3 * Avoid a potential Freelance deadlock during initial execution
4 of afsd_service.exe if the old ini file data has to be
7 * Three rx library corrections. (1) Idle data connection
8 processing could timeout if the send window filled and
9 took longer than the idle data timeout period for the
10 transmit window to re-open. (2) The transmit queue
11 could be emptied prematurely. A required check for the
12 queue being in use was forgotten. (3) The function that
13 is supposed to implement a wait for the transmit queue
14 to cease being busy failed to wait.
17 * Restore use of DNS AFSDB and SRV records which were
18 unintentionally disabled in 1.5.69
21 * Add a context menu to the NetIdMgr AFS Provider
24 * Prevent an empty directory Btree from being created and
25 marked as valid if cm_BPlusDirBuildTree fails.
27 * cm_BPlusEnumAlloc should not fail if the enumeration
28 contains zero entries.
30 * In cm_BPlusXXX functions, return ENOMEM if malloc() fails.
32 * Do not leak the cm_scache_t dirlock if cm_BPlusDirBuildTree
35 * fs examine should report owner and group ids as signed values.
37 * Protect buffers in smb_WriteData from simultaneous writes.
39 * Query network interface MTU values from the registry because
40 the IP Helper GetAdapterAddresses() fails to report a manually
43 * Improve performance by not dropping and reacquiring the rx_call
44 lock in rx_WriteProc* and rx_ReadProc*.
46 * When configuring the rx library, the network interface MTU
47 must be applied to both the maximum send size as well as the
48 maximum receive size as is done when rx_SetMaxMTU is called
51 There was no 1.5.67 release
54 * Convert afsauthent.dll to use the xdr_alloc and xdr_free
55 routines from afsrpc.dll in place of malloc and free.
56 This permits the two libraries to be built with different
57 C runtime libraries without crashing.
59 * Set the DOS Readonly attribute on a file/directory
60 whenever the unix mode combined with the mask 0200
61 is true. Previously there was a discrepency between
62 the mask used for testing for readonly behavior and
63 that used for setting the attribute.
65 * Disable AFSVolSync based .readonly "whole-volume callback"
66 support because the all file servers prior to 1.5.67
67 (and perhaps 1.4.12) do not properly assign a value to
68 the AFSVolSync structure in bulk status RPC responses.
70 * Improve the error output from aklog to output the value
71 from krb5 error_message() if the afs_com_err output
72 indicates an unknown value.
74 * Hold a lock on the cm_scache_t object when the smb
75 server is assigning a new clientModTime value.
77 * Permit custom build version numbers to be set
78 without modifying the src/config/NTMakefile.<platform>
81 * Add support for binary code signing with cross-signed
82 certificates and arbitrary versions of signtool.exe.
84 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and
85 do not permit them to be exposed to the smb redirector.
87 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT
88 to avoid confusion within the smb redirector.
90 * Fix the byte order assigned to port numbers associated
91 with AFSDB record lookups. They must be network byte
92 order not host byte order.
94 * Add dynamic server ranking based on RPC round trip
98 * Official support for Windows 7 and Server 2008 R2
100 * Improved Rx RTT computation following the lessons
103 * Prevent a file server bug (FetchData returning an
104 invalid length instead of zero) from causing an
105 "unexpected network error" when writing to files.
107 * Promote DNS SRV records as superior to DNS AFSDB
108 records. Support arbitrary port numbers for vldb
111 * Add "fs listacl -cmd" support.
113 * Add AFSVolSync based .readonly "whole-volume callback"
114 support. With this functionality, multiple objects from
115 a .readonly volume can have their status validated by
116 issuing a single RXAFS_FetchStatus RPC.
118 * Remove drive mapping functionality and service start/stop
121 * Remove drive mapping functionality from afs_config.exe.
123 * Adjust SMB error return codes to avoid returning
124 STATUS_TIMEOUT which results in the SMB redirector
127 * Network Identity Manager OpenAFS Provider now provides its
128 own "AFS lock" notification icon to report the status of
129 "have tokens, have no tokens, service not started, service
130 started but inaccessible". Hovering over the icon lists the
131 cells for which tokens exist (if any) and the OpenAFS version
132 number. Double-clicking executes the Network Identity Manager
135 * When merging the current status for an AFS object, ensure
136 that the volume status for that object is set to 'vl_online'.
138 * Permit cm_Analyze to request a retry even when the cm_req_t is
139 flagged CM_REQ_NORETRY if the RPC failed due to RX_CALL_DEAD.
140 The retry will force the use of a new connection.
142 * Add digital signatures to resource only dlls as required by
145 * Prevent pioctl calls from retrying indefinitely when a sharing
146 violation error occurs.
148 * Add {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab"
149 to restore access to drive mapping functionality in afscreds.exe
152 * Permit Rx minimum peer timeout, maximum receive window size and
153 maximum send window size to be configured via the registry.
155 * The MSI installer now sets the ARPINSTALLLOCATION as required by
158 * DNS lookups set the port numbers in network byte order within
161 * Replace use of the 16-bit compatibility API WinExec with
162 ShellExecuteEx in afs_config.exe and afscreds.exe. WinExec
163 is incompatible with Windows 7 / Server 2008 R2 UAC
167 * Short circuit background volume checks if a shutdown
168 or suspend request is received.
170 * Avoid contacting the file server in order to allocate
171 an empty buffer beyond the current length of the file.
173 * When a request for streams on a directory or mount
174 point object is received, do not offer a default stream.
175 This was affecting the saving of roaming profiles.
178 * Fix Windows 2000 compatibility. Do not call rand_s()
179 even if it is compatible with the C run time library
182 * Fix a data consistency error between the output of
183 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
184 workstation group name "AFS", and report server name as
185 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
187 * Enable executables to be run from \\AFS on Windows 7.
188 Return "Name not found" instead of "File not found" when
189 a directory or file name cannot be found.
191 * Prevent cache manager from marking file server "down" when
192 the data returned in response to either RXAFS_FetchData64
193 or RXAFS_StoreData64 is invalid.
195 * Add pioctl data validation to the AFS Explorer Shell extension.
198 * After 1.5.62 afslogon.dll would no longer crash but it
199 also would not load "domain" specific configurations.
202 * An uninitialized variable was present in the symlink
203 recursion detection loop.
205 * Reverse addition of unique per file GUID during
206 Create/Open responses. Inclusion of the GUID prevents
207 Cygwin applications (and others) from accessing //afs.
209 * Treat "filename::$DATA" as a synonym for "filename".
210 This addresses the recent report of roaming profiles
211 failing to save back to the file server.
213 * When given a choice between an error from rx_Write()
214 or RXAFS_EndStoreData(), use the rx_Write() error.
215 This ensures that over quota and access denied errors
218 * If an error occurs during RXAFS_StoreData operations,
219 apply that error to all of the buffers in the BIOD.
220 This ensures that failures will not be retried if they
221 will only fail again.
223 * Do not perform a DNS lookup for a cellname if the
224 cellname prefix is _._AFS_IOCTL_._. This will avoid
225 DNS lookups when the query is for
226 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
228 * Fix Freelance root.afs volume object callback processing.
229 Object synchronization and status merging bugs.
231 * Fix Explorer Shell "invalid parameter" bug introduced
234 * Log RXKAD errors to the trace log.
236 * Prevent simultaneous pioctl calls from multiple processes
237 from stepping on each other's data.
239 * Prevent simultaneous pioctl calls from crashing the
242 * Add data validation to pioctl output processing in
243 ktc_GetToken and ktc_ListTokens.
245 * Add data validation to pioctl output processing in
246 fs.exe and symlink.exe.
249 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
250 No longer will browsing \\AFS produce truncated share names.
252 * Registry specified server preferences did not result in
253 the CM_SERVERFLAG_PREF_SET flag being applied to the
256 * Avoid unnecessary DNS lookups of share names as cells
257 when it is known that the name cannot be a cell name.
258 Any name that does not contain a dot is skipped.
260 * When processing Dfs Referral requests, do not return an
261 error if the server is down or busy; if the volume is
262 offline; the cell vldb cannot be reached; etc.
263 These paths are still \\AFS paths and so the client
264 should not be sent to search elsewhere for resolution.
266 * CRITICAL: Prevent the cm_Daemon thread from terminating when
267 the machine enters the suspend state. IF the cm_Daemon thread
268 dies, there is nothing to execute down server checks.
270 * Prevent use of smb_StartedLock before initialization which
271 can result in a panic in the lock package.
273 * Improve error checking in afslogon.dll in order to prevent
274 NULL pointer dereferences within GetLogonDomainOptions()
275 if the Lsa operations fail.
277 * Add support to the build system to automatically update
278 a Microsoft Symbol Store.
280 * Do not access the cm_conn_t after executing the cm_Analyze
281 loops. Doing so can result in an access to memory that has
284 * CRITICAL: If a StoreData request offset is not aligned on the
285 buffer module blockSize, (offset % blockSize) bytes of dirty
286 data will not be written to the file server.
288 * CRITICAL: If a StoreData64 request is sent to a file server
289 that does not support large files (>2GB) and the client is
290 unaware that the server has no large file support, the first
291 chunksize worth of data will not be written to the file server.
292 This bug was introduced in 1.5.3. It affects all IBM AFS file
293 servers and all OpenAFS file servers older than 1.4.0.
296 * If a file server becomes inaccessible while the cache manager has
297 dirty buffers to write, the afsd_service buf_IncrSync thread can
298 attempt to use 100% of the cpu.
300 * Fix the locking that protects the cm_cell_t lists
301 (all, name, id, free).
303 * Remove cm_cell_t objects from the name and id hash tables before
304 placing them on the free list.
306 * Fix "fs newcell" which was broken in 1.5.60.
308 * Do not attempt to synchronize dirty buffers if the associated
309 volume is known to be unavailable.
311 * Log to the Windows Event Log dirty buffers that cannot be written
312 during the shutdown process.
314 * Modify behavior of a Freelance mountpoint target that does not
315 specify a cell. Instead of assuming the target volume is in the
316 Freelance.Local cell, use the workstation "Cell" specified in the
317 registry. A mountpoint target of "#root.cell." will now mean the
318 root.cell volume in the workstation cell for the current session.
319 If the workstation cell changes from "athena.mit.edu" to
320 "andrew.cmu.edu", the referenced volume will also change without
321 requiring that the mount point targets be altered.
323 * Avoid false infinite loop errors when validating the cm_cell_t
324 free list during startup.
326 * Fix two more locations where xdr_free() should be used instead of
327 free() to permit a checked built of afsd_service.exe to work with
328 release builds of afsrpc.dll.
330 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
331 to permit the server Uuid to be used to lookup the server object
332 and from that determine the cell. This permits callbacks that are
333 received from alternate addresses to be processed with a known server
334 object. Previously a request from an unknown server would clear all
335 callbacks from all cells.
337 * Fix a bug that prevented optimal performance when using a non-zero
338 value for 'daemonCheckVolCBInterval'. As a reminder, when
339 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
340 volume callbacks are automatically renewed 90 minutes before their
343 * Fix automatic ranking of vldb servers whose values are obtained from
346 * Add logging for RX CALL DEAD errors which are generated if the remote
349 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
350 or the call is to a vldb server.
352 * Add registry based cell search functionality to NetIdMgr,
353 afs_config.exe, and klog.exe.
355 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
356 lookups on the host names in the CellServDB instead of using the
357 specified IP addresses. This provides aklog, pts, vos, etc. the same
358 CellServDB behavior that the Windows Cache Manager uses.
360 * Completely a behind the scenes change. The snprintf() routines have
361 been completely rewritten.
363 * When updating the stat cache entry callback of a .readonly object
364 from the volume group object, update the file server reference to
365 ensure it matches the most update to date callback.
367 * Add proper support for processing callbacks from multi-homed file
368 servers. Instead of comparing servers by cm_server_t pointer,
369 compare them by UUID when the UUID is known.
371 * During a shutdown short circuit the offline volume check daemon
374 * Properly track link counts during cm_Link and cm_Unlink operations.
376 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
377 preference to an error code reported by rx_EndCall.
379 * Add "PerFileAccessCheck" registry value to permit testing against
380 experimental file servers that include per-file acl support. This
381 value is intentionally undocumented. It is not to be used by
382 production environment deployments.
384 * Log a warning to trace log if a volume group is marked as a DFS
385 File Set. The Windows OpenAFS client does not include DFS support.
387 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
388 provider from obtaining tokens when referrals are in play.
390 * Add "fs chown" and "fs chgrp" commands to permit the owner and
391 group of objects stored in AFS to be set from Windows.
393 * Avoid performing background daemon operations when the machine is
394 going into suspend mode.
396 * Perform offline volume checks in most recently used order.
398 * Prevent crash when a data version for a cache object goes backwards.
400 * fs setquota, fs setcachesize, vos setfields, and vos create now
401 accept human readable orders of magnitude. (K, M, G)
403 * fs listquota fixed to permit large quota sizes to be displayed.
405 * rx packet resend and data packets sent counts were incorrect.
407 * Multi-thread safe library versions are now being generated and used.
408 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
410 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
411 values are now available on XP through Windows 7. Add functionality
412 to autodetect if such support is present on the machine. If so,
413 configure it if necessary and dynamically adjust the AFS Rx timeout
416 * Correct another race condition in the Rx library that could result
417 in an unexpected panic while freeing the Rx call iovq.
420 * A fix to the pioctl library to support drive substitution
421 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
423 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
424 that corrects a deadlock in the smb redirector and also adds new
425 functionality that permits the AFS SMB server to be given a longer
426 timeout than is normally the case. New functionality has been added
427 to configure these additional LanmanWorkstation\Parameter values.
428 (This functionality has been backported to XP SP3 and is scheduled
429 to be released on June 5th.)
431 * The BackConnectionHostNames registry value configuration was broken
432 when dynamic re-establishment of Netbios Name registrations was added.
435 * Hidden vos commands are revealed.
437 * If the "DisableLoopbackCheck" registry value is set, do not unset
438 it during the same service session.
440 * Reorganize code that prevents multiple Store operations to the same
443 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
444 permit broken symlinks to be treated as being in AFS.
446 * Fix vos commands that output 64-bit integer values.
448 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
449 Permits building cygwin applications against OpenAFS libraries.
451 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
452 as Windows HTML Help (.CHM) files.
454 * NSIS installer does a much better job of cleaning up files left
455 over from previous installs.
457 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
458 or "fs flushall" and on-going directory operations that can result in
459 afsd_service.exe crashing.
461 * Add support for DNS SRV records in place of AFSDB records.Â
462 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
463 Weight is currently ignored.
465 * Add a method of specifying Client CellServDB information within the
466 registry that can be used to either override the CellServDB file or
467 force the use of DNS lookups for a given cell.
469 The registry schema is as follows:
471 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
472 "LinkedCell" REG_SZ "[cellname]"
473 "Description" REG_SZ "[comment]"
474 "ForceDNS" DWORD {0,1}
476 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
477 "HostName" REG_SZ "[hostname]" Default: [servername]
478 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
479 "IPv6Address" REG_SZ "[address]" <future>
480 "Comment" REG_SZ "[comment]"
481 "Rank" DWORD "0..65535" Default: 0
482 "Clone" DWORD "{0,1}" <future: server only>
483 "vlserver" DWORD "7003" <future>
484 "ptserver" DWORD ... <future>
486 ForceDNS is implied non-zero if there are no [servername]
487 keys under the [cellname] key. Otherwise, ForceDNS is zero.
488 If [servername] keys are specified and none of them evaluate
489 to a valid server configuration, the return code is success.
490 This prevents failover to the CellServDB file or DNS.
492 Only one of "HostName" or "IPv4Address" is required. "HostName"
493 is optional if [servername] is the host name.
495 * Extend registry based CellServDB functionality to
496 afsconf_GetCellInfo() interface used by aklog and many
497 other command line utilities.
499 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
500 as it is now used by both client and server components.
503 * PriorityClass of afsd_service.exe process raised to
504 "High" to match the priority of the system services
505 that are waiting for afsd_service.exe to respond.
506 Configurable using [TransarcAFSDaemon\Parameters]
507 DWORD "PriorityClass" value. (use values from Win32
510 * Change error reporting of CM_ERROR_BADSHARENAME to
511 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
512 The error we were sending would cause the smb
513 redirector to drop the connection.
515 * Change the error reporting of CM_ERROR_ALLOFFLINE
516 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
517 Previous we would return Bad Network Path which
518 would cause the smb redirector to cancel all file
519 handles on the same "share".
521 * A debug version of afsd_service.exe can now be used
522 with a release version of afsrpc.dll (or vice versa).
523 afsrpc.dll now implements and exports xdr_free().
525 * Removed all unprotected OutputDebugString() calls as
526 those calls can block for up to 10 seconds if there
527 is contention for the globally shared memory or if
528 the debugger fails to read data in a timely manner.
530 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
531 supported for builds. The production releases still
535 * Discovered that XP/2003 supports a variant on recvmsg
536 called WSARecvMsg. The input parameters are mostly
537 compatible with the Posix recvmsg. This will save
538 multiple memcpy() calls per received packet.
540 * Discovered that Vista/2008 supports a variant on
541 sendmsg called WSASendMsg. The input parameters
542 are mostly compatible with the Posix sendmsg. This
543 will save multiple memcpy() calls per sent packet.
545 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
546 and utilizing rx_ReadProc32 from within the cache manager
547 as part of the FetchData operations. rx_ReadProc32 and
548 rx_WriteProc32 are optimized versions of rx_ReadProc and
549 rx_WriteProc for use when reading 32-bit integer values.
551 * The Network Identity Manager AFS provider now reports
552 progress events. These provide a better user experience
553 with Network Identity Manager 2.0.
555 * [RT 124293] A race condition exists which permits the
556 scp field of the an smb_fid_t object to become invalid
557 after a request on the smb_fid_t has begun. Remove the
558 race by protecting all access to the scp field with the
559 smb_fid_t mx mutex and obtaining a local reference on
560 the cm_scache_t object for the length of the request.
562 * [RT 124276] If the vldb is out of sync with the contents
563 of the file servers, afsd_service will retry too many
564 times when a file server reports a volume as not being
565 present. Now if the list reported by the vldb is the
566 same as the previously seen list, then the retry is
569 * [RT 124276] Read-only volume failover was broken in
570 1.5.53 whenever accessing a volume results in VNOVOL or
573 * [RT 124276] Prior to 1.3.70 the volume server reference
574 list was not reference counted and would be prematurely
575 freed while in use. When reference counting was added
576 in 1.3.70 a bug was introduced that could result in
577 service reference list corruption.
579 * Add Windows Application Event Log warning messages for
580 "Client SMB MPX value too large" and
581 "Client SMB Buffer Size too small".
583 * Renaming of files across directory boundaries would result
584 in an invalid handle error when attempting to access the
585 files after the move.
587 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
588 when a smb file descriptor cannot be found. The whole point
589 of using a Path Info function is because an smb file
590 descriptor wasn't allocated.
592 * More edge cases in which dynamic addition of Freelance
593 root.afs entries would get the wrong FID or where the
594 root.afs directory would not be refreshed.
596 * Buffer overflow could occur if the workstation cell name was
597 longer than 64 characters. Crashes could occur in
598 afscreds.exe, afslogon.dll, and afsd_service.exe.
600 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
603 * Improve consistency of the bad data version value used for
606 * Log file server uuid values as part of the cm_server object
607 when available. Dump the cm_server object list in response
610 * Optimize the performance of resetting access control lists
611 when tokens are set or removed.
613 * Fix vos.exe so that it can work with very large volume
616 * Remove symlink recursion tests and increase max symlink count
620 * Undo the "UAC manifest fix" applied to afs_config.exe.
622 * Fixed some additional race conditions in the Rx library
624 * Ensure that Freelance allocation of vnodes follow
625 the AFS convention of odd vnodes are directories
626 and everything else is an even vnode.
628 * Add Freelance logic to mount point and symlink
629 evaluation functions.
631 * Enhance smb_ParseASCIIBlock() so that it can handle
632 all of the STRING formats defined by the CIFS Technical
635 * Validate the output of smb_ParseASCIIBlock() in all
636 callers. Return CM_ERROR_BADSMB if the STRING field
637 cannot be parsed. CM_ERROR_BADSMB will cause the
638 contents of the packet to be logged.
640 * If multiple SMB Raw Write operations were taking
641 place at the same time, there could be data corruption
642 because unique event objects were not generated for
643 each Netbios receive operation.
647 * Fix a race in rxi_WriteProc which could result
648 in a freed or in use packet being passed to
649 rxi_PrepareSendPacket if the call is currently
650 in a transmit wait state. Passing a freed packet
651 will result in a panic. Passing an in use packet
652 might result in a call failure or perhaps data
655 * Kerberos referrals broke the NIM AFS provider
656 logic for automated configuration of the workstation
657 cell with a new identity. Now if a referrals
658 response is returned, assume the cell belongs to the
661 * The interface to buf_CleanAsync() and
662 buf_CleanAsyncLocked() has been changed to permit
663 errors to be propagated to the caller.
665 * During an enumeration of a directory, if the btree
666 has not yet been generated, return an error.
668 * When mapping CM_ERROR values to NTSTATUS values
669 do not treat 0 as an error.
671 * Avoid a cm_scache_t object reference count underflow
672 when renaming via the SMB interface.
674 * smb_IsMask() can be called with a non-8.3 input buffer.
675 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
676 input buffers and smb_IsMask() for non-8.3 input buffers.
678 * Add cm_buf_t error state to the "fs memdump" output.
680 * pioctl drive mapping checks need to take into account
681 Global Drive Mappings and SUBST assignments.
683 * Rename operations should not be permitted if the
684 destination name collides with an existing name that
685 differs only by case UNLESS the collision is the object
686 that is being renamed.
688 * Increase the maximum number of background daemon threads
689 to 64. Leave the default at 4.
691 * Set the FILE_UNICODE_ON_DISK flag when reporting the
692 file system characteristics.
694 * Report the maximum component name length as 255 even
695 for Unicode as part of the file system characteristics.
697 * Add RX debugging information to "fs memdump" output
700 * Refcount the cm_server_t objects returned by
703 * Fix the UAC manifest applied to afs_config.exe
705 * rxdebug now indicates if any RX_CONN flag is set
706 not just some of them
708 * Change the mapping of CM_ERROR_WOULDBLOCK from
709 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
713 * Includes #GCO Public CellServDB 07 Nov 2008
715 * Prevent separate cell entries from being created for
716 "foobar.org" and "foobar.org." which can happen if
717 an AFSDB record is found instead of a CellServDB
720 * Fix a panic when checking the SMB virtual circuit list.
721 If a VC is destroyed while the list is being checked
722 it is possible for the next object to become invalid.
724 * Increased logging to afsd_init.log of netbios name
727 * Fix a DNLC panic when a UTF-16 string that cannot be
728 converted to a utf-8 string is provided as a filename.
730 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
731 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
732 location to refer directly to the new location in
733 %programfiles%\openafs\client\program
735 * Be more aggressive about responding to Netbios
736 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
737 the LANA within the listener thread without performing a
740 * Use the WNet Enumeration of all connected drives as a filter
741 in the pioctl GetIoctlHandle operation. This should remove
742 the last situations in which there was a delay when opening
743 an Explorer Shell properties menu with the AFS Shell
746 * Fix the parsing of CellServDB in Windows to handle linked
747 cells as is done on Unix.
749 * Add support for linked cells to the cache manager,
750 netidmgr afs provider, and aklog. When two cells are
751 linked, a volume lookup that fails in one cell is looked
752 up in the other. This can be used for construction of
753 test cells that override specific volumes; cell renaming;
756 * Increase the length of cell and realm names that can
757 be entered in the NIM AFS provider.
759 * Actively prevent a 32-bit afsd_service.exe from
760 executing in the WOW64 environment.
762 * Protect against invalid IP address values in CellServDB
767 * Fix an Rx panic caused by an attempt to double free
768 an rx_packet when an error occurs while writing the
771 * Fix an exception cause by a race condition when
772 processing multiple file change notifications registered
773 by the Explorer Shell or other applications.
775 * Modify the Rx debug interface to use var_args and
776 permit Rx debugging to be used on 64-bit Windows
777 which previously had it disabled.
781 * Automated lock hierarchy enforcement added to the
782 osi base lock package. This permits detection of lock
783 violations that might but often does not result in
784 deadlocks. Enabled by default in the checked builds.
785 May be activated in release builds via the registry
786 value "LockOrderValidation" in the TransarcAFSDaemon
789 When a violation is detection an assertion is thrown
790 generating a dump file and a message indicating which
791 locks were obtained out of order is logged to the
792 Windows Application Event Log.
794 * Attempt to ensure that a dump file will be generated
795 whenever the service panics.
797 * Removed unused lock cm_bufGetMutex
799 * Many out of order lock acquisitions were detected
800 using the lock order verifier and were corrected.
802 * A race condition within the function path cm_GetSCache() ->
803 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
804 cm_scache_t object to be issued simultaneously to two
805 threads. This would eventually result in a panic due
806 to the resulting undercount.
808 * Fix interpretation of the empty string as the ioctl path
809 to mean the current directory. "fs lsm", "symlink list",
810 etc. now return a "not a ..." error instead of "does not
813 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
814 to indicate that the input string is a nul terminated string.
815 Assign ti when input strings are nul terminated instead of
816 length counted. (from 1.5.50)
818 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
819 does not match the observed behavor. The 'nameLength' is specified
820 in bytes, not in characters. Fix the implementation to match
821 the observed behavior. (1.5.50)
823 * The cm_HaveAccessRights() attempt at deadlock avoidance by
824 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
825 does not avoid the deadlock. Avoid the deadlock by enforcing the
826 lock order of the lowest vnode first. Then remove the infinite
827 loop avoidance in cm_SyncOp() that was returning an unwarranted
830 * File Server lock synchronization was not propertly enforced in
831 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
833 * Another change to the short name generator. Plus and equal are
834 not valid characters in a short name according to the
835 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
836 with left and right brace in order to minimize the impact.
838 * Fix a race condition that can result in a heap overwrite error if
839 the number of known file servers or vldb servers increases during
840 the preparation phase of a cm_CheckServers() operation.
842 * If the SMB Authentication would fail, a LSA allocated memory
843 buffer would be leaked.
845 * Rewrite the cm_cell_t allocation code to permit cell vldb server
846 lookups to fail without leaking cm_cell_t objects.
848 * Add "fs setcrypt auth" to permit the use of rxkad authentication
849 only mode. Not documented.
851 * Initialize Rx mutex and lock objects once and only once as part of
852 the DLL load process.
854 * When reporting the Volume Name "AFS" do not include the terminating
855 NUL in the reported length.
857 * Refresh Volume Location information upon receipt of a VNOVOL error
858 from the file server and avoid a race condition that prevented the
859 updated volume location information from being destroyed immediately
860 after it was acquired.
862 * In response to VNOVOL and VMOVED, remove the volume id from the server's
863 volume list so that the volume's up and down status is not affected
864 when the machine is shutdown.
866 * Many fixes to the Rx stack related to packet allocation and deallocation;
867 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
868 by overwriting the rx_call currentPacket pointer when it was already
869 referring to a packet.
871 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
872 in the TransarcAFSDeamon/Parameters key.
874 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
875 constructed directory B+ trees for flushed directory objects.
877 * When cached pages for directory objects have been locally modified, do
878 not permit the locally modified pages to be mixed with pages subsequently
879 obtained from the file server. Doing so can result in directory
882 * If Ubix errors are received in response to VL_xxx operations record
883 the error on the server instance and permit the operation request to
884 be retried in case another server is able to answer the query. We have
885 seen at least one instance of a cell in which a VL server returned
886 UNOQUORUM (possible database truncation) when the other servers were
887 responding successfully.
890 * Optimize the assignment of write errors to dirty
891 buffers by applying the first error to all buffers
892 belonging to the same object instead of trying each
895 * Add resource version info to xstat_fs_test and xstat_cm_test
897 * Fix a delete vs delete[] usage problem in the WiX custom
900 * When creating a mount point, validate the volume name to
903 * When processing a mount point, pay attention to VL_BADNAME
906 * Remove a race condition from cm_volume_t recycling that could
907 have resulted in two threads making use of the same object at
910 * Fix a reference count leak on smb_vc_t objects introduced in
911 1.5.50. Permit the objects to be freed once again.
913 * Fix the output of "fs memdump" for smb_vc_t objects that was
916 * Add routines for debugging smb refcount errors.
918 * When shutting down the cache manager, do not permit AFS
919 callback requests to be processed. Doing so can result in
920 invalid memory accesses.
922 * Do not recycle a deleted cm_scache_t object unless the
923 refCount is 0. Doing so can result in a deadlock and
924 invalid memory access.
926 * Minimize the number of RPCs sent to the file server for
927 directory bulk status requests. Only request status for
928 objects that we are actually reporting on and only if we
929 do not already have status for the object and then make
930 sure that we return as much as we can at once via the SMB
931 interface but not so much that it exceeds the RDR timeout
933 * Enable bos restricted operations
935 * Fix the creation of submounts used by the afscreds and afs_config
938 * Avoid repeated retries when attempting to obtain access rights.
939 If the access rights were obtained successfully but do not
940 satisfy the request, fail the request.
942 * Fix error code reported when attempting delete a file on a
943 readonly volume or one that is marked with the readonly DOS
946 * Fix a short name truncation bug.
948 * Fix a heap corruption error when reading the CellServDB file
951 * Fix directory listings when the directory has "lookup" permission
952 only and the directory has never been previously accessed.
954 * Fix several tracking issues for cm_scache_t rw locks.
956 * Ensure that all buffers used to store cell names are the
959 * Restore Windows 2000 support (broken in 1.5.50).
960 LOCALE_INVARIANT was introduced in XP.
962 * Add RxUdpBufSize registry value. Default is 256K.
964 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
966 * Do not include trailing NULs when returning file names in
967 a directory search response.
969 * Prevent RX from crashing the afsd_service by attempting
970 to allocate to many buffers at once when memory is fragmented.
972 * Avoid a deadlock when freeing cm_server_t objects
974 * Optimize search cell file. Do not search to end of file
975 if the cell has already been found.
977 * Remove explorer shell extension delay on local disks.
979 * Automatically pre-allocate 64 Rx packets per thread
981 * Add smb lock requests and stat cache lock allocations to
982 the output from fs memdump.
985 * Fix Large File Support (broken in 1.5.50)
987 * Fix "fs setacl" (broken in 1.5.50)
989 * Fix pioctl path evaluation of drive letter
990 mapped paths. (broken in 1.5.50)
992 * Fix installation of VC8 Runtime library in the EXE
993 Retail installer. (broken in 1.5.50)
995 * Avoid a potential deadlock during Volume Location
996 updates if all the file servers are unreachable at
997 the time the VL update succeeds.
999 * Ensure that rx connection object is references
1002 * Fix propagation of over quota errors during cm_FSync()
1005 Since 1.5.39 [1.5.50 released 16 July 2008]
1006 * There were no releases between 1.5.39 and 1.5.50.
1008 * FIRST RELEASE WITH UNICODE SUPPORT
1010 * Pioctl interfaces to the cache manager refactored
1011 to provide layering between the SMB specific code
1012 and the general purpose ioctl operation.
1014 * Garbage collect dead SMB virtual circuits as soon
1015 as they are no longer being referenced. This avoids
1016 problems with outstanding locks not being dropped
1017 when the virtual circuit becomes invalid.
1019 * Remove the IBM Administration Reference documentation
1020 and replace it with the OpenAFS Command Reference
1023 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
1024 functions each time a connection is about to be used.
1025 Do not hold a lock on the rx connection object while
1026 it is being selected. This avoids a race between
1027 threads attempting to set the timeout values and
1028 removes a bottleneck that was hampering performance.
1030 * Ensure that the smb directory attribute is set
1031 for all directory objects.
1033 * Add vs2008 support to the NSIS installer scripts
1035 * Replace the VC Runtime EXE installer with the MSI
1036 installer in the NSIS installer scripts
1038 * Properly delete the VC7.1 C Runtime libraries.
1040 Since 1.5.36 [1.5.39 released 23 June 2008]
1041 * There were no 1.5.37 or 1.5.38 releases for Windows
1043 * Use Visual Studio version of vsprintf() instead of
1044 home grown version that resulted in linking conflicts
1046 * Prevent buserver, ptserver and fileserver from crashing
1047 due to non-portable strftime() format strings.
1049 * Prevent server crashes caused by rx_InitMorePackets()
1050 being called prior to allocation of all of the required
1051 synchronization objects.
1053 * Fixed a memory leak in the internal B+ tree directory.
1054 For objects with non-8.3 names, the 'longname' memory
1055 allocation would be leaked during object deletion or
1058 * Fixed a memory leak in the background daemon cm_CheckServer()
1059 routine. The memory allocated to store the server list
1062 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
1064 * Correct an abstraction layer violation. cm_scache_t objects
1065 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
1066 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
1067 performed in cm_ReleaseSCache() when the reference count hits zero.
1069 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
1071 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
1074 * Ensure that scache object 'nextp' pointers are set to NULL
1075 when the objects are removed from the hash table linked list.
1077 * If a dirty buffer cannot be written to the file server because
1078 of an access denied error, mark the error in the cm_buf_t
1079 structure so that it can be dealt with instead of retrying
1083 * Update CellServDB file to GCO Public 23 Apr 2008
1085 * Fix a cm_buf_t reference count leak when attempts to write
1086 dirty buffers to the file server from within cm_IncrSyncer()
1089 * Prevent udebug from crashing.
1091 * Activate RX Idle Timeouts. If the file server is busy for more
1092 then 30 seconds, attempt to failover to another server without
1093 marking the busy server down.
1095 * Another VNOVNODE issue fixed. When writing a dirty buffer
1096 to the file server, if VNOVNODE is received, mark all buffers
1097 as invalid without further attempts to contact the file server.
1099 * Improved performance on high latency links.
1102 * Fix a bug preventing the re-initialization of the Freelance
1103 root.cell contents when a change is made.
1105 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
1106 or mountpoint that does not exist.
1108 * aklog, afscreds, integrated logon, and the NIM AFS credential
1109 provider will all attempt to first acquire an afs service ticket
1110 using the service principal: afs/<cell>@<CLIENT_REALM>. This
1111 permits the proper detection of the cell realm when multiple
1112 realms are local realms for the cell.
1114 * Fix fs commands that re-use file identifiers instead of forcing
1115 a path evaluation each time. This was broken in 1.5.33 when
1116 hashes were added to the cm_fid_t structure.
1118 * Fix symlink and mountpoint removal and then recreation when
1119 applied to the fake Freelance root.afs volume.
1121 * Fix SMB change notifications for the Freelance root.afs volume
1122 when creating or removing symlinks and mount points.
1124 * Force DNS AFSDB cell name lookups to lower case if the DNS
1125 response is not all lower case.
1127 * 32-bit installers are now built with Visual Studio 2005
1129 * All vos.exe commands now support the "-noresolve" option which
1130 forces the display of IP addresses instead of DNS hostnames.
1132 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
1133 permits operations that do not require up to date volume
1134 location data to continue while waiting for the RPC to complete.
1135 Multiple threads waiting for a volume location update on the
1136 same volume set no longer trigger multiple calls VL_xxx RPCs.
1138 * Volume server references could be marked offline and never
1139 reset to the not_busy state.
1141 * Further optimizations to cm_Analyze() retry logic for
1142 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
1145 * Optimize the DNLC by applying Interlocked incrementing for
1146 statistics gathering and enabling greater use of read
1147 locks instead of write locks
1149 * Further optimize osi_Log operations when logging is disabled
1150 by removing the need for the function call overhead.
1152 * Add a new lock operation, Convert Read to Write, and apply
1153 it throughout the cache manager. This function provides a
1154 fast transition from a read lock to a write lock when the
1155 caller is the only reader.
1157 * Fix a timeout problem when opening files or request locks
1158 when the file in question is already write-locked by another
1161 * Fix a deadlock that can occur if "fs flushall" is executed
1162 while applications are writing to AFS.
1164 * The cache manager would re-read data from the file server
1165 after file truncation when obtaining buffers to write to the
1166 same file until the first StoreData RPC completes.
1168 * Fixes error handling in cm_NTCheckDelete()
1170 * Replaces cm_scache_t mutex with a read-write lock permitting
1171 additional parallel access.
1173 * Prevent an cm_scache_t reference undercount when evaluating
1174 symlinks containing @sys when none of the active sysnames
1175 match any of the available targets.
1177 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1180 * Set RxMaxMTU registry value to 0 which is the equivalent of
1181 it not being set at all.
1183 * Reduce contention between BkgDaemon threads
1185 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1188 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1189 objects to use the faster Interlocked operations which avoid
1190 the need to hold write locks.
1192 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1193 non-fatal error that forces a retry against another service
1196 * Fix a crash in smb_WriteData that would occur if an application
1197 opened the file as read-only and then sent a write request anyway.
1200 * Respond to VL_NOENT errors by removing cm_volume_t object and
1201 setting it to be the first object to be recycled. This is not
1202 only a more efficient use of resources but it also prevents
1203 repeated attempts to query the VLDB server for the volume from
1204 the CheckOfflineVolumes routine.
1206 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1209 * A first cut at a cache manager statistics monitor that can be used
1210 to determine the necessary cache parameters to support the
1211 application working set.
1213 Off by default, the performance package can be activated by setting
1214 "daemonPerformanceTuningInterval" in the service Parameters key.
1215 As with the other daemon interval values the unit is in seconds.
1217 At service start and each succeeding interval the cache manager
1218 will write statistics to %TEMP%\afsd_performance.txt showing the
1219 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1220 The FID statistics keep track of all FIDs seen by the cache manager
1221 during the service session whether or not they are backed by any
1222 live object in the cache.
1224 These statistics are not stored in the cache file.
1226 * Fix windows event logging broken in 1.5.33.
1228 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1229 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1230 as well as when computing the server preference list.
1232 * When merging status after a StoreData operation prevent data
1233 buffers from being discarded that we would prefer to keep.
1234 (Bug introduced in 1.5.33.)
1236 * Modify installers the set RxMaxMTU to 0.
1238 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1239 dialog, the Background Daemons and Service Threads fields were
1242 * Increase default number of Background Daemons to 4.
1244 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1245 cm_volume_t objects to be recycled.
1247 * Windows Error Reporting has shown a number of previously unseen
1248 stack corruption errors in the Explorer Shell extension. The
1249 minidump appeared to indicate the problem was in ParseAcl().
1250 Reviewing the function it appears that numerous things could go
1251 wrong if invalid input was provided. Added error checking and
1252 data validation. Hopefully this will address the problem.
1254 * Windows Error Reporting has shown crashes in aklog as a result
1255 of krb5_cc_get_default() failures not being handled. They are
1259 * The Rx library used a 32-bit type for sockets which was
1260 truncating the socket value on 64-bit Windows. This
1261 was not an issue in general but became an issue when
1262 Rx "hot thread" support was enabled.
1264 * Enable "Rx hot thread" support which ensures that while
1265 a received message is being processed that a new thread
1266 is allocated to receive the next message to arrive.
1268 * Increased parallel access to smb_rctLock protected
1269 resources. Reduce the number of times that the
1270 lock is dropped and re-obtained.
1272 * Increased parallel access to cm_scache_t and cm_buf_t
1273 resources by replacing the previous global write lock
1274 protected reference counting mechanism with one built
1275 upon the use of InterlockedIncrement and InterlockedDecrement.
1277 * Implement a hash for cm_fid_t to reduce comparison costs
1278 and inline the cm_FidCmp() function.
1280 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1281 hash which has a better distribution
1283 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1284 the most recent read request instead of the next chunkSize.
1285 cm_chunkSize can be dozens or hundreds of buffers. As a result
1286 too much time is spent performing the evaluation.
1288 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1289 lock is to prevent the creation of new buffers while a truncation
1290 is being performed. All references to bufCreateLock have been
1291 removed except in two places:
1292 i. a write-lock surrounding the function that calls buf_Truncate()
1293 ii. a read-lock within buf_GetNewLocked() that actually allocates
1296 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1297 CM_CONFIG_DATA_VERSION number which value be used to force the
1298 replacement of the cache file contents when incompatible changes
1299 are made between releases.
1301 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1302 cm_SyncOpCheckContinue
1304 * Avoid calls to multi_Rx if nconns == 0
1306 * Modify smb_WriteData to perform background writes based upon the
1307 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1308 boundaries. This will slow down writes from the SMB interface but
1309 will avoid the risk of the CIFS client disconnecting from the AFS
1310 client SMB server. The default value for 'smb_AsyncStoreSize'
1313 * SMB AsyncStore functionality is on by default but can be
1314 disabled using the registry value:
1316 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1318 smb_AsyncStoreSize can be adjusted with
1320 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1322 The value must be a multiple of the buffer block size
1323 and cannot be larger than the chunk size.
1325 * Do not leak a read lock on the bufCreateLock
1327 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1330 * In MergeStatus, only remove clean buffers from the hash table.
1332 * Avoid the need to update the data version number on each buffer
1333 associated with a scache when MergeStatus is called after a
1334 StoreData by maintaining a range of valid data versions as part
1335 of the cm_scache_t object.
1337 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1338 from cm_HaveAccess... We can safely avoid the call so do so.
1340 * Add "ForceAnonVLDB" service parameters registry value. Default
1341 is 0. Set to 1 to force the use of anonymous connections to
1342 the VLDB servers. This is useful when using krb5 tokens and the
1343 VLDB servers have not been updated.
1346 * Do not attempt to search for a cell name that matches one
1347 of the special share names "svrsvc", "wkssvc", "ipc$"
1349 * Issue permission denied warnings from afscreds if attempts
1350 to modify the startup shortcut or registry values fail.
1351 This happens under Vista UAC.
1353 * Do not return BAD_NETWORK_ERROR in response to attempts
1354 to create directories or files in the Freelance root.afs
1355 volume. Instead return ACCESS_DENIED.
1357 BAD_NETWORK_ERROR was being sent because the Freelance cell
1358 has no vldb servers. Short circuit the error by testing for
1359 freelance in CreateFile and MakeDir
1361 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1362 to be opened and closed without actually existing. For now
1363 we treat them as equivalent to ioctls but this should
1366 * Fix RAP processing to not reject names that exist in the
1367 root.afs volume for GetInfo queries.
1369 * In afscreds, use GetUserNameEx to obtain the username instead
1370 of searching the registry. Try UserPrincipalName first and
1371 fallback to SamCompatible.
1373 * Stop linking to rpcns4.dll which is no longer supported on
1374 Vista and 2008. TaAfsAdmSvr is going to have to be
1377 * When processing dfs referral requests, if the sharename is
1378 a partial match and therefore will not be added automatically
1379 to the share list, make sure that we return an error.
1381 * Avoid a crash when logging VNOVOL errors returned from
1382 cm_CheckOfflineVolume.
1384 * When trace logging is disabled, do not permit osi_LogSaveString
1385 to perform the string copying. By disabling the string
1386 copies, throughput of the service more than doubles.
1388 * when checking offline volumes, update the cell vlserver info
1389 if required and don't perform a check if there are no servers
1392 * Add volumes in the vl_unknown state to the list of objects to be
1393 checked as offline or retried when an allbusy error is received
1396 * Add support for configuring rx_ExtraPackets
1398 * All calls to buf_Get() must be protected by a read lock on the
1399 cm_scache_t bufCreateLock
1401 * When a MergeStatus results in the invalidation of buffers, do
1402 not remove buffers from the hash tables that have active references.
1404 * Upon receiving a VNOVNODE error from a file server, be sure to
1405 invalidate any file locks. This prevents on-going attempts to
1406 extend the locks and will cause subsequent attempts to access
1407 the file to receive a invalid handle error.
1409 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1410 to directory entry names.
1412 * No longer discard tokens when RXKADBADTICKET error is received.
1413 Instead return STATUS_NO_KERB_KEY error.
1416 * Add "cmdebug <host> -cellservdb" option and client service
1417 implementation permits querying the vldb server for each
1418 cell known to the cache manager. The output is in a form
1419 that permits it to be used as a CellServDB file.
1421 * Add Symlinks->Show option to the AFS Explorer Shell Context
1424 * Prevent buffer overrun in the cmd parser used by command
1425 line tools such as fs, vos, pts, etc.
1427 * The path ioctl operations have several issues:
1429 (1) the specified path for "fs examine, whereis, whichcell,
1430 flush" and so always has the follow symlinks and mount
1431 points semantics. This makes it impossible to determine
1432 what the FID of a symlink or mount point is.
1434 (2) "fs examine" out is not the result of a single pioctl
1435 operation but is actually the combined output of half
1436 a dozen operations. Path evaluation is an expensive
1437 operation. It would be faster if the caller could
1438 evaluate the FID first and then perform all of the
1439 rest of the operations by specifying the FID instead
1442 (3) fs output reports all objects as files. By adding a
1443 GetFileType pioctl more informative output can be provided
1444 that indicates what type of object the path evaluates to.
1446 (4) the Windows fs command includes a number of commands that
1447 do nothing but exist only because the Unix cache manager
1450 A new extendible data structure cm_ioctl_query_opts_t has been
1451 added which can be optionally specified with pioctls that do not
1452 already require input data. The first two fields of this structure
1453 are 'literal' and 'fid'. The literal field is used to indicate
1454 whether the last component of the path should be evaluated following
1455 symlinks and mount points.
1457 The fid field permits a fid to be specified.
1459 * A new GetFileType pioctl has been added. The type of objects are
1462 * A new "-literal" option is available for "fs examine, flush, whereis,
1465 * Unimplemented fs commands have been removed.
1467 * There are circumstances where a volume object is being accessed and
1468 the volume is marked indicating that the volume location information
1469 is out of date but where it is also pointless and perhaps dangerous
1470 to block waiting for the rpc to complete. One example is when
1471 processing the cmdebug requests. If we know that we are not going
1472 to use the volume object to contact a server, then we can now set
1473 the CM_GETVOL_FLAG_NO_RESET flag.
1475 * when moving up in the directory tree we search the recorded fid list
1476 to find a matching fid that we have already crossed. we must also
1477 reset the fid count based upon what we discover.
1479 * Cell names are published as share names. As a result they are searched
1480 for as part of the Dfs Referral evaluation. If share "foo" can not be
1481 reached, the CIFS client will fallback to searching for "fo" as well.
1483 Since the freelance client automatically adds a symlink for prefixes
1484 of the cell name, this results in both "foo" and "fo" being added to
1485 the freelance root.afs volume. This patch prevents that negative side
1488 * when tracking Kerberos credential cache names be sure to include
1489 the cache type prefix.
1491 * Microsoft has assigned OpenAFS a network type value
1493 * The algorithms that were used to produce a Netbios name were broken.
1494 The name that was produced when the hostname was too long was in fact
1495 longer than the maximum permitted netbios name.
1497 Also, the "NetbiosName" value was not used as a suffix when the loopback
1498 adapter was not installed. The hardcoded string "AFS" was used instead.
1500 * avoid another deadlock during server probes initiated by the ipaddr
1501 change daemon thread.
1503 * prevent the afs client service from crashing when fs newcell is
1506 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1507 to \Program Files\OpenAFS\Client\Program in order to get them out
1508 of the %windir% tree
1510 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1511 end STATUS_OUT_OF_QUOTA can be returned to the application.
1513 * Modify the search order for determining the location of CellServDB
1514 and other client configuration files.
1516 1. AFSCONF environment variable
1518 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1519 4. use Program Files\OpenAFS\Client
1521 * More Vista Power Management tweaks
1523 * Close a socket handle leak in the kauth routines
1525 * Don't hold a cm_volume_t mutex across RPCs
1527 * Convert downTimes to use clock time instead of relative times.
1528 This permits correct comparisons with expiration times.
1530 * Update the Wix installer to properly install the prerequisites
1531 for translate_et.exe
1533 * Do not translate pioctl paths as they are not converted by the
1536 * Implement multi_rx probes for checking whether servers are up
1537 or down. Turns probes into a constant time activity instead
1538 of one dependent upon the number of known servers.
1540 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1541 update of the volume location data and retry. VNOVOL means
1542 that the volume could not be attached. It doesn't mean that
1543 the volume is not on this server.
1545 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1546 not attempt to validate the target path. There is no reliable
1549 * Improve performance of \\afs\<cell><mount-type><volume>\
1550 access by eliminating extraneous attempts to search the
1551 CellServDB file and DNS AFSDB records. Share names
1552 containing '%' or '#' are not valid cell names.
1554 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1555 to unlock a range that is not currently locked.
1557 * Enable the 32-bit tools to read the 64-bit registry keys
1558 when executing in the WOW64 environment.
1562 * Fix buffer refcount leak introduced in 1.5.29
1564 * Prevent the NIM AFS Provider from crashing when configured to use
1565 Kerberos v4 for token acquisition in situations where no Kerberos v4
1566 support is available. For example, 64-bit KFW or sites that delete
1567 the Kerberos v4 DLL.
1569 * In afsd_init.log, support logging cache sizes greater than 2GB.
1571 * The volume status output of "fs examine" was being determined based
1572 upon the wrong error value. Instead of using the pioctl() return
1573 code, the errno value must be used.
1575 * Prevent invalid pts auto-registration attempts by aklog and the
1576 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1578 * Add synchronization protection to all of the SMB Listener State
1579 variables. This prevents race conditions when the addition or
1580 loss of a network adapter takes place.
1582 * On Vista, there is a race condition between the restart of the
1583 afs client service and the network adapters. Ensure that the
1584 afs client service can handle switching between loopback and
1585 non-loopback modes of operation.
1587 * Fix FollowBackupPath to work with mount points that are not
1588 explicitly read/write.
1590 * Fix a deadlock that has been observed on Vista while resuming
1591 after at least four hours of sleep.
1593 * Add new HKLM registry configuration options that can be used
1594 to provide configuration hints to the NIM AFS Provider when
1595 creating new identities. See release notes for details on
1596 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1599 * Speed up write buffer operations by holding a global read-lock
1600 instead of a write-lock when updating the buffer data version
1601 number. The data version is protect by the associated stat
1604 * Speed up read buffer operations by removing out of date buffers
1605 from the buffer hash tables.
1607 * Protect against a null pointer dereference in afslogon.dll
1608 if QueryAdHomePathFromSid() fails.
1610 * Remove a race condition while accessing private freelance data
1613 * Update the buffer data versions when Freelance directory data
1616 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1619 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1620 instead of returning zero length.
1622 * Return ERROR_RETRY when a server sends VBUSY.
1624 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1625 a volstat plug-in is in use.
1627 * Enhance the volstat interface to support Dfs path mappings.
1629 * Fix "symlink list" to do the right thing for Dfs links
1631 * Fix callback expiration that was broken in 1.5.27.
1633 * Fix a crash at shutdown caused by the lack of synchronization
1634 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1636 * Hold the cm_server_t mutex while walking or modifying the
1637 cm_server_vols_t list in order to avoid a race condition.
1639 * Ensure that the cm_server_t refCount is safely modified.
1641 * The pthread library implementations of pthread_cond_wait
1642 and pthread_cond_timedwait were failing to return with
1643 the mutex held when error conditions occurred. This could
1644 result in race conditions wherever they are used. One case
1645 in which problems were seen was the Rx event_handler().
1647 * Prevent cm_GetBuffer() from locking all buffers associated with
1648 a stat cache object before deciding if the current buffer is up
1651 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1652 object for each offset that we will attempt to reserve.
1654 * Reduce the default number of background daemon threads used to
1655 prefetch or store to one in order to prevent two or more threads
1656 from stepping on each other.
1658 * Reduce the default chunk size to 256KB in order to prevent read
1659 timeouts on low bandwidth connections.
1661 * Fix a race condition that could result in threads becoming stranded
1662 when synchronizing multiple requests on the same stat cache object.
1665 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1666 has been disabled. It has been disabled because it turns out
1667 that the RPC is improperly implemented in all file servers
1668 previous to 1.4.6. The RPC is executed without the proper
1669 locks being held resulting in data corruption and eventual file
1670 server failure. The functionality will be re-enabled in a
1673 * a couple of .readonly Volume Callback optimizations were
1675 - Apply most recent volume callback to all stat cache objects
1676 in the volume. This avoids premature callback expirations
1677 and unnecessary FetchStatus calls.
1678 - Add option to permit attempts to renew .readonly Volume
1679 Callbacks on a periodic basis to avoid invalidation of stat
1680 cache data. [Requires daemonCheckVolCBInterval be set to
1681 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1683 * Adds support for 64-bit data version values. Previous releases
1684 only supported 32-bit values. Larger values sent by the file server
1685 would result in rollover.
1687 * aklog and asetkey no longer generate an exception if Kerberos for
1688 Windows is not installed.
1690 * If the LSA authentication functions report a temporary out of memory
1691 condition, fail the current authentication and do not cause the AFS
1692 client service to terminate unexpectedly.
1694 * When the AFS client service is shutting down, prevent the daemon check
1695 thread from executing additional checks which might attempt to access
1696 discarded locks or unmapped memory.
1698 * Correct a double free condition in the NIM AFS credentials provider.
1700 * Add "FollowBackupPath" service parameter registry value.
1701 This provides equivalent functionality to the UNIX afsd -backuptree
1702 option. When set to a non-zero value, normal mount points originating
1703 in a .backup volume will prefer a .backup volume.
1706 * Fix NSIS uninstall from the .MSI installer.
1708 * Add Stack Frame logging for AMD64
1710 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1711 server returns EWOULDBLOCK, retry the request every two
1712 seconds for up to the RDRtimeout.
1714 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1715 is not assigned the same value as EIO.
1717 * Update CellServDB to GCO Public 25 Oct 2007
1719 * Add additional validation and error handling code after
1720 each call to getSlot(). If an invalid slot is returned,
1721 return NONODE. If the invalid slot is returned when
1722 extracting a data node, invalidate the tree.
1724 * Modify compareKeys() to always perform a case-insensitive
1725 comparison and only perform a case sensistive comparison
1726 if the case-insensitive one matches. This ensures the
1727 ordering is consistently reported.
1729 * Add lock assertions to ensure that all calls are being
1730 performed with the correct locks being held. There have been
1731 some crash reports that provide stack data that does not appear
1732 to be possible unless there is a race. However, there are no
1733 obvious locations where the race is taking place and the test
1734 suite indicates that all of the correct locks are being held.
1735 We shall see what happens in the field.
1737 * For consistency replace all calls to findKey in which the range i
1738 s (1,numentries) with calls to getSlot().
1740 * Optimize the depth search loop by testing the slot value in the
1741 for statement instead of forcing the loop to be broken later.
1743 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1744 There are a number of locations where locks are obtained, dropped,
1745 and reobtained. This reorganization attempts to accomplish several
1748 (1) be optimistic for the most common case so it will be fast
1750 (2) add consistency checks after each location where locks are
1751 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1752 and the bplus tree is out of date, retry until we get to a
1753 consistent state that we can use.
1755 (3) Ensure that all operations take place with the correct locks.
1757 * One of the issues that has become a serious problem since the
1758 addition of the local directory updates is that although cm_SyncOp
1759 synchronizes operations, it does not preserve the order of requests.
1760 This has always been a problem in that it has been possible for a
1761 request to fail to complete due to its worker thread's bad luck.
1762 When a request takes longer than the Windows SMB Redirector's timeout,
1763 the SMB Redirector tears down the SMB virtual circuit.
1765 When using the local directory updates it is really important that
1766 the directory update operations complete in the order that they were
1767 sent to the file server. If they don't, then the local directory state
1768 and the file server state will not match and the local directory state
1769 must be discarded which in turn forces a new read of the entire
1770 directory contents over the network.
1772 This patch adds a new cm_scache_waiter_t object that is used to store
1773 the current thread, buffer, and syncop flags within a waiters queue
1774 on each cm_scache_t object. If a thread is forced to sleep in
1775 cm_SyncOp, upon waking it will check to see if there are any other
1776 threads waiting that are attempting to perform a similar task ahead
1777 of it in the queue. If yes, the thread goes back to sleep. If not,
1778 it goes ahead and enters the cm_SyncOp conflict resolution block.
1780 This patch has the additional side effect of reducing the number of
1781 competing threads that must obtain the cm_scache_t mutex and process
1782 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1783 utilization of the service and the clock time associated with processing
1784 requests will be reduced.
1786 * assert that the cm_scache_t mutex is held when calling
1787 buf_ForceDataVersion().
1789 * rename findNode to leafNode in bplus_Lookup
1791 * replace all OutputDebugString calls with osi_LogX calls
1793 * modify bestMatch to special case the return values for leaf nodes.
1794 If an entry is above or below the values available in the leaf node
1795 return BTLOWER or BTUPPER instead of BTERROR.
1797 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1798 if true convert to either slot 0 or Max and perform the insertion.
1799 This produces easier to read code when performing lookups.
1801 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1802 descriptive message.
1804 * the difference between osi_assert() and osi_assertx() is whether or not
1805 a description is specified as part of the call. When no message is
1806 specified afsd_notifier() is called with a NULL msgp parameter. This
1807 results in a null pointer dereference during the EventLog() call.
1809 As a result, none of the information describing the assertion is written
1812 This commit sets a default message for use when no other message is
1813 provided and it replaces all calls to osi_assert with osi_assertx and
1816 * Migrate B+ tree search key into thread local storage
1819 * Fix an uninitialized variable in lana_IsLoopback() which
1820 could result in a random determination that a loopback
1821 adapter is not a loopback adapter. This would result in
1822 a computed Netbios service name of %MACHINE%-AFS instead
1823 of AFS. This is a serious problem if the value computed
1824 by the AFS Client Service and the client process pioctl
1825 calls produce different results as it will result in a
1826 communication failure.
1828 * Fix a crash in the B+ Tree code that results when a file
1829 rename only results in a change of case.
1831 * Fix the pioctl path parsing. The previous algorithm
1832 broke the current directory of the client into two parts,
1833 the TIDpath which represented the device root and the
1834 relative path. The problem is that this is not sufficient
1835 from proper processing of symlinks and mount points.
1837 What is required is that the last component of the path
1838 must not have the symlink or mount point be evaluated
1839 whereas symlink and mount point evaluation must be performed
1840 for the rest of the relative path. Therefore, the path is
1841 now split into three parts: TIDpath, intermediate path,
1844 "fs examine" now reports the data of the symlink or mount
1845 point object instead of the target and "symlink list" and
1846 "symlink remove" now works. The "Explorer Shell Extension"
1847 menu now is displayed for symlink objects.
1849 * More informative symlink.exe error messages.
1851 * When the cmd.exe "del" command is used to remove a symlink,
1852 not only was the link object removed, but the stat cache
1853 object for the target was marked deleted. This resulted
1854 in cache corruption which would require the destruction
1855 of the cache to correct.
1857 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1858 which results in cached contents from read only volumes
1859 to be treated as valid when there is no callback registered
1860 and all servers are down or all instances are offline.
1862 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1864 * Add documented registry value, "AcceptDottedPrincipalNames"
1865 which is used to permit the acquisition of AFS tokens
1866 using Kerberos v5 principal names which include a dot in
1867 the first component.
1869 {HKLM\HKCU}\Software\OpenAFS\Client
1871 * "fs uuid" does not require "AFS Client Administrator" group
1872 membership to display the current UUID value.
1875 * Update NSIS installer to version 2.30
1877 * Update Wix installer to version 2.0.5325
1879 * 64-bit KFW support:
1880 - Network Identity Manager plug-in
1882 - 64-bit krb5 enable afscreds.exe, afs server manager,
1885 * Enable B+ tree Directory Search by default.
1887 * NIM plug-in is now compatible with FILE: ccaches
1889 * Install instloop.exe into Client/Programs/ even though
1890 it isn't required. Permit it to be used by organizations
1891 that clone Windows systems.
1893 * On a suspend, only drop callbacks for servers that
1894 are online. If the server is offline, maintain the callback
1895 until the server is once again available.
1897 * Prevent against acquisition of a directory write lock when
1898 there are no updates to be performed. This was resulting
1899 in a write-lock leak which later on would lead to a deadlock.
1901 * Add configurable file pre-fetch mode. For files that match
1902 the specified file extensions, pre-fetch the entire file into
1905 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1906 (MULTI_SZ) PrefetchExecutableExtensions
1908 * Fix the client_osi large integer support so that return values
1909 can be properly assigned to large integer variables.
1911 * Do not hold cm_scache_t mutex across cm_GetCell() calls
1912 since cm_GetCell() can block in a RPC.
1914 * Add short file name support to B+ tree searches.
1916 * Resolve a conflict between the desire to notify file servers
1917 that the client workstation is suspending and the desire to
1918 permit objects with valid callbacks when the client suspends
1919 to continue to use those objects from the cache if the client
1920 resumes without network connectivity.
1922 * Fix stack variables to ensure that all buffers storing paths
1923 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
1924 would cause a stack overflow condition.
1926 * On Vista, force afs_config.exe to run with the highest privileges
1927 available to the end user. For a user in the Administrators
1928 group running under UAC, this will cause afs_config.exe to be
1929 be run with full Administrators privileges after the user
1930 approves the privilege escalation.
1932 * More VNOVNODE handling issues. If a dirty buffer cannot be
1933 written due to VNOVNODE, clear the buffer and ignore the error.
1934 If a cm_scache_t object has been marked deleted due to a VNOVNODE
1935 error, the next time the smb_fid_t object that references it is
1936 used return either NOSUCHFILE or NOSUCHPATH and close the
1937 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
1938 result in an invalid handle error.
1940 * For Vista, set the RequestExecutionLevel of the NSIS installer
1943 * For the build environment, support the generation of manifests
1944 when using the CL1310 (VS .NET 2003) compiler/linker.
1946 * There was a bug in OutputDebugHexString() which could overflow
1947 a stack buffer resulting in unexpected process termination
1948 of afsd_service.exe if the input Security Token contained
1949 many bytes equivalent to the '%' char.
1951 * Add directory lookup statistics to "fs memdump" output.
1955 * Windows uses case-insensitive file name pattern matching
1956 but AFS is a case sensitive file system. The AFS3 directory
1957 format is block based, uses network byte order and
1958 includes a hash table for fast case sensitive lookups.
1959 This causes several problems for the Windows AFS client.
1960 (1) Traversing the directory blocks is cpu expensive
1961 (2) A hash table miss does not indicate that the desired
1962 entry does not exist.
1963 (3) Determining whether a non-ambiguous inexact match or
1964 the entry does not exist requires a linear traversal
1965 of the entire directory.
1966 These issues often result in 100% CPU utilization.
1968 These issues are addressed by building a modified B+ tree for
1969 each directory and then using the B+ tree for searches.
1971 This feature is disabled by default. To enable it set
1972 the "BPlusTrees" registry value to 0x01 in the AFS Client
1973 Service Parameters key.
1975 * Correct a null pointer dereference in the symlink recursion
1976 detection code. (Introduced in 1.5.22)
1978 * Add the Network Identity Manager AFS Provider to AMD64
1983 * A new registry value, "BlockSize", can now be used to adjust
1984 the size of the internal buffers used to store file data.
1985 The default is 4KB. When set, the values 1..1024 specify
1986 multiples of 4KB. Values > 1024 specify actual block sizes
1987 and must be multiples of 4KB. The actual block size in KB
1988 can not be larger than the chunk size which defaults to 1MB.
1990 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1993 * pts.exe and ptserver.exe built with -DSUPERGROUPS
1995 * Service Manager shutdown hints have been added. The hint
1996 specifies that it may take the AFS client service up to two
1997 minutes to shutdown.
1999 * Fix the recursion detection feature introduced in 1.5.22.
2000 The previous implementation broke the evaluation of paths
2001 using ../path or ./path. It also broke the evaluation of
2002 symlinks to absolute paths. For example, /afs/cell/path.
2004 * Fix an uninitialized variable in cm_MakeDir that could result
2005 in crash if the requested directory could not be created.
2007 * Fix a the VolumeStatusNotification log messages to save
2008 the input string in the circular log buffer so garbarge
2009 won't be output to afsd.log when the log is dumped.
2012 * Increase the SMB Maximum Path length to 260 characters
2013 from 256 as that is what CIFS supports.
2015 * On 64-bit Windows, use a unique description string for
2016 the 32-bit AFS Shell Extension.
2020 * In the AFS Admin Library, replaced all of the
2021 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
2022 allows parameter checking and proper type conversions to be
2023 performed. This is crucial when passing parameters that are
2024 64-bit wide on systems in which 'long' is a 32-bit value.
2025 Otherwise, the parameter list on the stack is interpretted
2026 incorrectly and the function not only doesn't succeed but it
2027 might corrupt memory and if lucky crash the application.
2029 The only program that we care about that uses this library
2030 is the AFS Server Manager. Now it doesn't crash.
2032 * Fixed smb directory searches for non-wildcard names. The
2033 optimized function should not have set the bulk in progress
2034 flag. This would adversely affect later directory searches
2035 that required a bulk search.
2037 * Detect infinite recursions (up to 512 stat cache objects
2038 in a chain.) If an infinite recursion is detected, return
2039 an error to stop the CIFS client from querying us forever.
2041 * Executables executed out of AFS name space can have their
2042 pages swapped out by Windows. Those pages will be reloaded
2043 when they are required. However, if the cache manager no
2044 longer has a callback and all of the file servers with
2045 the volume data are down, then the application will crash
2046 due to the inability to read the required page.
2048 We now track the down time of each server. If the callback
2049 expires after the server that issued it went down and all
2050 servers are down, we delay the expiration to permit the data
2051 in the cache to be continue to be used by Windows or an
2054 * The Windows AFS client has always had very poor performance
2055 with regards to directory management. In the case of
2056 directory updates caused by create, rename,and delete operations,
2057 the AFS cache manager will now perform local directory updates
2058 when possible in order to avoid re-reading the directory from
2059 the file server. This will be done in whenever both of the
2060 following conditions are true:
2062 - the data version on the directory increased by one as a
2063 result of the operation
2065 - all of the directory buffers necessary to make the change
2066 are present in the cache
2068 * The error message in the AFS Control Panel describing the
2069 need to be a Windows Administrator has been replaced with
2070 a message describing the need to be in the AFS Admins Group.
2072 * If the vldb server list is empty when attempting to evaluate
2073 a cell mountpoint it was possible for a crash to occur. This
2074 can only happen if there is an existing mount point and the
2075 is a failure attempting to load the server list from the
2076 CellServDB file or via AFSDB queries.
2078 * dirty buffers only have their dirty bytes stored to the file
2079 server instead of the entire chunksize
2081 * the default chunksize has been increased from 128KB to 1MB
2086 * changed the enum values for cm_serverRef_t state info to use a
2087 private name space to avoid collisions (srv_)
2089 * added a srv_deleted state for cm_serverRef_t objects. This
2090 state is set when cm_FreeServerList() is called with the
2091 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
2092 not always delete the cm_serverRef_t from the list if it is
2093 still in use by another thread. the srv_deleted state means
2094 the object's contents are no longer valid and it must be
2095 skipped. It will be deleted the next time the object is
2096 freed and the refcount hits zero.
2098 * the srv_deleted state is also used when a file server reports
2099 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
2100 as offline. This is done to prevent additional usage of the
2101 stale vldb data while waiting for the update volume request
2104 * added a state field to the cm_volume_t object (enum volstate
2105 vl_ name space) that maintains the state of the volume based
2106 upon the states of all of the cm_serverRef_t and cm_server_t
2109 * modified cm_UpdateVolume() to set the state of the cm_volume_t
2110 RW, RO, and BK to either vl_alldown or vl_online. There can't
2111 be any other states because cm_UpdateVolume() destroys any
2112 previous knowledge we might have had regarding busy or offline
2115 * modified cm_UpdateVolume() to update the volume name in the
2116 cm_volume_t to the volume base name if the previous value was
2119 * modified cm_FollowMountPoint() to check to see if the volume
2120 name is a volume ID and if so call cm_GetVolumeByID instead
2121 of cm_GetVolumeByName. This ensures that volume IDs are always
2122 looked up as numeric values. There is no longer a need to
2123 maintain a separate cm_volume_t containing the string representation
2126 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
2127 The first flag is a "CREATE" flag which is set by all existing
2128 calls. The flag is not set by calls to cm_GetVolumeByID() from
2129 the server probe code when volume status is being updated. We
2130 do not want the server probe operation to result in additional
2131 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
2132 which is set when the server probe code updates the volume status.
2133 This flag will be used to prevent the server probe operation from
2134 changing the order of the least recently used queue.
2136 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
2137 allocated for a given set of normal, readonly, and backup volumes
2138 regardless of whether or not the volume is accessed via name or
2139 ID number. The cm_volume_t namep field is always the base name
2142 * Added a new volume state, vl_unknown. This state is used as
2143 the initial state for all cm_volume_t when the cache manager starts,
2144 for each cm_volume_t at creation, and for each cm_volume_t when
2145 recycling. The cache manager does not know the state of all
2146 volumes in the world, only those that are in the cache and for
2147 which it has queried the VLDB and hosting file servers.
2149 * modified cm_GetVolumeByName() to initialize the state of a
2150 volume to vl_unknown. The actual state will be set when a
2151 cm_VolumeUpdate() call completes successfully.
2153 * changed name of scache hash table variables to avoid ambiguity
2154 when adding hash tables for volumes
2156 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
2158 * modified cm_UpdateVolume() to handle the case in which there is
2159 no RW volume but there is are RO volumes for a given base name.
2160 This is done by querying for the ".readonly" volume name if the
2161 base name does not exist in the VLDB. We never query for the
2162 .backup name because under the current usage model a .backup
2163 volume may only exist on the server that the read-write volume
2164 is located. If there is no RW volume, there can be no .backup.
2166 * Added four hash tables for cm_volume_t objects to improve the
2167 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2168 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2169 BKID. Three ID hash tables are necessary as long as it is
2170 desireable to maintain a single cm_volume_t containing all
2171 of the related RW, RO, and BK volume data. Having the RW and
2172 RO volume data in the same object is necessary for the
2173 implementation of cm_GetROVolumeID() which returns either the
2174 RO or RW ID depending upon the existence of RO volume instances.
2176 * Added a volume LRU queue so that volume reuse becomes fairer.
2177 This does not replace the all Volumes list which is used when
2178 it is desireable to walk a list of all the volumes whose order
2179 is not going to change out from underneath you which makes it
2180 safe to drop the cm_volumeLock.
2182 * handles volume hash table updates where volume name to
2183 volume ID number changes. The volume name remains
2184 constant in the cm_volume_t. if a vos rename is performed,
2185 the name of the volume will change and the volume IDs will be
2186 updated. Subsequent access to the old volume ID will create a
2187 new cm_volume_t with the new name.
2189 * Added a daemon thread operation to query the state of volumes
2190 listed as busy or offline. cm_CheckBusyVolumes() calls
2191 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2192 or vl_offline. If the volume is now online, the status on the
2193 volume is updated. The default period is 600 seconds. This can
2194 be configured with the BusyVolumeCheckInterval registry value.
2196 * Added prototype for smb_IoctlPrepareRead() which was missing a
2197 return type in the function definition.
2199 * Added volume id lists to the cm_server_t. These lists are
2200 allocated in blocks of ~32 IDs. When a cm_PingServer()
2201 detects a change in server state, the state of the cm_volume_t
2204 * Added volID to the cm_serverRef_t object. volID is used
2205 to identify the volume for which the object is a referral.
2206 cm_FreeServerList() uses the volID to remove the volume
2207 from the cm_server_t.
2209 * In cm_Analyze, when VNOVOL or VMOVED are received,
2210 call cm_ForceVolumeUpdate() to force a refresh of the volume
2213 * Added cm_VolumeStatusNotification() which is used at the moment
2214 to log volume status changes to the trace log. It will also
2215 be used as the access point to the File System Filter driver
2216 notification engine.
2218 * Added an all cm_scache_t list to cm_data. This replaces the use
2219 of the stat cache LRU queue when we need to enumerate all
2220 entries. The LRU list order is not static and when using it to
2221 enumerate all entries it can result in items being missed or
2222 items being processed more than once.
2224 * Modified cm_Analyze(). Instead of reseting the busy or offline
2225 state of a volume and forcing a retry of the operation
2226 cm_Analyze will defer to the background daemon thread that will
2227 update the state once every 600 seconds.
2229 * Added the automatic generation of a Freelance ".root" read-write
2230 mountpoint that refers to the root.afs volume of the workstation
2231 cellname at the time the mountpoint is created.
2233 * re-write cm_Analyze to make better use of the known volume
2234 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2237 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2238 busy volumes will be reset to srv_non_busy by the function
2239 but there is no mechanism for querying the busy state other
2240 than by attempting to access the resource.
2242 * cm_Analyze will query the state of an offline volume before
2243 deciding whether or not to retry when all volume instances
2246 * the algorithm used after an IP address list change detection
2247 has been modified. we now check the state of the VLDB servers,
2248 then reset the rx connections, and finally check the state of
2249 the file servers. Previously we flushed the rx connections
2250 and then checked the servers in whatever random order they
2251 were first discovered.
2253 * fix the volume status reported by "fs examine"
2255 * various improvements to the AFS Network Identity Manager Provider.
2256 - Don't access freed objects after the window has been destroyed.
2257 - Write back the AFS cell list in the identity configuration
2258 even when its empty.
2259 - Disable all use of the provider when AFS is disabled for a
2260 particular identity.
2262 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2263 alias instead of the full cellname, double check the full cell
2264 name and see if we have a matching cell before allocating a new
2267 * Dump more data with "fs memdump". All volumes and scaches and
2270 * Reverse the order of the all cells list. This ensures that
2271 the workstation cell is always first in the list.
2273 * If we recognize the server from which a callback revoke was
2274 received, then we can ensure that we only invalidate the status
2275 information for cm_scache_t objects belonging to that cell.
2277 * When building for amd64, use APPVER 5.02 as we do not support
2278 64-bit Windows 2000 and explicitly state the machine type in
2281 * no longer treat VMWare network adapters as equivalent to a
2282 loopback adapter. For some unknown reason, publishing the "AFS"
2283 netbios name on them frequently results in a name collision
2286 * Use read locks whenever possible instead of write locks when
2287 walking the allCellsp list. This permits better use of multiple
2290 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2291 is not present on the directory optimization. It would report
2292 mount points and directory objects for which there was no
2293 status information as a zero length file. Instead we can use
2294 the vnode number of the object to determine if it should be
2295 considered to be a directory or a file. If it is odd it is a
2296 directory; even it is a file.
2298 * Give Back Callbacks to the file servers in response to suspend
2299 and shutdown power management events. Only give up callbacks
2300 to UP servers and do so with a very short RX dead timeout.
2303 * Fix a race condition in the SMB FID cleanup code that could
2304 result in a stat cache mutex never being released.
2306 * Fix the graphics files that are installed as part of the
2307 IBM Administrators Guide.
2309 * Change the behavior of pioctl path parsing to not follow
2310 symlinks or mount points during evaluation. This results in
2311 the FID being the symlink or mount point and not the object
2312 they point to. This is consistent with the behavior of the
2315 * Enable AFSDB record lookup for command line tools such as
2318 * CellServDB update 2007-07-05
2320 * Added a Volume Status Plug-in Interface that permits third parties
2321 to receive notifications when the volume state changes occur.
2323 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2325 * Prevent the afsprocmgmt.dll from being unloaded while the
2326 background thread it starts is running.
2328 * Avoid a race condition when closing smb_fid_t objects.
2331 * fix install path of IBM Quick Start Guide when installed by
2334 * fix the client so that it can write into directories with
2335 only 'li' permissions.
2337 * fix installation of 32-bit Shell Extension for 64-bit systems.
2338 Install it under its own HKCR keys.
2340 * modify the wix installer to remove and upgrade install packages
2341 with the same version number.
2343 * when flushing volume location data every two hours or in
2344 response to "fs checkvolumes", also flush the mount point
2345 links in order to force their re-evaluation.
2347 * aklog was copying an uninitialized buffer as part of the
2348 foreign cell automatic pts registration
2350 * when obtaining directory listings, check for read permission
2351 on the directory. If read permission does not exist, fake the
2352 status information for the directory entries. This improves
2353 performance, prevents the client from pounding the file server,
2354 and prevents leakage of cached status information.
2356 * "fs getcacheparms" output was incorrect.
2358 * NetIdMgr AFS Provider should only automatically add token
2359 acquisition for cell foo.com if the realm of the cell matches
2360 the realm of the identity.
2362 * Fix a null pointer dereference in cm_MergeStatus()
2366 * network identity manager provider:
2368 - Correct for the fact that pr_CreateUser tampers with the
2369 contents of the username input buffer.
2371 - Take into account that the krb4 credential data structure
2372 does not have separate realm fields for the user and service.
2373 When krb524 is used with a cross-realm ticket, the realm
2374 describes the service not the user.
2376 * Many integrated logon changes:
2378 - Fix eventlog reporting. Do not attempt to log an event
2379 if the event source registration fails. Use DebugEvent0
2380 instead of DebugEvent when there are no parameters.
2382 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2385 - Fix the reading of Domain specific configuration for
2386 LogonScript and TheseCells. Previously the dwSize value was
2387 being overwritten so that subsequent RegQueryValueEx call
2390 - Fix a memory leak in the TheseCells reading code.
2392 - Add support for Domain specific "Realm" specification. The
2393 realm is the realm to be appended to the username. When
2394 logging in as a domain or to the local machine, the specified
2395 "Domain" name is not going to be a valid realm name.
2397 - Construct a proper principal name based upon the domain
2398 specified realm for use in obtaining tokens with KFW.
2400 - If the domain specified "TheseCells" list includes the
2401 default cell, do not obtain tokens twice.
2403 * An unprotected call to the Kerberos v5 function,
2404 krb5_free_context(), can result in an invalid memory access
2405 exception if the contents of the krb5 profile, krb5.ini, cannot
2406 be parsed correctly. The error is located within the afskfw
2407 library which is used by afscreds.exe, afssvrmgr.exe, and
2408 afslogon.dll. When the error is triggered within afslogon.dll
2409 a blue screen condition can occur when the winlogon.exe system
2410 process terminates unexpectedly.
2414 * Do not stop the afs client service with an assert
2415 if the cell name cannot be determined when responding
2416 to a "fs whereis" command.
2418 * Integrated Logon (afslogon.dll). No longer attempt
2419 to preserve Kerberos 5 tickets for the logon session
2420 as that job is now supported by KFW 3.1 and above.
2422 * The Network Identity Manager plug-in will attempt to
2423 remove expired service tickets from the cache.
2427 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2429 * AFS Server Manager and AFS Account Manager data cache
2430 moved from HKCU\Software\TransarcCorporation to
2431 HKCU\Software\OpenAFS to prevent crashes due to data
2434 * Reduce the default UP server probe time from 10 minutes
2435 to 5 minutes to match UNIX. This will not solve the NAT
2436 port mapping timeout issue because most NATs default to
2437 a value less that 5 minutes and this value ensures that
2438 the afs client server will start probes 5 minutes after
2439 the last probe set ended. The probe period for any one
2440 server is therefore longer than 5 minutes. The more
2441 servers in the list the longer the time between probes.
2443 * Before marking a server as down in cm_Analyze, retry the
2444 request using a new rx connection to the server. If the
2445 cache manager is behind a NAT, it is possible that the
2446 response from the server is being sent to the original
2447 address:port from which the rx connection was established.
2448 This will appear to the cache manager as the server being
2449 down when in fact it is not. Forcing the use of a new rx
2450 connection will result in the server responding to the
2451 most recent address:port assigned to the client by the NAT.
2453 If the server is in fact down, it will simply take a little
2454 longer to notice it.
2458 * Removing a directory that contained entries was failing with
2459 the wrong error code being returned to the CIFS client.
2463 * rxdebug exponential back-off. don't take down a server that
2464 is already struggling to respond.
2466 * Network Identity Manager plug-in registers help on the Help
2467 menu if it is supported by NIM. (KFW 3.2 and above)
2469 * Plug-n-play networking did not always work after coming back
2470 from a Windows Suspend. The AFS client service failed to
2471 re-compute the LAN adapter to bind to and instead attempted
2472 to bind to all LAN adapters. This would polute the public
2473 network with the name "AFS" which could lead to conflicts
2476 * Build installers using Microsoft Hotfixes provided at
2477 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2478 Daylight Savings Time issues.
2480 * "fs checkserver" does not require admin privileges unless
2481 the probe interval is being changed.
2485 * A reference leak was discovered in the mountpoint traversal
2486 functionality which resulted in cm_volume_t objects not becoming
2487 eligible for re-use. This will trigger a panic condition when
2488 more than the maximum number of volumes is accessed.
2490 * Fixed the success test for the "fs memdump" operation. Success
2491 is now reported as such. Added cm_volume_t object information
2494 * A data corruption error has been fixed. Write requests
2495 queued for background processing were not retried upon failure.
2496 This leads to inconsistencies between the contents of the file
2497 in the local cache and the contents of the file on the file
2500 The implemented fix is to re-queue any request that has failed
2501 due to timeout, volume busy, volume offline, server busy, or
2502 server offline. Requests are not processed while all servers
2503 for a volume are inaccessible. When servers become accessible
2504 the background request will be retried.
2506 In addition, if a background write fails, do not mark the buffers
2507 as clean. Failure to preserve the dirty flag prevents the buffers
2508 from being written to the file server in the future.
2510 This fix does not address the possibility that the credentials
2511 necessary to perform a write request might expire between the
2512 time the request is queued and when the request can be processed
2513 due to server accessibility.
2515 * The Freelance volume does not support per user ACLs. The
2516 cm_HaveAccessRights function did not handle this special case.
2517 Requests for privileges greater than read|lookup would result
2518 in afsd_service.exe entering an infinite loop. ISOBuster is
2519 known to cause this behavior.
2521 * When deleting files, the SMB Server would attempt to obtain
2522 the access permissions for the parent directory instead of
2523 for the file itself. This would result in an incorrect
2524 assessment of whether or not the file can be deleted.
2526 * The Network Identity Manager AFS plug-in has had its string
2527 tables updated for consistency with the language used by NIM.
2528 All credentials are referred to as credentials instead of tokens.
2529 Drop-down combo boxes now in fact drop down.
2531 * The AFS Salvager was not properly being built on Windows.
2532 This is relevant only if the AFS servers are in use.
2534 * The default signal handlers for the AFS servers incorrectly
2535 processed SIGQUIT. Instead of stopping the server, an
2536 exception was raised.
2538 * Fix another volume refcount leak. This one was triggered by
2539 an inability to contact the vlservers in order to update the
2542 * Fix checklist control used by the drive mappings tabs in
2543 afscreds.exe and afs_control.exe.
2545 * Improve output of "fs memdump" related to buffer management.
2547 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2548 already in the process of being fetched.
2550 * Improve cache performance by computing the buffer hash table
2551 size based upon the number of buffers so that the average
2552 number of entries in a bucket is seven.
2554 * Improve cache performance of read/write data by locally
2555 updating the dataVersion of all buffers associated with the
2556 file on each StoreData or StoreStatus operation provided that
2557 the dataVersion was incremented by one and the dataVersion of
2558 the buffer matched the original dataVersion of the cached
2559 status information for the file.
2561 * When recycling cm_scache_t objects, make sure that we
2562 clear all of the previous fields. It is no longer the
2563 case that we can assume that the stat cache object is being
2564 obtained only after a successful FetchStatus call.
2565 InlineBulkStatus will allocate stat cache objects for which
2566 the no status information is available.
2568 * When a file is being opened with the OPEN_ALWAYS disposition,
2569 do not require write permissions when the file already exists
2570 unless of course the file is being opened requesting write
2573 * When a file has been opened for read-only access, if the
2574 application requests an exclusive lock, obtain a read-lock
2575 instead of a write-lock. The purpose of the exclusive lock
2576 in this situation is to prevent changes to the file and the
2577 full file read lock will provide this guarrantee.
2582 * Enforce Unix Mode Write bit to prevent file deletion of read only
2585 * The CIFS READ and LIST permissions use the same value 0x0001.
2586 When a directory is being accessed, check for PRSFS_LOOKUP
2587 instead of PRSFS_READ.
2591 * Fix computations of short names of mount points. They were
2592 broken as part of the directory search optimization. (1.5.13a)
2594 * When all buffers are in use, periodically sleep the threads
2595 waiting for a buffer to become free while the global buffer
2596 lock is not being held in order to permit those threads that
2597 can free buffers to do so.
2599 * Power Management improvements. Maintain a global flag that
2600 specifies whether or not the service is in a suspend state.
2601 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2602 lana is no longer valid. Instead, stop the listener threads
2603 and if all listener threads are stopped, reset the lana_list.
2604 Allow the cm_Daemon() thread to periodically check the state
2605 of the smb listeners. If they are all stopped and the service
2606 is not suspended, attempt to restart them. If there are no valid
2607 lanas, return to the stopped state.
2609 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2610 the user is permitted to obtain read or write locks. This function
2611 would obtain the lock and then drop it returning whether or not
2612 the lock could be obtained. If the lock was in fact required,
2613 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2614 of course being that this pattern results in three RPCs to the
2615 file server (lock, unlock, lock). This is reduced to one RPC by
2616 implementing cm_CheckNTOpenDone() which frees the allocated byte
2617 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2620 * Optimize the SMB FindFirst processing for the case in which there
2621 are no wildcards. If a cm_Lookup succeeds, generate a response
2622 from the stat cache entry we have a callback for. Its cheaper
2623 than walking each buffer of directory data associated with the
2624 directory stat cache entry when we have thousands of directory
2627 * Increase the SMB Ioctl MaxData size to support larger token sizes
2629 * Disable autorun of the AFS Server Wizard when the AFS Server is
2632 * Add AFS Server service dependencies on tcpip and PNP_TDI
2634 * Remove the AFS Server Configuration Tool shortcut
2636 * Add Freelance mode awareness to AFS Server Configuration Tool
2640 * Unlike previous versions of the OS, Vista performs a shutdown on
2641 the Microsoft Loopback adapter just like it would on a real adapter.
2642 This causes the smb_Listener threads to trigger a panic during a
2643 suspend/hibernate power event.
2645 The fix is to unbind from the network adapters in response to a
2646 suspend/hibernate power event and then rebind to the adapters
2647 when a resume power event is received. Note that the resume events
2648 are not reliably delivered so it is possible the afs service will
2649 not be accessible. However, this is the best we can do.
2651 * don't allow environment variables we set in afslogon.dll to be
2652 inherited by child processes
2654 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2657 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2658 thread walking the all buffer list periodically searching for dirty
2659 buffers to write to the file server. This had several negative
2660 results. The algorithm ate up ever increasing amounts of CPU time
2661 even when AFS is idle as the size of the cache increases. Also,
2662 buffers were written to the file server in an order based upon the
2663 original buffer allocation which has nothing to do with the order
2664 in which the buffers became dirty.
2666 The new algorithm maintains a dirty buffer list. Items are added
2667 when the buffer is originally marked dirty. A buffer is only
2668 removed from the list by the buf_IncrSyncer when the buffer is no
2669 longer dirty. If the list is empty the thread goes back to sleep
2670 immediately without additional processing requirements.
2674 * Enable renaming of directories from the Explorer Shell when
2675 Offline Folders is enabled
2677 * An error is reported in response to the SMB ReceiveCoreFlush
2678 request when an error has not occurred.
2682 * optimize dirty buffer flushes in response to suspend operations
2683 do not flush if only loopback adapter is installed
2684 do no retry dirty buffer writes
2686 * do not reset netbios when performing loopback checks. doing
2687 so will break the existing lana bind
2689 * remove a deadlock in smb3.c QPathInfo
2693 * Add NetIDMgr AFS Plug-in
2695 * "Show Tray Icon" checkbox removed from afs_config.exe
2697 * Disable DNS Registrations for the Loopback Adapter and make sure
2698 that Netbios is turned on.
2700 * A reference count leak on stat cache objects was fixed
2702 * For debugging, tread ids of writers are now tracked on locks
2704 * Corrected an incorrect lock that was being held but not
2705 freed in a rare error case
2707 * Changed trace log end of line to CR-LF so it can be read
2710 * modify the buffer synchronization thread to write all dirty
2711 buffers and sleep only when there is no work to be done
2713 * improve behavior of cache manager when large numbers of
2714 temporary files are being created by multiple clients within
2717 * do not permit dirty buffers that belong to deleted files
2718 to be written to the file server. its a waste of bandwidth
2720 * do not leak stat cache references when the file server
2723 * afscreds: do not permit an attempt to obtain tokens if the
2724 afs service could not be started. this was causing stack
2725 overflows on Vista after resuming from a suspend
2729 * better handling of dot directory
2731 * fs sq and fs mkm now behave as on UNIX
2733 * add license to installers
2735 * Return Invalid Parameter error for unknown InfoLevels in
2736 SMB Tran2 Search Directory.
2738 * Correctly use WSAStartup and WSACleanup within the RX library
2740 * Check that afsd_service.exe is running before performing
2743 * Force AFS Shell Extension to load resource DLL using
2744 afs_shl_ext.dll as a base name.
2746 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2747 OpenAFS\Client\Program
2749 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2751 * Do not panic if the maximum number of volume entries are in use
2752 and one of them can be recycled.
2756 * Fix the code that reads the backconnnectionhostnames value
2757 so that there is not an uninitialized variable
2761 * integrated logon could cause winlogin.exe to blue screen windows
2762 if it was called with a domain that does not exist on the machine.
2763 this is possible during a sysprep.
2765 * corrections to cm_buf_t handling to ensure that buffers are
2766 removed from queues before they are freed and that holds are
2767 obtained on the next buffer in a chain before releasing the
2770 * do not generate MessageBox dialogs on network errors. they
2771 cause more trouble then they are worth.
2773 * move GlobalAutoMapper into a separate thread so that it does
2774 not prevent service startup if the WNetAddConnection2 call
2777 * restrict hard dead and connection timeouts to the CIFS Session
2778 Timeout value. This is being done to try to prevent virtual
2779 circuits from being destroyed prematurely.
2781 * Fix the code that writes the backconnectionhostnames value
2782 to ensure that the data buffer is written with the correct
2785 (there were no windows specific changes between 1.5.4 and 1.5.5)
2789 * Corrections to the Integrated Logon functionality that permits
2790 Kerberos 5 TGTs to be communicated into the user session. The
2791 transfer is now more secure and succeeds when the user does not
2792 have read permission to the %SystemRoot%\TEMP directory.
2794 * NSIS installer updated to version 2.18
2796 * WiX installer updated to version 2.0.4221.0
2798 * Updates to the IFS functionality for 64-bit correctness and
2799 permit compilation. The IFS functionality works for pioctl()
2800 operations but not much else at this point.
2802 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2803 query and set functions.
2805 * Ignore the Windows CIFS client session timeout entirely. Timeout
2806 only after the AFS Hard Dead Timeout period expires.
2810 * move headers, libraries, and samples from OpenAFS\Client to
2813 * remove the auto-registration of AFS IDs for foreign cells
2814 from integrated logon. this permits afslogon.dll to avoid
2815 initializing the RX library. afslogon.dll is a network
2816 provider library and as such it is loaded into all applications
2817 which perform a CreateProcessAsUser call. Initializing the
2818 RX library spawns background threads to listen for inbound
2819 packets. When the network provider is unloaded from the
2820 process, the RX library does not cleanup the threads. If they
2821 continue to execute after the library is unloaded, the process
2824 * implement inline bulk status calls. this will significantly
2825 improve the performance of directory listings.
2827 * prevent a crash when evaluating mountpoints to volumes that do not
2830 * new command: fs uuid [-generate]
2831 displays the current uuid and allows a new one to be generated
2832 on the fly without restarting the afs client service.
2834 * fix the 64-bit installers to properly document the install
2837 * move help files to the proper location so they can be located by
2838 the system tray tool and the control panel.
2840 * "over quota" and "disk full" errors were not being reported to the
2843 * largefile support has been added. Files greater than 31-bits can no
2844 longer be written to file servers which do not support large files.
2845 When file servers support large files, files up to 64-bits can be
2848 * cm_SyncOp could return with the input buffer unlocked upon error
2849 which would result in a panic condition. This bug was introduced
2852 * vlserver failover when freelance mode did not work. This prevented
2853 the root.afs volume from being loaded if the first vlserver could
2858 * A bug in the CIFS Standard Info response packet caused directories
2859 to appear as standard files. This prevented Microsoft Vista from
2860 being able to walk the AFS directory tree.
2862 * Mount points must be reported as directories in response to
2863 Query Path Info operations.
2865 * A deadlock was detected when performing "fs flushall" if the
2866 file server reports VNOVNODE. The scp->createBufferLock is
2867 already held by the current thread and there is no mechanism
2868 to propagate the knowledge. Therefore, an alternate mechanism
2869 for clearing the cache must be developed. A new function
2870 cm_RecycleSCache(scp,flags) has been extracted from
2871 cm_GetNewSCache(). This function performs the task of recycling
2872 an cm_scache_t object. When called from cm_GetNewSCache() with
2873 no flags the expectation is that there are no associated buffers
2874 that are queued to be read or written. When called from
2875 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2876 flag, any queued buffers will be de-queued and marked as if
2877 the operations were performed so the data can be discarded.
2879 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2881 * In response to VNOVNODE, the parent is only discarded if the current
2882 cm_scache_t is not a directory.
2884 * In the Ioctl FlushFile and FlushVolume functions, there were no
2885 checks to protect against flushing the Freelance SCache entries.
2887 * In FlushFile, the wrong cm_scache_t object was being released.
2889 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2890 recycled. Choose a new entry if cm_RecycleSCache fails.
2892 * Speed up the performance of the cache manager by not
2893 holding the smb_fid_t mutex across calls to cm_SyncOp
2896 * Ensure that all smb_fid_t flag references are protected.
2898 * Remove a deadlock between smb_fid_t mutex and
2899 smb_rctLock introduced in 1.4.1.
2901 * Fix aklog AFSID lookup to use not send the realm
2902 name when the realm is the local realm for the
2903 cell. The PT Server doesn't strip the realm
2904 name when it is the local realm.
2906 * Treat "\\afs\*." as an alias for "\\afs\all"
2908 * Dynamically adjust the priority of server threads
2909 based upon the age of the CIFS request that is
2910 being processed. The priority is bumped one level
2911 for every 15 seconds of processing time.
2913 * The original openafs contribution mis-used the
2914 cm_GetCallback function. This function should only
2915 be called as a side effect of a call to
2916 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
2917 was being called results in a FetchStatus RPC being
2918 performed even when the client already has a valid
2919 callback and does so in a manner that destroys the
2920 synchronization of the threads that are calling
2921 cm_SyncOp. If nothing else this results in a
2922 significant performance penalty.
2924 * More changes to cleanup of smb_vc_t objects in
2925 order to prevent race conditions.
2927 * Reduce the amount of Power Management event logging
2930 * If the client detects that the IP address has
2931 changed, force the use of new RX connections. This
2932 enables the file server to recognize that the client
2935 * Remove a race condition and deadlock associated
2936 with cleanup of smb_vc_t objects.
2938 * Fix the ANSI filename support. One function that
2939 required it was missing the translations.
2941 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
2944 * Add a DllMain() function to afspthread.dll. Perform
2945 cleanup of thread local storage, rx queues, critical
2946 sections, event handles, etc. This was implemented
2947 to avoid crashes in afspthread.dll when applications
2948 which load/unload the dll attempt to terminate.
2950 * Lock request behavior has been altered. As of this
2951 release a locally allocated lock will be used whenever
2952 the volume is RO or the requesting user's permissions
2953 are 'rl'. This will allow end users to execute programs
2954 from RW volumes when they do not have the 'k' privilege.
2956 Note that a recent discovery indicates that the 'k' bit
2957 is only supposed to control the ability to obtain read
2958 locks. Write locks are to be controlled by the 'w' and
2961 * Code has been added to enable the cache manager to be
2962 able to obtain the AFSID of the authenticated user.
2963 This is required to be able to determine when "Insert"
2964 privileges should apply. AFS considers a file to be
2965 a new file whenever the owner's AFSID matches the
2966 AFSID of the user. The current implementation relies
2967 on the name specified during the SetToken call to be
2968 correct. This will be replaced in the future with a
2969 new RPC that allows the AFSID of the authenticated
2970 user to be obtained.
2972 * A 32-bit-tools installer for use on 64-bit systems
2973 is now available. These tools are for use with 32-bit
2974 versions of Kerberos for Windows or JPSoftware's 4NT.
2976 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
2978 * Remove the use of the "active_vcp" variable. This variable
2979 no longer serves any useful purpose and was the cause of
2980 race conditions that produced under counts of the smb_vc_t
2983 * Change default minidump output to include data segments
2985 * restore access to cm_IoctlGetVolumeStatus and
2986 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
2987 Still need to figure out why they are being called
2988 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
2990 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
2992 * Update build system to include the "sys_name" in the
2995 * optimize the number of calls to obtain the smb_user_t for a
2996 smb virtual circuit.
2998 * add logging of server addresses, status and capabilities when
3001 * prevent multiple threads from pinging the same server simultaneously.
3002 instead, threads block and then use the result of the ping that was
3003 already taking place.
3005 * Add registry values to allow the defaults for daemon checks to be altered
3006 at startup. Log values to afsd_init.log.
3008 daemonCheckDownInterval
3009 daemonCheckUpInterval
3010 daemonCheckVolInterval
3011 daemonCheckCBInterval
3012 daemonCheckLockInterval
3013 daemonCheckTokenInterval
3015 * Add EventLog messages for transitions in fs crypt state.
3017 * Byte range locking (when locks will be obtained from the
3018 is still under discussion)
3020 * 64-bit Windows operating systems now supported
3021 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
3022 This is a native 64-bit implementation of OpenAFS for Windows.
3023 There is currently no 64-bit version of KFW and therefore
3024 there is no (easy) method for obtaining AFS credentials.
3025 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
3026 to obtain AFS tokens but there is not for the feint of
3029 * A faster implementation of the fcrypt algorithms improves
3032 * The afsd_service.exe now supplies Event Log Message Text.
3033 The event log no longer generates warning text due to lack
3034 of message configuration.
3036 * When afs servers are unreachable and network stack returns
3037 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
3038 mark the attempted server as being down. This allows for
3039 very fast response to the calling application. This change
3040 has a pronounced improvement when the Windows Explorer is
3041 used on disconnected mobile computers.
3043 * Added DebugOutputString debugging of the RX library
3045 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
3047 * Fix the rx-lwp implementation to randomly generate rx call
3048 identifiers which are supposed to be unique at a given point
3049 in time. This affects some of the command line tools used
3050 for debugging but not the AFS Client Service itself which
3051 uses the rx-pthreads implementation.
3053 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
3054 Windows although they are not distributed
3056 * don't generate an exception when logging unexpected SMB packets
3058 * the data structures used by cm_IoctlGetVolumeStatus and
3059 cm_IoctlSetVolumeStatus are not compatible with the CIFS
3060 Get/Set File Information calls so don't bother executing them.
3064 * Fix data written to registry as part of BackConnectionHostnames
3067 * Fix initialization bug when allocating cm_server_t objects that
3068 would cause servers to appear to be down after each allocation
3069 or renewal. Server would appear up after the next check servers
3074 * Windows will return WSAECONNRESET to the next receive packet
3075 operation on a socket if an ICMP packet is received in response
3076 to a packet that was sent and could not be delivered because
3077 the port was unreachable. We have to treat this as
3078 WSAEWOULDBLOCK so that we can retry the select() check and
3079 not mark the connection as bad.
3081 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
3083 * The client manager maintains a least recently used list of
3084 all the ACLs. If an ACL is already the most recently used
3085 do not bother to remove it from the list and place it on the
3088 * As part of the process of adding a new server, set the
3089 server is down flag and then perform a ping. If the server
3090 is up, the flag will be reset. If the server is down, the
3091 client will not timeout unnecessarily while attempting to
3094 * If a Mountpoint String is the empty string, return PATH NOT
3097 * Modify behavior of the Network Provider to avoid querying the
3098 profile location if integrated logon is known to be disabled.
3100 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
3101 will always be preserved regardless of whether or not integrated
3102 logon was used to obtain tokens. When unset, tokens will be
3103 destroyed regardless of whether integrated logon was used to
3104 obtain the tokens unless the profile is stored in AFS.
3106 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
3107 message is being delivered even when the user is not logging
3108 off of the computer.
3110 * Fix a deadlock condition introduced in rc6.
3112 * Cell names are case-insensitive. Since the founding of OpenAFS
3113 there has been a bug that treated "cell", "Cell", and "CELL" as
3114 different cells. This would result in what appeared to be
3115 random disappearance of user tokens or a failure to authenticate
3116 even when tokens were listed by the "tokens" command.
3118 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
3120 * Additional auditing of the smb code revealed that smb ioctl
3121 operations were failing to free smb_fid_t objects
3123 * The File Identifiers assigned to smb_fid_t objects could be
3124 given the value 65535 which is 0xFFFF which is also the
3125 INVALID_HANDLE value. Both 0 and 65535 are now treated as
3128 * Audited all assignments to fields and ensured that they are
3129 protected by obtaining the relevant mutex object.
3131 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
3132 if both the current cm_scache_t object and its parent are
3133 both returning VNOVNODE.
3135 * Removed all the logic associated with the 'dead_vcp' variable
3136 which must have been added for debugging because its used
3137 was never thread safe and it truly served no purpose.
3139 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
3141 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
3142 registry settings. Default is ON and 120 seconds. If set to
3143 OFF, tokens will not be destroyed.
3145 * Fixed a number of race conditions and reference count problems
3146 associated with global queues.
3148 * Added a method to detect abandoned smb virtual circuits so their
3149 associated file handles, locks, user tokens, and smb session
3150 resources can be released.
3152 * Added a mechanism to allow user tokens to be preserved
3153 during integrated login and be freed after logoff.
3155 * Check access permissions to the file before the file is
3158 * More improvements to the handling of SMB Virtual Circuits
3159 when SMB sessions are closed.
3161 * Correct a race condition introduced in rc4 that could result
3162 in a crash of afsd_service.exe.
3164 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3166 * Properly handle pre-mature SMB session closures.
3167 File handles, tree connections, and user sessions are now
3168 freed when the virtual circuit is closed.
3170 * smb_username_t and smb_user_t objects are now properly reference
3171 counted and freed as appropriate.
3173 * disable to use of AFS file locks. byte range locking is still
3174 implemented but all locks are managed locally. This prevents
3175 multiple processes from editing the same file on the same
3176 machine but it does prevent multiple processes on separate
3177 machines from modifying the file.
3179 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3181 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3182 the same server parameter would result in a panic condition
3183 if the server had not already been contacted by the AFS
3186 * Better warnings are written to the afsd.log file when
3187 the Windows RPC Protocols are not properly configured as
3188 is often the case on Windows 2000.
3190 * fix the procmgmt library to restore C RTL signal handlers when
3191 it is being detached from the process prior to process termination.
3193 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3195 * close a lock leak in the rx library that could be triggered while
3198 * prevent removal of drive mappings in response to afscreds.exe -M
3200 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3201 strings and flush the stat cache entry.
3203 * prevent server objects from being freed if user preferences are set.
3205 * Update WIX support to 2.0.2419.0
3207 * Release all locks on the server at service shutdown.
3209 * A down server should be marked up as long as it responds to the client.
3210 It doesn't matter if the server returned an error as long as the error
3211 is something other than VRESTARTING. If there is such an error, all
3212 connections on that server should be marked for replacement.
3214 * Change the check "up" servers period from once per hour to once every
3215 ten minutes. This matches the behavior of the Unix cache manager.
3217 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3218 * Byte Range Locking support has been added. Architectural documentation
3219 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3220 made to enforce Windows mandatory locking model using locally managed
3221 byte range locks with full file locks obtained from the file server.
3222 Read-only volumes provide the user with an implied locking privilege.
3223 Read-write volumes require that a user be granted the locking privilege
3224 in order for a lock to be obtained. As most Windows applications use
3225 Shared file access modes, they require the ability to obtain locks in
3226 order to access files.
3228 * Implement error translation functions to convert Universal Error Codes
3229 output by modern servers to local errno and winsock error values.
3230 This fixes the error message problems using pioctl() calls from fs.exe
3231 and the AFS Shell Extension.
3233 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3234 * Allow the AFS Client Service to set the firewall configuration by
3235 retrying until after Windows determines the boot period is over.
3237 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3240 * Detect IP address changes from within the AFS Client Service for the
3241 purpose of checking down servers sooner
3243 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3244 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3245 user that there is a Kerberos key problem. Previously the user would
3246 receive no feedback when there was something wrong.
3248 * Fix the large integer library support to properly handle values greater
3251 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3252 * Updated CellServDB
3254 * Fixed file name pattern matching. Question marks can match nothing
3255 if found at the end of a component.
3257 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3258 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3260 * The MSI installer now has a new Product ID for each build. This
3261 is to allow for automated upgrades. The Upgrade ID is remaining
3264 * Volumes that were cached and then subsequently were salvaged,
3265 moved or otherwise taken temporarily offline (as opposed to being
3266 busy) would become unreachable. A distinction is now made between
3267 a volume being offline and a server being down.
3269 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3271 * Fixed several AFS Shell Extension issues:
3272 - Checking Server Status would only work if a specific cell was
3274 - If servers were down, the Server Down dialog would write to
3275 invalid memory and cause explorer.exe to crash
3276 - Default Push Buttons were not specified for many dialogs
3277 - The extension would be unloaded by the Explorer Shell and
3278 never be reloaded under some circumstances.
3280 * Fixed the AFS System Tray menu to disappear if another Window
3281 was selected before a menu item was selected
3283 * Added volume owner and group information to "fs examine" output
3285 * Added fs and registry support for enabling/disabling/clearing
3286 rx statistics gathering.
3288 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3290 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3291 GUIDs for each platform.
3293 * Documentation directory structure and default installation rules
3296 * New OpenAFS for Windows HTML Release Notes installed
3297 as part of the documentation set.
3299 * Add thread id to rx debugging messages
3301 * When all servers are offline, return a bad network path error
3302 immediately. This is necessary to allow Explorer to do the
3303 right thing and not hang. The background daemon thread will
3304 check the down servers every 30 seconds to see if they are
3307 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3309 * The AFSCache file has been modified to store both the
3310 serial number of the volume on which the file resides
3311 and the SID of the local machine. If either change,
3312 a new UUID is generated.
3314 * add "fs flushall" command that forces all data buffers
3317 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3319 * add support for '/' instead of '\' in pioctl() calls
3321 * Apply AFS Client Admins group protect to AFS Shell Extension
3323 * Add support for \\afs\<name> to most AFS Shell Extension
3324 commands except the symlink methods.
3326 * when installing the Microsoft Loopback Adapter, enable
3327 MS Client for Networks and prevent an install failure
3328 by not calling CoInitialize twice in the same thread.
3330 * reload cell vldb values from the CellServDB every two
3331 hours in case it changes
3333 * When updating cell information from DNS, be sure to set
3336 * Add support to allow use of \\AFS\<foo> where <foo> is
3337 either a mount point or symlink. As <foo> is normally
3338 treated as a share name, we transform it into \\AFS\all\<foo>
3341 * Init 'code' to prevent false errors when integrated logon
3342 is disabled and the service is not running
3344 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3346 * Removed trace log messages that were hampering performance.
3348 * Fixed a deadlock that was being triggered by editing Word
3349 documents stored within AFS with WinWord 2003.
3351 * Bit 3 of the TraceOption registry value is now used to set the
3352 default for "fs trace" in the non-Debug builds. The new default
3353 is off for release builds and on for debug builds.
3355 * Bit 2 of the TraceOption registry value can be used to turn on
3356 real-time output of debug log entries to the Windows Debug
3357 Output monitor interface. This data can be viewed with tools
3358 such as Sysinternal's DbgView.
3360 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3361 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3362 a resource we should not be returning status codes such
3363 as Network Busy or Remote Host Down because these errors
3364 will imply that the AFS Client Service is Busy or Down
3365 and that is simply not the case. Instead we will return
3366 Bad Network Path as the path is temporarily not available.
3368 Instead of returning Sharing Paused when there is a
3369 timeout error, return Timeout. Once again, the AFS Client
3370 Service is not paused.
3372 * afscreds.exe would display an Obtain Creds dialog when
3373 the expired credentials reminder was triggered even
3374 if there was no network path to the KDC. This is prevented
3375 by adding KDC probe logic to the reminder thread.
3377 * afscreds.exe would display expired tokens no differently
3378 than unexpired ones. This would make it difficult for the
3379 user to distiguish when the tokens were expired. For the
3380 English build added a new resource string "(expired) that
3381 is displayed instead of the expiration time.
3384 * "fs wscell" when executed in freelance mode will return the
3385 name of the cell configured in the registry. The root.afs
3386 volume is not loaded from this cell, but it is used for the
3387 default for aklog and integrated login.
3389 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3390 no longer work on \\AFS\all when freelance is being used
3391 unless the user is a member of the "AFS Client Admins" Windows
3394 * some more checks performed during persistent cache validation.
3397 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3398 down servers for the current cell not all cells.
3400 * If the scache LRU list becomes corrupted, the AFS Client Service
3401 might enter an infinite loop at startup. Detect the corruption
3402 and discard the cache contents in this case.
3404 * Fixed the Explorer Shell's remove mountpoint functionality.
3406 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3407 obtain a mutex lock on a cm_scache_t object whose lock is held by
3408 a thread attempting to obtain the cm_aclLock.
3411 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3412 misinterpretted as "csail.mit.edu" when the resolver
3413 is configured to append ".mit.edu" to failed DNS queries.
3415 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3416 that can be used to force the preservation of user tokens upon logout.
3418 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3419 recommended installation categories: AFS Client, AFS Administrator,
3420 AFS Server, AFS Development Kit. Each category includes a different
3421 default subset of the OpenAFS components.
3423 The OpenAFS logo is now associated with the NSIS uninstall entry
3424 in the Add/Remove Programs control panel.
3426 * The user name associated with AFS tokens when obtained with
3427 integrated login, the afs systray tool, or aklog will always
3428 include the full Kerberos 5 user principal name regardless of
3429 whether or not the cell is local to the realm.
3431 * Modify integrated login so that it does not enter an infinite
3432 loop if the service is not set to auto start.
3434 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3437 * Added uninstall.exe to Wix installation
3439 * More modifications to algorithms used to wake sleeping threads.
3443 * Changes to the algorithms used to wake threads when they are
3444 put to sleep because of conflicts over buffer operations.
3447 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3448 the process of producing AFS tokens. The tokens are stored within
3449 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3450 New functionality has been added to temporarily copy the tickets to
3451 a file ccache during the short window between the NPLogonNotify and
3452 WinLogon Logon Event. The Logon Event handler starts a process as
3453 user, afscpcc.exe, that copies the tickets into the default ccache.
3454 The file ccache is then destroyed by the event handler.
3456 * During synchronization operations on stat cache entries, if threads
3457 are forced to sleep it was possible that the threads would not be
3458 woken under heavy load.
3460 * Enforce the delete on close flag specified by the NTCreateX()
3463 * A race condition in the rx library was fixed that resulted in
3464 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3465 was cleared when threads were still waiting to use the
3466 connection. Also, fixed a potential case in which the
3467 connection could be destroyed while threads were waiting to use it.
3469 * The DNLC freelist has been seen to become corrupted with still
3470 active entries being stored on the freelist. Changes were made
3471 to perform a better job of cleaning entries before placing them
3472 onto the freelist; marking them with the correct magic value when
3473 purging the list; and allowing the list to be purged on startup
3474 if the cache validation fails.
3476 * Windows Crash Reporting does not get triggered for afsd_service.exe
3477 because it provides its own top level exception handler. This patch
3478 places an explicit call to ReportCrash() on platforms which support
3481 If you configure Crash Reporting via AD Policy to capture crash
3482 reports within the domain, then you will receive this output.
3484 * Generate MiniDumps independent of the Windows Crash Reporting.
3486 (a) If an exception occurs,a minidump will be generated at the path
3487 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3488 can be set to any other type via use of the
3491 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3493 (b) MiniDumps can be generated at any time using the "fs minidump"
3494 command. This allows you to generate a minidump without
3495 stopping the process or even requiring the use of a debugger
3498 * an RX connection which reports bad ticket now treats it the same way
3499 the expired ticket is treated. The ticket is removed from the
3500 connection and a new anonymous connection is established.
3502 * fs memdump now requires that the caller be in the AFS Client Admins
3505 * additional debugging information is provided in cm_Analyze when the
3506 error code is ignored.
3508 * Fix a regression introduced into the pioctl() client code which
3509 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3510 MIT KFW is not installed. (1.3.82a)
3512 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3515 * Fix a race condition in cm_EndCallbackGrantingCall() that
3516 could leave 'nrevp' pointing to freed memory. In the process,
3517 optimize the processing to ensure that the applications monitoring
3518 the status of the stat cache entry are only notified at most
3521 * Conditionalize access to the Explorer Shell AFS->Mount Point
3522 ->Remove menu item. Disable the item if the selected item(s)
3523 are not mount points.
3525 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3526 Local Storage. This eliminates thread contention on the
3527 global RX free packet queue. This should improve performance
3528 on multi-CPU systems.
3530 * Fix the IP address filtering code to properly load/unload
3531 the IPHelper DLL so that Windows XP/2003 and beyond systems
3532 do not have to manually probe the registry.
3534 * CleanupACLEnt() was not being called consistently with
3535 the cm_scache_t object referred to by the back pointer
3536 mutex locked. This could in very rare conditions lead to
3537 an invalid memory access.
3539 * Added a script command to the msi installer "afs_replace"
3540 which can be used to replace one version of OpenAFS with
3541 another without requiring an uninstall and reinstall if
3542 the installed msi is no longer in its original location.
3544 * Apparently the problem with multi-domain forests with cross-
3545 realm trusts to non-Windows realms was not entirely solved.
3546 The authentication to the AFS SMB service failed because
3547 the wrong name was being used. Using ASU as an example,
3548 the authentication was being performed with the name
3549 "QAAD\user" (an account in the forest root) and not
3550 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3552 The solution was to add an additional dependency on KFW
3553 in order or to be able to easily obtain the client principal
3554 name stored in the MSLSA ccache TGT. This information is
3555 used in two locations:
3557 - the pioctl() function
3559 - a new WinLogon Event Handler for the "logon" event.
3561 The pioctl function will now be able to use the correct
3562 name when calling WNetAddConnection2() and the "logon"
3563 event handler will now be able to call WNetAddConnection2().
3564 The hope is that the "logon" event handler will be called
3565 before the profile is loaded but I have not guarrantee
3569 * Fixed a locking error in cm_TryBulkStat() which had the
3570 potential of deadlocking the system for the length of time
3571 it takes to perform a bulk status fetch operation.
3573 * Replaced time conversion code (UnixTime <-> FILETIME) to
3574 be completely arithmetic instead of relying on a bizarre
3575 algorithm involving a variety of C RTL time functions.
3576 This has the side effect that UnixTime and FILETIME which
3577 are both stored in UTC are interpretted as UTC throughout
3578 the year. Windows will apply the same localization to AFS
3579 as it does to NTFS. Applications which rely on the ability
3580 to sync files between the two file systems will no longer
3581 see the timestamps of files in AFS change an hour relative
3582 to the files stored in NTFS or Windows based backup devices.
3584 * Fixed a invalid memory access under a bizarre circumstance.
3585 Windows will allow a physical mass media device to be
3586 installed via Plug N Play to the system and will assign it
3587 the lowest available drive letter. This is true even if the
3588 drive letter is currently assigned in the user session to
3589 a network device via NET USE (or its equivalent.) When this
3590 happens, queries sent to the CIFS server will contain invalid
3591 data. This invalid data was not being caught by the AFS
3592 Client Service and was resulting in a crash when Freelance
3595 * Fixed a reference count error when registering callbacks
3596 on a stat cache object if the callback was already registered
3598 * Add a case to cm_Analyze() in order to handle the case where
3599 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3600 the server list for the volume is empty. In this case, force
3601 an update of the volume info and retry.
3603 * Insert a missing cm_EndCallbackGrantingCall() which could
3604 result in threads waiting for a callback to complete to never
3607 * In the persistent cache, there is a maximum number of volume
3608 entries. Allow volume entries to be reused if the maximum
3609 number have been allocated and their reference count is zero.
3611 * If we already have a dead virtual connection object, don't
3612 ignore the fact that additional dead vc objects must be taken
3615 * Removed a deadlock condition introduced in the summer of 2004.
3616 Do not hold mx locks around calls to RX functions including
3617 rx_NewCall(). That is what reference counts on the rx_connection
3620 * Fixed an initialization error in afslogon.dll which could
3621 result in random behavior including a failure to terminate
3622 the NPLogon function. (1.3.80b)
3624 * Fixed an error preventing the use of SMB authentication
3625 on Windows 2000. (1.3.80a)
3628 * Updated CellServDB to Public CellServDB 16 Mar 2005
3630 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3631 misinterpretted as "csail.mit.edu" when the resolver
3632 is configured to append ".mit.edu" to failed DNS queries.
3634 * Fixed another case in which the client would replace connections
3635 to the host when it was not necessary. In this case the
3636 situation would occur if cryptall was on and the connections
3637 were unauthenticated due to lack of an appropriate token.
3639 * OpenAFS for Windows has failed to work at sites which are
3640 utilizing a cross-realm trust between an MIT/Heimdal realm
3641 and a multi-domain Windows forest when the workstation being
3642 accessed is not located in the root domain. This is caused
3643 by a bug in the workstation which was triggered after the
3644 introduction of Windows 2003 Server. When the bug is triggered,
3645 the workstation attempts to authenticate users to afsd_service.exe
3646 by contacting the Domain Controller instead of using the
3647 LSA loopback authentication mechanism.
3649 One of the reasons this bug occurs is because the workstation
3650 does not have a reliable way of knowing that the service whose
3651 netbios name is "AFS" is located on the workstation. This will
3652 be fixed starting in Longhorn Beta 1 by Microsoft. The
3653 "BackConnectionHostNames" registry value will be used to
3654 indicate that the authentications to that service name should
3655 be performed using the loopback authentication mechanism.
3657 In the meantime, when Logon Caching is enabled, we can force
3658 afsd_service.exe to authenticate using the logon cache before
3659 contacting the Domain Controller. This will work with both
3660 password and smart card based logons.
3662 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3663 flag should not be cleared if a volume's server reference is
3664 marked as "offline". In the case where all of the volume's
3665 servers are either "down" or the volumes are "offline", we want
3666 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3667 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3668 should never occur in practice.
3670 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3671 5 seconds, clear the server down and volume busy flags, and
3672 then force an update from the VLDB. This allows the client to
3673 update the location of a volume if the reason for it being
3674 marked offline is because it is being moved. Calls to
3675 cm_ConnByMServers() will be retried either until success or
3676 the RDRtimeout period is reached.
3678 * Correct the Power Management code behavior in response to
3679 Standby, Suspend, and Shutdown events. Instead of flushing the
3680 buffers associated with the mounted SMB submounts, simply write
3681 all buffers which are marked dirty.
3683 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3684 provides a list of cells other than the default cell for which
3685 tokens should be obtained using the default Kerberos principal.
3686 This functionality is logon domain specific and is only available
3687 when using KFW for authentication.
3689 * Fixed FindNext Invalid Handle error caused by over agressive
3690 attempts at garbage collecting dirSearch entries when the
3691 dirSearch ID wraps from the maximum value to 1.
3693 * Add support for registry defined server preferences for VLDB and
3694 File servers. See registry.txt for details.
3696 * Increased default cache size to 96MB and default number of cache
3699 * Fixed refCount leaks related to directory lookups and pioctl calls.
3701 * Callbacks revoked during a race condition with an attempt to
3702 obtain the same callback no longer result in an inaccessible
3705 * New command line tool:
3707 afsdacl : Set or reset the DACL to allow starting or stopping
3708 the afsd service by any ordinary user.
3710 Usage : afsdacl [-set | -reset] [-show]
3711 -set : Sets the DACL
3712 -reset : Reset the DACL
3713 -show : Show current DACL (SDSF)
3715 * IP addresses are no longer queried once at startup. Instead IP
3716 addresses are obtained as needed. Loopback adapter addresses are
3717 no longer published to the server.
3719 * Pay attention to the MIT KFW registry configuration for automatic
3720 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3722 * Fix (once again) case-sensitive comparisons which was apparently
3725 * Activate support for vos listvol -format
3727 * Population of the "AFS Client Admins" group failed on non-English
3728 versions of Windows because the name of the "Administrators" group
3729 is localized. Now we lookup the name of the group by using the
3730 Administrators Group Alias Relative ID.
3732 * Multiple cell token acquisition within afscreds.exe was broken in the
3733 case of Kerberos 5 cross realm authentication. Instead of contacting
3734 the KDC associated with the cell's realm, afscreds would obtain a
3735 token from the KDC of the user's realm. This would result in
3736 "invalid kvno errors" while authenticating to the AFS servers.
3737 Unauthenticated access would work. The symptoms would vary based upon
3738 whether or not the VLDB servers had been contacted using unauthenticated
3739 connections prior to the user obtaining tokens.
3741 * The list of ACL entries was becoming corrupted. It appears as if
3742 an ACL when it expires was not being appropriately cleaned up. In
3743 fact, it was left in the list of ACLs associated with the scache
3746 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3747 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3748 and "amd64_win64" for amd 64-bit processors.
3750 * When executing executables, Windows provides a mechanisms by which
3751 the normal search for DLLs can be bypassed. If a file foo.exe is
3752 being executed and there exists a file or directory "foo.exe.local"
3753 (the contents are ignored if a file), then Windows will search for
3754 DLLs first in the "foo.exe.local" directory and second in the
3755 directory in which the "foo.exe" file is located.
3757 Previous releases of OAFW would improperly return
3758 STATUS_NOT_A_DIRECTORY instead of the expected
3759 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3760 to terminate the search for the DLL causing the execution of the
3761 application to fail.
3763 * Rework the reference counting of smb_vc_t objects. The references
3764 stored in the waiting locks were not counted. This could result in
3765 an assertion if the reference count drops to zero.
3767 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3768 when running in freelance mode.
3770 * Added support for persistent caching of file contents, stat entries,
3771 volume data, cell data, and name lookup data. When the registry
3772 value, "NonPersistentCaching", is not defined or set to 0 the file
3773 specified by "CachePath" becomes a persistent cache file. The
3774 size of the cache file is computed at run time. It has a maximum
3775 size of approximately 1.2GB. The contents of the file will be
3776 validated according to the rules specified by the "ValidateCache"
3779 * Added support for UUIDs. UUIDs are kept for as long as the
3780 cache file is intact.
3782 * Added cmdebug.exe and support cache manager debugging callback
3783 interfaces in afsd_service.exe.
3785 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3786 [-addrs] [-cache] [-help]
3787 Where: -long print all info
3788 -addrs print only host interfaces
3789 -cache print only cache configuration
3791 * Symlinks to \\AFS\[all\]... will now be treated the same as
3792 symlinks to /afs/.... However, please use /afs/... as otherwise
3793 the symlinks won't work on Unix.
3795 * Correct a problem with local CellID allocation for cells whose
3796 server list is obtained via DNS instead of a CellServDB file. If
3797 the DNS information expires the CellID assigned to the entry will
3798 be changed. This causes all of the cm_scache_t objects which refer
3799 to the old cellID number to become useless. Attempts to access
3800 files or directories with cm_scache_t objects using the old cellID
3801 will fail since the server list cannot be obtained.
3803 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3804 be triggered if the volume referenced by the cm_scache_t object is
3807 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3808 in addition to SERVICE_ACCEPT_STOP.
3810 * Move RPC shutdown until after the SMB and RX shutdown procedures
3811 complete. Block until RPC shutdown is complete.
3813 * Modify afslogon.dll (integrated logon) to wait for service if its
3814 state is START_PENDING. If the timeout period occurs, reset to
3815 the retry interval and not the sleep interval.
3817 * When renewing the server list for a cell obtained via DNS AFSDB
3818 records, the cm_cell_t entry must be removed from the list of all
3819 cells. Otherwise, the list of cells will be corrupted.
3821 * In the dcache and scache modules, use the cm_scache_t dataVersion
3822 instead of the cm_fakeDirVersion.
3824 * Synchronize fs.c with the unix version.
3826 * The variable used to determine whether a file or virtual memory
3827 mapped cache is used was not properly initialized to a default
3828 value. If the registry setting "NonPersistentCaching" was not
3829 set, the choice would be random. Properly initialized to be
3832 * The memory mapped view was never unmapped before closing the file
3833 at service shutdown. This is now properly cleaned up.
3835 * Default location of Cache file is now %TEMP%\AFSCache
3839 * A bug affecting new installations of 1.3.75/76 would result in
3840 the creating of incorrect mountpoints in the freelance root.afs
3841 volume for the default cell.
3843 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3844 of "root.cell", you have been affected by the bug. To correct the
3845 problem, execute the following commands:
3847 fs rmmount \\afs\all\<cellname>
3848 fs rmmount \\afs\all\.<cellname>
3850 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3851 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3855 * A bug has been fixed in the auto-generation of Freelance root.afs
3856 symlinks which produced random entries in the \\AFS\all directory.
3858 * Support has been added for multi-homed servers
3861 * Added a new registry value, "StoreAnsiFilenames", which can be used
3862 to force the use of ANSI character sets instead of OEM Code Pages.
3863 This feature is useful when users require the ability to create
3864 filenames with 8-bit characters and need to access the files from
3865 both Latin-1 based Unix systems as well as from Windows.
3867 Activation of this feature will prevent access to files stored with
3868 8-bit OEM characters.
3870 * Shutdown all SMB threads in a synchronized manner when stopping the
3873 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3874 by the largest contiguous block of unused memory within the 2GB process
3875 space which can be assigned to the memory mapped file. Unfortunately,
3876 when the executable digital signature verification code is activated
3877 Windows sees fit to further segment the process memory which in turn
3878 reduces the size of the maximum cache file to less then 800MB. If
3879 larger cache sizes are desired, a new registry value should be set:
3881 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3883 Setting this value will disable the runtime verification of digital
3884 signatures on afsd_service.exe and the afs dlls which it loads. It
3885 will not disable the the version number check on those same files.
3886 The signature verification is not a security messure and is only meant
3887 to enhance the ability to afsd_service.exe to detect potential
3888 destablizing mixtures of DLLs from incompatible distributions.
3890 Added code to auto-disable the signature verification check if
3891 the desired cache size is greater then 700MB.
3893 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3894 used to verify the validity of the afsd_service.exe binary
3895 as well as each of the AFS DLLs loaded by the service. Not only
3896 must the digital signature be valid but the signatures of the
3897 DLL must be signed by the same entity as the service.
3899 * Implement new functions: cm_freelanceMountPointExists and
3900 cm_freelanceSymlinkExists. Use them along with other validity
3901 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3902 ensure that name collisions do not occur and that empty strings
3903 are not valid file names.
3905 A symlink may not have a name which would resolve to a valid
3906 cell name. Doing so would prevent access to the cell.
3908 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3909 re-initialization code. The reference counts of the fake root.afs
3910 volume scache object(s) would become invalid when the mount point
3911 or symlink lists were altered.
3913 * Add registry entries to provide mappings from the afsdsbmt.ini
3914 to the new locations for applications which count on the use
3915 of the old Profile file APIs. These apps are likely to fail
3916 if the user does not have administrator privileges and the
3917 registry is locked down.
3919 * The afs_config.exe submounts dialog had two errors.
3920 First, attempts to remove entries failed because the registry
3921 key was being opened without KEY_WRITE privileges.
3922 Second, when editing a submount entry, changing the name
3923 would add a new key and leave the original one in place.
3924 Now the original submount will be removed if its name is
3927 * In recent months there have been several incidents in which
3928 users have experienced problems starting or accessing
3929 afsd_service.exe and after significant effort has been spent
3930 it has turned out that they have two versions of AFS on the
3931 machine or an inconsistent set of DLLs.
3933 Code has now been added to afsd_service.exe which will walk
3934 the list of modules loaded by afsd_service.exe and validate
3935 that the version of the AFS DLLs matches the version of the
3936 afsd_service.exe executable. If they do not match the service
3939 * When Freelance mode is enabled and there is no registry
3940 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
3941 will attempt to import the afs_freelance.ini file contents.
3942 If the file does not exist, it was creating a dummy file
3943 with a r/o and r/w entry for the default cell and then
3944 importing those values.
3946 This process has been changed. The temporary file is no
3947 longer created. Also, both the OpenAFS Client install
3948 directory as well as %WINDIR% are checked for previous
3949 afs_freelance.ini files.
3951 * Added support for VL_GetEntryByNameN(). Still need to add
3952 support for VL_GetEntryByNameU() for multi-homed support.
3954 * Fix a deadlock situation in afscreds.exe when canceling an
3955 auto-generated Obtain Tokens dialog
3958 * Fix the locking of objects during Directory Searches in the
3959 SMB/CIFS server. The failure to properly lock the reference
3960 counts was resulting in the premature freeing of smb_dirSearch_t
3961 objects while they were still in use by the SMB/CIFS client.
3962 This does not solve the "Invalid Handle" problem.
3964 * Fix Find Cell By Name pioctl call to return a valid cell
3965 name for the Freelance fake root.afs volume.
3966 "Freelance.Local.Root".
3968 * Fix the Explorer Shell Extension Symlinks->Add operation.
3969 The dialog template was missing and the link destination
3970 string was too short.
3972 * Add support for symlinks to Freelance root.afs volume
3973 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
3974 <number> = "<linkname>:<relative-path>."
3975 Use symlink.exe to create, list, or remove
3977 * Remove the fallback to the use of KFW's KRB4 library when obtaining
3978 tokens. We never obtain KRB4 tickets.
3980 * Fix AFS Client Configuration Control Panel to support new SysName
3983 * Fix a bug in afsd_service.exe which could result in the SysName
3984 not being read from the registry.
3987 * Add code to block the issuance of AFS tokens by aklog.exe or
3988 afscreds.exe when the Kerberos 5 principal name contains a dot.
3990 * Modify the IsAdmin() function to always treat the local SYSTEM
3991 account as an AFS client administrator. Affects fs.exe and
3994 * Modify the internal handling of Quota Exceeded errors
3996 * Upgrade all reference count fields in the Windows cache manager
3997 and the osi library to use unsigned long instead of signed short.
3998 A similar fix has been applied to the afs rpc (rx) library.
4000 * fix the Windows cache manager to prevent it from replacing the
4001 rx_connection object associated with the cm_conn_t object on each
4002 and every operation if "fs crypt" was set. This explains the
4003 dramatic performance difference when crypt is used vs clear.
4004 The problem: 'cryptall', a boolean flag indicating whether or not
4005 "fs crypt" is set, was being compared to the rx_connection
4006 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
4007 1 != 2 and therefore the rx_connection was always destroyed
4008 and replaced on each and every operation.
4010 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
4011 It is not safe for the cm_conn_t object to not be locked because
4012 rx_DestroyConnection might be called from another thread if:
4013 - the user's tokens have changed or expired
4014 - the crypt mode has changed
4016 This fix appears to have also taken care of the problems associated
4017 with Overlapped Writes resulting in Delayed Write errors.
4019 * fix NSIS installer's AdminGroup.exe to properly create and
4020 remove groups when given -create or -remove. The string comparison
4023 * fs sysname now accepts a list of sysname values
4025 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
4026 DWORD which when set to a non-zero value will cause error message
4027 text to be output to stderr from the pioctl() routine. Useful in
4028 debugging failures of fs.exe, tokens.exe, etc.
4030 * added a test to the power management code to only perform a
4031 flush operation if there is at least one network adapter which
4032 is not a loopback adapter.
4034 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
4035 "EnableKFW". This value will not be read if the key
4036 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
4037 value under that key does not.
4039 * provide mechanisms to force the use of krb524d for Kerberos 5
4040 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
4041 there is a new registry value "Use524" and for aklog.exe a new
4042 command line parameter "-m".
4044 * Fix the pattern matching algorithm to properly match patterns
4047 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
4048 which is used by both the original function and the new
4049 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
4050 creation of HardLinks in addition to Renaming. smb_Link()
4051 is a new function which creates HardLinks via cm_Link().
4052 cm_Link() is a new vnodeops function which creates links
4055 smb_ReceiveNTRename() does not support the File Copy and
4056 Move Cluster Information operations described in its interface.
4057 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
4059 * When opening files via symlinks, we should follow the symlinks
4060 until we reach the actual file stat cache entry. The stat cache
4061 entry of the file should then be stored in the FID instead of
4062 stat scache entry of the symlink.
4064 * return bad operation errors for all unimplemented functions
4065 even if we do not know the functions exist.
4067 * Log bad packets and unknown operation packets to the trace log
4069 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
4072 * Update list of known CIFS operations to include all those listed
4073 in CIFS-TR-1p00_FINAL.pdf.
4075 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
4076 to support the REG_EXPAND_SZ type.
4079 * A new Windows authorization group "AFS Client Admins" is now
4080 created and populated with the members of the "Administrators"
4081 group. The group is used to determine which accounts on the
4082 machine may be used to modify the AFS Client Configuration via
4083 the UI and command line tools. afs_config.exe, fs.exe,
4085 * Modify the WinLogon Logoff Event Handler to query NT4 domain
4086 controllers for the remote profile path if Active Directory
4087 services are not available.
4089 * Fix aklog.exe to not add the AFS ID to the username
4091 * PTS registration of new users to foreign cells has been added to
4094 * The cm_Daemon thread is used to perform checks for
4095 down servers, up servers, volumes, callback expirations,
4096 lock maintenance and token expiration. Due to a gaff in
4097 larger integer division the thread never performed any
4098 work. Instead the current time computation would always
4099 be less then the trigger times. This had an adverse affect
4100 on the client's ability to maintain communication with servers,
4101 keep volumes up to date, and flush user tokens and acls
4102 when they have expired. This was broken when the 1.3 branch
4103 was modified to support VC7 which no longer included
4106 * An initialization problem with the Freelance code was
4107 detected while fixing the callbackRequest. The cm_rootSCachep
4108 object is obtained during afsd_InitDaemons() but the callback
4109 information is incomplete. The callback information will not
4110 be obtained until cm_MergeStatus is called from within
4111 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
4112 test for the conditions under which the callback information
4115 * Reports have been filed indicating that callbacks were
4116 being lost. An examination of the code indicated that the
4117 cm_server_t objects were not being properly reference
4118 counted by the cm_scache_t and cm_callbackRequest_t objects.
4119 In particular, the cm_server_t objects may have been freed
4120 from beneath the cm_conn_t objects.
4122 All of the reference counting is now done via the functions:
4127 this improves the ability to track the referrals.
4129 Each cm_BeginCallbackGranting Call now allocates a reference
4130 to the cm_server_t. The cm_EndCallbackGrantingCall either
4131 frees the reference or transfers it to the cm_scache_t
4132 cbServerp field. These are then appropriately tracked
4133 through the cm_Analyze call.
4135 * Ensure that the dnlc hash table is the same size as the
4136 dir name hash table (as per original author's note).
4137 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
4138 for compatibility with 64-bit systems.
4140 * fix smb_ApplyV3DirListPatches to properly apply the hidden
4141 attribute to dotfiles when the infoLevel < 0x101 and
4142 cm_SyncOp has failed.
4144 * Fix the Freelance registry initialization code. There
4145 was a possibility that some systems could end up with
4146 garbage in the registry during a clean install.
4149 * file and directory names beginning with "." will now be given the
4150 hidden attribute when the volume access is anonymous. this matches
4151 the behavior when the volume access is via an authenticated user.
4153 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
4154 key. When a change occurs mark the root.afs data as invalid and
4155 for it to be reloaded on the next access. This allows administrators
4156 to modify the mount point list without restarting the service.
4158 The freelance client used to provide a fake modification time for
4159 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
4160 Added code to extract the last modification time of the Freelance
4161 registry key and use that instead. The time now represents the
4162 most recent mount point change.
4164 * PTS registration of new users to foreign cells has been added to
4167 * Additional Cache Control and Credential Manager options have been
4168 added to the WiX installer. See deployment guide for details.
4170 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4172 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4175 * The loopback installation code may have had a problem updating the
4176 %ETC%\HOSTS file which could have resulted in a premature failure.
4177 Work around code has been added for the case where the file cannot
4180 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4181 because Windows sends 64K blocks when using overlapped writes.
4183 * The default number of server threads was increased from 4 to 25 to
4184 better handle overlapped writes.
4186 * The "AfscredsShortcutParams" registry value was not being properly
4187 loaded by afscreds.exe. Therefore, the default value was always being
4188 used instead of the value set by the installer.
4190 * Windows XP provides downgrade attack detection to prevent an attacker
4191 from being able to force the use of NTLM simply by disrupting
4192 communication with the KDC. This attack cannot exist between the
4193 Windows CIFS client and the AFS Client Service. Therefore, when a
4194 downgrade has been detected the afs pioctl library will force the
4195 establishment of a new CIFS connection using NTLM.
4197 * A locking error was discovered surrounding all references to volume
4198 server lists within the cm_cell.c source file.
4200 * The logged into Windows username was incorrect on Terminal Server
4203 * A new registry value "NonPersistentCaching" was added to the service
4204 parameters key. When set to a non-zero value, the afs cache is stored
4205 in the Windows paging file. There are two limitations to choosing
4207 1. when persistent caching is implemented it won't work with
4208 this flag set since there will be nothing to persist.
4209 2. with this flag set the initial paging allocation cannot be
4210 changed while the service is running
4212 * An initialization bug was discovered in aklog.exe which affected users
4213 who have a domain name for their afs servers which could not be mapped
4217 * afs_config.exe now validates cell names against DNS in addition
4218 to the CellServDB file.
4220 * In order to allow the freelance client to connect to a volume with ID
4221 equal to 1 on the default cell we changed the fake root.afs volume ID
4222 once again. This time we choose 0xFFFFFFFF. In addition, we change
4223 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4224 It will now be impossible for a volume ID to match that of another
4225 cell unless the client is connected to 0xFFFFFFFD cells. That should
4226 be enough room for growth.
4228 * Fix "fs mkmount" command to work with UNC paths and when
4229 started from non-AFS drives. It is now possible to create a mount
4230 point in the freelance fake root.afs volume with the command
4232 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4236 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4237 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4239 * The algorithm used to re-attempt access to the servers associated with
4240 a volume has been altered to properly address the case in which all
4241 servers have been marked down. The previous algorithm did not reset
4242 the server's down flags so the servers were never actually retried.
4243 This caused a problem with active volumes if the network connectivity
4244 was lost as could be the case with a network cable removal, wireless
4245 drop, or laptop hibernation. With the fix volume access is restored
4246 almost instantenously when network connectivity becomes available.
4248 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4249 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4250 listing of AFS submounts and freelance mount points. Support for
4251 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4252 significantly enhances the behavior of AFS volumes within the Explorer
4253 Shell. For instance, "AFS" now shows up as server in the Explorer
4254 with each submount or freelance mount point visible as a share.
4255 The right click menu in each folder now works with full functionality
4256 on a consistent basis.
4258 * The network provider can be configured to have different behavior
4259 depending on the domain that the user logs into. These settings are
4260 only relevant when using integrated login. A domain refers to an
4261 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4262 local machine (i.e. local account logins). The domain name that is
4263 used for selecting the domain would be the domain that is passed into
4264 the NPLogonNotify function of the network provider. (see registry.txt
4267 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4268 "Authentication Cell" which may be used to specify a default
4269 authentication cell for afscreds.exe which is different from
4270 the default cell for the AFS Client Service daemon.
4272 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4273 afslogon.dll moved to %WINDIR%\System32\.
4274 New registry entries added to register the dll for Winlogon events.
4276 The logoff event will now force a call to ktc_ForgetAllTokens()
4277 using the context of the user being logged off as long as the
4278 user's profile is not loaded from within AFS. If the profile
4279 was loaded from AFS we can't release the tokens since the Logoff
4280 event is triggered prior to the profile being written back to
4281 the its source location. This is now performed in an XP SP2
4284 * Windows XP SP2 Internet Connection Firewall interoperability
4287 * The %WINDIR%\afsdsbmt.ini contains four sections:
4288 Submounts, Drive Mappings, Active Maps and CSC Policies.
4289 The Submounts and CSC policies are now stored in the registry under
4290 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4291 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4292 The Drive Mappings and Active Maps are stored in the registry under
4293 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4294 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4296 There is no automatic migration of this data as it would be impossible
4297 to consistently migrate data to user profiles which may not be active
4298 when the machine is updated.
4300 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4301 fake root.afs volume. For the same reasons as for the cellservdb file,
4302 this information should not be in %WINDIR%. This information is now
4303 kept under the registry key
4304 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4306 The data from the afs_freelance.ini file will be automatically
4307 migrated to the registry on first execution of afsd_service.exe
4309 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4310 troublesome for several reasons. One, it is confusing for those who
4311 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4312 Two, this file is not a Windows Profile formatted file. Three,
4313 applications should not be reading or writing to %WINDIR%. It causes
4314 problems for Windows Terminal Server.
4316 The new location of CellServDB will be the OpenAFS Client install
4317 directory which is by default C:\Program Files\OpenAFS\Client and can
4318 be determined by querying the registry for
4319 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4321 The existing afsdcell.ini will be migrated by the NSIS installer.
4322 The Wix installer must still be updated to do the same.
4324 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4325 High Security mode; and to add Terminal Services compatibility registry
4326 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4327 configuration files in %WINDIR%.
4329 * Add support for authenticated SMB connections. This will remove
4330 the need for high security mode in most situations. Both NTLM
4331 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4332 only NTLM can be used even though Kerberos is now supported. The
4333 reason is that it is not possible to construct a service principal
4334 which is unique to each individual machine.
4336 SMB Extended Auth does not work on XP SP2 unless one of two registry
4337 modifications are made:
4339 (1) To disable the check for matching host names on loopback connections
4340 set this key. This does not require a reboot:
4342 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4343 "DisableLoopbackCheck"=dword:00000001
4345 (2) To add the AFS SMB/CIFS service name to an approved list. This
4346 does require a reboot:
4348 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4349 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4351 afsd_service.exe will automatically add the current Netbios Name
4352 to the BackConnectionHostNames list and then temporarily disable
4353 the loopback check for one cycle of startup/shutdown of the service.
4354 We assume most folks do not start/stop without a reboot so this
4355 will be adequate in most cases.
4357 * Fix security hole in afslogon.dll which allowed passwords to be
4358 sent in clear text to the KDC in a misformed principal name.
4360 * Fix cm_GetCell() to properly handle expired dns entries
4363 * If Freelance mode is active and the afs_freelance.ini
4364 file does not exist, do not create an empty file.
4365 Instead create a file containing ro and rw mountpoints
4366 to the default cell using the standard conventions.
4368 * Modify the Freelance support to handle the ability
4369 to create rw mount points in the fake root.afs volume.
4371 * Changed the RPC mechanism used for token setting from
4372 named pipes to local. Use of named pipes can be restored
4373 by setting the environment variable AFS_RPC_PROTSEQ to
4376 Named pipes were required when a Windows 9x system was
4377 using a NT system in gateway mode which is incompatible
4378 with our use of local loopback adapters.
4380 * In afscreds.exe, if a username of the form user@REALM is
4381 specified and no password is specified, do not perform a
4382 kinit operation. Only perform the aklog functionality.
4384 * Add a new registry value which allows the number of processors
4385 on which afsd_service.exe executes to be restricted. Valid
4386 values are 1..numOfProcessors
4388 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4392 * A second MSI based installer option is now available.
4394 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4395 request tickets for user/instance@REALM instead of just user@REALM
4397 * Modify the Power Management Notify routine to wait for the Hard Dead
4398 timeout period instead of a fixed 19 seconds. With the longer timeout
4399 periods Hibernation and Standby could never succeed when network
4400 connectivity is not available.
4402 * The following fs.exe commands are now restricted to Administrator:
4403 - checkservers with a non-zero timer value
4406 - sysname with a new sysname list
4415 setting the default sysname for a machine should be done via the
4416 registry and not via "fs sysname".
4418 * NSIS installer adds options to install Debugging Symbols
4419 and the Microsoft Loopback Adapter; the user is now also
4420 given the ability to select the afscreds.exe startup options.
4422 * Build system modified to generate symbols for FREE (aka RELEASE)
4423 builds as well as CHECKED (aka DEBUG) builds
4425 * Sites which have a volume ID of 0x20000001 assigned to their
4426 root.cell volumes have been experiencing problems with accessing
4427 the root.cell volume of their cell when Freelance mode has been
4428 active. This was because 0x20000001 was assigned to the fake
4429 root.afs volume created by freelance. The fake volume id is
4430 now set to 0x00000001 to prevent conflicts.
4432 * The timeout logic in the AFS Client Service has been wrong
4433 for sometime. It is based on two different assumptions.
4434 First, the SMB client timeout is a fix value as was the case
4435 with OS/2 Lan Manager. This assumption is incorrect. The
4436 SMB timeout in Windows is a dynamic value computed based upon
4437 a fixed minimum timeout to which is added time based upon the
4438 size of the request and the performance characteristics of
4439 the connection. Second, it is the responsibility of the
4440 SMB Server to enforce the timeout requirements of the client.
4441 This is untrue. The SMB Server cannot be expected to know
4442 the requirements of the client. More importantly, if the
4443 SMB server uses the SMB client timeout as a value to restrict
4444 its behavior as an RX client, the performance characteristics
4445 of the local SMB session would be used to prematurely terminate
4446 WAN connections with significantly different performance
4449 The timeout logic has therefore been modified in the following
4451 . the Lan Manager Workstation (SMB) Session Timeout is used only
4452 as a basis for configuring the Connection Dead Timeout
4453 and Hard Dead Timeout values. The Connection Dead Timeout
4454 must be at least 15 seconds longer than the SMB Timeout
4455 and the Hard Dead Timeout must be at least double the
4456 Connection Dead Timeout.
4457 . New registry entries have been added to allow the Connection
4458 Dead Timeout and Hard Dead Timeout values independent of the
4459 Lan Manager Workstation Session Timeout
4460 . The test to enforce the SMB Client Timeout has been removed.
4462 One of the side-effects of removing the enforcement of the SMB
4463 Client Timeout is that regardless of whether or not the SMB client
4464 is available to receive the response (and how would the SMB server
4465 know) the RX protocol response can be used to update the AFS
4466 Client Service state for ready access by future SMB client
4469 This should be the end of the "Server paused or restarting messages"
4471 * Add support for arbitrary UNC paths to the pioctl() support.
4472 This enables the fs commands as well as the AFS Shell Extension
4473 to work correctly when UNC paths are being used.
4475 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4476 DNS if the cell configuration cannot be determined via CellServDB
4478 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4479 are ever set more than once at a time
4481 * Fix the management of lists of cm_cell_t structures when using
4482 DNS to lookup cell information. The previous code would fail to
4483 reuse the same cellID for a cell if DNS was used more than once
4484 for a given cell name. When the ttl expired, a single cm_cell_t
4485 could be inserted into the cm_allCellsp list more than once
4486 producing a loop. In addition, the vlServerp list belonging to
4487 the cell was not freed resulting in improper refCounting of the
4490 * Add DNS support to cm_IoctlNewCell() which previous only examined
4493 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4494 to properly garbage collect cm_server_t objects
4496 * Add numVCs variable to smb.c to track the number of smb_vc_t
4497 objects created and use it to initialize the vcID field which
4498 previously was set to 0 in all objects resulting in FindByID
4501 * Fixed DNS lookups to work consistently throughout the OpenAFS
4502 product instead of just from within the afsd_service.exe
4504 * Added a runtime check to ensure that AFS Client Service SMB
4505 Server is accepting connections before attempting to mount
4508 * Read IP addresses for volume servers out of the CellServDB
4509 file if gethostbyname() on the hostname fails.
4511 * Fix getcellconfig() to populate both the Hostnames as well
4512 as IP addresses when loading cell data via DNS
4514 * Increase the Connection Dead Time to 50 from 20 seconds
4515 Increase the Hard Dead Time to 120 from 40 seconds
4516 (matches the Unix values)
4518 * Fixed an assertion validating the number of allocated NCBs
4520 * Fixed the build environment to consistently build for
4521 Windows 2000 and above. (APPVER = 5.0)
4523 * Fixed rx_debug to properly validate the receipt in incoming
4524 data with select() and recvfrom(). Do not copy data out of
4525 the socket buffer unless success is indicated.
4528 * afsd_service.exe will now display a message box to the
4529 desktop when it terminates due to an IP Address Change.
4531 * installer no longer deletes AFS Server configuration data
4534 * installer generates a warning dialog if the RPC service
4535 is not properly configured
4537 * installer compressed with lzma instead of bzip2
4539 * afsd_service.exe shutdown crash solved once and for all
4541 * reference counting of smb_vc_t data structures improved
4543 * name space collision of smb_fid_t event objects corrected
4545 * the output of "fs memdump" is now written to
4546 %WINDIR%\TEMP\afsd_alloc.log
4548 * the file TaAfsApp_1033.dll is now properly installed allowing
4549 the User Manager to start
4551 * a new algorithm is used for computing filename pattern matches
4553 * afscreds.exe now accepts user names containing instance
4556 * Fix the Directory Name Lookup Cache to be case-sensitive.
4557 This is crucial in environments in which a Windows client
4558 is accessing a directory with more than one filename that
4559 differs only by case. If the directory contains "FOO"
4560 and "Foo". You want "DEL Foo" to delete the correct one.
4561 We still have a problem in that "DEL foo" will delete a
4562 random filename. This will be addressed in a future release.
4564 * Fix afscreds.exe -M option (renewMaps) to work when High
4565 Security mode is off. Also, remember to disable the ActiveMap
4566 flag in afsdsbmt.ini when a drive mapping is removed.
4568 * Updates to NSIS installer script. AFS Server configuration
4569 data will not be destroyed on un-install or re-install.
4570 Use a better compression algorithm.
4572 * afslogon.dll now uses KFW to obtain tokens when available
4574 * afslogon.dll when given an all uppercase username will
4575 attempt to authenticate with both the uppercase name
4576 and an all lowercase variation
4578 * DST modification removed. The fix appears to make things
4579 worse after a reboot of the machine.
4581 * fs.exe: added "cscpolicy" which is used to
4582 change client side caching policy for AFS shares
4584 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4585 [-documents] [-disable] [-help]
4587 * Several uninitialized variables have been initialized
4589 * It is now possible to obtain tokens using cross realm
4590 Kerberos within afscreds.exe:
4592 user: jaltman@ATHENA.MIT.EDU
4594 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4595 will will in turn be used to obtain afs@DEMENTIA.ORG.
4596 The resulting token will be stored with the display name
4597 jaltman@ATHENA.MIT.EDU@dementia.org
4599 * aklog.exe has been added to the client
4601 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4602 [[-p | -path] pathname]
4606 -d gives debugging information.
4607 krb_realm is the kerberos realm of a cell.
4608 pathname is the name of a directory to which you wish to authenticate.
4609 -noprdb means don't try to determine AFS ID.
4610 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4611 (default is Kerberos V)
4612 No commandline arguments means authenticate to the local cell.
4615 * All of the resource files have been restructured to adhere to
4616 a set of rules IBM implemented for loading string resources.
4617 These rules had either been forgotten or were not discovered
4618 by folks working on the OpenAFS sources. The end result was
4619 memory corruption. This is primary item which was preventing
4620 the AFS Server from working.
4622 * Increased the size of the maximum ticket size stored in a token
4623 from 344 bytes to 12,000. Increased the buffers used to convey
4624 messages between the pioctl() caller and the SMB Server from
4625 1000 bytes to 12,512. The code appeared to have been writing
4626 above the top of the stack by quite a few number of bytes.
4627 (The increased ticket size is necessary for the next item.)
4629 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4630 Instead the raw Kerberos 5 ticket is used in its entirety. This
4631 is extremely important as it allows us to use pure Kerberos 5 KDCs
4632 as the source of the AFS authentication. The use of up to 12,000 byte
4633 tickets will allow tickets produced by all versions of Microsoft
4634 Active Directory to be used.
4635 - create a user account.
4636 - designate it DES only
4638 - specify its UPN to be "afs@realm"
4639 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4641 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4642 is special when using long file names. (you must use "*.*" and not "*")
4643 Instead only enforce it when performing 8dot3 searches.
4645 * Fixed the DST problem with creation times being set one hour ahead
4647 * Fixed the problem when using \\afs\cell-alias. For example,
4648 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4649 for the alias name; instead simply expand the name. One of the
4650 symptoms of this problem was a loss of acquired tokens.
4652 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4653 of strings. (Only English strings provided.)
4655 * Fixed the installer to properly replace in use files.
4657 * Fixed the build system to cleanup generated component version files
4659 * The release build compiled with MSVC 6.0 compiler to avoid the
4660 afsd_service.exe shutdown crash. This does not solve the problem
4661 but simply avoids it for the time being.
4665 * fix afslogon.dll to not corrupt memory when High Security mode
4668 * fix afsd_service.exe to not attempt to restore the stack when
4669 an exception occurs. (not safe in multi-threaded programs)
4671 * fix uninstaller to properly remove the CRT and MFC DLLs
4673 * remove a Message Box from afscreds.exe when getcellconfig()
4674 fails on a kerberos realm which is not a cell
4676 The following is a list of changes to the OpenAFS for Window client
4679 * "fs setserverprefs" will leave afsd service deadlocked
4681 * "vos listaddrs" will core dump
4683 * installer sets the appropriate keys to support Integrated Logon
4685 * installer disables the "Find Lana by Name" functionality as it
4686 was causing headaches for many users
4688 * fix the intermittent crash of the power management thread when
4689 shutting down the AFS Client Service
4691 * optimizes the obtain drive mount list functionality which is
4692 executed every time the mount tab in afscreds.exe and afs_config.exe
4693 are refreshed. (this happens a lot)
4695 * fix the service shutdown logic. add the STOP_PENDING state
4696 and do not accept additional service events after we declare
4699 The following is a list of changes to the OpenAFS for Window client
4702 * flexelint was run against the source tree and hundreds (perhaps
4703 thousands) of corrections were applied to ensure prototypes
4704 were in use; types were used consistently; variables were
4705 initialized; unused variables were removed; etc.
4707 * A wide variety of instrumentation was added including the
4708 ability to produce a stack trace from within afsd_service.exe
4711 * Dynamic configuration of the RDRtimeout value based upon the
4712 LanMan Workstation Session Timeout
4714 * The mount root no longer needs to be called "/afs". This
4715 is now set by a registry value "MountRoot" within the key
4716 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4718 * The cell list is now only read out of afsdcell.ini when the
4719 file changes instead of each time a cell is resolved.
4721 * Thread synchronization was added to cm_server.c and ktc_nt.c
4723 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4724 calloc()/free(). The Global functions were needed on Windows 3.x
4725 but have caused a variety of problems on the Win32 platforms.
4726 Avoiding them is highly recommended by several Microsoft
4727 Knowledgebase articles
4729 * Support for Symbolic Links added to the AFS Shell Extension
4731 * Added a registry value "OverlayEnabled" to determine if
4732 Shell Extension Overlays should be enabled.
4733 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4735 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4736 separate trees for checked and free builds. Build system supports
4737 a custom directory src\WINNT\extra which can be used as a grafting
4738 location of organization specific additions to the build tree.
4740 * New installer built using NSIS 2.0.
4742 * Named all kernel objects in order to allow them to be monitored
4743 with tools such as SysInternals' ProcExp.exe.
4745 * Introduced new EventLog framework for AFSD
4747 * Introduced Power Management interface to AFSD for Standby and
4748 Hibernate modes to allow cache to be flushed prior to network
4751 * Utilize Win32 DNSQuery API instead of internal routines. This
4752 allows DNS SRV queries to be sent to all current domain name
4753 servers. Not just one specified in an INI file. DNS is now
4756 * "NetbiosName" registry value may be used to specify a fixed
4757 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4758 when the loopback adapter is in use. If you need to use the
4759 old notation with a loopback adapter installed specify a registry
4762 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4764 * Refactor all modules which depend on LAN Adapter and NetbiosName
4765 determination in a new library: lanahelper.lib. This allows for
4766 consistent behavior throughout the product.
4768 * Move the afsd.log and afsd_init.log files to the directory specified
4769 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4770 for services. Added the Date to the log entries.
4772 * New registry value "RxMaxMTU" used to limit the size of the RX
4773 packets sent by the AFS Client Service to the Server. In order
4774 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4775 size must be restricted to 1264 or smaller. The latest NSIS
4776 installer sets a value of 1260 by default.
4778 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4779 packets. This is not needed in order to work across the Cisco
4780 VPN but might be needed for other network environments. This
4781 value is not set by the NSIS installer.
4783 * New registry value "HideDotFiles" is used to apply the Hidden
4784 attribute to files whose names begin with a '.'. This value
4785 is set by the NSIS installer.
4787 * New registry value "MaxMpxRequests" allows the maximum number
4788 of multiplexed sessions to be configured at run time. This
4789 value is not set by the NSIS installer. The default value is
4792 * New registry value "MaxVCPerServer" allows the maxmimum number
4793 of VCs per server to be configured at run time. This value is
4794 not set by the NSIS installer. The default value is 100.
4796 * New registry value "AllSubmount" allows the "all" submount to
4797 be disabled by setting its value to 0x00.
4799 * Allow cells names to be valid mount points
4800 \\<netbiosName>\<cellname>
4802 * Store the active state of drive mappings in order for afscreds.exe
4803 to restore them upon startup
4805 * Add exception handling to generate a Stack Trace to the afsd_init.log
4806 file if one happens to occur.
4808 * Add lots of logging to help detect the cause of invalid SMB packets
4810 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4811 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4812 renewal of expiring tokens as long as afscreds.exe is running.
4814 * New afscreds.exe command line options:
4816 -M = renew drive maps
4817 -N = ip address change detection
4820 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4821 determines whether or not MIT Kerberos for Windows should be used
4822 to obtain tokens via Kerberos 5 tickets.
4824 * New registry value "AfscredsShortcutParams" in
4825 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4826 determines the command line parameters to be specified when "fixing"
4827 the AFS Shortcut in the user's startup folder.
4829 * The "ShowTrayIcon" registry value has been moved from
4830 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4831 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4833 * The <cell name> registry values used to store the token expiration
4834 reminders have been moved from
4835 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4836 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4838 * Obtain the Logon User Name from the Explorer key when available
4840 * new text document doc\txt\winnotes\registry.txt lists all registry
4841 values used by OpenAFS (excluding the AFS Server)
4843 * BUG: rx_securityClass objects were not properly reference
4844 counted and were never freed.
4846 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4847 would be generated. The existence of a server does not imply
4848 that it is not down. If all of the servers for a cell are down
4849 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4852 * BUG: the directory name lookup cache failed to free the entries
4853 in the cache when the name cache entries cycled. The entries
4854 in the cache would become dereferenced without being freed.
4856 * BUG: fs setserverprefs could be executed without Administrator
4859 * BUG: the number of allocated NCB objects (100) exceeded the number
4860 which could actually be waited upon by the kernel (64). Any objects
4861 which were utilized above the limit could never have event completions
4864 * BUG: smb_username_t objects were not being reference counted and
4865 were not properly freed.
4867 * BUG: smb_tid_t objects could under unusual circumstances be freed
4868 before they were no longer referenced.
4870 * BUG: smb_fid_t object pointer were frequently used even when
4871 their value could be NULL. They were not properly released and
4872 therefore they were never freed.
4874 * BUG: smb_packet_t data structures were not completely initialized
4877 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4878 Denied" because that causes the Explorer Shell to try again until
4879 access is obtained. Instead return "Remote Resources" which allows
4880 the shell to move on and treat the error as transient.
4882 * BUG: when initializing the NCBreturns structure, separate Event objects
4883 were created for each NCB although a single Event object was supposed
4884 to be shared by all.
4886 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4889 * BUG: smb_tran2Packet_t objects were not being properly referenced
4892 * BUG: directory path creation did not handle the case of multiple
4893 directories requiring creation in one attempt
4895 * BUG: SMB requests which required an Extended Response were ignored.
4896 This prevented some files from being written to AFS volumes.
4898 * BUG: character strings were being freed even after they were
4899 inserted into in use data structures
4901 * BUG: inconsistent usernames were used when High Security mode was
4902 enabled. (there is still much to do in this area)
4904 * BUG: pioctl() calls which require out of band RPC operations were
4905 susceptible to race conditions when performed by multiple processes
4907 * BUG: memory allocation and deallocation crossed instances of the
4908 C Runtime Library producing memory leakage and corruption in
4909 afscreds and the client configurator.