdmin() function to always treat the local SYSTEM
3939 account as an AFS client administrator. Affects fs.exe and
3942 * Modify the internal handling of Quota Exceeded errors
3944 * Upgrade all reference count fields in the Windows cache manager
3945 and the osi library to use unsigned long instead of signed short.
3946 A similar fix has been applied to the afs rpc (rx) library.
3948 * fix the Windows cache manager to prevent it from replacing the
3949 rx_connection object associated with the cm_conn_t object on each
3950 and every operation if "fs crypt" was set. This explains the
3951 dramatic performance difference when crypt is used vs clear.
3952 The problem: 'cryptall', a boolean flag indicating whether or not
3953 "fs crypt" is set, was being compared to the rx_connection
3954 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
3955 1 != 2 and therefore the rx_connection was always destroyed
3956 and replaced on each and every operation.
3958 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
3959 It is not safe for the cm_conn_t object to not be locked because
3960 rx_DestroyConnection might be called from another thread if:
3961 - the user's tokens have changed or expired
3962 - the crypt mode has changed
3964 This fix appears to have also taken care of the problems associated
3965 with Overlapped Writes resulting in Delayed Write errors.
3967 * fix NSIS installer's AdminGroup.exe to properly create and
3968 remove groups when given -create or -remove. The string comparison
3971 * fs sysname now accepts a list of sysname values
3973 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
3974 DWORD which when set to a non-zero value will cause error message
3975 text to be output to stderr from the pioctl() routine. Useful in
3976 debugging failures of fs.exe, tokens.exe, etc.
3978 * added a test to the power management code to only perform a
3979 flush operation if there is at least one network adapter which
3980 is not a loopback adapter.
3982 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
3983 "EnableKFW". This value will not be read if the key
3984 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
3985 value under that key does not.
3987 * provide mechanisms to force the use of krb524d for Kerberos 5
3988 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
3989 there is a new registry value "Use524" and for aklog.exe a new
3990 command line parameter "-m".
3992 * Fix the pattern matching algorithm to properly match patterns
3995 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
3996 which is used by both the original function and the new
3997 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
3998 creation of HardLinks in addition to Renaming. smb_Link()
3999 is a new function which creates HardLinks via cm_Link().
4000 cm_Link() is a new vnodeops function which creates links
4003 smb_ReceiveNTRename() does not support the File Copy and
4004 Move Cluster Information operations described in its interface.
4005 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
4007 * When opening files via symlinks, we should follow the symlinks
4008 until we reach the actual file stat cache entry. The stat cache
4009 entry of the file should then be stored in the FID instead of
4010 stat scache entry of the symlink.
4012 * return bad operation errors for all unimplemented functions
4013 even if we do not know the functions exist.
4015 * Log bad packets and unknown operation packets to the trace log
4017 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
4020 * Update list of known CIFS operations to include all those listed
4021 in CIFS-TR-1p00_FINAL.pdf.
4023 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
4024 to support the REG_EXPAND_SZ type.
4027 * A new Windows authorization group "AFS Client Admins" is now
4028 created and populated with the members of the "Administrators"
4029 group. The group is used to determine which accounts on the
4030 machine may be used to modify the AFS Client Configuration via
4031 the UI and command line tools. afs_config.exe, fs.exe,
4033 * Modify the WinLogon Logoff Event Handler to query NT4 domain
4034 controllers for the remote profile path if Active Directory
4035 services are not available.
4037 * Fix aklog.exe to not add the AFS ID to the username
4039 * PTS registration of new users to foreign cells has been added to
4042 * The cm_Daemon thread is used to perform checks for
4043 down servers, up servers, volumes, callback expirations,
4044 lock maintenance and token expiration. Due to a gaff in
4045 larger integer division the thread never performed any
4046 work. Instead the current time computation would always
4047 be less then the trigger times. This had an adverse affect
4048 on the client's ability to maintain communication with servers,
4049 keep volumes up to date, and flush user tokens and acls
4050 when they have expired. This was broken when the 1.3 branch
4051 was modified to support VC7 which no longer included
4054 * An initialization problem with the Freelance code was
4055 detected while fixing the callbackRequest. The cm_rootSCachep
4056 object is obtained during afsd_InitDaemons() but the callback
4057 information is incomplete. The callback information will not
4058 be obtained until cm_MergeStatus is called from within
4059 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
4060 test for the conditions under which the callback information
4063 * Reports have been filed indicating that callbacks were
4064 being lost. An examination of the code indicated that the
4065 cm_server_t objects were not being properly reference
4066 counted by the cm_scache_t and cm_callbackRequest_t objects.
4067 In particular, the cm_server_t objects may have been freed
4068 from beneath the cm_conn_t objects.
4070 All of the reference counting is now done via the functions:
4075 this improves the ability to track the referrals.
4077 Each cm_BeginCallbackGranting Call now allocates a reference
4078 to the cm_server_t. The cm_EndCallbackGrantingCall either
4079 frees the reference or transfers it to the cm_scache_t
4080 cbServerp field. These are then appropriately tracked
4081 through the cm_Analyze call.
4083 * Ensure that the dnlc hash table is the same size as the
4084 dir name hash table (as per original author's note).
4085 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
4086 for compatibility with 64-bit systems.
4088 * fix smb_ApplyV3DirListPatches to properly apply the hidden
4089 attribute to dotfiles when the infoLevel < 0x101 and
4090 cm_SyncOp has failed.
4092 * Fix the Freelance registry initialization code. There
4093 was a possibility that some systems could end up with
4094 garbage in the registry during a clean install.
4097 * file and directory names beginning with "." will now be given the
4098 hidden attribute when the volume access is anonymous. this matches
4099 the behavior when the volume access is via an authenticated user.
4101 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
4102 key. When a change occurs mark the root.afs data as invalid and
4103 for it to be reloaded on the next access. This allows administrators
4104 to modify the mount point list without restarting the service.
4106 The freelance client used to provide a fake modification time for
4107 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
4108 Added code to extract the last modification time of the Freelance
4109 registry key and use that instead. The time now represents the
4110 most recent mount point change.
4112 * PTS registration of new users to foreign cells has been added to
4115 * Additional Cache Control and Credential Manager options have been
4116 added to the WiX installer. See deployment guide for details.
4118 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4120 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4123 * The loopback installation code may have had a problem updating the
4124 %ETC%\HOSTS file which could have resulted in a premature failure.
4125 Work around code has been added for the case where the file cannot
4128 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4129 because Windows sends 64K blocks when using overlapped writes.
4131 * The default number of server threads was increased from 4 to 25 to
4132 better handle overlapped writes.
4134 * The "AfscredsShortcutParams" registry value was not being properly
4135 loaded by afscreds.exe. Therefore, the default value was always being
4136 used instead of the value set by the installer.
4138 * Windows XP provides downgrade attack detection to prevent an attacker
4139 from being able to force the use of NTLM simply by disrupting
4140 communication with the KDC. This attack cannot exist between the
4141 Windows CIFS client and the AFS Client Service. Therefore, when a
4142 downgrade has been detected the afs pioctl library will force the
4143 establishment of a new CIFS connection using NTLM.
4145 * A locking error was discovered surrounding all references to volume
4146 server lists within the cm_cell.c source file.
4148 * The logged into Windows username was incorrect on Terminal Server
4151 * A new registry value "NonPersistentCaching" was added to the service
4152 parameters key. When set to a non-zero value, the afs cache is stored
4153 in the Windows paging file. There are two limitations to choosing
4155 1. when persistent caching is implemented it won't work with
4156 this flag set since there will be nothing to persist.
4157 2. with this flag set the initial paging allocation cannot be
4158 changed while the service is running
4160 * An initialization bug was discovered in aklog.exe which affected users
4161 who have a domain name for their afs servers which could not be mapped
4165 * afs_config.exe now validates cell names against DNS in addition
4166 to the CellServDB file.
4168 * In order to allow the freelance client to connect to a volume with ID
4169 equal to 1 on the default cell we changed the fake root.afs volume ID
4170 once again. This time we choose 0xFFFFFFFF. In addition, we change
4171 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4172 It will now be impossible for a volume ID to match that of another
4173 cell unless the client is connected to 0xFFFFFFFD cells. That should
4174 be enough room for growth.
4176 * Fix "fs mkmount" command to work with UNC paths and when
4177 started from non-AFS drives. It is now possible to create a mount
4178 point in the freelance fake root.afs volume with the command
4180 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4184 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4185 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4187 * The algorithm used to re-attempt access to the servers associated with
4188 a volume has been altered to properly address the case in which all
4189 servers have been marked down. The previous algorithm did not reset
4190 the server's down flags so the servers were never actually retried.
4191 This caused a problem with active volumes if the network connectivity
4192 was lost as could be the case with a network cable removal, wireless
4193 drop, or laptop hibernation. With the fix volume access is restored
4194 almost instantenously when network connectivity becomes available.
4196 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4197 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4198 listing of AFS submounts and freelance mount points. Support for
4199 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4200 significantly enhances the behavior of AFS volumes within the Explorer
4201 Shell. For instance, "AFS" now shows up as server in the Explorer
4202 with each submount or freelance mount point visible as a share.
4203 The right click menu in each folder now works with full functionality
4204 on a consistent basis.
4206 * The network provider can be configured to have different behavior
4207 depending on the domain that the user logs into. These settings are
4208 only relevant when using integrated login. A domain refers to an
4209 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4210 local machine (i.e. local account logins). The domain name that is
4211 used for selecting the domain would be the domain that is passed into
4212 the NPLogonNotify function of the network provider. (see registry.txt
4215 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4216 "Authentication Cell" which may be used to specify a default
4217 authentication cell for afscreds.exe which is different from
4218 the default cell for the AFS Client Service daemon.
4220 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4221 afslogon.dll moved to %WINDIR%\System32\.
4222 New registry entries added to register the dll for Winlogon events.
4224 The logoff event will now force a call to ktc_ForgetAllTokens()
4225 using the context of the user being logged off as long as the
4226 user's profile is not loaded from within AFS. If the profile
4227 was loaded from AFS we can't release the tokens since the Logoff
4228 event is triggered prior to the profile being written back to
4229 the its source location. This is now performed in an XP SP2
4232 * Windows XP SP2 Internet Connection Firewall interoperability
4235 * The %WINDIR%\afsdsbmt.ini contains four sections:
4236 Submounts, Drive Mappings, Active Maps and CSC Policies.
4237 The Submounts and CSC policies are now stored in the registry under
4238 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4239 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4240 The Drive Mappings and Active Maps are stored in the registry under
4241 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4242 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4244 There is no automatic migration of this data as it would be impossible
4245 to consistently migrate data to user profiles which may not be active
4246 when the machine is updated.
4248 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4249 fake root.afs volume. For the same reasons as for the cellservdb file,
4250 this information should not be in %WINDIR%. This information is now
4251 kept under the registry key
4252 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4254 The data from the afs_freelance.ini file will be automatically
4255 migrated to the registry on first execution of afsd_service.exe
4257 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4258 troublesome for several reasons. One, it is confusing for those who
4259 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4260 Two, this file is not a Windows Profile formatted file. Three,
4261 applications should not be reading or writing to %WINDIR%. It causes
4262 problems for Windows Terminal Server.
4264 The new location of CellServDB will be the OpenAFS Client install
4265 directory which is by default C:\Program Files\OpenAFS\Client and can
4266 be determined by querying the registry for
4267 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4269 The existing afsdcell.ini will be migrated by the NSIS installer.
4270 The Wix installer must still be updated to do the same.
4272 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4273 High Security mode; and to add Terminal Services compatibility registry
4274 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4275 configuration files in %WINDIR%.
4277 * Add support for authenticated SMB connections. This will remove
4278 the need for high security mode in most situations. Both NTLM
4279 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4280 only NTLM can be used even though Kerberos is now supported. The
4281 reason is that it is not possible to construct a service principal
4282 which is unique to each individual machine.
4284 SMB Extended Auth does not work on XP SP2 unless one of two registry
4285 modifications are made:
4287 (1) To disable the check for matching host names on loopback connections
4288 set this key. This does not require a reboot:
4290 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4291 "DisableLoopbackCheck"=dword:00000001
4293 (2) To add the AFS SMB/CIFS service name to an approved list. This
4294 does require a reboot:
4296 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4297 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4299 afsd_service.exe will automatically add the current Netbios Name
4300 to the BackConnectionHostNames list and then temporarily disable
4301 the loopback check for one cycle of startup/shutdown of the service.
4302 We assume most folks do not start/stop without a reboot so this
4303 will be adequate in most cases.
4305 * Fix security hole in afslogon.dll which allowed passwords to be
4306 sent in clear text to the KDC in a misformed principal name.
4308 * Fix cm_GetCell() to properly handle expired dns entries
4311 * If Freelance mode is active and the afs_freelance.ini
4312 file does not exist, do not create an empty file.
4313 Instead create a file containing ro and rw mountpoints
4314 to the default cell using the standard conventions.
4316 * Modify the Freelance support to handle the ability
4317 to create rw mount points in the fake root.afs volume.
4319 * Changed the RPC mechanism used for token setting from
4320 named pipes to local. Use of named pipes can be restored
4321 by setting the environment variable AFS_RPC_PROTSEQ to
4324 Named pipes were required when a Windows 9x system was
4325 using a NT system in gateway mode which is incompatible
4326 with our use of local loopback adapters.
4328 * In afscreds.exe, if a username of the form user@REALM is
4329 specified and no password is specified, do not perform a
4330 kinit operation. Only perform the aklog functionality.
4332 * Add a new registry value which allows the number of processors
4333 on which afsd_service.exe executes to be restricted. Valid
4334 values are 1..numOfProcessors
4336 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4340 * A second MSI based installer option is now available.
4342 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4343 request tickets for user/instance@REALM instead of just user@REALM
4345 * Modify the Power Management Notify routine to wait for the Hard Dead
4346 timeout period instead of a fixed 19 seconds. With the longer timeout
4347 periods Hibernation and Standby could never succeed when network
4348 connectivity is not available.
4350 * The following fs.exe commands are now restricted to Administrator:
4351 - checkservers with a non-zero timer value
4354 - sysname with a new sysname list
4363 setting the default sysname for a machine should be done via the
4364 registry and not via "fs sysname".
4366 * NSIS installer adds options to install Debugging Symbols
4367 and the Microsoft Loopback Adapter; the user is now also
4368 given the ability to select the afscreds.exe startup options.
4370 * Build system modified to generate symbols for FREE (aka RELEASE)
4371 builds as well as CHECKED (aka DEBUG) builds
4373 * Sites which have a volume ID of 0x20000001 assigned to their
4374 root.cell volumes have been experiencing problems with accessing
4375 the root.cell volume of their cell when Freelance mode has been
4376 active. This was because 0x20000001 was assigned to the fake
4377 root.afs volume created by freelance. The fake volume id is
4378 now set to 0x00000001 to prevent conflicts.
4380 * The timeout logic in the AFS Client Service has been wrong
4381 for sometime. It is based on two different assumptions.
4382 First, the SMB client timeout is a fix value as was the case
4383 with OS/2 Lan Manager. This assumption is incorrect. The
4384 SMB timeout in Windows is a dynamic value computed based upon
4385 a fixed minimum timeout to which is added time based upon the
4386 size of the request and the performance characteristics of
4387 the connection. Second, it is the responsibility of the
4388 SMB Server to enforce the timeout requirements of the client.
4389 This is untrue. The SMB Server cannot be expected to know
4390 the requirements of the client. More importantly, if the
4391 SMB server uses the SMB client timeout as a value to restrict
4392 its behavior as an RX client, the performance characteristics
4393 of the local SMB session would be used to prematurely terminate
4394 WAN connections with significantly different performance
4397 The timeout logic has therefore been modified in the following
4399 . the Lan Manager Workstation (SMB) Session Timeout is used only
4400 as a basis for configuring the Connection Dead Timeout
4401 and Hard Dead Timeout values. The Connection Dead Timeout
4402 must be at least 15 seconds longer than the SMB Timeout
4403 and the Hard Dead Timeout must be at least double the
4404 Connection Dead Timeout.
4405 . New registry entries have been added to allow the Connection
4406 Dead Timeout and Hard Dead Timeout values independent of the
4407 Lan Manager Workstation Session Timeout
4408 . The test to enforce the SMB Client Timeout has been removed.
4410 One of the side-effects of removing the enforcement of the SMB
4411 Client Timeout is that regardless of whether or not the SMB client
4412 is available to receive the response (and how would the SMB server
4413 know) the RX protocol response can be used to update the AFS
4414 Client Service state for ready access by future SMB client
4417 This should be the end of the "Server paused or restarting messages"
4419 * Add support for arbitrary UNC paths to the pioctl() support.
4420 This enables the fs commands as well as the AFS Shell Extension
4421 to work correctly when UNC paths are being used.
4423 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4424 DNS if the cell configuration cannot be determined via CellServDB
4426 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4427 are ever set more than once at a time
4429 * Fix the management of lists of cm_cell_t structures when using
4430 DNS to lookup cell information. The previous code would fail to
4431 reuse the same cellID for a cell if DNS was used more than once
4432 for a given cell name. When the ttl expired, a single cm_cell_t
4433 could be inserted into the cm_allCellsp list more than once
4434 producing a loop. In addition, the vlServerp list belonging to
4435 the cell was not freed resulting in improper refCounting of the
4438 * Add DNS support to cm_IoctlNewCell() which previous only examined
4441 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4442 to properly garbage collect cm_server_t objects
4444 * Add numVCs variable to smb.c to track the number of smb_vc_t
4445 objects created and use it to initialize the vcID field which
4446 previously was set to 0 in all objects resulting in FindByID
4449 * Fixed DNS lookups to work consistently throughout the OpenAFS
4450 product instead of just from within the afsd_service.exe
4452 * Added a runtime check to ensure that AFS Client Service SMB
4453 Server is accepting connections before attempting to mount
4456 * Read IP addresses for volume servers out of the CellServDB
4457 file if gethostbyname() on the hostname fails.
4459 * Fix getcellconfig() to populate both the Hostnames as well
4460 as IP addresses when loading cell data via DNS
4462 * Increase the Connection Dead Time to 50 from 20 seconds
4463 Increase the Hard Dead Time to 120 from 40 seconds
4464 (matches the Unix values)
4466 * Fixed an assertion validating the number of allocated NCBs
4468 * Fixed the build environment to consistently build for
4469 Windows 2000 and above. (APPVER = 5.0)
4471 * Fixed rx_debug to properly validate the receipt in incoming
4472 data with select() and recvfrom(). Do not copy data out of
4473 the socket buffer unless success is indicated.
4476 * afsd_service.exe will now display a message box to the
4477 desktop when it terminates due to an IP Address Change.
4479 * installer no longer deletes AFS Server configuration data
4482 * installer generates a warning dialog if the RPC service
4483 is not properly configured
4485 * installer compressed with lzma instead of bzip2
4487 * afsd_service.exe shutdown crash solved once and for all
4489 * reference counting of smb_vc_t data structures improved
4491 * name space collision of smb_fid_t event objects corrected
4493 * the output of "fs memdump" is now written to
4494 %WINDIR%\TEMP\afsd_alloc.log
4496 * the file TaAfsApp_1033.dll is now properly installed allowing
4497 the User Manager to start
4499 * a new algorithm is used for computing filename pattern matches
4501 * afscreds.exe now accepts user names containing instance
4504 * Fix the Directory Name Lookup Cache to be case-sensitive.
4505 This is crucial in environments in which a Windows client
4506 is accessing a directory with more than one filename that
4507 differs only by case. If the directory contains "FOO"
4508 and "Foo". You want "DEL Foo" to delete the correct one.
4509 We still have a problem in that "DEL foo" will delete a
4510 random filename. This will be addressed in a future release.
4512 * Fix afscreds.exe -M option (renewMaps) to work when High
4513 Security mode is off. Also, remember to disable the ActiveMap
4514 flag in afsdsbmt.ini when a drive mapping is removed.
4516 * Updates to NSIS installer script. AFS Server configuration
4517 data will not be destroyed on un-install or re-install.
4518 Use a better compression algorithm.
4520 * afslogon.dll now uses KFW to obtain tokens when available
4522 * afslogon.dll when given an all uppercase username will
4523 attempt to authenticate with both the uppercase name
4524 and an all lowercase variation
4526 * DST modification removed. The fix appears to make things
4527 worse after a reboot of the machine.
4529 * fs.exe: added "cscpolicy" which is used to
4530 change client side caching policy for AFS shares
4532 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4533 [-documents] [-disable] [-help]
4535 * Several uninitialized variables have been initialized
4537 * It is now possible to obtain tokens using cross realm
4538 Kerberos within afscreds.exe:
4540 user: jaltman@ATHENA.MIT.EDU
4542 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4543 will will in turn be used to obtain afs@DEMENTIA.ORG.
4544 The resulting token will be stored with the display name
4545 jaltman@ATHENA.MIT.EDU@dementia.org
4547 * aklog.exe has been added to the client
4549 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4550 [[-p | -path] pathname]
4554 -d gives debugging information.
4555 krb_realm is the kerberos realm of a cell.
4556 pathname is the name of a directory to which you wish to authenticate.
4557 -noprdb means don't try to determine AFS ID.
4558 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4559 (default is Kerberos V)
4560 No commandline arguments means authenticate to the local cell.
4563 * All of the resource files have been restructured to adhere to
4564 a set of rules IBM implemented for loading string resources.
4565 These rules had either been forgotten or were not discovered
4566 by folks working on the OpenAFS sources. The end result was
4567 memory corruption. This is primary item which was preventing
4568 the AFS Server from working.
4570 * Increased the size of the maximum ticket size stored in a token
4571 from 344 bytes to 12,000. Increased the buffers used to convey
4572 messages between the pioctl() caller and the SMB Server from
4573 1000 bytes to 12,512. The code appeared to have been writing
4574 above the top of the stack by quite a few number of bytes.
4575 (The increased ticket size is necessary for the next item.)
4577 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4578 Instead the raw Kerberos 5 ticket is used in its entirety. This
4579 is extremely important as it allows us to use pure Kerberos 5 KDCs
4580 as the source of the AFS authentication. The use of up to 12,000 byte
4581 tickets will allow tickets produced by all versions of Microsoft
4582 Active Directory to be used.
4583 - create a user account.
4584 - designate it DES only
4586 - specify its UPN to be "afs@realm"
4587 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4589 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4590 is special when using long file names. (you must use "*.*" and not "*")
4591 Instead only enforce it when performing 8dot3 searches.
4593 * Fixed the DST problem with creation times being set one hour ahead
4595 * Fixed the problem when using \\afs\cell-alias. For example,
4596 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4597 for the alias name; instead simply expand the name. One of the
4598 symptoms of this problem was a loss of acquired tokens.
4600 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4601 of strings. (Only English strings provided.)
4603 * Fixed the installer to properly replace in use files.
4605 * Fixed the build system to cleanup generated component version files
4607 * The release build compiled with MSVC 6.0 compiler to avoid the
4608 afsd_service.exe shutdown crash. This does not solve the problem
4609 but simply avoids it for the time being.
4613 * fix afslogon.dll to not corrupt memory when High Security mode
4616 * fix afsd_service.exe to not attempt to restore the stack when
4617 an exception occurs. (not safe in multi-threaded programs)
4619 * fix uninstaller to properly remove the CRT and MFC DLLs
4621 * remove a Message Box from afscreds.exe when getcellconfig()
4622 fails on a kerberos realm which is not a cell
4624 The following is a list of changes to the OpenAFS for Window client
4627 * "fs setserverprefs" will leave afsd service deadlocked
4629 * "vos listaddrs" will core dump
4631 * installer sets the appropriate keys to support Integrated Logon
4633 * installer disables the "Find Lana by Name" functionality as it
4634 was causing headaches for many users
4636 * fix the intermittent crash of the power management thread when
4637 shutting down the AFS Client Service
4639 * optimizes the obtain drive mount list functionality which is
4640 executed every time the mount tab in afscreds.exe and afs_config.exe
4641 are refreshed. (this happens a lot)
4643 * fix the service shutdown logic. add the STOP_PENDING state
4644 and do not accept additional service events after we declare
4647 The following is a list of changes to the OpenAFS for Window client
4650 * flexelint was run against the source tree and hundreds (perhaps
4651 thousands) of corrections were applied to ensure prototypes
4652 were in use; types were used consistently; variables were
4653 initialized; unused variables were removed; etc.
4655 * A wide variety of instrumentation was added including the
4656 ability to produce a stack trace from within afsd_service.exe
4659 * Dynamic configuration of the RDRtimeout value based upon the
4660 LanMan Workstation Session Timeout
4662 * The mount root no longer needs to be called "/afs". This
4663 is now set by a registry value "MountRoot" within the key
4664 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4666 * The cell list is now only read out of afsdcell.ini when the
4667 file changes instead of each time a cell is resolved.
4669 * Thread synchronization was added to cm_server.c and ktc_nt.c
4671 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4672 calloc()/free(). The Global functions were needed on Windows 3.x
4673 but have caused a variety of problems on the Win32 platforms.
4674 Avoiding them is highly recommended by several Microsoft
4675 Knowledgebase articles
4677 * Support for Symbolic Links added to the AFS Shell Extension
4679 * Added a registry value "OverlayEnabled" to determine if
4680 Shell Extension Overlays should be enabled.
4681 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4683 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4684 separate trees for checked and free builds. Build system supports
4685 a custom directory src\WINNT\extra which can be used as a grafting
4686 location of organization specific additions to the build tree.
4688 * New installer built using NSIS 2.0.
4690 * Named all kernel objects in order to allow them to be monitored
4691 with tools such as SysInternals' ProcExp.exe.
4693 * Introduced new EventLog framework for AFSD
4695 * Introduced Power Management interface to AFSD for Standby and
4696 Hibernate modes to allow cache to be flushed prior to network
4699 * Utilize Win32 DNSQuery API instead of internal routines. This
4700 allows DNS SRV queries to be sent to all current domain name
4701 servers. Not just one specified in an INI file. DNS is now
4704 * "NetbiosName" registry value may be used to specify a fixed
4705 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4706 when the loopback adapter is in use. If you need to use the
4707 old notation with a loopback adapter installed specify a registry
4710 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4712 * Refactor all modules which depend on LAN Adapter and NetbiosName
4713 determination in a new library: lanahelper.lib. This allows for
4714 consistent behavior throughout the product.
4716 * Move the afsd.log and afsd_init.log files to the directory specified
4717 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4718 for services. Added the Date to the log entries.
4720 * New registry value "RxMaxMTU" used to limit the size of the RX
4721 packets sent by the AFS Client Service to the Server. In order
4722 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4723 size must be restricted to 1264 or smaller. The latest NSIS
4724 installer sets a value of 1260 by default.
4726 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4727 packets. This is not needed in order to work across the Cisco
4728 VPN but might be needed for other network environments. This
4729 value is not set by the NSIS installer.
4731 * New registry value "HideDotFiles" is used to apply the Hidden
4732 attribute to files whose names begin with a '.'. This value
4733 is set by the NSIS installer.
4735 * New registry value "MaxMpxRequests" allows the maximum number
4736 of multiplexed sessions to be configured at run time. This
4737 value is not set by the NSIS installer. The default value is
4740 * New registry value "MaxVCPerServer" allows the maxmimum number
4741 of VCs per server to be configured at run time. This value is
4742 not set by the NSIS installer. The default value is 100.
4744 * New registry value "AllSubmount" allows the "all" submount to
4745 be disabled by setting its value to 0x00.
4747 * Allow cells names to be valid mount points
4748 \\<netbiosName>\<cellname>
4750 * Store the active state of drive mappings in order for afscreds.exe
4751 to restore them upon startup
4753 * Add exception handling to generate a Stack Trace to the afsd_init.log
4754 file if one happens to occur.
4756 * Add lots of logging to help detect the cause of invalid SMB packets
4758 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4759 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4760 renewal of expiring tokens as long as afscreds.exe is running.
4762 * New afscreds.exe command line options:
4764 -M = renew drive maps
4765 -N = ip address change detection
4768 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4769 determines whether or not MIT Kerberos for Windows should be used
4770 to obtain tokens via Kerberos 5 tickets.
4772 * New registry value "AfscredsShortcutParams" in
4773 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4774 determines the command line parameters to be specified when "fixing"
4775 the AFS Shortcut in the user's startup folder.
4777 * The "ShowTrayIcon" registry value has been moved from
4778 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4779 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4781 * The <cell name> registry values used to store the token expiration
4782 reminders have been moved from
4783 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4784 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4786 * Obtain the Logon User Name from the Explorer key when available
4788 * new text document doc\txt\winnotes\registry.txt lists all registry
4789 values used by OpenAFS (excluding the AFS Server)
4791 * BUG: rx_securityClass objects were not properly reference
4792 counted and were never freed.
4794 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4795 would be generated. The existence of a server does not imply
4796 that it is not down. If all of the servers for a cell are down
4797 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4800 * BUG: the directory name lookup cache failed to free the entries
4801 in the cache when the name cache entries cycled. The entries
4802 in the cache would become dereferenced without being freed.
4804 * BUG: fs setserverprefs could be executed without Administrator
4807 * BUG: the number of allocated NCB objects (100) exceeded the number
4808 which could actually be waited upon by the kernel (64). Any objects
4809 which were utilized above the limit could never have event completions
4812 * BUG: smb_username_t objects were not being reference counted and
4813 were not properly freed.
4815 * BUG: smb_tid_t objects could under unusual circumstances be freed
4816 before they were no longer referenced.
4818 * BUG: smb_fid_t object pointer were frequently used even when
4819 their value could be NULL. They were not properly released and
4820 therefore they were never freed.
4822 * BUG: smb_packet_t data structures were not completely initialized
4825 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4826 Denied" because that causes the Explorer Shell to try again until
4827 access is obtained. Instead return "Remote Resources" which allows
4828 the shell to move on and treat the error as transient.
4830 * BUG: when initializing the NCBreturns structure, separate Event objects
4831 were created for each NCB although a single Event object was supposed
4832 to be shared by all.
4834 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4837 * BUG: smb_tran2Packet_t objects were not being properly referenced
4840 * BUG: directory path creation did not handle the case of multiple
4841 directories requiring creation in one attempt
4843 * BUG: SMB requests which required an Extended Response were ignored.
4844 This prevented some files from being written to AFS volumes.
4846 * BUG: character strings were being freed even after they were
4847 inserted into in use data structures
4849 * BUG: inconsistent usernames were used when High Security mode was
4850 enabled. (there is still much to do in this area)
4852 * BUG: pioctl() calls which require out of band RPC operations were
4853 susceptible to race conditions when performed by multiple processes
4855 * BUG: memory allocation and deallocation crossed instances of the
4856 C Runtime Library producing memory leakage and corruption in
4857 afscreds and the client configurator.