3 * Prevent the Explorer Shell extension from crashing if
4 symlink creation failed. (126406)
6 * A Rx level NAT ping has been implemented.
7 Add NatPingInterval registry value to
8 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
9 to permit Nat Ping to be enabled. The default value is 0 seconds.
11 * When a re-initialization is taking place, be sure to reset
12 cm_noLocalMountPoints to 0 in case someone deletes the "Freelance"
13 registry key out from underneath the service.
15 * Add krb5 error message translation to aklog, afscreds,
16 afslogon.dll, the network identity manager afs provider
19 * Mode bits aren't directly exposed by the Win32 API. We were leaving
20 them to default to 0777 when creating new files and directories.
21 This version introduces two configuration parameters;
22 'UnixModeFileDefault' and 'UnixModeDirDefault' which are DWORD
23 registry entries that are used to set the initial mode bits.
24 If the values are set to 0, then the behavior is identical to what we
27 * Minidump files are now produced with a timestamp appended
30 * An SMB request debugging monitor has been added. When activated
31 the monitor will automatically turn on trace logging if any SMB
32 request has required longer than 60 seconds to complete and will
33 then create a minidump every 60 seconds thereafter until the
38 * Restore use of DNS AFSDB and SRV records by kaserver clients.
42 * Avoid a potential Freelance deadlock during initial execution
43 of afsd_service.exe if the old ini file data has to be
46 * Three rx library corrections. (1) Idle data connection
47 processing could timeout if the send window filled and
48 took longer than the idle data timeout period for the
49 transmit window to re-open. (2) The transmit queue
50 could be emptied prematurely. A required check for the
51 queue being in use was forgotten. (3) The function that
52 is supposed to implement a wait for the transmit queue
53 to cease being busy failed to wait.
56 * Restore use of DNS AFSDB and SRV records which were
57 unintentionally disabled in 1.5.69
60 * Add a context menu to the NetIdMgr AFS Provider
63 * Prevent an empty directory Btree from being created and
64 marked as valid if cm_BPlusDirBuildTree fails.
66 * cm_BPlusEnumAlloc should not fail if the enumeration
67 contains zero entries.
69 * In cm_BPlusXXX functions, return ENOMEM if malloc() fails.
71 * Do not leak the cm_scache_t dirlock if cm_BPlusDirBuildTree
74 * fs examine should report owner and group ids as signed values.
76 * Protect buffers in smb_WriteData from simultaneous writes.
78 * Query network interface MTU values from the registry because
79 the IP Helper GetAdapterAddresses() fails to report a manually
82 * Improve performance by not dropping and reacquiring the rx_call
83 lock in rx_WriteProc* and rx_ReadProc*.
85 * When configuring the rx library, the network interface MTU
86 must be applied to both the maximum send size as well as the
87 maximum receive size as is done when rx_SetMaxMTU is called
90 There was no 1.5.67 release
93 * Convert afsauthent.dll to use the xdr_alloc and xdr_free
94 routines from afsrpc.dll in place of malloc and free.
95 This permits the two libraries to be built with different
96 C runtime libraries without crashing.
98 * Set the DOS Readonly attribute on a file/directory
99 whenever the unix mode combined with the mask 0200
100 is true. Previously there was a discrepency between
101 the mask used for testing for readonly behavior and
102 that used for setting the attribute.
104 * Disable AFSVolSync based .readonly "whole-volume callback"
105 support because the all file servers prior to 1.5.67
106 (and perhaps 1.4.12) do not properly assign a value to
107 the AFSVolSync structure in bulk status RPC responses.
109 * Improve the error output from aklog to output the value
110 from krb5 error_message() if the afs_com_err output
111 indicates an unknown value.
113 * Hold a lock on the cm_scache_t object when the smb
114 server is assigning a new clientModTime value.
116 * Permit custom build version numbers to be set
117 without modifying the src/config/NTMakefile.<platform>
120 * Add support for binary code signing with cross-signed
121 certificates and arbitrary versions of signtool.exe.
123 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and
124 do not permit them to be exposed to the smb redirector.
126 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT
127 to avoid confusion within the smb redirector.
129 * Fix the byte order assigned to port numbers associated
130 with AFSDB record lookups. They must be network byte
131 order not host byte order.
133 * Add dynamic server ranking based on RPC round trip
137 * Official support for Windows 7 and Server 2008 R2
139 * Improved Rx RTT computation following the lessons
142 * Prevent a file server bug (FetchData returning an
143 invalid length instead of zero) from causing an
144 "unexpected network error" when writing to files.
146 * Promote DNS SRV records as superior to DNS AFSDB
147 records. Support arbitrary port numbers for vldb
150 * Add "fs listacl -cmd" support.
152 * Add AFSVolSync based .readonly "whole-volume callback"
153 support. With this functionality, multiple objects from
154 a .readonly volume can have their status validated by
155 issuing a single RXAFS_FetchStatus RPC.
157 * Remove drive mapping functionality and service start/stop
160 * Remove drive mapping functionality from afs_config.exe.
162 * Adjust SMB error return codes to avoid returning
163 STATUS_TIMEOUT which results in the SMB redirector
166 * Network Identity Manager OpenAFS Provider now provides its
167 own "AFS lock" notification icon to report the status of
168 "have tokens, have no tokens, service not started, service
169 started but inaccessible". Hovering over the icon lists the
170 cells for which tokens exist (if any) and the OpenAFS version
171 number. Double-clicking executes the Network Identity Manager
174 * When merging the current status for an AFS object, ensure
175 that the volume status for that object is set to 'vl_online'.
177 * Permit cm_Analyze to request a retry even when the cm_req_t is
178 flagged CM_REQ_NORETRY if the RPC failed due to RX_CALL_DEAD.
179 The retry will force the use of a new connection.
181 * Add digital signatures to resource only dlls as required by
184 * Prevent pioctl calls from retrying indefinitely when a sharing
185 violation error occurs.
187 * Add {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab"
188 to restore access to drive mapping functionality in afscreds.exe
191 * Permit Rx minimum peer timeout, maximum receive window size and
192 maximum send window size to be configured via the registry.
194 * The MSI installer now sets the ARPINSTALLLOCATION as required by
197 * DNS lookups set the port numbers in network byte order within
200 * Replace use of the 16-bit compatibility API WinExec with
201 ShellExecuteEx in afs_config.exe and afscreds.exe. WinExec
202 is incompatible with Windows 7 / Server 2008 R2 UAC
206 * Short circuit background volume checks if a shutdown
207 or suspend request is received.
209 * Avoid contacting the file server in order to allocate
210 an empty buffer beyond the current length of the file.
212 * When a request for streams on a directory or mount
213 point object is received, do not offer a default stream.
214 This was affecting the saving of roaming profiles.
217 * Fix Windows 2000 compatibility. Do not call rand_s()
218 even if it is compatible with the C run time library
221 * Fix a data consistency error between the output of
222 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
223 workstation group name "AFS", and report server name as
224 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
226 * Enable executables to be run from \\AFS on Windows 7.
227 Return "Name not found" instead of "File not found" when
228 a directory or file name cannot be found.
230 * Prevent cache manager from marking file server "down" when
231 the data returned in response to either RXAFS_FetchData64
232 or RXAFS_StoreData64 is invalid.
234 * Add pioctl data validation to the AFS Explorer Shell extension.
237 * After 1.5.62 afslogon.dll would no longer crash but it
238 also would not load "domain" specific configurations.
241 * An uninitialized variable was present in the symlink
242 recursion detection loop.
244 * Reverse addition of unique per file GUID during
245 Create/Open responses. Inclusion of the GUID prevents
246 Cygwin applications (and others) from accessing //afs.
248 * Treat "filename::$DATA" as a synonym for "filename".
249 This addresses the recent report of roaming profiles
250 failing to save back to the file server.
252 * When given a choice between an error from rx_Write()
253 or RXAFS_EndStoreData(), use the rx_Write() error.
254 This ensures that over quota and access denied errors
257 * If an error occurs during RXAFS_StoreData operations,
258 apply that error to all of the buffers in the BIOD.
259 This ensures that failures will not be retried if they
260 will only fail again.
262 * Do not perform a DNS lookup for a cellname if the
263 cellname prefix is _._AFS_IOCTL_._. This will avoid
264 DNS lookups when the query is for
265 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
267 * Fix Freelance root.afs volume object callback processing.
268 Object synchronization and status merging bugs.
270 * Fix Explorer Shell "invalid parameter" bug introduced
273 * Log RXKAD errors to the trace log.
275 * Prevent simultaneous pioctl calls from multiple processes
276 from stepping on each other's data.
278 * Prevent simultaneous pioctl calls from crashing the
281 * Add data validation to pioctl output processing in
282 ktc_GetToken and ktc_ListTokens.
284 * Add data validation to pioctl output processing in
285 fs.exe and symlink.exe.
288 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
289 No longer will browsing \\AFS produce truncated share names.
291 * Registry specified server preferences did not result in
292 the CM_SERVERFLAG_PREF_SET flag being applied to the
295 * Avoid unnecessary DNS lookups of share names as cells
296 when it is known that the name cannot be a cell name.
297 Any name that does not contain a dot is skipped.
299 * When processing Dfs Referral requests, do not return an
300 error if the server is down or busy; if the volume is
301 offline; the cell vldb cannot be reached; etc.
302 These paths are still \\AFS paths and so the client
303 should not be sent to search elsewhere for resolution.
305 * CRITICAL: Prevent the cm_Daemon thread from terminating when
306 the machine enters the suspend state. IF the cm_Daemon thread
307 dies, there is nothing to execute down server checks.
309 * Prevent use of smb_StartedLock before initialization which
310 can result in a panic in the lock package.
312 * Improve error checking in afslogon.dll in order to prevent
313 NULL pointer dereferences within GetLogonDomainOptions()
314 if the Lsa operations fail.
316 * Add support to the build system to automatically update
317 a Microsoft Symbol Store.
319 * Do not access the cm_conn_t after executing the cm_Analyze
320 loops. Doing so can result in an access to memory that has
323 * CRITICAL: If a StoreData request offset is not aligned on the
324 buffer module blockSize, (offset % blockSize) bytes of dirty
325 data will not be written to the file server.
327 * CRITICAL: If a StoreData64 request is sent to a file server
328 that does not support large files (>2GB) and the client is
329 unaware that the server has no large file support, the first
330 chunksize worth of data will not be written to the file server.
331 This bug was introduced in 1.5.3. It affects all IBM AFS file
332 servers and all OpenAFS file servers older than 1.4.0.
335 * If a file server becomes inaccessible while the cache manager has
336 dirty buffers to write, the afsd_service buf_IncrSync thread can
337 attempt to use 100% of the cpu.
339 * Fix the locking that protects the cm_cell_t lists
340 (all, name, id, free).
342 * Remove cm_cell_t objects from the name and id hash tables before
343 placing them on the free list.
345 * Fix "fs newcell" which was broken in 1.5.60.
347 * Do not attempt to synchronize dirty buffers if the associated
348 volume is known to be unavailable.
350 * Log to the Windows Event Log dirty buffers that cannot be written
351 during the shutdown process.
353 * Modify behavior of a Freelance mountpoint target that does not
354 specify a cell. Instead of assuming the target volume is in the
355 Freelance.Local cell, use the workstation "Cell" specified in the
356 registry. A mountpoint target of "#root.cell." will now mean the
357 root.cell volume in the workstation cell for the current session.
358 If the workstation cell changes from "athena.mit.edu" to
359 "andrew.cmu.edu", the referenced volume will also change without
360 requiring that the mount point targets be altered.
362 * Avoid false infinite loop errors when validating the cm_cell_t
363 free list during startup.
365 * Fix two more locations where xdr_free() should be used instead of
366 free() to permit a checked built of afsd_service.exe to work with
367 release builds of afsrpc.dll.
369 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
370 to permit the server Uuid to be used to lookup the server object
371 and from that determine the cell. This permits callbacks that are
372 received from alternate addresses to be processed with a known server
373 object. Previously a request from an unknown server would clear all
374 callbacks from all cells.
376 * Fix a bug that prevented optimal performance when using a non-zero
377 value for 'daemonCheckVolCBInterval'. As a reminder, when
378 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
379 volume callbacks are automatically renewed 90 minutes before their
382 * Fix automatic ranking of vldb servers whose values are obtained from
385 * Add logging for RX CALL DEAD errors which are generated if the remote
388 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
389 or the call is to a vldb server.
391 * Add registry based cell search functionality to NetIdMgr,
392 afs_config.exe, and klog.exe.
394 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
395 lookups on the host names in the CellServDB instead of using the
396 specified IP addresses. This provides aklog, pts, vos, etc. the same
397 CellServDB behavior that the Windows Cache Manager uses.
399 * Completely a behind the scenes change. The snprintf() routines have
400 been completely rewritten.
402 * When updating the stat cache entry callback of a .readonly object
403 from the volume group object, update the file server reference to
404 ensure it matches the most update to date callback.
406 * Add proper support for processing callbacks from multi-homed file
407 servers. Instead of comparing servers by cm_server_t pointer,
408 compare them by UUID when the UUID is known.
410 * During a shutdown short circuit the offline volume check daemon
413 * Properly track link counts during cm_Link and cm_Unlink operations.
415 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
416 preference to an error code reported by rx_EndCall.
418 * Add "PerFileAccessCheck" registry value to permit testing against
419 experimental file servers that include per-file acl support. This
420 value is intentionally undocumented. It is not to be used by
421 production environment deployments.
423 * Log a warning to trace log if a volume group is marked as a DFS
424 File Set. The Windows OpenAFS client does not include DFS support.
426 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
427 provider from obtaining tokens when referrals are in play.
429 * Add "fs chown" and "fs chgrp" commands to permit the owner and
430 group of objects stored in AFS to be set from Windows.
432 * Avoid performing background daemon operations when the machine is
433 going into suspend mode.
435 * Perform offline volume checks in most recently used order.
437 * Prevent crash when a data version for a cache object goes backwards.
439 * fs setquota, fs setcachesize, vos setfields, and vos create now
440 accept human readable orders of magnitude. (K, M, G)
442 * fs listquota fixed to permit large quota sizes to be displayed.
444 * rx packet resend and data packets sent counts were incorrect.
446 * Multi-thread safe library versions are now being generated and used.
447 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
449 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
450 values are now available on XP through Windows 7. Add functionality
451 to autodetect if such support is present on the machine. If so,
452 configure it if necessary and dynamically adjust the AFS Rx timeout
455 * Correct another race condition in the Rx library that could result
456 in an unexpected panic while freeing the Rx call iovq.
459 * A fix to the pioctl library to support drive substitution
460 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
462 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
463 that corrects a deadlock in the smb redirector and also adds new
464 functionality that permits the AFS SMB server to be given a longer
465 timeout than is normally the case. New functionality has been added
466 to configure these additional LanmanWorkstation\Parameter values.
467 (This functionality has been backported to XP SP3 and is scheduled
468 to be released on June 5th.)
470 * The BackConnectionHostNames registry value configuration was broken
471 when dynamic re-establishment of Netbios Name registrations was added.
474 * Hidden vos commands are revealed.
476 * If the "DisableLoopbackCheck" registry value is set, do not unset
477 it during the same service session.
479 * Reorganize code that prevents multiple Store operations to the same
482 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
483 permit broken symlinks to be treated as being in AFS.
485 * Fix vos commands that output 64-bit integer values.
487 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
488 Permits building cygwin applications against OpenAFS libraries.
490 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
491 as Windows HTML Help (.CHM) files.
493 * NSIS installer does a much better job of cleaning up files left
494 over from previous installs.
496 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
497 or "fs flushall" and on-going directory operations that can result in
498 afsd_service.exe crashing.
500 * Add support for DNS SRV records in place of AFSDB records.Â
501 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
502 Weight is currently ignored.
504 * Add a method of specifying Client CellServDB information within the
505 registry that can be used to either override the CellServDB file or
506 force the use of DNS lookups for a given cell.
508 The registry schema is as follows:
510 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
511 "LinkedCell" REG_SZ "[cellname]"
512 "Description" REG_SZ "[comment]"
513 "ForceDNS" DWORD {0,1}
515 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
516 "HostName" REG_SZ "[hostname]" Default: [servername]
517 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
518 "IPv6Address" REG_SZ "[address]" <future>
519 "Comment" REG_SZ "[comment]"
520 "Rank" DWORD "0..65535" Default: 0
521 "Clone" DWORD "{0,1}" <future: server only>
522 "vlserver" DWORD "7003" <future>
523 "ptserver" DWORD ... <future>
525 ForceDNS is implied non-zero if there are no [servername]
526 keys under the [cellname] key. Otherwise, ForceDNS is zero.
527 If [servername] keys are specified and none of them evaluate
528 to a valid server configuration, the return code is success.
529 This prevents failover to the CellServDB file or DNS.
531 Only one of "HostName" or "IPv4Address" is required. "HostName"
532 is optional if [servername] is the host name.
534 * Extend registry based CellServDB functionality to
535 afsconf_GetCellInfo() interface used by aklog and many
536 other command line utilities.
538 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
539 as it is now used by both client and server components.
542 * PriorityClass of afsd_service.exe process raised to
543 "High" to match the priority of the system services
544 that are waiting for afsd_service.exe to respond.
545 Configurable using [TransarcAFSDaemon\Parameters]
546 DWORD "PriorityClass" value. (use values from Win32
549 * Change error reporting of CM_ERROR_BADSHARENAME to
550 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
551 The error we were sending would cause the smb
552 redirector to drop the connection.
554 * Change the error reporting of CM_ERROR_ALLOFFLINE
555 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
556 Previous we would return Bad Network Path which
557 would cause the smb redirector to cancel all file
558 handles on the same "share".
560 * A debug version of afsd_service.exe can now be used
561 with a release version of afsrpc.dll (or vice versa).
562 afsrpc.dll now implements and exports xdr_free().
564 * Removed all unprotected OutputDebugString() calls as
565 those calls can block for up to 10 seconds if there
566 is contention for the globally shared memory or if
567 the debugger fails to read data in a timely manner.
569 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
570 supported for builds. The production releases still
574 * Discovered that XP/2003 supports a variant on recvmsg
575 called WSARecvMsg. The input parameters are mostly
576 compatible with the Posix recvmsg. This will save
577 multiple memcpy() calls per received packet.
579 * Discovered that Vista/2008 supports a variant on
580 sendmsg called WSASendMsg. The input parameters
581 are mostly compatible with the Posix sendmsg. This
582 will save multiple memcpy() calls per sent packet.
584 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
585 and utilizing rx_ReadProc32 from within the cache manager
586 as part of the FetchData operations. rx_ReadProc32 and
587 rx_WriteProc32 are optimized versions of rx_ReadProc and
588 rx_WriteProc for use when reading 32-bit integer values.
590 * The Network Identity Manager AFS provider now reports
591 progress events. These provide a better user experience
592 with Network Identity Manager 2.0.
594 * [RT 124293] A race condition exists which permits the
595 scp field of the an smb_fid_t object to become invalid
596 after a request on the smb_fid_t has begun. Remove the
597 race by protecting all access to the scp field with the
598 smb_fid_t mx mutex and obtaining a local reference on
599 the cm_scache_t object for the length of the request.
601 * [RT 124276] If the vldb is out of sync with the contents
602 of the file servers, afsd_service will retry too many
603 times when a file server reports a volume as not being
604 present. Now if the list reported by the vldb is the
605 same as the previously seen list, then the retry is
608 * [RT 124276] Read-only volume failover was broken in
609 1.5.53 whenever accessing a volume results in VNOVOL or
612 * [RT 124276] Prior to 1.3.70 the volume server reference
613 list was not reference counted and would be prematurely
614 freed while in use. When reference counting was added
615 in 1.3.70 a bug was introduced that could result in
616 service reference list corruption.
618 * Add Windows Application Event Log warning messages for
619 "Client SMB MPX value too large" and
620 "Client SMB Buffer Size too small".
622 * Renaming of files across directory boundaries would result
623 in an invalid handle error when attempting to access the
624 files after the move.
626 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
627 when a smb file descriptor cannot be found. The whole point
628 of using a Path Info function is because an smb file
629 descriptor wasn't allocated.
631 * More edge cases in which dynamic addition of Freelance
632 root.afs entries would get the wrong FID or where the
633 root.afs directory would not be refreshed.
635 * Buffer overflow could occur if the workstation cell name was
636 longer than 64 characters. Crashes could occur in
637 afscreds.exe, afslogon.dll, and afsd_service.exe.
639 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
642 * Improve consistency of the bad data version value used for
645 * Log file server uuid values as part of the cm_server object
646 when available. Dump the cm_server object list in response
649 * Optimize the performance of resetting access control lists
650 when tokens are set or removed.
652 * Fix vos.exe so that it can work with very large volume
655 * Remove symlink recursion tests and increase max symlink count
659 * Undo the "UAC manifest fix" applied to afs_config.exe.
661 * Fixed some additional race conditions in the Rx library
663 * Ensure that Freelance allocation of vnodes follow
664 the AFS convention of odd vnodes are directories
665 and everything else is an even vnode.
667 * Add Freelance logic to mount point and symlink
668 evaluation functions.
670 * Enhance smb_ParseASCIIBlock() so that it can handle
671 all of the STRING formats defined by the CIFS Technical
674 * Validate the output of smb_ParseASCIIBlock() in all
675 callers. Return CM_ERROR_BADSMB if the STRING field
676 cannot be parsed. CM_ERROR_BADSMB will cause the
677 contents of the packet to be logged.
679 * If multiple SMB Raw Write operations were taking
680 place at the same time, there could be data corruption
681 because unique event objects were not generated for
682 each Netbios receive operation.
686 * Fix a race in rxi_WriteProc which could result
687 in a freed or in use packet being passed to
688 rxi_PrepareSendPacket if the call is currently
689 in a transmit wait state. Passing a freed packet
690 will result in a panic. Passing an in use packet
691 might result in a call failure or perhaps data
694 * Kerberos referrals broke the NIM AFS provider
695 logic for automated configuration of the workstation
696 cell with a new identity. Now if a referrals
697 response is returned, assume the cell belongs to the
700 * The interface to buf_CleanAsync() and
701 buf_CleanAsyncLocked() has been changed to permit
702 errors to be propagated to the caller.
704 * During an enumeration of a directory, if the btree
705 has not yet been generated, return an error.
707 * When mapping CM_ERROR values to NTSTATUS values
708 do not treat 0 as an error.
710 * Avoid a cm_scache_t object reference count underflow
711 when renaming via the SMB interface.
713 * smb_IsMask() can be called with a non-8.3 input buffer.
714 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
715 input buffers and smb_IsMask() for non-8.3 input buffers.
717 * Add cm_buf_t error state to the "fs memdump" output.
719 * pioctl drive mapping checks need to take into account
720 Global Drive Mappings and SUBST assignments.
722 * Rename operations should not be permitted if the
723 destination name collides with an existing name that
724 differs only by case UNLESS the collision is the object
725 that is being renamed.
727 * Increase the maximum number of background daemon threads
728 to 64. Leave the default at 4.
730 * Set the FILE_UNICODE_ON_DISK flag when reporting the
731 file system characteristics.
733 * Report the maximum component name length as 255 even
734 for Unicode as part of the file system characteristics.
736 * Add RX debugging information to "fs memdump" output
739 * Refcount the cm_server_t objects returned by
742 * Fix the UAC manifest applied to afs_config.exe
744 * rxdebug now indicates if any RX_CONN flag is set
745 not just some of them
747 * Change the mapping of CM_ERROR_WOULDBLOCK from
748 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
752 * Includes #GCO Public CellServDB 07 Nov 2008
754 * Prevent separate cell entries from being created for
755 "foobar.org" and "foobar.org." which can happen if
756 an AFSDB record is found instead of a CellServDB
759 * Fix a panic when checking the SMB virtual circuit list.
760 If a VC is destroyed while the list is being checked
761 it is possible for the next object to become invalid.
763 * Increased logging to afsd_init.log of netbios name
766 * Fix a DNLC panic when a UTF-16 string that cannot be
767 converted to a utf-8 string is provided as a filename.
769 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
770 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
771 location to refer directly to the new location in
772 %programfiles%\openafs\client\program
774 * Be more aggressive about responding to Netbios
775 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
776 the LANA within the listener thread without performing a
779 * Use the WNet Enumeration of all connected drives as a filter
780 in the pioctl GetIoctlHandle operation. This should remove
781 the last situations in which there was a delay when opening
782 an Explorer Shell properties menu with the AFS Shell
785 * Fix the parsing of CellServDB in Windows to handle linked
786 cells as is done on Unix.
788 * Add support for linked cells to the cache manager,
789 netidmgr afs provider, and aklog. When two cells are
790 linked, a volume lookup that fails in one cell is looked
791 up in the other. This can be used for construction of
792 test cells that override specific volumes; cell renaming;
795 * Increase the length of cell and realm names that can
796 be entered in the NIM AFS provider.
798 * Actively prevent a 32-bit afsd_service.exe from
799 executing in the WOW64 environment.
801 * Protect against invalid IP address values in CellServDB
806 * Fix an Rx panic caused by an attempt to double free
807 an rx_packet when an error occurs while writing the
810 * Fix an exception cause by a race condition when
811 processing multiple file change notifications registered
812 by the Explorer Shell or other applications.
814 * Modify the Rx debug interface to use var_args and
815 permit Rx debugging to be used on 64-bit Windows
816 which previously had it disabled.
820 * Automated lock hierarchy enforcement added to the
821 osi base lock package. This permits detection of lock
822 violations that might but often does not result in
823 deadlocks. Enabled by default in the checked builds.
824 May be activated in release builds via the registry
825 value "LockOrderValidation" in the TransarcAFSDaemon
828 When a violation is detection an assertion is thrown
829 generating a dump file and a message indicating which
830 locks were obtained out of order is logged to the
831 Windows Application Event Log.
833 * Attempt to ensure that a dump file will be generated
834 whenever the service panics.
836 * Removed unused lock cm_bufGetMutex
838 * Many out of order lock acquisitions were detected
839 using the lock order verifier and were corrected.
841 * A race condition within the function path cm_GetSCache() ->
842 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
843 cm_scache_t object to be issued simultaneously to two
844 threads. This would eventually result in a panic due
845 to the resulting undercount.
847 * Fix interpretation of the empty string as the ioctl path
848 to mean the current directory. "fs lsm", "symlink list",
849 etc. now return a "not a ..." error instead of "does not
852 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
853 to indicate that the input string is a nul terminated string.
854 Assign ti when input strings are nul terminated instead of
855 length counted. (from 1.5.50)
857 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
858 does not match the observed behavor. The 'nameLength' is specified
859 in bytes, not in characters. Fix the implementation to match
860 the observed behavior. (1.5.50)
862 * The cm_HaveAccessRights() attempt at deadlock avoidance by
863 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
864 does not avoid the deadlock. Avoid the deadlock by enforcing the
865 lock order of the lowest vnode first. Then remove the infinite
866 loop avoidance in cm_SyncOp() that was returning an unwarranted
869 * File Server lock synchronization was not propertly enforced in
870 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
872 * Another change to the short name generator. Plus and equal are
873 not valid characters in a short name according to the
874 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
875 with left and right brace in order to minimize the impact.
877 * Fix a race condition that can result in a heap overwrite error if
878 the number of known file servers or vldb servers increases during
879 the preparation phase of a cm_CheckServers() operation.
881 * If the SMB Authentication would fail, a LSA allocated memory
882 buffer would be leaked.
884 * Rewrite the cm_cell_t allocation code to permit cell vldb server
885 lookups to fail without leaking cm_cell_t objects.
887 * Add "fs setcrypt auth" to permit the use of rxkad authentication
888 only mode. Not documented.
890 * Initialize Rx mutex and lock objects once and only once as part of
891 the DLL load process.
893 * When reporting the Volume Name "AFS" do not include the terminating
894 NUL in the reported length.
896 * Refresh Volume Location information upon receipt of a VNOVOL error
897 from the file server and avoid a race condition that prevented the
898 updated volume location information from being destroyed immediately
899 after it was acquired.
901 * In response to VNOVOL and VMOVED, remove the volume id from the server's
902 volume list so that the volume's up and down status is not affected
903 when the machine is shutdown.
905 * Many fixes to the Rx stack related to packet allocation and deallocation;
906 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
907 by overwriting the rx_call currentPacket pointer when it was already
908 referring to a packet.
910 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
911 in the TransarcAFSDeamon/Parameters key.
913 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
914 constructed directory B+ trees for flushed directory objects.
916 * When cached pages for directory objects have been locally modified, do
917 not permit the locally modified pages to be mixed with pages subsequently
918 obtained from the file server. Doing so can result in directory
921 * If Ubix errors are received in response to VL_xxx operations record
922 the error on the server instance and permit the operation request to
923 be retried in case another server is able to answer the query. We have
924 seen at least one instance of a cell in which a VL server returned
925 UNOQUORUM (possible database truncation) when the other servers were
926 responding successfully.
929 * Optimize the assignment of write errors to dirty
930 buffers by applying the first error to all buffers
931 belonging to the same object instead of trying each
934 * Add resource version info to xstat_fs_test and xstat_cm_test
936 * Fix a delete vs delete[] usage problem in the WiX custom
939 * When creating a mount point, validate the volume name to
942 * When processing a mount point, pay attention to VL_BADNAME
945 * Remove a race condition from cm_volume_t recycling that could
946 have resulted in two threads making use of the same object at
949 * Fix a reference count leak on smb_vc_t objects introduced in
950 1.5.50. Permit the objects to be freed once again.
952 * Fix the output of "fs memdump" for smb_vc_t objects that was
955 * Add routines for debugging smb refcount errors.
957 * When shutting down the cache manager, do not permit AFS
958 callback requests to be processed. Doing so can result in
959 invalid memory accesses.
961 * Do not recycle a deleted cm_scache_t object unless the
962 refCount is 0. Doing so can result in a deadlock and
963 invalid memory access.
965 * Minimize the number of RPCs sent to the file server for
966 directory bulk status requests. Only request status for
967 objects that we are actually reporting on and only if we
968 do not already have status for the object and then make
969 sure that we return as much as we can at once via the SMB
970 interface but not so much that it exceeds the RDR timeout
972 * Enable bos restricted operations
974 * Fix the creation of submounts used by the afscreds and afs_config
977 * Avoid repeated retries when attempting to obtain access rights.
978 If the access rights were obtained successfully but do not
979 satisfy the request, fail the request.
981 * Fix error code reported when attempting delete a file on a
982 readonly volume or one that is marked with the readonly DOS
985 * Fix a short name truncation bug.
987 * Fix a heap corruption error when reading the CellServDB file
990 * Fix directory listings when the directory has "lookup" permission
991 only and the directory has never been previously accessed.
993 * Fix several tracking issues for cm_scache_t rw locks.
995 * Ensure that all buffers used to store cell names are the
998 * Restore Windows 2000 support (broken in 1.5.50).
999 LOCALE_INVARIANT was introduced in XP.
1001 * Add RxUdpBufSize registry value. Default is 256K.
1003 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
1005 * Do not include trailing NULs when returning file names in
1006 a directory search response.
1008 * Prevent RX from crashing the afsd_service by attempting
1009 to allocate to many buffers at once when memory is fragmented.
1011 * Avoid a deadlock when freeing cm_server_t objects
1013 * Optimize search cell file. Do not search to end of file
1014 if the cell has already been found.
1016 * Remove explorer shell extension delay on local disks.
1018 * Automatically pre-allocate 64 Rx packets per thread
1020 * Add smb lock requests and stat cache lock allocations to
1021 the output from fs memdump.
1024 * Fix Large File Support (broken in 1.5.50)
1026 * Fix "fs setacl" (broken in 1.5.50)
1028 * Fix pioctl path evaluation of drive letter
1029 mapped paths. (broken in 1.5.50)
1031 * Fix installation of VC8 Runtime library in the EXE
1032 Retail installer. (broken in 1.5.50)
1034 * Avoid a potential deadlock during Volume Location
1035 updates if all the file servers are unreachable at
1036 the time the VL update succeeds.
1038 * Ensure that rx connection object is references
1041 * Fix propagation of over quota errors during cm_FSync()
1044 Since 1.5.39 [1.5.50 released 16 July 2008]
1045 * There were no releases between 1.5.39 and 1.5.50.
1047 * FIRST RELEASE WITH UNICODE SUPPORT
1049 * Pioctl interfaces to the cache manager refactored
1050 to provide layering between the SMB specific code
1051 and the general purpose ioctl operation.
1053 * Garbage collect dead SMB virtual circuits as soon
1054 as they are no longer being referenced. This avoids
1055 problems with outstanding locks not being dropped
1056 when the virtual circuit becomes invalid.
1058 * Remove the IBM Administration Reference documentation
1059 and replace it with the OpenAFS Command Reference
1062 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
1063 functions each time a connection is about to be used.
1064 Do not hold a lock on the rx connection object while
1065 it is being selected. This avoids a race between
1066 threads attempting to set the timeout values and
1067 removes a bottleneck that was hampering performance.
1069 * Ensure that the smb directory attribute is set
1070 for all directory objects.
1072 * Add vs2008 support to the NSIS installer scripts
1074 * Replace the VC Runtime EXE installer with the MSI
1075 installer in the NSIS installer scripts
1077 * Properly delete the VC7.1 C Runtime libraries.
1079 Since 1.5.36 [1.5.39 released 23 June 2008]
1080 * There were no 1.5.37 or 1.5.38 releases for Windows
1082 * Use Visual Studio version of vsprintf() instead of
1083 home grown version that resulted in linking conflicts
1085 * Prevent buserver, ptserver and fileserver from crashing
1086 due to non-portable strftime() format strings.
1088 * Prevent server crashes caused by rx_InitMorePackets()
1089 being called prior to allocation of all of the required
1090 synchronization objects.
1092 * Fixed a memory leak in the internal B+ tree directory.
1093 For objects with non-8.3 names, the 'longname' memory
1094 allocation would be leaked during object deletion or
1097 * Fixed a memory leak in the background daemon cm_CheckServer()
1098 routine. The memory allocated to store the server list
1101 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
1103 * Correct an abstraction layer violation. cm_scache_t objects
1104 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
1105 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
1106 performed in cm_ReleaseSCache() when the reference count hits zero.
1108 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
1110 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
1113 * Ensure that scache object 'nextp' pointers are set to NULL
1114 when the objects are removed from the hash table linked list.
1116 * If a dirty buffer cannot be written to the file server because
1117 of an access denied error, mark the error in the cm_buf_t
1118 structure so that it can be dealt with instead of retrying
1122 * Update CellServDB file to GCO Public 23 Apr 2008
1124 * Fix a cm_buf_t reference count leak when attempts to write
1125 dirty buffers to the file server from within cm_IncrSyncer()
1128 * Prevent udebug from crashing.
1130 * Activate RX Idle Timeouts. If the file server is busy for more
1131 then 30 seconds, attempt to failover to another server without
1132 marking the busy server down.
1134 * Another VNOVNODE issue fixed. When writing a dirty buffer
1135 to the file server, if VNOVNODE is received, mark all buffers
1136 as invalid without further attempts to contact the file server.
1138 * Improved performance on high latency links.
1141 * Fix a bug preventing the re-initialization of the Freelance
1142 root.cell contents when a change is made.
1144 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
1145 or mountpoint that does not exist.
1147 * aklog, afscreds, integrated logon, and the NIM AFS credential
1148 provider will all attempt to first acquire an afs service ticket
1149 using the service principal: afs/<cell>@<CLIENT_REALM>. This
1150 permits the proper detection of the cell realm when multiple
1151 realms are local realms for the cell.
1153 * Fix fs commands that re-use file identifiers instead of forcing
1154 a path evaluation each time. This was broken in 1.5.33 when
1155 hashes were added to the cm_fid_t structure.
1157 * Fix symlink and mountpoint removal and then recreation when
1158 applied to the fake Freelance root.afs volume.
1160 * Fix SMB change notifications for the Freelance root.afs volume
1161 when creating or removing symlinks and mount points.
1163 * Force DNS AFSDB cell name lookups to lower case if the DNS
1164 response is not all lower case.
1166 * 32-bit installers are now built with Visual Studio 2005
1168 * All vos.exe commands now support the "-noresolve" option which
1169 forces the display of IP addresses instead of DNS hostnames.
1171 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
1172 permits operations that do not require up to date volume
1173 location data to continue while waiting for the RPC to complete.
1174 Multiple threads waiting for a volume location update on the
1175 same volume set no longer trigger multiple calls VL_xxx RPCs.
1177 * Volume server references could be marked offline and never
1178 reset to the not_busy state.
1180 * Further optimizations to cm_Analyze() retry logic for
1181 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
1184 * Optimize the DNLC by applying Interlocked incrementing for
1185 statistics gathering and enabling greater use of read
1186 locks instead of write locks
1188 * Further optimize osi_Log operations when logging is disabled
1189 by removing the need for the function call overhead.
1191 * Add a new lock operation, Convert Read to Write, and apply
1192 it throughout the cache manager. This function provides a
1193 fast transition from a read lock to a write lock when the
1194 caller is the only reader.
1196 * Fix a timeout problem when opening files or request locks
1197 when the file in question is already write-locked by another
1200 * Fix a deadlock that can occur if "fs flushall" is executed
1201 while applications are writing to AFS.
1203 * The cache manager would re-read data from the file server
1204 after file truncation when obtaining buffers to write to the
1205 same file until the first StoreData RPC completes.
1207 * Fixes error handling in cm_NTCheckDelete()
1209 * Replaces cm_scache_t mutex with a read-write lock permitting
1210 additional parallel access.
1212 * Prevent an cm_scache_t reference undercount when evaluating
1213 symlinks containing @sys when none of the active sysnames
1214 match any of the available targets.
1216 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1219 * Set RxMaxMTU registry value to 0 which is the equivalent of
1220 it not being set at all.
1222 * Reduce contention between BkgDaemon threads
1224 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1227 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1228 objects to use the faster Interlocked operations which avoid
1229 the need to hold write locks.
1231 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1232 non-fatal error that forces a retry against another service
1235 * Fix a crash in smb_WriteData that would occur if an application
1236 opened the file as read-only and then sent a write request anyway.
1239 * Respond to VL_NOENT errors by removing cm_volume_t object and
1240 setting it to be the first object to be recycled. This is not
1241 only a more efficient use of resources but it also prevents
1242 repeated attempts to query the VLDB server for the volume from
1243 the CheckOfflineVolumes routine.
1245 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1248 * A first cut at a cache manager statistics monitor that can be used
1249 to determine the necessary cache parameters to support the
1250 application working set.
1252 Off by default, the performance package can be activated by setting
1253 "daemonPerformanceTuningInterval" in the service Parameters key.
1254 As with the other daemon interval values the unit is in seconds.
1256 At service start and each succeeding interval the cache manager
1257 will write statistics to %TEMP%\afsd_performance.txt showing the
1258 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1259 The FID statistics keep track of all FIDs seen by the cache manager
1260 during the service session whether or not they are backed by any
1261 live object in the cache.
1263 These statistics are not stored in the cache file.
1265 * Fix windows event logging broken in 1.5.33.
1267 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1268 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1269 as well as when computing the server preference list.
1271 * When merging status after a StoreData operation prevent data
1272 buffers from being discarded that we would prefer to keep.
1273 (Bug introduced in 1.5.33.)
1275 * Modify installers the set RxMaxMTU to 0.
1277 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1278 dialog, the Background Daemons and Service Threads fields were
1281 * Increase default number of Background Daemons to 4.
1283 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1284 cm_volume_t objects to be recycled.
1286 * Windows Error Reporting has shown a number of previously unseen
1287 stack corruption errors in the Explorer Shell extension. The
1288 minidump appeared to indicate the problem was in ParseAcl().
1289 Reviewing the function it appears that numerous things could go
1290 wrong if invalid input was provided. Added error checking and
1291 data validation. Hopefully this will address the problem.
1293 * Windows Error Reporting has shown crashes in aklog as a result
1294 of krb5_cc_get_default() failures not being handled. They are
1298 * The Rx library used a 32-bit type for sockets which was
1299 truncating the socket value on 64-bit Windows. This
1300 was not an issue in general but became an issue when
1301 Rx "hot thread" support was enabled.
1303 * Enable "Rx hot thread" support which ensures that while
1304 a received message is being processed that a new thread
1305 is allocated to receive the next message to arrive.
1307 * Increased parallel access to smb_rctLock protected
1308 resources. Reduce the number of times that the
1309 lock is dropped and re-obtained.
1311 * Increased parallel access to cm_scache_t and cm_buf_t
1312 resources by replacing the previous global write lock
1313 protected reference counting mechanism with one built
1314 upon the use of InterlockedIncrement and InterlockedDecrement.
1316 * Implement a hash for cm_fid_t to reduce comparison costs
1317 and inline the cm_FidCmp() function.
1319 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1320 hash which has a better distribution
1322 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1323 the most recent read request instead of the next chunkSize.
1324 cm_chunkSize can be dozens or hundreds of buffers. As a result
1325 too much time is spent performing the evaluation.
1327 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1328 lock is to prevent the creation of new buffers while a truncation
1329 is being performed. All references to bufCreateLock have been
1330 removed except in two places:
1331 i. a write-lock surrounding the function that calls buf_Truncate()
1332 ii. a read-lock within buf_GetNewLocked() that actually allocates
1335 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1336 CM_CONFIG_DATA_VERSION number which value be used to force the
1337 replacement of the cache file contents when incompatible changes
1338 are made between releases.
1340 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1341 cm_SyncOpCheckContinue
1343 * Avoid calls to multi_Rx if nconns == 0
1345 * Modify smb_WriteData to perform background writes based upon the
1346 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1347 boundaries. This will slow down writes from the SMB interface but
1348 will avoid the risk of the CIFS client disconnecting from the AFS
1349 client SMB server. The default value for 'smb_AsyncStoreSize'
1352 * SMB AsyncStore functionality is on by default but can be
1353 disabled using the registry value:
1355 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1357 smb_AsyncStoreSize can be adjusted with
1359 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1361 The value must be a multiple of the buffer block size
1362 and cannot be larger than the chunk size.
1364 * Do not leak a read lock on the bufCreateLock
1366 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1369 * In MergeStatus, only remove clean buffers from the hash table.
1371 * Avoid the need to update the data version number on each buffer
1372 associated with a scache when MergeStatus is called after a
1373 StoreData by maintaining a range of valid data versions as part
1374 of the cm_scache_t object.
1376 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1377 from cm_HaveAccess... We can safely avoid the call so do so.
1379 * Add "ForceAnonVLDB" service parameters registry value. Default
1380 is 0. Set to 1 to force the use of anonymous connections to
1381 the VLDB servers. This is useful when using krb5 tokens and the
1382 VLDB servers have not been updated.
1385 * Do not attempt to search for a cell name that matches one
1386 of the special share names "svrsvc", "wkssvc", "ipc$"
1388 * Issue permission denied warnings from afscreds if attempts
1389 to modify the startup shortcut or registry values fail.
1390 This happens under Vista UAC.
1392 * Do not return BAD_NETWORK_ERROR in response to attempts
1393 to create directories or files in the Freelance root.afs
1394 volume. Instead return ACCESS_DENIED.
1396 BAD_NETWORK_ERROR was being sent because the Freelance cell
1397 has no vldb servers. Short circuit the error by testing for
1398 freelance in CreateFile and MakeDir
1400 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1401 to be opened and closed without actually existing. For now
1402 we treat them as equivalent to ioctls but this should
1405 * Fix RAP processing to not reject names that exist in the
1406 root.afs volume for GetInfo queries.
1408 * In afscreds, use GetUserNameEx to obtain the username instead
1409 of searching the registry. Try UserPrincipalName first and
1410 fallback to SamCompatible.
1412 * Stop linking to rpcns4.dll which is no longer supported on
1413 Vista and 2008. TaAfsAdmSvr is going to have to be
1416 * When processing dfs referral requests, if the sharename is
1417 a partial match and therefore will not be added automatically
1418 to the share list, make sure that we return an error.
1420 * Avoid a crash when logging VNOVOL errors returned from
1421 cm_CheckOfflineVolume.
1423 * When trace logging is disabled, do not permit osi_LogSaveString
1424 to perform the string copying. By disabling the string
1425 copies, throughput of the service more than doubles.
1427 * when checking offline volumes, update the cell vlserver info
1428 if required and don't perform a check if there are no servers
1431 * Add volumes in the vl_unknown state to the list of objects to be
1432 checked as offline or retried when an allbusy error is received
1435 * Add support for configuring rx_ExtraPackets
1437 * All calls to buf_Get() must be protected by a read lock on the
1438 cm_scache_t bufCreateLock
1440 * When a MergeStatus results in the invalidation of buffers, do
1441 not remove buffers from the hash tables that have active references.
1443 * Upon receiving a VNOVNODE error from a file server, be sure to
1444 invalidate any file locks. This prevents on-going attempts to
1445 extend the locks and will cause subsequent attempts to access
1446 the file to receive a invalid handle error.
1448 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1449 to directory entry names.
1451 * No longer discard tokens when RXKADBADTICKET error is received.
1452 Instead return STATUS_NO_KERB_KEY error.
1455 * Add "cmdebug <host> -cellservdb" option and client service
1456 implementation permits querying the vldb server for each
1457 cell known to the cache manager. The output is in a form
1458 that permits it to be used as a CellServDB file.
1460 * Add Symlinks->Show option to the AFS Explorer Shell Context
1463 * Prevent buffer overrun in the cmd parser used by command
1464 line tools such as fs, vos, pts, etc.
1466 * The path ioctl operations have several issues:
1468 (1) the specified path for "fs examine, whereis, whichcell,
1469 flush" and so always has the follow symlinks and mount
1470 points semantics. This makes it impossible to determine
1471 what the FID of a symlink or mount point is.
1473 (2) "fs examine" out is not the result of a single pioctl
1474 operation but is actually the combined output of half
1475 a dozen operations. Path evaluation is an expensive
1476 operation. It would be faster if the caller could
1477 evaluate the FID first and then perform all of the
1478 rest of the operations by specifying the FID instead
1481 (3) fs output reports all objects as files. By adding a
1482 GetFileType pioctl more informative output can be provided
1483 that indicates what type of object the path evaluates to.
1485 (4) the Windows fs command includes a number of commands that
1486 do nothing but exist only because the Unix cache manager
1489 A new extendible data structure cm_ioctl_query_opts_t has been
1490 added which can be optionally specified with pioctls that do not
1491 already require input data. The first two fields of this structure
1492 are 'literal' and 'fid'. The literal field is used to indicate
1493 whether the last component of the path should be evaluated following
1494 symlinks and mount points.
1496 The fid field permits a fid to be specified.
1498 * A new GetFileType pioctl has been added. The type of objects are
1501 * A new "-literal" option is available for "fs examine, flush, whereis,
1504 * Unimplemented fs commands have been removed.
1506 * There are circumstances where a volume object is being accessed and
1507 the volume is marked indicating that the volume location information
1508 is out of date but where it is also pointless and perhaps dangerous
1509 to block waiting for the rpc to complete. One example is when
1510 processing the cmdebug requests. If we know that we are not going
1511 to use the volume object to contact a server, then we can now set
1512 the CM_GETVOL_FLAG_NO_RESET flag.
1514 * when moving up in the directory tree we search the recorded fid list
1515 to find a matching fid that we have already crossed. we must also
1516 reset the fid count based upon what we discover.
1518 * Cell names are published as share names. As a result they are searched
1519 for as part of the Dfs Referral evaluation. If share "foo" can not be
1520 reached, the CIFS client will fallback to searching for "fo" as well.
1522 Since the freelance client automatically adds a symlink for prefixes
1523 of the cell name, this results in both "foo" and "fo" being added to
1524 the freelance root.afs volume. This patch prevents that negative side
1527 * when tracking Kerberos credential cache names be sure to include
1528 the cache type prefix.
1530 * Microsoft has assigned OpenAFS a network type value
1532 * The algorithms that were used to produce a Netbios name were broken.
1533 The name that was produced when the hostname was too long was in fact
1534 longer than the maximum permitted netbios name.
1536 Also, the "NetbiosName" value was not used as a suffix when the loopback
1537 adapter was not installed. The hardcoded string "AFS" was used instead.
1539 * avoid another deadlock during server probes initiated by the ipaddr
1540 change daemon thread.
1542 * prevent the afs client service from crashing when fs newcell is
1545 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1546 to \Program Files\OpenAFS\Client\Program in order to get them out
1547 of the %windir% tree
1549 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1550 end STATUS_OUT_OF_QUOTA can be returned to the application.
1552 * Modify the search order for determining the location of CellServDB
1553 and other client configuration files.
1555 1. AFSCONF environment variable
1557 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1558 4. use Program Files\OpenAFS\Client
1560 * More Vista Power Management tweaks
1562 * Close a socket handle leak in the kauth routines
1564 * Don't hold a cm_volume_t mutex across RPCs
1566 * Convert downTimes to use clock time instead of relative times.
1567 This permits correct comparisons with expiration times.
1569 * Update the Wix installer to properly install the prerequisites
1570 for translate_et.exe
1572 * Do not translate pioctl paths as they are not converted by the
1575 * Implement multi_rx probes for checking whether servers are up
1576 or down. Turns probes into a constant time activity instead
1577 of one dependent upon the number of known servers.
1579 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1580 update of the volume location data and retry. VNOVOL means
1581 that the volume could not be attached. It doesn't mean that
1582 the volume is not on this server.
1584 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1585 not attempt to validate the target path. There is no reliable
1588 * Improve performance of \\afs\<cell><mount-type><volume>\
1589 access by eliminating extraneous attempts to search the
1590 CellServDB file and DNS AFSDB records. Share names
1591 containing '%' or '#' are not valid cell names.
1593 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1594 to unlock a range that is not currently locked.
1596 * Enable the 32-bit tools to read the 64-bit registry keys
1597 when executing in the WOW64 environment.
1601 * Fix buffer refcount leak introduced in 1.5.29
1603 * Prevent the NIM AFS Provider from crashing when configured to use
1604 Kerberos v4 for token acquisition in situations where no Kerberos v4
1605 support is available. For example, 64-bit KFW or sites that delete
1606 the Kerberos v4 DLL.
1608 * In afsd_init.log, support logging cache sizes greater than 2GB.
1610 * The volume status output of "fs examine" was being determined based
1611 upon the wrong error value. Instead of using the pioctl() return
1612 code, the errno value must be used.
1614 * Prevent invalid pts auto-registration attempts by aklog and the
1615 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1617 * Add synchronization protection to all of the SMB Listener State
1618 variables. This prevents race conditions when the addition or
1619 loss of a network adapter takes place.
1621 * On Vista, there is a race condition between the restart of the
1622 afs client service and the network adapters. Ensure that the
1623 afs client service can handle switching between loopback and
1624 non-loopback modes of operation.
1626 * Fix FollowBackupPath to work with mount points that are not
1627 explicitly read/write.
1629 * Fix a deadlock that has been observed on Vista while resuming
1630 after at least four hours of sleep.
1632 * Add new HKLM registry configuration options that can be used
1633 to provide configuration hints to the NIM AFS Provider when
1634 creating new identities. See release notes for details on
1635 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1638 * Speed up write buffer operations by holding a global read-lock
1639 instead of a write-lock when updating the buffer data version
1640 number. The data version is protect by the associated stat
1643 * Speed up read buffer operations by removing out of date buffers
1644 from the buffer hash tables.
1646 * Protect against a null pointer dereference in afslogon.dll
1647 if QueryAdHomePathFromSid() fails.
1649 * Remove a race condition while accessing private freelance data
1652 * Update the buffer data versions when Freelance directory data
1655 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1658 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1659 instead of returning zero length.
1661 * Return ERROR_RETRY when a server sends VBUSY.
1663 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1664 a volstat plug-in is in use.
1666 * Enhance the volstat interface to support Dfs path mappings.
1668 * Fix "symlink list" to do the right thing for Dfs links
1670 * Fix callback expiration that was broken in 1.5.27.
1672 * Fix a crash at shutdown caused by the lack of synchronization
1673 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1675 * Hold the cm_server_t mutex while walking or modifying the
1676 cm_server_vols_t list in order to avoid a race condition.
1678 * Ensure that the cm_server_t refCount is safely modified.
1680 * The pthread library implementations of pthread_cond_wait
1681 and pthread_cond_timedwait were failing to return with
1682 the mutex held when error conditions occurred. This could
1683 result in race conditions wherever they are used. One case
1684 in which problems were seen was the Rx event_handler().
1686 * Prevent cm_GetBuffer() from locking all buffers associated with
1687 a stat cache object before deciding if the current buffer is up
1690 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1691 object for each offset that we will attempt to reserve.
1693 * Reduce the default number of background daemon threads used to
1694 prefetch or store to one in order to prevent two or more threads
1695 from stepping on each other.
1697 * Reduce the default chunk size to 256KB in order to prevent read
1698 timeouts on low bandwidth connections.
1700 * Fix a race condition that could result in threads becoming stranded
1701 when synchronizing multiple requests on the same stat cache object.
1704 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1705 has been disabled. It has been disabled because it turns out
1706 that the RPC is improperly implemented in all file servers
1707 previous to 1.4.6. The RPC is executed without the proper
1708 locks being held resulting in data corruption and eventual file
1709 server failure. The functionality will be re-enabled in a
1712 * a couple of .readonly Volume Callback optimizations were
1714 - Apply most recent volume callback to all stat cache objects
1715 in the volume. This avoids premature callback expirations
1716 and unnecessary FetchStatus calls.
1717 - Add option to permit attempts to renew .readonly Volume
1718 Callbacks on a periodic basis to avoid invalidation of stat
1719 cache data. [Requires daemonCheckVolCBInterval be set to
1720 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1722 * Adds support for 64-bit data version values. Previous releases
1723 only supported 32-bit values. Larger values sent by the file server
1724 would result in rollover.
1726 * aklog and asetkey no longer generate an exception if Kerberos for
1727 Windows is not installed.
1729 * If the LSA authentication functions report a temporary out of memory
1730 condition, fail the current authentication and do not cause the AFS
1731 client service to terminate unexpectedly.
1733 * When the AFS client service is shutting down, prevent the daemon check
1734 thread from executing additional checks which might attempt to access
1735 discarded locks or unmapped memory.
1737 * Correct a double free condition in the NIM AFS credentials provider.
1739 * Add "FollowBackupPath" service parameter registry value.
1740 This provides equivalent functionality to the UNIX afsd -backuptree
1741 option. When set to a non-zero value, normal mount points originating
1742 in a .backup volume will prefer a .backup volume.
1745 * Fix NSIS uninstall from the .MSI installer.
1747 * Add Stack Frame logging for AMD64
1749 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1750 server returns EWOULDBLOCK, retry the request every two
1751 seconds for up to the RDRtimeout.
1753 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1754 is not assigned the same value as EIO.
1756 * Update CellServDB to GCO Public 25 Oct 2007
1758 * Add additional validation and error handling code after
1759 each call to getSlot(). If an invalid slot is returned,
1760 return NONODE. If the invalid slot is returned when
1761 extracting a data node, invalidate the tree.
1763 * Modify compareKeys() to always perform a case-insensitive
1764 comparison and only perform a case sensistive comparison
1765 if the case-insensitive one matches. This ensures the
1766 ordering is consistently reported.
1768 * Add lock assertions to ensure that all calls are being
1769 performed with the correct locks being held. There have been
1770 some crash reports that provide stack data that does not appear
1771 to be possible unless there is a race. However, there are no
1772 obvious locations where the race is taking place and the test
1773 suite indicates that all of the correct locks are being held.
1774 We shall see what happens in the field.
1776 * For consistency replace all calls to findKey in which the range i
1777 s (1,numentries) with calls to getSlot().
1779 * Optimize the depth search loop by testing the slot value in the
1780 for statement instead of forcing the loop to be broken later.
1782 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1783 There are a number of locations where locks are obtained, dropped,
1784 and reobtained. This reorganization attempts to accomplish several
1787 (1) be optimistic for the most common case so it will be fast
1789 (2) add consistency checks after each location where locks are
1790 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1791 and the bplus tree is out of date, retry until we get to a
1792 consistent state that we can use.
1794 (3) Ensure that all operations take place with the correct locks.
1796 * One of the issues that has become a serious problem since the
1797 addition of the local directory updates is that although cm_SyncOp
1798 synchronizes operations, it does not preserve the order of requests.
1799 This has always been a problem in that it has been possible for a
1800 request to fail to complete due to its worker thread's bad luck.
1801 When a request takes longer than the Windows SMB Redirector's timeout,
1802 the SMB Redirector tears down the SMB virtual circuit.
1804 When using the local directory updates it is really important that
1805 the directory update operations complete in the order that they were
1806 sent to the file server. If they don't, then the local directory state
1807 and the file server state will not match and the local directory state
1808 must be discarded which in turn forces a new read of the entire
1809 directory contents over the network.
1811 This patch adds a new cm_scache_waiter_t object that is used to store
1812 the current thread, buffer, and syncop flags within a waiters queue
1813 on each cm_scache_t object. If a thread is forced to sleep in
1814 cm_SyncOp, upon waking it will check to see if there are any other
1815 threads waiting that are attempting to perform a similar task ahead
1816 of it in the queue. If yes, the thread goes back to sleep. If not,
1817 it goes ahead and enters the cm_SyncOp conflict resolution block.
1819 This patch has the additional side effect of reducing the number of
1820 competing threads that must obtain the cm_scache_t mutex and process
1821 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1822 utilization of the service and the clock time associated with processing
1823 requests will be reduced.
1825 * assert that the cm_scache_t mutex is held when calling
1826 buf_ForceDataVersion().
1828 * rename findNode to leafNode in bplus_Lookup
1830 * replace all OutputDebugString calls with osi_LogX calls
1832 * modify bestMatch to special case the return values for leaf nodes.
1833 If an entry is above or below the values available in the leaf node
1834 return BTLOWER or BTUPPER instead of BTERROR.
1836 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1837 if true convert to either slot 0 or Max and perform the insertion.
1838 This produces easier to read code when performing lookups.
1840 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1841 descriptive message.
1843 * the difference between osi_assert() and osi_assertx() is whether or not
1844 a description is specified as part of the call. When no message is
1845 specified afsd_notifier() is called with a NULL msgp parameter. This
1846 results in a null pointer dereference during the EventLog() call.
1848 As a result, none of the information describing the assertion is written
1851 This commit sets a default message for use when no other message is
1852 provided and it replaces all calls to osi_assert with osi_assertx and
1855 * Migrate B+ tree search key into thread local storage
1858 * Fix an uninitialized variable in lana_IsLoopback() which
1859 could result in a random determination that a loopback
1860 adapter is not a loopback adapter. This would result in
1861 a computed Netbios service name of %MACHINE%-AFS instead
1862 of AFS. This is a serious problem if the value computed
1863 by the AFS Client Service and the client process pioctl
1864 calls produce different results as it will result in a
1865 communication failure.
1867 * Fix a crash in the B+ Tree code that results when a file
1868 rename only results in a change of case.
1870 * Fix the pioctl path parsing. The previous algorithm
1871 broke the current directory of the client into two parts,
1872 the TIDpath which represented the device root and the
1873 relative path. The problem is that this is not sufficient
1874 from proper processing of symlinks and mount points.
1876 What is required is that the last component of the path
1877 must not have the symlink or mount point be evaluated
1878 whereas symlink and mount point evaluation must be performed
1879 for the rest of the relative path. Therefore, the path is
1880 now split into three parts: TIDpath, intermediate path,
1883 "fs examine" now reports the data of the symlink or mount
1884 point object instead of the target and "symlink list" and
1885 "symlink remove" now works. The "Explorer Shell Extension"
1886 menu now is displayed for symlink objects.
1888 * More informative symlink.exe error messages.
1890 * When the cmd.exe "del" command is used to remove a symlink,
1891 not only was the link object removed, but the stat cache
1892 object for the target was marked deleted. This resulted
1893 in cache corruption which would require the destruction
1894 of the cache to correct.
1896 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1897 which results in cached contents from read only volumes
1898 to be treated as valid when there is no callback registered
1899 and all servers are down or all instances are offline.
1901 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1903 * Add documented registry value, "AcceptDottedPrincipalNames"
1904 which is used to permit the acquisition of AFS tokens
1905 using Kerberos v5 principal names which include a dot in
1906 the first component.
1908 {HKLM\HKCU}\Software\OpenAFS\Client
1910 * "fs uuid" does not require "AFS Client Administrator" group
1911 membership to display the current UUID value.
1914 * Update NSIS installer to version 2.30
1916 * Update Wix installer to version 2.0.5325
1918 * 64-bit KFW support:
1919 - Network Identity Manager plug-in
1921 - 64-bit krb5 enable afscreds.exe, afs server manager,
1924 * Enable B+ tree Directory Search by default.
1926 * NIM plug-in is now compatible with FILE: ccaches
1928 * Install instloop.exe into Client/Programs/ even though
1929 it isn't required. Permit it to be used by organizations
1930 that clone Windows systems.
1932 * On a suspend, only drop callbacks for servers that
1933 are online. If the server is offline, maintain the callback
1934 until the server is once again available.
1936 * Prevent against acquisition of a directory write lock when
1937 there are no updates to be performed. This was resulting
1938 in a write-lock leak which later on would lead to a deadlock.
1940 * Add configurable file pre-fetch mode. For files that match
1941 the specified file extensions, pre-fetch the entire file into
1944 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1945 (MULTI_SZ) PrefetchExecutableExtensions
1947 * Fix the client_osi large integer support so that return values
1948 can be properly assigned to large integer variables.
1950 * Do not hold cm_scache_t mutex across cm_GetCell() calls
1951 since cm_GetCell() can block in a RPC.
1953 * Add short file name support to B+ tree searches.
1955 * Resolve a conflict between the desire to notify file servers
1956 that the client workstation is suspending and the desire to
1957 permit objects with valid callbacks when the client suspends
1958 to continue to use those objects from the cache if the client
1959 resumes without network connectivity.
1961 * Fix stack variables to ensure that all buffers storing paths
1962 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
1963 would cause a stack overflow condition.
1965 * On Vista, force afs_config.exe to run with the highest privileges
1966 available to the end user. For a user in the Administrators
1967 group running under UAC, this will cause afs_config.exe to be
1968 be run with full Administrators privileges after the user
1969 approves the privilege escalation.
1971 * More VNOVNODE handling issues. If a dirty buffer cannot be
1972 written due to VNOVNODE, clear the buffer and ignore the error.
1973 If a cm_scache_t object has been marked deleted due to a VNOVNODE
1974 error, the next time the smb_fid_t object that references it is
1975 used return either NOSUCHFILE or NOSUCHPATH and close the
1976 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
1977 result in an invalid handle error.
1979 * For Vista, set the RequestExecutionLevel of the NSIS installer
1982 * For the build environment, support the generation of manifests
1983 when using the CL1310 (VS .NET 2003) compiler/linker.
1985 * There was a bug in OutputDebugHexString() which could overflow
1986 a stack buffer resulting in unexpected process termination
1987 of afsd_service.exe if the input Security Token contained
1988 many bytes equivalent to the '%' char.
1990 * Add directory lookup statistics to "fs memdump" output.
1994 * Windows uses case-insensitive file name pattern matching
1995 but AFS is a case sensitive file system. The AFS3 directory
1996 format is block based, uses network byte order and
1997 includes a hash table for fast case sensitive lookups.
1998 This causes several problems for the Windows AFS client.
1999 (1) Traversing the directory blocks is cpu expensive
2000 (2) A hash table miss does not indicate that the desired
2001 entry does not exist.
2002 (3) Determining whether a non-ambiguous inexact match or
2003 the entry does not exist requires a linear traversal
2004 of the entire directory.
2005 These issues often result in 100% CPU utilization.
2007 These issues are addressed by building a modified B+ tree for
2008 each directory and then using the B+ tree for searches.
2010 This feature is disabled by default. To enable it set
2011 the "BPlusTrees" registry value to 0x01 in the AFS Client
2012 Service Parameters key.
2014 * Correct a null pointer dereference in the symlink recursion
2015 detection code. (Introduced in 1.5.22)
2017 * Add the Network Identity Manager AFS Provider to AMD64
2022 * A new registry value, "BlockSize", can now be used to adjust
2023 the size of the internal buffers used to store file data.
2024 The default is 4KB. When set, the values 1..1024 specify
2025 multiples of 4KB. Values > 1024 specify actual block sizes
2026 and must be multiples of 4KB. The actual block size in KB
2027 can not be larger than the chunk size which defaults to 1MB.
2029 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
2032 * pts.exe and ptserver.exe built with -DSUPERGROUPS
2034 * Service Manager shutdown hints have been added. The hint
2035 specifies that it may take the AFS client service up to two
2036 minutes to shutdown.
2038 * Fix the recursion detection feature introduced in 1.5.22.
2039 The previous implementation broke the evaluation of paths
2040 using ../path or ./path. It also broke the evaluation of
2041 symlinks to absolute paths. For example, /afs/cell/path.
2043 * Fix an uninitialized variable in cm_MakeDir that could result
2044 in crash if the requested directory could not be created.
2046 * Fix a the VolumeStatusNotification log messages to save
2047 the input string in the circular log buffer so garbarge
2048 won't be output to afsd.log when the log is dumped.
2051 * Increase the SMB Maximum Path length to 260 characters
2052 from 256 as that is what CIFS supports.
2054 * On 64-bit Windows, use a unique description string for
2055 the 32-bit AFS Shell Extension.
2059 * In the AFS Admin Library, replaced all of the
2060 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
2061 allows parameter checking and proper type conversions to be
2062 performed. This is crucial when passing parameters that are
2063 64-bit wide on systems in which 'long' is a 32-bit value.
2064 Otherwise, the parameter list on the stack is interpretted
2065 incorrectly and the function not only doesn't succeed but it
2066 might corrupt memory and if lucky crash the application.
2068 The only program that we care about that uses this library
2069 is the AFS Server Manager. Now it doesn't crash.
2071 * Fixed smb directory searches for non-wildcard names. The
2072 optimized function should not have set the bulk in progress
2073 flag. This would adversely affect later directory searches
2074 that required a bulk search.
2076 * Detect infinite recursions (up to 512 stat cache objects
2077 in a chain.) If an infinite recursion is detected, return
2078 an error to stop the CIFS client from querying us forever.
2080 * Executables executed out of AFS name space can have their
2081 pages swapped out by Windows. Those pages will be reloaded
2082 when they are required. However, if the cache manager no
2083 longer has a callback and all of the file servers with
2084 the volume data are down, then the application will crash
2085 due to the inability to read the required page.
2087 We now track the down time of each server. If the callback
2088 expires after the server that issued it went down and all
2089 servers are down, we delay the expiration to permit the data
2090 in the cache to be continue to be used by Windows or an
2093 * The Windows AFS client has always had very poor performance
2094 with regards to directory management. In the case of
2095 directory updates caused by create, rename,and delete operations,
2096 the AFS cache manager will now perform local directory updates
2097 when possible in order to avoid re-reading the directory from
2098 the file server. This will be done in whenever both of the
2099 following conditions are true:
2101 - the data version on the directory increased by one as a
2102 result of the operation
2104 - all of the directory buffers necessary to make the change
2105 are present in the cache
2107 * The error message in the AFS Control Panel describing the
2108 need to be a Windows Administrator has been replaced with
2109 a message describing the need to be in the AFS Admins Group.
2111 * If the vldb server list is empty when attempting to evaluate
2112 a cell mountpoint it was possible for a crash to occur. This
2113 can only happen if there is an existing mount point and the
2114 is a failure attempting to load the server list from the
2115 CellServDB file or via AFSDB queries.
2117 * dirty buffers only have their dirty bytes stored to the file
2118 server instead of the entire chunksize
2120 * the default chunksize has been increased from 128KB to 1MB
2125 * changed the enum values for cm_serverRef_t state info to use a
2126 private name space to avoid collisions (srv_)
2128 * added a srv_deleted state for cm_serverRef_t objects. This
2129 state is set when cm_FreeServerList() is called with the
2130 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
2131 not always delete the cm_serverRef_t from the list if it is
2132 still in use by another thread. the srv_deleted state means
2133 the object's contents are no longer valid and it must be
2134 skipped. It will be deleted the next time the object is
2135 freed and the refcount hits zero.
2137 * the srv_deleted state is also used when a file server reports
2138 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
2139 as offline. This is done to prevent additional usage of the
2140 stale vldb data while waiting for the update volume request
2143 * added a state field to the cm_volume_t object (enum volstate
2144 vl_ name space) that maintains the state of the volume based
2145 upon the states of all of the cm_serverRef_t and cm_server_t
2148 * modified cm_UpdateVolume() to set the state of the cm_volume_t
2149 RW, RO, and BK to either vl_alldown or vl_online. There can't
2150 be any other states because cm_UpdateVolume() destroys any
2151 previous knowledge we might have had regarding busy or offline
2154 * modified cm_UpdateVolume() to update the volume name in the
2155 cm_volume_t to the volume base name if the previous value was
2158 * modified cm_FollowMountPoint() to check to see if the volume
2159 name is a volume ID and if so call cm_GetVolumeByID instead
2160 of cm_GetVolumeByName. This ensures that volume IDs are always
2161 looked up as numeric values. There is no longer a need to
2162 maintain a separate cm_volume_t containing the string representation
2165 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
2166 The first flag is a "CREATE" flag which is set by all existing
2167 calls. The flag is not set by calls to cm_GetVolumeByID() from
2168 the server probe code when volume status is being updated. We
2169 do not want the server probe operation to result in additional
2170 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
2171 which is set when the server probe code updates the volume status.
2172 This flag will be used to prevent the server probe operation from
2173 changing the order of the least recently used queue.
2175 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
2176 allocated for a given set of normal, readonly, and backup volumes
2177 regardless of whether or not the volume is accessed via name or
2178 ID number. The cm_volume_t namep field is always the base name
2181 * Added a new volume state, vl_unknown. This state is used as
2182 the initial state for all cm_volume_t when the cache manager starts,
2183 for each cm_volume_t at creation, and for each cm_volume_t when
2184 recycling. The cache manager does not know the state of all
2185 volumes in the world, only those that are in the cache and for
2186 which it has queried the VLDB and hosting file servers.
2188 * modified cm_GetVolumeByName() to initialize the state of a
2189 volume to vl_unknown. The actual state will be set when a
2190 cm_VolumeUpdate() call completes successfully.
2192 * changed name of scache hash table variables to avoid ambiguity
2193 when adding hash tables for volumes
2195 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
2197 * modified cm_UpdateVolume() to handle the case in which there is
2198 no RW volume but there is are RO volumes for a given base name.
2199 This is done by querying for the ".readonly" volume name if the
2200 base name does not exist in the VLDB. We never query for the
2201 .backup name because under the current usage model a .backup
2202 volume may only exist on the server that the read-write volume
2203 is located. If there is no RW volume, there can be no .backup.
2205 * Added four hash tables for cm_volume_t objects to improve the
2206 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2207 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2208 BKID. Three ID hash tables are necessary as long as it is
2209 desireable to maintain a single cm_volume_t containing all
2210 of the related RW, RO, and BK volume data. Having the RW and
2211 RO volume data in the same object is necessary for the
2212 implementation of cm_GetROVolumeID() which returns either the
2213 RO or RW ID depending upon the existence of RO volume instances.
2215 * Added a volume LRU queue so that volume reuse becomes fairer.
2216 This does not replace the all Volumes list which is used when
2217 it is desireable to walk a list of all the volumes whose order
2218 is not going to change out from underneath you which makes it
2219 safe to drop the cm_volumeLock.
2221 * handles volume hash table updates where volume name to
2222 volume ID number changes. The volume name remains
2223 constant in the cm_volume_t. if a vos rename is performed,
2224 the name of the volume will change and the volume IDs will be
2225 updated. Subsequent access to the old volume ID will create a
2226 new cm_volume_t with the new name.
2228 * Added a daemon thread operation to query the state of volumes
2229 listed as busy or offline. cm_CheckBusyVolumes() calls
2230 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2231 or vl_offline. If the volume is now online, the status on the
2232 volume is updated. The default period is 600 seconds. This can
2233 be configured with the BusyVolumeCheckInterval registry value.
2235 * Added prototype for smb_IoctlPrepareRead() which was missing a
2236 return type in the function definition.
2238 * Added volume id lists to the cm_server_t. These lists are
2239 allocated in blocks of ~32 IDs. When a cm_PingServer()
2240 detects a change in server state, the state of the cm_volume_t
2243 * Added volID to the cm_serverRef_t object. volID is used
2244 to identify the volume for which the object is a referral.
2245 cm_FreeServerList() uses the volID to remove the volume
2246 from the cm_server_t.
2248 * In cm_Analyze, when VNOVOL or VMOVED are received,
2249 call cm_ForceVolumeUpdate() to force a refresh of the volume
2252 * Added cm_VolumeStatusNotification() which is used at the moment
2253 to log volume status changes to the trace log. It will also
2254 be used as the access point to the File System Filter driver
2255 notification engine.
2257 * Added an all cm_scache_t list to cm_data. This replaces the use
2258 of the stat cache LRU queue when we need to enumerate all
2259 entries. The LRU list order is not static and when using it to
2260 enumerate all entries it can result in items being missed or
2261 items being processed more than once.
2263 * Modified cm_Analyze(). Instead of reseting the busy or offline
2264 state of a volume and forcing a retry of the operation
2265 cm_Analyze will defer to the background daemon thread that will
2266 update the state once every 600 seconds.
2268 * Added the automatic generation of a Freelance ".root" read-write
2269 mountpoint that refers to the root.afs volume of the workstation
2270 cellname at the time the mountpoint is created.
2272 * re-write cm_Analyze to make better use of the known volume
2273 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2276 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2277 busy volumes will be reset to srv_non_busy by the function
2278 but there is no mechanism for querying the busy state other
2279 than by attempting to access the resource.
2281 * cm_Analyze will query the state of an offline volume before
2282 deciding whether or not to retry when all volume instances
2285 * the algorithm used after an IP address list change detection
2286 has been modified. we now check the state of the VLDB servers,
2287 then reset the rx connections, and finally check the state of
2288 the file servers. Previously we flushed the rx connections
2289 and then checked the servers in whatever random order they
2290 were first discovered.
2292 * fix the volume status reported by "fs examine"
2294 * various improvements to the AFS Network Identity Manager Provider.
2295 - Don't access freed objects after the window has been destroyed.
2296 - Write back the AFS cell list in the identity configuration
2297 even when its empty.
2298 - Disable all use of the provider when AFS is disabled for a
2299 particular identity.
2301 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2302 alias instead of the full cellname, double check the full cell
2303 name and see if we have a matching cell before allocating a new
2306 * Dump more data with "fs memdump". All volumes and scaches and
2309 * Reverse the order of the all cells list. This ensures that
2310 the workstation cell is always first in the list.
2312 * If we recognize the server from which a callback revoke was
2313 received, then we can ensure that we only invalidate the status
2314 information for cm_scache_t objects belonging to that cell.
2316 * When building for amd64, use APPVER 5.02 as we do not support
2317 64-bit Windows 2000 and explicitly state the machine type in
2320 * no longer treat VMWare network adapters as equivalent to a
2321 loopback adapter. For some unknown reason, publishing the "AFS"
2322 netbios name on them frequently results in a name collision
2325 * Use read locks whenever possible instead of write locks when
2326 walking the allCellsp list. This permits better use of multiple
2329 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2330 is not present on the directory optimization. It would report
2331 mount points and directory objects for which there was no
2332 status information as a zero length file. Instead we can use
2333 the vnode number of the object to determine if it should be
2334 considered to be a directory or a file. If it is odd it is a
2335 directory; even it is a file.
2337 * Give Back Callbacks to the file servers in response to suspend
2338 and shutdown power management events. Only give up callbacks
2339 to UP servers and do so with a very short RX dead timeout.
2342 * Fix a race condition in the SMB FID cleanup code that could
2343 result in a stat cache mutex never being released.
2345 * Fix the graphics files that are installed as part of the
2346 IBM Administrators Guide.
2348 * Change the behavior of pioctl path parsing to not follow
2349 symlinks or mount points during evaluation. This results in
2350 the FID being the symlink or mount point and not the object
2351 they point to. This is consistent with the behavior of the
2354 * Enable AFSDB record lookup for command line tools such as
2357 * CellServDB update 2007-07-05
2359 * Added a Volume Status Plug-in Interface that permits third parties
2360 to receive notifications when the volume state changes occur.
2362 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2364 * Prevent the afsprocmgmt.dll from being unloaded while the
2365 background thread it starts is running.
2367 * Avoid a race condition when closing smb_fid_t objects.
2370 * fix install path of IBM Quick Start Guide when installed by
2373 * fix the client so that it can write into directories with
2374 only 'li' permissions.
2376 * fix installation of 32-bit Shell Extension for 64-bit systems.
2377 Install it under its own HKCR keys.
2379 * modify the wix installer to remove and upgrade install packages
2380 with the same version number.
2382 * when flushing volume location data every two hours or in
2383 response to "fs checkvolumes", also flush the mount point
2384 links in order to force their re-evaluation.
2386 * aklog was copying an uninitialized buffer as part of the
2387 foreign cell automatic pts registration
2389 * when obtaining directory listings, check for read permission
2390 on the directory. If read permission does not exist, fake the
2391 status information for the directory entries. This improves
2392 performance, prevents the client from pounding the file server,
2393 and prevents leakage of cached status information.
2395 * "fs getcacheparms" output was incorrect.
2397 * NetIdMgr AFS Provider should only automatically add token
2398 acquisition for cell foo.com if the realm of the cell matches
2399 the realm of the identity.
2401 * Fix a null pointer dereference in cm_MergeStatus()
2405 * network identity manager provider:
2407 - Correct for the fact that pr_CreateUser tampers with the
2408 contents of the username input buffer.
2410 - Take into account that the krb4 credential data structure
2411 does not have separate realm fields for the user and service.
2412 When krb524 is used with a cross-realm ticket, the realm
2413 describes the service not the user.
2415 * Many integrated logon changes:
2417 - Fix eventlog reporting. Do not attempt to log an event
2418 if the event source registration fails. Use DebugEvent0
2419 instead of DebugEvent when there are no parameters.
2421 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2424 - Fix the reading of Domain specific configuration for
2425 LogonScript and TheseCells. Previously the dwSize value was
2426 being overwritten so that subsequent RegQueryValueEx call
2429 - Fix a memory leak in the TheseCells reading code.
2431 - Add support for Domain specific "Realm" specification. The
2432 realm is the realm to be appended to the username. When
2433 logging in as a domain or to the local machine, the specified
2434 "Domain" name is not going to be a valid realm name.
2436 - Construct a proper principal name based upon the domain
2437 specified realm for use in obtaining tokens with KFW.
2439 - If the domain specified "TheseCells" list includes the
2440 default cell, do not obtain tokens twice.
2442 * An unprotected call to the Kerberos v5 function,
2443 krb5_free_context(), can result in an invalid memory access
2444 exception if the contents of the krb5 profile, krb5.ini, cannot
2445 be parsed correctly. The error is located within the afskfw
2446 library which is used by afscreds.exe, afssvrmgr.exe, and
2447 afslogon.dll. When the error is triggered within afslogon.dll
2448 a blue screen condition can occur when the winlogon.exe system
2449 process terminates unexpectedly.
2453 * Do not stop the afs client service with an assert
2454 if the cell name cannot be determined when responding
2455 to a "fs whereis" command.
2457 * Integrated Logon (afslogon.dll). No longer attempt
2458 to preserve Kerberos 5 tickets for the logon session
2459 as that job is now supported by KFW 3.1 and above.
2461 * The Network Identity Manager plug-in will attempt to
2462 remove expired service tickets from the cache.
2466 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2468 * AFS Server Manager and AFS Account Manager data cache
2469 moved from HKCU\Software\TransarcCorporation to
2470 HKCU\Software\OpenAFS to prevent crashes due to data
2473 * Reduce the default UP server probe time from 10 minutes
2474 to 5 minutes to match UNIX. This will not solve the NAT
2475 port mapping timeout issue because most NATs default to
2476 a value less that 5 minutes and this value ensures that
2477 the afs client server will start probes 5 minutes after
2478 the last probe set ended. The probe period for any one
2479 server is therefore longer than 5 minutes. The more
2480 servers in the list the longer the time between probes.
2482 * Before marking a server as down in cm_Analyze, retry the
2483 request using a new rx connection to the server. If the
2484 cache manager is behind a NAT, it is possible that the
2485 response from the server is being sent to the original
2486 address:port from which the rx connection was established.
2487 This will appear to the cache manager as the server being
2488 down when in fact it is not. Forcing the use of a new rx
2489 connection will result in the server responding to the
2490 most recent address:port assigned to the client by the NAT.
2492 If the server is in fact down, it will simply take a little
2493 longer to notice it.
2497 * Removing a directory that contained entries was failing with
2498 the wrong error code being returned to the CIFS client.
2502 * rxdebug exponential back-off. don't take down a server that
2503 is already struggling to respond.
2505 * Network Identity Manager plug-in registers help on the Help
2506 menu if it is supported by NIM. (KFW 3.2 and above)
2508 * Plug-n-play networking did not always work after coming back
2509 from a Windows Suspend. The AFS client service failed to
2510 re-compute the LAN adapter to bind to and instead attempted
2511 to bind to all LAN adapters. This would polute the public
2512 network with the name "AFS" which could lead to conflicts
2515 * Build installers using Microsoft Hotfixes provided at
2516 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2517 Daylight Savings Time issues.
2519 * "fs checkserver" does not require admin privileges unless
2520 the probe interval is being changed.
2524 * A reference leak was discovered in the mountpoint traversal
2525 functionality which resulted in cm_volume_t objects not becoming
2526 eligible for re-use. This will trigger a panic condition when
2527 more than the maximum number of volumes is accessed.
2529 * Fixed the success test for the "fs memdump" operation. Success
2530 is now reported as such. Added cm_volume_t object information
2533 * A data corruption error has been fixed. Write requests
2534 queued for background processing were not retried upon failure.
2535 This leads to inconsistencies between the contents of the file
2536 in the local cache and the contents of the file on the file
2539 The implemented fix is to re-queue any request that has failed
2540 due to timeout, volume busy, volume offline, server busy, or
2541 server offline. Requests are not processed while all servers
2542 for a volume are inaccessible. When servers become accessible
2543 the background request will be retried.
2545 In addition, if a background write fails, do not mark the buffers
2546 as clean. Failure to preserve the dirty flag prevents the buffers
2547 from being written to the file server in the future.
2549 This fix does not address the possibility that the credentials
2550 necessary to perform a write request might expire between the
2551 time the request is queued and when the request can be processed
2552 due to server accessibility.
2554 * The Freelance volume does not support per user ACLs. The
2555 cm_HaveAccessRights function did not handle this special case.
2556 Requests for privileges greater than read|lookup would result
2557 in afsd_service.exe entering an infinite loop. ISOBuster is
2558 known to cause this behavior.
2560 * When deleting files, the SMB Server would attempt to obtain
2561 the access permissions for the parent directory instead of
2562 for the file itself. This would result in an incorrect
2563 assessment of whether or not the file can be deleted.
2565 * The Network Identity Manager AFS plug-in has had its string
2566 tables updated for consistency with the language used by NIM.
2567 All credentials are referred to as credentials instead of tokens.
2568 Drop-down combo boxes now in fact drop down.
2570 * The AFS Salvager was not properly being built on Windows.
2571 This is relevant only if the AFS servers are in use.
2573 * The default signal handlers for the AFS servers incorrectly
2574 processed SIGQUIT. Instead of stopping the server, an
2575 exception was raised.
2577 * Fix another volume refcount leak. This one was triggered by
2578 an inability to contact the vlservers in order to update the
2581 * Fix checklist control used by the drive mappings tabs in
2582 afscreds.exe and afs_control.exe.
2584 * Improve output of "fs memdump" related to buffer management.
2586 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2587 already in the process of being fetched.
2589 * Improve cache performance by computing the buffer hash table
2590 size based upon the number of buffers so that the average
2591 number of entries in a bucket is seven.
2593 * Improve cache performance of read/write data by locally
2594 updating the dataVersion of all buffers associated with the
2595 file on each StoreData or StoreStatus operation provided that
2596 the dataVersion was incremented by one and the dataVersion of
2597 the buffer matched the original dataVersion of the cached
2598 status information for the file.
2600 * When recycling cm_scache_t objects, make sure that we
2601 clear all of the previous fields. It is no longer the
2602 case that we can assume that the stat cache object is being
2603 obtained only after a successful FetchStatus call.
2604 InlineBulkStatus will allocate stat cache objects for which
2605 the no status information is available.
2607 * When a file is being opened with the OPEN_ALWAYS disposition,
2608 do not require write permissions when the file already exists
2609 unless of course the file is being opened requesting write
2612 * When a file has been opened for read-only access, if the
2613 application requests an exclusive lock, obtain a read-lock
2614 instead of a write-lock. The purpose of the exclusive lock
2615 in this situation is to prevent changes to the file and the
2616 full file read lock will provide this guarrantee.
2621 * Enforce Unix Mode Write bit to prevent file deletion of read only
2624 * The CIFS READ and LIST permissions use the same value 0x0001.
2625 When a directory is being accessed, check for PRSFS_LOOKUP
2626 instead of PRSFS_READ.
2630 * Fix computations of short names of mount points. They were
2631 broken as part of the directory search optimization. (1.5.13a)
2633 * When all buffers are in use, periodically sleep the threads
2634 waiting for a buffer to become free while the global buffer
2635 lock is not being held in order to permit those threads that
2636 can free buffers to do so.
2638 * Power Management improvements. Maintain a global flag that
2639 specifies whether or not the service is in a suspend state.
2640 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2641 lana is no longer valid. Instead, stop the listener threads
2642 and if all listener threads are stopped, reset the lana_list.
2643 Allow the cm_Daemon() thread to periodically check the state
2644 of the smb listeners. If they are all stopped and the service
2645 is not suspended, attempt to restart them. If there are no valid
2646 lanas, return to the stopped state.
2648 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2649 the user is permitted to obtain read or write locks. This function
2650 would obtain the lock and then drop it returning whether or not
2651 the lock could be obtained. If the lock was in fact required,
2652 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2653 of course being that this pattern results in three RPCs to the
2654 file server (lock, unlock, lock). This is reduced to one RPC by
2655 implementing cm_CheckNTOpenDone() which frees the allocated byte
2656 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2659 * Optimize the SMB FindFirst processing for the case in which there
2660 are no wildcards. If a cm_Lookup succeeds, generate a response
2661 from the stat cache entry we have a callback for. Its cheaper
2662 than walking each buffer of directory data associated with the
2663 directory stat cache entry when we have thousands of directory
2666 * Increase the SMB Ioctl MaxData size to support larger token sizes
2668 * Disable autorun of the AFS Server Wizard when the AFS Server is
2671 * Add AFS Server service dependencies on tcpip and PNP_TDI
2673 * Remove the AFS Server Configuration Tool shortcut
2675 * Add Freelance mode awareness to AFS Server Configuration Tool
2679 * Unlike previous versions of the OS, Vista performs a shutdown on
2680 the Microsoft Loopback adapter just like it would on a real adapter.
2681 This causes the smb_Listener threads to trigger a panic during a
2682 suspend/hibernate power event.
2684 The fix is to unbind from the network adapters in response to a
2685 suspend/hibernate power event and then rebind to the adapters
2686 when a resume power event is received. Note that the resume events
2687 are not reliably delivered so it is possible the afs service will
2688 not be accessible. However, this is the best we can do.
2690 * don't allow environment variables we set in afslogon.dll to be
2691 inherited by child processes
2693 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2696 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2697 thread walking the all buffer list periodically searching for dirty
2698 buffers to write to the file server. This had several negative
2699 results. The algorithm ate up ever increasing amounts of CPU time
2700 even when AFS is idle as the size of the cache increases. Also,
2701 buffers were written to the file server in an order based upon the
2702 original buffer allocation which has nothing to do with the order
2703 in which the buffers became dirty.
2705 The new algorithm maintains a dirty buffer list. Items are added
2706 when the buffer is originally marked dirty. A buffer is only
2707 removed from the list by the buf_IncrSyncer when the buffer is no
2708 longer dirty. If the list is empty the thread goes back to sleep
2709 immediately without additional processing requirements.
2713 * Enable renaming of directories from the Explorer Shell when
2714 Offline Folders is enabled
2716 * An error is reported in response to the SMB ReceiveCoreFlush
2717 request when an error has not occurred.
2721 * optimize dirty buffer flushes in response to suspend operations
2722 do not flush if only loopback adapter is installed
2723 do no retry dirty buffer writes
2725 * do not reset netbios when performing loopback checks. doing
2726 so will break the existing lana bind
2728 * remove a deadlock in smb3.c QPathInfo
2732 * Add NetIDMgr AFS Plug-in
2734 * "Show Tray Icon" checkbox removed from afs_config.exe
2736 * Disable DNS Registrations for the Loopback Adapter and make sure
2737 that Netbios is turned on.
2739 * A reference count leak on stat cache objects was fixed
2741 * For debugging, tread ids of writers are now tracked on locks
2743 * Corrected an incorrect lock that was being held but not
2744 freed in a rare error case
2746 * Changed trace log end of line to CR-LF so it can be read
2749 * modify the buffer synchronization thread to write all dirty
2750 buffers and sleep only when there is no work to be done
2752 * improve behavior of cache manager when large numbers of
2753 temporary files are being created by multiple clients within
2756 * do not permit dirty buffers that belong to deleted files
2757 to be written to the file server. its a waste of bandwidth
2759 * do not leak stat cache references when the file server
2762 * afscreds: do not permit an attempt to obtain tokens if the
2763 afs service could not be started. this was causing stack
2764 overflows on Vista after resuming from a suspend
2768 * better handling of dot directory
2770 * fs sq and fs mkm now behave as on UNIX
2772 * add license to installers
2774 * Return Invalid Parameter error for unknown InfoLevels in
2775 SMB Tran2 Search Directory.
2777 * Correctly use WSAStartup and WSACleanup within the RX library
2779 * Check that afsd_service.exe is running before performing
2782 * Force AFS Shell Extension to load resource DLL using
2783 afs_shl_ext.dll as a base name.
2785 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2786 OpenAFS\Client\Program
2788 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2790 * Do not panic if the maximum number of volume entries are in use
2791 and one of them can be recycled.
2795 * Fix the code that reads the backconnnectionhostnames value
2796 so that there is not an uninitialized variable
2800 * integrated logon could cause winlogin.exe to blue screen windows
2801 if it was called with a domain that does not exist on the machine.
2802 this is possible during a sysprep.
2804 * corrections to cm_buf_t handling to ensure that buffers are
2805 removed from queues before they are freed and that holds are
2806 obtained on the next buffer in a chain before releasing the
2809 * do not generate MessageBox dialogs on network errors. they
2810 cause more trouble then they are worth.
2812 * move GlobalAutoMapper into a separate thread so that it does
2813 not prevent service startup if the WNetAddConnection2 call
2816 * restrict hard dead and connection timeouts to the CIFS Session
2817 Timeout value. This is being done to try to prevent virtual
2818 circuits from being destroyed prematurely.
2820 * Fix the code that writes the backconnectionhostnames value
2821 to ensure that the data buffer is written with the correct
2824 (there were no windows specific changes between 1.5.4 and 1.5.5)
2828 * Corrections to the Integrated Logon functionality that permits
2829 Kerberos 5 TGTs to be communicated into the user session. The
2830 transfer is now more secure and succeeds when the user does not
2831 have read permission to the %SystemRoot%\TEMP directory.
2833 * NSIS installer updated to version 2.18
2835 * WiX installer updated to version 2.0.4221.0
2837 * Updates to the IFS functionality for 64-bit correctness and
2838 permit compilation. The IFS functionality works for pioctl()
2839 operations but not much else at this point.
2841 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2842 query and set functions.
2844 * Ignore the Windows CIFS client session timeout entirely. Timeout
2845 only after the AFS Hard Dead Timeout period expires.
2849 * move headers, libraries, and samples from OpenAFS\Client to
2852 * remove the auto-registration of AFS IDs for foreign cells
2853 from integrated logon. this permits afslogon.dll to avoid
2854 initializing the RX library. afslogon.dll is a network
2855 provider library and as such it is loaded into all applications
2856 which perform a CreateProcessAsUser call. Initializing the
2857 RX library spawns background threads to listen for inbound
2858 packets. When the network provider is unloaded from the
2859 process, the RX library does not cleanup the threads. If they
2860 continue to execute after the library is unloaded, the process
2863 * implement inline bulk status calls. this will significantly
2864 improve the performance of directory listings.
2866 * prevent a crash when evaluating mountpoints to volumes that do not
2869 * new command: fs uuid [-generate]
2870 displays the current uuid and allows a new one to be generated
2871 on the fly without restarting the afs client service.
2873 * fix the 64-bit installers to properly document the install
2876 * move help files to the proper location so they can be located by
2877 the system tray tool and the control panel.
2879 * "over quota" and "disk full" errors were not being reported to the
2882 * largefile support has been added. Files greater than 31-bits can no
2883 longer be written to file servers which do not support large files.
2884 When file servers support large files, files up to 64-bits can be
2887 * cm_SyncOp could return with the input buffer unlocked upon error
2888 which would result in a panic condition. This bug was introduced
2891 * vlserver failover when freelance mode did not work. This prevented
2892 the root.afs volume from being loaded if the first vlserver could
2897 * A bug in the CIFS Standard Info response packet caused directories
2898 to appear as standard files. This prevented Microsoft Vista from
2899 being able to walk the AFS directory tree.
2901 * Mount points must be reported as directories in response to
2902 Query Path Info operations.
2904 * A deadlock was detected when performing "fs flushall" if the
2905 file server reports VNOVNODE. The scp->createBufferLock is
2906 already held by the current thread and there is no mechanism
2907 to propagate the knowledge. Therefore, an alternate mechanism
2908 for clearing the cache must be developed. A new function
2909 cm_RecycleSCache(scp,flags) has been extracted from
2910 cm_GetNewSCache(). This function performs the task of recycling
2911 an cm_scache_t object. When called from cm_GetNewSCache() with
2912 no flags the expectation is that there are no associated buffers
2913 that are queued to be read or written. When called from
2914 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2915 flag, any queued buffers will be de-queued and marked as if
2916 the operations were performed so the data can be discarded.
2918 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2920 * In response to VNOVNODE, the parent is only discarded if the current
2921 cm_scache_t is not a directory.
2923 * In the Ioctl FlushFile and FlushVolume functions, there were no
2924 checks to protect against flushing the Freelance SCache entries.
2926 * In FlushFile, the wrong cm_scache_t object was being released.
2928 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2929 recycled. Choose a new entry if cm_RecycleSCache fails.
2931 * Speed up the performance of the cache manager by not
2932 holding the smb_fid_t mutex across calls to cm_SyncOp
2935 * Ensure that all smb_fid_t flag references are protected.
2937 * Remove a deadlock between smb_fid_t mutex and
2938 smb_rctLock introduced in 1.4.1.
2940 * Fix aklog AFSID lookup to use not send the realm
2941 name when the realm is the local realm for the
2942 cell. The PT Server doesn't strip the realm
2943 name when it is the local realm.
2945 * Treat "\\afs\*." as an alias for "\\afs\all"
2947 * Dynamically adjust the priority of server threads
2948 based upon the age of the CIFS request that is
2949 being processed. The priority is bumped one level
2950 for every 15 seconds of processing time.
2952 * The original openafs contribution mis-used the
2953 cm_GetCallback function. This function should only
2954 be called as a side effect of a call to
2955 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
2956 was being called results in a FetchStatus RPC being
2957 performed even when the client already has a valid
2958 callback and does so in a manner that destroys the
2959 synchronization of the threads that are calling
2960 cm_SyncOp. If nothing else this results in a
2961 significant performance penalty.
2963 * More changes to cleanup of smb_vc_t objects in
2964 order to prevent race conditions.
2966 * Reduce the amount of Power Management event logging
2969 * If the client detects that the IP address has
2970 changed, force the use of new RX connections. This
2971 enables the file server to recognize that the client
2974 * Remove a race condition and deadlock associated
2975 with cleanup of smb_vc_t objects.
2977 * Fix the ANSI filename support. One function that
2978 required it was missing the translations.
2980 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
2983 * Add a DllMain() function to afspthread.dll. Perform
2984 cleanup of thread local storage, rx queues, critical
2985 sections, event handles, etc. This was implemented
2986 to avoid crashes in afspthread.dll when applications
2987 which load/unload the dll attempt to terminate.
2989 * Lock request behavior has been altered. As of this
2990 release a locally allocated lock will be used whenever
2991 the volume is RO or the requesting user's permissions
2992 are 'rl'. This will allow end users to execute programs
2993 from RW volumes when they do not have the 'k' privilege.
2995 Note that a recent discovery indicates that the 'k' bit
2996 is only supposed to control the ability to obtain read
2997 locks. Write locks are to be controlled by the 'w' and
3000 * Code has been added to enable the cache manager to be
3001 able to obtain the AFSID of the authenticated user.
3002 This is required to be able to determine when "Insert"
3003 privileges should apply. AFS considers a file to be
3004 a new file whenever the owner's AFSID matches the
3005 AFSID of the user. The current implementation relies
3006 on the name specified during the SetToken call to be
3007 correct. This will be replaced in the future with a
3008 new RPC that allows the AFSID of the authenticated
3009 user to be obtained.
3011 * A 32-bit-tools installer for use on 64-bit systems
3012 is now available. These tools are for use with 32-bit
3013 versions of Kerberos for Windows or JPSoftware's 4NT.
3015 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
3017 * Remove the use of the "active_vcp" variable. This variable
3018 no longer serves any useful purpose and was the cause of
3019 race conditions that produced under counts of the smb_vc_t
3022 * Change default minidump output to include data segments
3024 * restore access to cm_IoctlGetVolumeStatus and
3025 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
3026 Still need to figure out why they are being called
3027 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
3029 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
3031 * Update build system to include the "sys_name" in the
3034 * optimize the number of calls to obtain the smb_user_t for a
3035 smb virtual circuit.
3037 * add logging of server addresses, status and capabilities when
3040 * prevent multiple threads from pinging the same server simultaneously.
3041 instead, threads block and then use the result of the ping that was
3042 already taking place.
3044 * Add registry values to allow the defaults for daemon checks to be altered
3045 at startup. Log values to afsd_init.log.
3047 daemonCheckDownInterval
3048 daemonCheckUpInterval
3049 daemonCheckVolInterval
3050 daemonCheckCBInterval
3051 daemonCheckLockInterval
3052 daemonCheckTokenInterval
3054 * Add EventLog messages for transitions in fs crypt state.
3056 * Byte range locking (when locks will be obtained from the
3057 is still under discussion)
3059 * 64-bit Windows operating systems now supported
3060 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
3061 This is a native 64-bit implementation of OpenAFS for Windows.
3062 There is currently no 64-bit version of KFW and therefore
3063 there is no (easy) method for obtaining AFS credentials.
3064 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
3065 to obtain AFS tokens but there is not for the feint of
3068 * A faster implementation of the fcrypt algorithms improves
3071 * The afsd_service.exe now supplies Event Log Message Text.
3072 The event log no longer generates warning text due to lack
3073 of message configuration.
3075 * When afs servers are unreachable and network stack returns
3076 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
3077 mark the attempted server as being down. This allows for
3078 very fast response to the calling application. This change
3079 has a pronounced improvement when the Windows Explorer is
3080 used on disconnected mobile computers.
3082 * Added DebugOutputString debugging of the RX library
3084 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
3086 * Fix the rx-lwp implementation to randomly generate rx call
3087 identifiers which are supposed to be unique at a given point
3088 in time. This affects some of the command line tools used
3089 for debugging but not the AFS Client Service itself which
3090 uses the rx-pthreads implementation.
3092 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
3093 Windows although they are not distributed
3095 * don't generate an exception when logging unexpected SMB packets
3097 * the data structures used by cm_IoctlGetVolumeStatus and
3098 cm_IoctlSetVolumeStatus are not compatible with the CIFS
3099 Get/Set File Information calls so don't bother executing them.
3103 * Fix data written to registry as part of BackConnectionHostnames
3106 * Fix initialization bug when allocating cm_server_t objects that
3107 would cause servers to appear to be down after each allocation
3108 or renewal. Server would appear up after the next check servers
3113 * Windows will return WSAECONNRESET to the next receive packet
3114 operation on a socket if an ICMP packet is received in response
3115 to a packet that was sent and could not be delivered because
3116 the port was unreachable. We have to treat this as
3117 WSAEWOULDBLOCK so that we can retry the select() check and
3118 not mark the connection as bad.
3120 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
3122 * The client manager maintains a least recently used list of
3123 all the ACLs. If an ACL is already the most recently used
3124 do not bother to remove it from the list and place it on the
3127 * As part of the process of adding a new server, set the
3128 server is down flag and then perform a ping. If the server
3129 is up, the flag will be reset. If the server is down, the
3130 client will not timeout unnecessarily while attempting to
3133 * If a Mountpoint String is the empty string, return PATH NOT
3136 * Modify behavior of the Network Provider to avoid querying the
3137 profile location if integrated logon is known to be disabled.
3139 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
3140 will always be preserved regardless of whether or not integrated
3141 logon was used to obtain tokens. When unset, tokens will be
3142 destroyed regardless of whether integrated logon was used to
3143 obtain the tokens unless the profile is stored in AFS.
3145 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
3146 message is being delivered even when the user is not logging
3147 off of the computer.
3149 * Fix a deadlock condition introduced in rc6.
3151 * Cell names are case-insensitive. Since the founding of OpenAFS
3152 there has been a bug that treated "cell", "Cell", and "CELL" as
3153 different cells. This would result in what appeared to be
3154 random disappearance of user tokens or a failure to authenticate
3155 even when tokens were listed by the "tokens" command.
3157 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
3159 * Additional auditing of the smb code revealed that smb ioctl
3160 operations were failing to free smb_fid_t objects
3162 * The File Identifiers assigned to smb_fid_t objects could be
3163 given the value 65535 which is 0xFFFF which is also the
3164 INVALID_HANDLE value. Both 0 and 65535 are now treated as
3167 * Audited all assignments to fields and ensured that they are
3168 protected by obtaining the relevant mutex object.
3170 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
3171 if both the current cm_scache_t object and its parent are
3172 both returning VNOVNODE.
3174 * Removed all the logic associated with the 'dead_vcp' variable
3175 which must have been added for debugging because its used
3176 was never thread safe and it truly served no purpose.
3178 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
3180 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
3181 registry settings. Default is ON and 120 seconds. If set to
3182 OFF, tokens will not be destroyed.
3184 * Fixed a number of race conditions and reference count problems
3185 associated with global queues.
3187 * Added a method to detect abandoned smb virtual circuits so their
3188 associated file handles, locks, user tokens, and smb session
3189 resources can be released.
3191 * Added a mechanism to allow user tokens to be preserved
3192 during integrated login and be freed after logoff.
3194 * Check access permissions to the file before the file is
3197 * More improvements to the handling of SMB Virtual Circuits
3198 when SMB sessions are closed.
3200 * Correct a race condition introduced in rc4 that could result
3201 in a crash of afsd_service.exe.
3203 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3205 * Properly handle pre-mature SMB session closures.
3206 File handles, tree connections, and user sessions are now
3207 freed when the virtual circuit is closed.
3209 * smb_username_t and smb_user_t objects are now properly reference
3210 counted and freed as appropriate.
3212 * disable to use of AFS file locks. byte range locking is still
3213 implemented but all locks are managed locally. This prevents
3214 multiple processes from editing the same file on the same
3215 machine but it does prevent multiple processes on separate
3216 machines from modifying the file.
3218 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3220 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3221 the same server parameter would result in a panic condition
3222 if the server had not already been contacted by the AFS
3225 * Better warnings are written to the afsd.log file when
3226 the Windows RPC Protocols are not properly configured as
3227 is often the case on Windows 2000.
3229 * fix the procmgmt library to restore C RTL signal handlers when
3230 it is being detached from the process prior to process termination.
3232 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3234 * close a lock leak in the rx library that could be triggered while
3237 * prevent removal of drive mappings in response to afscreds.exe -M
3239 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3240 strings and flush the stat cache entry.
3242 * prevent server objects from being freed if user preferences are set.
3244 * Update WIX support to 2.0.2419.0
3246 * Release all locks on the server at service shutdown.
3248 * A down server should be marked up as long as it responds to the client.
3249 It doesn't matter if the server returned an error as long as the error
3250 is something other than VRESTARTING. If there is such an error, all
3251 connections on that server should be marked for replacement.
3253 * Change the check "up" servers period from once per hour to once every
3254 ten minutes. This matches the behavior of the Unix cache manager.
3256 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3257 * Byte Range Locking support has been added. Architectural documentation
3258 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3259 made to enforce Windows mandatory locking model using locally managed
3260 byte range locks with full file locks obtained from the file server.
3261 Read-only volumes provide the user with an implied locking privilege.
3262 Read-write volumes require that a user be granted the locking privilege
3263 in order for a lock to be obtained. As most Windows applications use
3264 Shared file access modes, they require the ability to obtain locks in
3265 order to access files.
3267 * Implement error translation functions to convert Universal Error Codes
3268 output by modern servers to local errno and winsock error values.
3269 This fixes the error message problems using pioctl() calls from fs.exe
3270 and the AFS Shell Extension.
3272 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3273 * Allow the AFS Client Service to set the firewall configuration by
3274 retrying until after Windows determines the boot period is over.
3276 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3279 * Detect IP address changes from within the AFS Client Service for the
3280 purpose of checking down servers sooner
3282 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3283 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3284 user that there is a Kerberos key problem. Previously the user would
3285 receive no feedback when there was something wrong.
3287 * Fix the large integer library support to properly handle values greater
3290 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3291 * Updated CellServDB
3293 * Fixed file name pattern matching. Question marks can match nothing
3294 if found at the end of a component.
3296 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3297 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3299 * The MSI installer now has a new Product ID for each build. This
3300 is to allow for automated upgrades. The Upgrade ID is remaining
3303 * Volumes that were cached and then subsequently were salvaged,
3304 moved or otherwise taken temporarily offline (as opposed to being
3305 busy) would become unreachable. A distinction is now made between
3306 a volume being offline and a server being down.
3308 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3310 * Fixed several AFS Shell Extension issues:
3311 - Checking Server Status would only work if a specific cell was
3313 - If servers were down, the Server Down dialog would write to
3314 invalid memory and cause explorer.exe to crash
3315 - Default Push Buttons were not specified for many dialogs
3316 - The extension would be unloaded by the Explorer Shell and
3317 never be reloaded under some circumstances.
3319 * Fixed the AFS System Tray menu to disappear if another Window
3320 was selected before a menu item was selected
3322 * Added volume owner and group information to "fs examine" output
3324 * Added fs and registry support for enabling/disabling/clearing
3325 rx statistics gathering.
3327 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3329 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3330 GUIDs for each platform.
3332 * Documentation directory structure and default installation rules
3335 * New OpenAFS for Windows HTML Release Notes installed
3336 as part of the documentation set.
3338 * Add thread id to rx debugging messages
3340 * When all servers are offline, return a bad network path error
3341 immediately. This is necessary to allow Explorer to do the
3342 right thing and not hang. The background daemon thread will
3343 check the down servers every 30 seconds to see if they are
3346 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3348 * The AFSCache file has been modified to store both the
3349 serial number of the volume on which the file resides
3350 and the SID of the local machine. If either change,
3351 a new UUID is generated.
3353 * add "fs flushall" command that forces all data buffers
3356 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3358 * add support for '/' instead of '\' in pioctl() calls
3360 * Apply AFS Client Admins group protect to AFS Shell Extension
3362 * Add support for \\afs\<name> to most AFS Shell Extension
3363 commands except the symlink methods.
3365 * when installing the Microsoft Loopback Adapter, enable
3366 MS Client for Networks and prevent an install failure
3367 by not calling CoInitialize twice in the same thread.
3369 * reload cell vldb values from the CellServDB every two
3370 hours in case it changes
3372 * When updating cell information from DNS, be sure to set
3375 * Add support to allow use of \\AFS\<foo> where <foo> is
3376 either a mount point or symlink. As <foo> is normally
3377 treated as a share name, we transform it into \\AFS\all\<foo>
3380 * Init 'code' to prevent false errors when integrated logon
3381 is disabled and the service is not running
3383 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3385 * Removed trace log messages that were hampering performance.
3387 * Fixed a deadlock that was being triggered by editing Word
3388 documents stored within AFS with WinWord 2003.
3390 * Bit 3 of the TraceOption registry value is now used to set the
3391 default for "fs trace" in the non-Debug builds. The new default
3392 is off for release builds and on for debug builds.
3394 * Bit 2 of the TraceOption registry value can be used to turn on
3395 real-time output of debug log entries to the Windows Debug
3396 Output monitor interface. This data can be viewed with tools
3397 such as Sysinternal's DbgView.
3399 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3400 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3401 a resource we should not be returning status codes such
3402 as Network Busy or Remote Host Down because these errors
3403 will imply that the AFS Client Service is Busy or Down
3404 and that is simply not the case. Instead we will return
3405 Bad Network Path as the path is temporarily not available.
3407 Instead of returning Sharing Paused when there is a
3408 timeout error, return Timeout. Once again, the AFS Client
3409 Service is not paused.
3411 * afscreds.exe would display an Obtain Creds dialog when
3412 the expired credentials reminder was triggered even
3413 if there was no network path to the KDC. This is prevented
3414 by adding KDC probe logic to the reminder thread.
3416 * afscreds.exe would display expired tokens no differently
3417 than unexpired ones. This would make it difficult for the
3418 user to distiguish when the tokens were expired. For the
3419 English build added a new resource string "(expired) that
3420 is displayed instead of the expiration time.
3423 * "fs wscell" when executed in freelance mode will return the
3424 name of the cell configured in the registry. The root.afs
3425 volume is not loaded from this cell, but it is used for the
3426 default for aklog and integrated login.
3428 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3429 no longer work on \\AFS\all when freelance is being used
3430 unless the user is a member of the "AFS Client Admins" Windows
3433 * some more checks performed during persistent cache validation.
3436 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3437 down servers for the current cell not all cells.
3439 * If the scache LRU list becomes corrupted, the AFS Client Service
3440 might enter an infinite loop at startup. Detect the corruption
3441 and discard the cache contents in this case.
3443 * Fixed the Explorer Shell's remove mountpoint functionality.
3445 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3446 obtain a mutex lock on a cm_scache_t object whose lock is held by
3447 a thread attempting to obtain the cm_aclLock.
3450 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3451 misinterpretted as "csail.mit.edu" when the resolver
3452 is configured to append ".mit.edu" to failed DNS queries.
3454 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3455 that can be used to force the preservation of user tokens upon logout.
3457 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3458 recommended installation categories: AFS Client, AFS Administrator,
3459 AFS Server, AFS Development Kit. Each category includes a different
3460 default subset of the OpenAFS components.
3462 The OpenAFS logo is now associated with the NSIS uninstall entry
3463 in the Add/Remove Programs control panel.
3465 * The user name associated with AFS tokens when obtained with
3466 integrated login, the afs systray tool, or aklog will always
3467 include the full Kerberos 5 user principal name regardless of
3468 whether or not the cell is local to the realm.
3470 * Modify integrated login so that it does not enter an infinite
3471 loop if the service is not set to auto start.
3473 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3476 * Added uninstall.exe to Wix installation
3478 * More modifications to algorithms used to wake sleeping threads.
3482 * Changes to the algorithms used to wake threads when they are
3483 put to sleep because of conflicts over buffer operations.
3486 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3487 the process of producing AFS tokens. The tokens are stored within
3488 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3489 New functionality has been added to temporarily copy the tickets to
3490 a file ccache during the short window between the NPLogonNotify and
3491 WinLogon Logon Event. The Logon Event handler starts a process as
3492 user, afscpcc.exe, that copies the tickets into the default ccache.
3493 The file ccache is then destroyed by the event handler.
3495 * During synchronization operations on stat cache entries, if threads
3496 are forced to sleep it was possible that the threads would not be
3497 woken under heavy load.
3499 * Enforce the delete on close flag specified by the NTCreateX()
3502 * A race condition in the rx library was fixed that resulted in
3503 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3504 was cleared when threads were still waiting to use the
3505 connection. Also, fixed a potential case in which the
3506 connection could be destroyed while threads were waiting to use it.
3508 * The DNLC freelist has been seen to become corrupted with still
3509 active entries being stored on the freelist. Changes were made
3510 to perform a better job of cleaning entries before placing them
3511 onto the freelist; marking them with the correct magic value when
3512 purging the list; and allowing the list to be purged on startup
3513 if the cache validation fails.
3515 * Windows Crash Reporting does not get triggered for afsd_service.exe
3516 because it provides its own top level exception handler. This patch
3517 places an explicit call to ReportCrash() on platforms which support
3520 If you configure Crash Reporting via AD Policy to capture crash
3521 reports within the domain, then you will receive this output.
3523 * Generate MiniDumps independent of the Windows Crash Reporting.
3525 (a) If an exception occurs,a minidump will be generated at the path
3526 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3527 can be set to any other type via use of the
3530 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3532 (b) MiniDumps can be generated at any time using the "fs minidump"
3533 command. This allows you to generate a minidump without
3534 stopping the process or even requiring the use of a debugger
3537 * an RX connection which reports bad ticket now treats it the same way
3538 the expired ticket is treated. The ticket is removed from the
3539 connection and a new anonymous connection is established.
3541 * fs memdump now requires that the caller be in the AFS Client Admins
3544 * additional debugging information is provided in cm_Analyze when the
3545 error code is ignored.
3547 * Fix a regression introduced into the pioctl() client code which
3548 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3549 MIT KFW is not installed. (1.3.82a)
3551 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3554 * Fix a race condition in cm_EndCallbackGrantingCall() that
3555 could leave 'nrevp' pointing to freed memory. In the process,
3556 optimize the processing to ensure that the applications monitoring
3557 the status of the stat cache entry are only notified at most
3560 * Conditionalize access to the Explorer Shell AFS->Mount Point
3561 ->Remove menu item. Disable the item if the selected item(s)
3562 are not mount points.
3564 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3565 Local Storage. This eliminates thread contention on the
3566 global RX free packet queue. This should improve performance
3567 on multi-CPU systems.
3569 * Fix the IP address filtering code to properly load/unload
3570 the IPHelper DLL so that Windows XP/2003 and beyond systems
3571 do not have to manually probe the registry.
3573 * CleanupACLEnt() was not being called consistently with
3574 the cm_scache_t object referred to by the back pointer
3575 mutex locked. This could in very rare conditions lead to
3576 an invalid memory access.
3578 * Added a script command to the msi installer "afs_replace"
3579 which can be used to replace one version of OpenAFS with
3580 another without requiring an uninstall and reinstall if
3581 the installed msi is no longer in its original location.
3583 * Apparently the problem with multi-domain forests with cross-
3584 realm trusts to non-Windows realms was not entirely solved.
3585 The authentication to the AFS SMB service failed because
3586 the wrong name was being used. Using ASU as an example,
3587 the authentication was being performed with the name
3588 "QAAD\user" (an account in the forest root) and not
3589 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3591 The solution was to add an additional dependency on KFW
3592 in order or to be able to easily obtain the client principal
3593 name stored in the MSLSA ccache TGT. This information is
3594 used in two locations:
3596 - the pioctl() function
3598 - a new WinLogon Event Handler for the "logon" event.
3600 The pioctl function will now be able to use the correct
3601 name when calling WNetAddConnection2() and the "logon"
3602 event handler will now be able to call WNetAddConnection2().
3603 The hope is that the "logon" event handler will be called
3604 before the profile is loaded but I have not guarrantee
3608 * Fixed a locking error in cm_TryBulkStat() which had the
3609 potential of deadlocking the system for the length of time
3610 it takes to perform a bulk status fetch operation.
3612 * Replaced time conversion code (UnixTime <-> FILETIME) to
3613 be completely arithmetic instead of relying on a bizarre
3614 algorithm involving a variety of C RTL time functions.
3615 This has the side effect that UnixTime and FILETIME which
3616 are both stored in UTC are interpretted as UTC throughout
3617 the year. Windows will apply the same localization to AFS
3618 as it does to NTFS. Applications which rely on the ability
3619 to sync files between the two file systems will no longer
3620 see the timestamps of files in AFS change an hour relative
3621 to the files stored in NTFS or Windows based backup devices.
3623 * Fixed a invalid memory access under a bizarre circumstance.
3624 Windows will allow a physical mass media device to be
3625 installed via Plug N Play to the system and will assign it
3626 the lowest available drive letter. This is true even if the
3627 drive letter is currently assigned in the user session to
3628 a network device via NET USE (or its equivalent.) When this
3629 happens, queries sent to the CIFS server will contain invalid
3630 data. This invalid data was not being caught by the AFS
3631 Client Service and was resulting in a crash when Freelance
3634 * Fixed a reference count error when registering callbacks
3635 on a stat cache object if the callback was already registered
3637 * Add a case to cm_Analyze() in order to handle the case where
3638 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3639 the server list for the volume is empty. In this case, force
3640 an update of the volume info and retry.
3642 * Insert a missing cm_EndCallbackGrantingCall() which could
3643 result in threads waiting for a callback to complete to never
3646 * In the persistent cache, there is a maximum number of volume
3647 entries. Allow volume entries to be reused if the maximum
3648 number have been allocated and their reference count is zero.
3650 * If we already have a dead virtual connection object, don't
3651 ignore the fact that additional dead vc objects must be taken
3654 * Removed a deadlock condition introduced in the summer of 2004.
3655 Do not hold mx locks around calls to RX functions including
3656 rx_NewCall(). That is what reference counts on the rx_connection
3659 * Fixed an initialization error in afslogon.dll which could
3660 result in random behavior including a failure to terminate
3661 the NPLogon function. (1.3.80b)
3663 * Fixed an error preventing the use of SMB authentication
3664 on Windows 2000. (1.3.80a)
3667 * Updated CellServDB to Public CellServDB 16 Mar 2005
3669 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3670 misinterpretted as "csail.mit.edu" when the resolver
3671 is configured to append ".mit.edu" to failed DNS queries.
3673 * Fixed another case in which the client would replace connections
3674 to the host when it was not necessary. In this case the
3675 situation would occur if cryptall was on and the connections
3676 were unauthenticated due to lack of an appropriate token.
3678 * OpenAFS for Windows has failed to work at sites which are
3679 utilizing a cross-realm trust between an MIT/Heimdal realm
3680 and a multi-domain Windows forest when the workstation being
3681 accessed is not located in the root domain. This is caused
3682 by a bug in the workstation which was triggered after the
3683 introduction of Windows 2003 Server. When the bug is triggered,
3684 the workstation attempts to authenticate users to afsd_service.exe
3685 by contacting the Domain Controller instead of using the
3686 LSA loopback authentication mechanism.
3688 One of the reasons this bug occurs is because the workstation
3689 does not have a reliable way of knowing that the service whose
3690 netbios name is "AFS" is located on the workstation. This will
3691 be fixed starting in Longhorn Beta 1 by Microsoft. The
3692 "BackConnectionHostNames" registry value will be used to
3693 indicate that the authentications to that service name should
3694 be performed using the loopback authentication mechanism.
3696 In the meantime, when Logon Caching is enabled, we can force
3697 afsd_service.exe to authenticate using the logon cache before
3698 contacting the Domain Controller. This will work with both
3699 password and smart card based logons.
3701 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3702 flag should not be cleared if a volume's server reference is
3703 marked as "offline". In the case where all of the volume's
3704 servers are either "down" or the volumes are "offline", we want
3705 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3706 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3707 should never occur in practice.
3709 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3710 5 seconds, clear the server down and volume busy flags, and
3711 then force an update from the VLDB. This allows the client to
3712 update the location of a volume if the reason for it being
3713 marked offline is because it is being moved. Calls to
3714 cm_ConnByMServers() will be retried either until success or
3715 the RDRtimeout period is reached.
3717 * Correct the Power Management code behavior in response to
3718 Standby, Suspend, and Shutdown events. Instead of flushing the
3719 buffers associated with the mounted SMB submounts, simply write
3720 all buffers which are marked dirty.
3722 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3723 provides a list of cells other than the default cell for which
3724 tokens should be obtained using the default Kerberos principal.
3725 This functionality is logon domain specific and is only available
3726 when using KFW for authentication.
3728 * Fixed FindNext Invalid Handle error caused by over agressive
3729 attempts at garbage collecting dirSearch entries when the
3730 dirSearch ID wraps from the maximum value to 1.
3732 * Add support for registry defined server preferences for VLDB and
3733 File servers. See registry.txt for details.
3735 * Increased default cache size to 96MB and default number of cache
3738 * Fixed refCount leaks related to directory lookups and pioctl calls.
3740 * Callbacks revoked during a race condition with an attempt to
3741 obtain the same callback no longer result in an inaccessible
3744 * New command line tool:
3746 afsdacl : Set or reset the DACL to allow starting or stopping
3747 the afsd service by any ordinary user.
3749 Usage : afsdacl [-set | -reset] [-show]
3750 -set : Sets the DACL
3751 -reset : Reset the DACL
3752 -show : Show current DACL (SDSF)
3754 * IP addresses are no longer queried once at startup. Instead IP
3755 addresses are obtained as needed. Loopback adapter addresses are
3756 no longer published to the server.
3758 * Pay attention to the MIT KFW registry configuration for automatic
3759 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3761 * Fix (once again) case-sensitive comparisons which was apparently
3764 * Activate support for vos listvol -format
3766 * Population of the "AFS Client Admins" group failed on non-English
3767 versions of Windows because the name of the "Administrators" group
3768 is localized. Now we lookup the name of the group by using the
3769 Administrators Group Alias Relative ID.
3771 * Multiple cell token acquisition within afscreds.exe was broken in the
3772 case of Kerberos 5 cross realm authentication. Instead of contacting
3773 the KDC associated with the cell's realm, afscreds would obtain a
3774 token from the KDC of the user's realm. This would result in
3775 "invalid kvno errors" while authenticating to the AFS servers.
3776 Unauthenticated access would work. The symptoms would vary based upon
3777 whether or not the VLDB servers had been contacted using unauthenticated
3778 connections prior to the user obtaining tokens.
3780 * The list of ACL entries was becoming corrupted. It appears as if
3781 an ACL when it expires was not being appropriately cleaned up. In
3782 fact, it was left in the list of ACLs associated with the scache
3785 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3786 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3787 and "amd64_win64" for amd 64-bit processors.
3789 * When executing executables, Windows provides a mechanisms by which
3790 the normal search for DLLs can be bypassed. If a file foo.exe is
3791 being executed and there exists a file or directory "foo.exe.local"
3792 (the contents are ignored if a file), then Windows will search for
3793 DLLs first in the "foo.exe.local" directory and second in the
3794 directory in which the "foo.exe" file is located.
3796 Previous releases of OAFW would improperly return
3797 STATUS_NOT_A_DIRECTORY instead of the expected
3798 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3799 to terminate the search for the DLL causing the execution of the
3800 application to fail.
3802 * Rework the reference counting of smb_vc_t objects. The references
3803 stored in the waiting locks were not counted. This could result in
3804 an assertion if the reference count drops to zero.
3806 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3807 when running in freelance mode.
3809 * Added support for persistent caching of file contents, stat entries,
3810 volume data, cell data, and name lookup data. When the registry
3811 value, "NonPersistentCaching", is not defined or set to 0 the file
3812 specified by "CachePath" becomes a persistent cache file. The
3813 size of the cache file is computed at run time. It has a maximum
3814 size of approximately 1.2GB. The contents of the file will be
3815 validated according to the rules specified by the "ValidateCache"
3818 * Added support for UUIDs. UUIDs are kept for as long as the
3819 cache file is intact.
3821 * Added cmdebug.exe and support cache manager debugging callback
3822 interfaces in afsd_service.exe.
3824 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3825 [-addrs] [-cache] [-help]
3826 Where: -long print all info
3827 -addrs print only host interfaces
3828 -cache print only cache configuration
3830 * Symlinks to \\AFS\[all\]... will now be treated the same as
3831 symlinks to /afs/.... However, please use /afs/... as otherwise
3832 the symlinks won't work on Unix.
3834 * Correct a problem with local CellID allocation for cells whose
3835 server list is obtained via DNS instead of a CellServDB file. If
3836 the DNS information expires the CellID assigned to the entry will
3837 be changed. This causes all of the cm_scache_t objects which refer
3838 to the old cellID number to become useless. Attempts to access
3839 files or directories with cm_scache_t objects using the old cellID
3840 will fail since the server list cannot be obtained.
3842 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3843 be triggered if the volume referenced by the cm_scache_t object is
3846 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3847 in addition to SERVICE_ACCEPT_STOP.
3849 * Move RPC shutdown until after the SMB and RX shutdown procedures
3850 complete. Block until RPC shutdown is complete.
3852 * Modify afslogon.dll (integrated logon) to wait for service if its
3853 state is START_PENDING. If the timeout period occurs, reset to
3854 the retry interval and not the sleep interval.
3856 * When renewing the server list for a cell obtained via DNS AFSDB
3857 records, the cm_cell_t entry must be removed from the list of all
3858 cells. Otherwise, the list of cells will be corrupted.
3860 * In the dcache and scache modules, use the cm_scache_t dataVersion
3861 instead of the cm_fakeDirVersion.
3863 * Synchronize fs.c with the unix version.
3865 * The variable used to determine whether a file or virtual memory
3866 mapped cache is used was not properly initialized to a default
3867 value. If the registry setting "NonPersistentCaching" was not
3868 set, the choice would be random. Properly initialized to be
3871 * The memory mapped view was never unmapped before closing the file
3872 at service shutdown. This is now properly cleaned up.
3874 * Default location of Cache file is now %TEMP%\AFSCache
3878 * A bug affecting new installations of 1.3.75/76 would result in
3879 the creating of incorrect mountpoints in the freelance root.afs
3880 volume for the default cell.
3882 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3883 of "root.cell", you have been affected by the bug. To correct the
3884 problem, execute the following commands:
3886 fs rmmount \\afs\all\<cellname>
3887 fs rmmount \\afs\all\.<cellname>
3889 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3890 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3894 * A bug has been fixed in the auto-generation of Freelance root.afs
3895 symlinks which produced random entries in the \\AFS\all directory.
3897 * Support has been added for multi-homed servers
3900 * Added a new registry value, "StoreAnsiFilenames", which can be used
3901 to force the use of ANSI character sets instead of OEM Code Pages.
3902 This feature is useful when users require the ability to create
3903 filenames with 8-bit characters and need to access the files from
3904 both Latin-1 based Unix systems as well as from Windows.
3906 Activation of this feature will prevent access to files stored with
3907 8-bit OEM characters.
3909 * Shutdown all SMB threads in a synchronized manner when stopping the
3912 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3913 by the largest contiguous block of unused memory within the 2GB process
3914 space which can be assigned to the memory mapped file. Unfortunately,
3915 when the executable digital signature verification code is activated
3916 Windows sees fit to further segment the process memory which in turn
3917 reduces the size of the maximum cache file to less then 800MB. If
3918 larger cache sizes are desired, a new registry value should be set:
3920 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3922 Setting this value will disable the runtime verification of digital
3923 signatures on afsd_service.exe and the afs dlls which it loads. It
3924 will not disable the the version number check on those same files.
3925 The signature verification is not a security messure and is only meant
3926 to enhance the ability to afsd_service.exe to detect potential
3927 destablizing mixtures of DLLs from incompatible distributions.
3929 Added code to auto-disable the signature verification check if
3930 the desired cache size is greater then 700MB.
3932 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3933 used to verify the validity of the afsd_service.exe binary
3934 as well as each of the AFS DLLs loaded by the service. Not only
3935 must the digital signature be valid but the signatures of the
3936 DLL must be signed by the same entity as the service.
3938 * Implement new functions: cm_freelanceMountPointExists and
3939 cm_freelanceSymlinkExists. Use them along with other validity
3940 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3941 ensure that name collisions do not occur and that empty strings
3942 are not valid file names.
3944 A symlink may not have a name which would resolve to a valid
3945 cell name. Doing so would prevent access to the cell.
3947 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3948 re-initialization code. The reference counts of the fake root.afs
3949 volume scache object(s) would become invalid when the mount point
3950 or symlink lists were altered.
3952 * Add registry entries to provide mappings from the afsdsbmt.ini
3953 to the new locations for applications which count on the use
3954 of the old Profile file APIs. These apps are likely to fail
3955 if the user does not have administrator privileges and the
3956 registry is locked down.
3958 * The afs_config.exe submounts dialog had two errors.
3959 First, attempts to remove entries failed because the registry
3960 key was being opened without KEY_WRITE privileges.
3961 Second, when editing a submount entry, changing the name
3962 would add a new key and leave the original one in place.
3963 Now the original submount will be removed if its name is
3966 * In recent months there have been several incidents in which
3967 users have experienced problems starting or accessing
3968 afsd_service.exe and after significant effort has been spent
3969 it has turned out that they have two versions of AFS on the
3970 machine or an inconsistent set of DLLs.
3972 Code has now been added to afsd_service.exe which will walk
3973 the list of modules loaded by afsd_service.exe and validate
3974 that the version of the AFS DLLs matches the version of the
3975 afsd_service.exe executable. If they do not match the service
3978 * When Freelance mode is enabled and there is no registry
3979 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
3980 will attempt to import the afs_freelance.ini file contents.
3981 If the file does not exist, it was creating a dummy file
3982 with a r/o and r/w entry for the default cell and then
3983 importing those values.
3985 This process has been changed. The temporary file is no
3986 longer created. Also, both the OpenAFS Client install
3987 directory as well as %WINDIR% are checked for previous
3988 afs_freelance.ini files.
3990 * Added support for VL_GetEntryByNameN(). Still need to add
3991 support for VL_GetEntryByNameU() for multi-homed support.
3993 * Fix a deadlock situation in afscreds.exe when canceling an
3994 auto-generated Obtain Tokens dialog
3997 * Fix the locking of objects during Directory Searches in the
3998 SMB/CIFS server. The failure to properly lock the reference
3999 counts was resulting in the premature freeing of smb_dirSearch_t
4000 objects while they were still in use by the SMB/CIFS client.
4001 This does not solve the "Invalid Handle" problem.
4003 * Fix Find Cell By Name pioctl call to return a valid cell
4004 name for the Freelance fake root.afs volume.
4005 "Freelance.Local.Root".
4007 * Fix the Explorer Shell Extension Symlinks->Add operation.
4008 The dialog template was missing and the link destination
4009 string was too short.
4011 * Add support for symlinks to Freelance root.afs volume
4012 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
4013 <number> = "<linkname>:<relative-path>."
4014 Use symlink.exe to create, list, or remove
4016 * Remove the fallback to the use of KFW's KRB4 library when obtaining
4017 tokens. We never obtain KRB4 tickets.
4019 * Fix AFS Client Configuration Control Panel to support new SysName
4022 * Fix a bug in afsd_service.exe which could result in the SysName
4023 not being read from the registry.
4026 * Add code to block the issuance of AFS tokens by aklog.exe or
4027 afscreds.exe when the Kerberos 5 principal name contains a dot.
4029 * Modify the IsAdmin() function to always treat the local SYSTEM
4030 account as an AFS client administrator. Affects fs.exe and
4033 * Modify the internal handling of Quota Exceeded errors
4035 * Upgrade all reference count fields in the Windows cache manager
4036 and the osi library to use unsigned long instead of signed short.
4037 A similar fix has been applied to the afs rpc (rx) library.
4039 * fix the Windows cache manager to prevent it from replacing the
4040 rx_connection object associated with the cm_conn_t object on each
4041 and every operation if "fs crypt" was set. This explains the
4042 dramatic performance difference when crypt is used vs clear.
4043 The problem: 'cryptall', a boolean flag indicating whether or not
4044 "fs crypt" is set, was being compared to the rx_connection
4045 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
4046 1 != 2 and therefore the rx_connection was always destroyed
4047 and replaced on each and every operation.
4049 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
4050 It is not safe for the cm_conn_t object to not be locked because
4051 rx_DestroyConnection might be called from another thread if:
4052 - the user's tokens have changed or expired
4053 - the crypt mode has changed
4055 This fix appears to have also taken care of the problems associated
4056 with Overlapped Writes resulting in Delayed Write errors.
4058 * fix NSIS installer's AdminGroup.exe to properly create and
4059 remove groups when given -create or -remove. The string comparison
4062 * fs sysname now accepts a list of sysname values
4064 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
4065 DWORD which when set to a non-zero value will cause error message
4066 text to be output to stderr from the pioctl() routine. Useful in
4067 debugging failures of fs.exe, tokens.exe, etc.
4069 * added a test to the power management code to only perform a
4070 flush operation if there is at least one network adapter which
4071 is not a loopback adapter.
4073 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
4074 "EnableKFW". This value will not be read if the key
4075 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
4076 value under that key does not.
4078 * provide mechanisms to force the use of krb524d for Kerberos 5
4079 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
4080 there is a new registry value "Use524" and for aklog.exe a new
4081 command line parameter "-m".
4083 * Fix the pattern matching algorithm to properly match patterns
4086 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
4087 which is used by both the original function and the new
4088 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
4089 creation of HardLinks in addition to Renaming. smb_Link()
4090 is a new function which creates HardLinks via cm_Link().
4091 cm_Link() is a new vnodeops function which creates links
4094 smb_ReceiveNTRename() does not support the File Copy and
4095 Move Cluster Information operations described in its interface.
4096 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
4098 * When opening files via symlinks, we should follow the symlinks
4099 until we reach the actual file stat cache entry. The stat cache
4100 entry of the file should then be stored in the FID instead of
4101 stat scache entry of the symlink.
4103 * return bad operation errors for all unimplemented functions
4104 even if we do not know the functions exist.
4106 * Log bad packets and unknown operation packets to the trace log
4108 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
4111 * Update list of known CIFS operations to include all those listed
4112 in CIFS-TR-1p00_FINAL.pdf.
4114 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
4115 to support the REG_EXPAND_SZ type.
4118 * A new Windows authorization group "AFS Client Admins" is now
4119 created and populated with the members of the "Administrators"
4120 group. The group is used to determine which accounts on the
4121 machine may be used to modify the AFS Client Configuration via
4122 the UI and command line tools. afs_config.exe, fs.exe,
4124 * Modify the WinLogon Logoff Event Handler to query NT4 domain
4125 controllers for the remote profile path if Active Directory
4126 services are not available.
4128 * Fix aklog.exe to not add the AFS ID to the username
4130 * PTS registration of new users to foreign cells has been added to
4133 * The cm_Daemon thread is used to perform checks for
4134 down servers, up servers, volumes, callback expirations,
4135 lock maintenance and token expiration. Due to a gaff in
4136 larger integer division the thread never performed any
4137 work. Instead the current time computation would always
4138 be less then the trigger times. This had an adverse affect
4139 on the client's ability to maintain communication with servers,
4140 keep volumes up to date, and flush user tokens and acls
4141 when they have expired. This was broken when the 1.3 branch
4142 was modified to support VC7 which no longer included
4145 * An initialization problem with the Freelance code was
4146 detected while fixing the callbackRequest. The cm_rootSCachep
4147 object is obtained during afsd_InitDaemons() but the callback
4148 information is incomplete. The callback information will not
4149 be obtained until cm_MergeStatus is called from within
4150 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
4151 test for the conditions under which the callback information
4154 * Reports have been filed indicating that callbacks were
4155 being lost. An examination of the code indicated that the
4156 cm_server_t objects were not being properly reference
4157 counted by the cm_scache_t and cm_callbackRequest_t objects.
4158 In particular, the cm_server_t objects may have been freed
4159 from beneath the cm_conn_t objects.
4161 All of the reference counting is now done via the functions:
4166 this improves the ability to track the referrals.
4168 Each cm_BeginCallbackGranting Call now allocates a reference
4169 to the cm_server_t. The cm_EndCallbackGrantingCall either
4170 frees the reference or transfers it to the cm_scache_t
4171 cbServerp field. These are then appropriately tracked
4172 through the cm_Analyze call.
4174 * Ensure that the dnlc hash table is the same size as the
4175 dir name hash table (as per original author's note).
4176 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
4177 for compatibility with 64-bit systems.
4179 * fix smb_ApplyV3DirListPatches to properly apply the hidden
4180 attribute to dotfiles when the infoLevel < 0x101 and
4181 cm_SyncOp has failed.
4183 * Fix the Freelance registry initialization code. There
4184 was a possibility that some systems could end up with
4185 garbage in the registry during a clean install.
4188 * file and directory names beginning with "." will now be given the
4189 hidden attribute when the volume access is anonymous. this matches
4190 the behavior when the volume access is via an authenticated user.
4192 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
4193 key. When a change occurs mark the root.afs data as invalid and
4194 for it to be reloaded on the next access. This allows administrators
4195 to modify the mount point list without restarting the service.
4197 The freelance client used to provide a fake modification time for
4198 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
4199 Added code to extract the last modification time of the Freelance
4200 registry key and use that instead. The time now represents the
4201 most recent mount point change.
4203 * PTS registration of new users to foreign cells has been added to
4206 * Additional Cache Control and Credential Manager options have been
4207 added to the WiX installer. See deployment guide for details.
4209 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4211 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4214 * The loopback installation code may have had a problem updating the
4215 %ETC%\HOSTS file which could have resulted in a premature failure.
4216 Work around code has been added for the case where the file cannot
4219 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4220 because Windows sends 64K blocks when using overlapped writes.
4222 * The default number of server threads was increased from 4 to 25 to
4223 better handle overlapped writes.
4225 * The "AfscredsShortcutParams" registry value was not being properly
4226 loaded by afscreds.exe. Therefore, the default value was always being
4227 used instead of the value set by the installer.
4229 * Windows XP provides downgrade attack detection to prevent an attacker
4230 from being able to force the use of NTLM simply by disrupting
4231 communication with the KDC. This attack cannot exist between the
4232 Windows CIFS client and the AFS Client Service. Therefore, when a
4233 downgrade has been detected the afs pioctl library will force the
4234 establishment of a new CIFS connection using NTLM.
4236 * A locking error was discovered surrounding all references to volume
4237 server lists within the cm_cell.c source file.
4239 * The logged into Windows username was incorrect on Terminal Server
4242 * A new registry value "NonPersistentCaching" was added to the service
4243 parameters key. When set to a non-zero value, the afs cache is stored
4244 in the Windows paging file. There are two limitations to choosing
4246 1. when persistent caching is implemented it won't work with
4247 this flag set since there will be nothing to persist.
4248 2. with this flag set the initial paging allocation cannot be
4249 changed while the service is running
4251 * An initialization bug was discovered in aklog.exe which affected users
4252 who have a domain name for their afs servers which could not be mapped
4256 * afs_config.exe now validates cell names against DNS in addition
4257 to the CellServDB file.
4259 * In order to allow the freelance client to connect to a volume with ID
4260 equal to 1 on the default cell we changed the fake root.afs volume ID
4261 once again. This time we choose 0xFFFFFFFF. In addition, we change
4262 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4263 It will now be impossible for a volume ID to match that of another
4264 cell unless the client is connected to 0xFFFFFFFD cells. That should
4265 be enough room for growth.
4267 * Fix "fs mkmount" command to work with UNC paths and when
4268 started from non-AFS drives. It is now possible to create a mount
4269 point in the freelance fake root.afs volume with the command
4271 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4275 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4276 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4278 * The algorithm used to re-attempt access to the servers associated with
4279 a volume has been altered to properly address the case in which all
4280 servers have been marked down. The previous algorithm did not reset
4281 the server's down flags so the servers were never actually retried.
4282 This caused a problem with active volumes if the network connectivity
4283 was lost as could be the case with a network cable removal, wireless
4284 drop, or laptop hibernation. With the fix volume access is restored
4285 almost instantenously when network connectivity becomes available.
4287 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4288 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4289 listing of AFS submounts and freelance mount points. Support for
4290 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4291 significantly enhances the behavior of AFS volumes within the Explorer
4292 Shell. For instance, "AFS" now shows up as server in the Explorer
4293 with each submount or freelance mount point visible as a share.
4294 The right click menu in each folder now works with full functionality
4295 on a consistent basis.
4297 * The network provider can be configured to have different behavior
4298 depending on the domain that the user logs into. These settings are
4299 only relevant when using integrated login. A domain refers to an
4300 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4301 local machine (i.e. local account logins). The domain name that is
4302 used for selecting the domain would be the domain that is passed into
4303 the NPLogonNotify function of the network provider. (see registry.txt
4306 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4307 "Authentication Cell" which may be used to specify a default
4308 authentication cell for afscreds.exe which is different from
4309 the default cell for the AFS Client Service daemon.
4311 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4312 afslogon.dll moved to %WINDIR%\System32\.
4313 New registry entries added to register the dll for Winlogon events.
4315 The logoff event will now force a call to ktc_ForgetAllTokens()
4316 using the context of the user being logged off as long as the
4317 user's profile is not loaded from within AFS. If the profile
4318 was loaded from AFS we can't release the tokens since the Logoff
4319 event is triggered prior to the profile being written back to
4320 the its source location. This is now performed in an XP SP2
4323 * Windows XP SP2 Internet Connection Firewall interoperability
4326 * The %WINDIR%\afsdsbmt.ini contains four sections:
4327 Submounts, Drive Mappings, Active Maps and CSC Policies.
4328 The Submounts and CSC policies are now stored in the registry under
4329 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4330 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4331 The Drive Mappings and Active Maps are stored in the registry under
4332 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4333 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4335 There is no automatic migration of this data as it would be impossible
4336 to consistently migrate data to user profiles which may not be active
4337 when the machine is updated.
4339 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4340 fake root.afs volume. For the same reasons as for the cellservdb file,
4341 this information should not be in %WINDIR%. This information is now
4342 kept under the registry key
4343 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4345 The data from the afs_freelance.ini file will be automatically
4346 migrated to the registry on first execution of afsd_service.exe
4348 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4349 troublesome for several reasons. One, it is confusing for those who
4350 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4351 Two, this file is not a Windows Profile formatted file. Three,
4352 applications should not be reading or writing to %WINDIR%. It causes
4353 problems for Windows Terminal Server.
4355 The new location of CellServDB will be the OpenAFS Client install
4356 directory which is by default C:\Program Files\OpenAFS\Client and can
4357 be determined by querying the registry for
4358 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4360 The existing afsdcell.ini will be migrated by the NSIS installer.
4361 The Wix installer must still be updated to do the same.
4363 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4364 High Security mode; and to add Terminal Services compatibility registry
4365 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4366 configuration files in %WINDIR%.
4368 * Add support for authenticated SMB connections. This will remove
4369 the need for high security mode in most situations. Both NTLM
4370 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4371 only NTLM can be used even though Kerberos is now supported. The
4372 reason is that it is not possible to construct a service principal
4373 which is unique to each individual machine.
4375 SMB Extended Auth does not work on XP SP2 unless one of two registry
4376 modifications are made:
4378 (1) To disable the check for matching host names on loopback connections
4379 set this key. This does not require a reboot:
4381 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4382 "DisableLoopbackCheck"=dword:00000001
4384 (2) To add the AFS SMB/CIFS service name to an approved list. This
4385 does require a reboot:
4387 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4388 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4390 afsd_service.exe will automatically add the current Netbios Name
4391 to the BackConnectionHostNames list and then temporarily disable
4392 the loopback check for one cycle of startup/shutdown of the service.
4393 We assume most folks do not start/stop without a reboot so this
4394 will be adequate in most cases.
4396 * Fix security hole in afslogon.dll which allowed passwords to be
4397 sent in clear text to the KDC in a misformed principal name.
4399 * Fix cm_GetCell() to properly handle expired dns entries
4402 * If Freelance mode is active and the afs_freelance.ini
4403 file does not exist, do not create an empty file.
4404 Instead create a file containing ro and rw mountpoints
4405 to the default cell using the standard conventions.
4407 * Modify the Freelance support to handle the ability
4408 to create rw mount points in the fake root.afs volume.
4410 * Changed the RPC mechanism used for token setting from
4411 named pipes to local. Use of named pipes can be restored
4412 by setting the environment variable AFS_RPC_PROTSEQ to
4415 Named pipes were required when a Windows 9x system was
4416 using a NT system in gateway mode which is incompatible
4417 with our use of local loopback adapters.
4419 * In afscreds.exe, if a username of the form user@REALM is
4420 specified and no password is specified, do not perform a
4421 kinit operation. Only perform the aklog functionality.
4423 * Add a new registry value which allows the number of processors
4424 on which afsd_service.exe executes to be restricted. Valid
4425 values are 1..numOfProcessors
4427 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4431 * A second MSI based installer option is now available.
4433 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4434 request tickets for user/instance@REALM instead of just user@REALM
4436 * Modify the Power Management Notify routine to wait for the Hard Dead
4437 timeout period instead of a fixed 19 seconds. With the longer timeout
4438 periods Hibernation and Standby could never succeed when network
4439 connectivity is not available.
4441 * The following fs.exe commands are now restricted to Administrator:
4442 - checkservers with a non-zero timer value
4445 - sysname with a new sysname list
4454 setting the default sysname for a machine should be done via the
4455 registry and not via "fs sysname".
4457 * NSIS installer adds options to install Debugging Symbols
4458 and the Microsoft Loopback Adapter; the user is now also
4459 given the ability to select the afscreds.exe startup options.
4461 * Build system modified to generate symbols for FREE (aka RELEASE)
4462 builds as well as CHECKED (aka DEBUG) builds
4464 * Sites which have a volume ID of 0x20000001 assigned to their
4465 root.cell volumes have been experiencing problems with accessing
4466 the root.cell volume of their cell when Freelance mode has been
4467 active. This was because 0x20000001 was assigned to the fake
4468 root.afs volume created by freelance. The fake volume id is
4469 now set to 0x00000001 to prevent conflicts.
4471 * The timeout logic in the AFS Client Service has been wrong
4472 for sometime. It is based on two different assumptions.
4473 First, the SMB client timeout is a fix value as was the case
4474 with OS/2 Lan Manager. This assumption is incorrect. The
4475 SMB timeout in Windows is a dynamic value computed based upon
4476 a fixed minimum timeout to which is added time based upon the
4477 size of the request and the performance characteristics of
4478 the connection. Second, it is the responsibility of the
4479 SMB Server to enforce the timeout requirements of the client.
4480 This is untrue. The SMB Server cannot be expected to know
4481 the requirements of the client. More importantly, if the
4482 SMB server uses the SMB client timeout as a value to restrict
4483 its behavior as an RX client, the performance characteristics
4484 of the local SMB session would be used to prematurely terminate
4485 WAN connections with significantly different performance
4488 The timeout logic has therefore been modified in the following
4490 . the Lan Manager Workstation (SMB) Session Timeout is used only
4491 as a basis for configuring the Connection Dead Timeout
4492 and Hard Dead Timeout values. The Connection Dead Timeout
4493 must be at least 15 seconds longer than the SMB Timeout
4494 and the Hard Dead Timeout must be at least double the
4495 Connection Dead Timeout.
4496 . New registry entries have been added to allow the Connection
4497 Dead Timeout and Hard Dead Timeout values independent of the
4498 Lan Manager Workstation Session Timeout
4499 . The test to enforce the SMB Client Timeout has been removed.
4501 One of the side-effects of removing the enforcement of the SMB
4502 Client Timeout is that regardless of whether or not the SMB client
4503 is available to receive the response (and how would the SMB server
4504 know) the RX protocol response can be used to update the AFS
4505 Client Service state for ready access by future SMB client
4508 This should be the end of the "Server paused or restarting messages"
4510 * Add support for arbitrary UNC paths to the pioctl() support.
4511 This enables the fs commands as well as the AFS Shell Extension
4512 to work correctly when UNC paths are being used.
4514 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4515 DNS if the cell configuration cannot be determined via CellServDB
4517 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4518 are ever set more than once at a time
4520 * Fix the management of lists of cm_cell_t structures when using
4521 DNS to lookup cell information. The previous code would fail to
4522 reuse the same cellID for a cell if DNS was used more than once
4523 for a given cell name. When the ttl expired, a single cm_cell_t
4524 could be inserted into the cm_allCellsp list more than once
4525 producing a loop. In addition, the vlServerp list belonging to
4526 the cell was not freed resulting in improper refCounting of the
4529 * Add DNS support to cm_IoctlNewCell() which previous only examined
4532 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4533 to properly garbage collect cm_server_t objects
4535 * Add numVCs variable to smb.c to track the number of smb_vc_t
4536 objects created and use it to initialize the vcID field which
4537 previously was set to 0 in all objects resulting in FindByID
4540 * Fixed DNS lookups to work consistently throughout the OpenAFS
4541 product instead of just from within the afsd_service.exe
4543 * Added a runtime check to ensure that AFS Client Service SMB
4544 Server is accepting connections before attempting to mount
4547 * Read IP addresses for volume servers out of the CellServDB
4548 file if gethostbyname() on the hostname fails.
4550 * Fix getcellconfig() to populate both the Hostnames as well
4551 as IP addresses when loading cell data via DNS
4553 * Increase the Connection Dead Time to 50 from 20 seconds
4554 Increase the Hard Dead Time to 120 from 40 seconds
4555 (matches the Unix values)
4557 * Fixed an assertion validating the number of allocated NCBs
4559 * Fixed the build environment to consistently build for
4560 Windows 2000 and above. (APPVER = 5.0)
4562 * Fixed rx_debug to properly validate the receipt in incoming
4563 data with select() and recvfrom(). Do not copy data out of
4564 the socket buffer unless success is indicated.
4567 * afsd_service.exe will now display a message box to the
4568 desktop when it terminates due to an IP Address Change.
4570 * installer no longer deletes AFS Server configuration data
4573 * installer generates a warning dialog if the RPC service
4574 is not properly configured
4576 * installer compressed with lzma instead of bzip2
4578 * afsd_service.exe shutdown crash solved once and for all
4580 * reference counting of smb_vc_t data structures improved
4582 * name space collision of smb_fid_t event objects corrected
4584 * the output of "fs memdump" is now written to
4585 %WINDIR%\TEMP\afsd_alloc.log
4587 * the file TaAfsApp_1033.dll is now properly installed allowing
4588 the User Manager to start
4590 * a new algorithm is used for computing filename pattern matches
4592 * afscreds.exe now accepts user names containing instance
4595 * Fix the Directory Name Lookup Cache to be case-sensitive.
4596 This is crucial in environments in which a Windows client
4597 is accessing a directory with more than one filename that
4598 differs only by case. If the directory contains "FOO"
4599 and "Foo". You want "DEL Foo" to delete the correct one.
4600 We still have a problem in that "DEL foo" will delete a
4601 random filename. This will be addressed in a future release.
4603 * Fix afscreds.exe -M option (renewMaps) to work when High
4604 Security mode is off. Also, remember to disable the ActiveMap
4605 flag in afsdsbmt.ini when a drive mapping is removed.
4607 * Updates to NSIS installer script. AFS Server configuration
4608 data will not be destroyed on un-install or re-install.
4609 Use a better compression algorithm.
4611 * afslogon.dll now uses KFW to obtain tokens when available
4613 * afslogon.dll when given an all uppercase username will
4614 attempt to authenticate with both the uppercase name
4615 and an all lowercase variation
4617 * DST modification removed. The fix appears to make things
4618 worse after a reboot of the machine.
4620 * fs.exe: added "cscpolicy" which is used to
4621 change client side caching policy for AFS shares
4623 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4624 [-documents] [-disable] [-help]
4626 * Several uninitialized variables have been initialized
4628 * It is now possible to obtain tokens using cross realm
4629 Kerberos within afscreds.exe:
4631 user: jaltman@ATHENA.MIT.EDU
4633 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4634 will will in turn be used to obtain afs@DEMENTIA.ORG.
4635 The resulting token will be stored with the display name
4636 jaltman@ATHENA.MIT.EDU@dementia.org
4638 * aklog.exe has been added to the client
4640 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4641 [[-p | -path] pathname]
4645 -d gives debugging information.
4646 krb_realm is the kerberos realm of a cell.
4647 pathname is the name of a directory to which you wish to authenticate.
4648 -noprdb means don't try to determine AFS ID.
4649 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4650 (default is Kerberos V)
4651 No commandline arguments means authenticate to the local cell.
4654 * All of the resource files have been restructured to adhere to
4655 a set of rules IBM implemented for loading string resources.
4656 These rules had either been forgotten or were not discovered
4657 by folks working on the OpenAFS sources. The end result was
4658 memory corruption. This is primary item which was preventing
4659 the AFS Server from working.
4661 * Increased the size of the maximum ticket size stored in a token
4662 from 344 bytes to 12,000. Increased the buffers used to convey
4663 messages between the pioctl() caller and the SMB Server from
4664 1000 bytes to 12,512. The code appeared to have been writing
4665 above the top of the stack by quite a few number of bytes.
4666 (The increased ticket size is necessary for the next item.)
4668 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4669 Instead the raw Kerberos 5 ticket is used in its entirety. This
4670 is extremely important as it allows us to use pure Kerberos 5 KDCs
4671 as the source of the AFS authentication. The use of up to 12,000 byte
4672 tickets will allow tickets produced by all versions of Microsoft
4673 Active Directory to be used.
4674 - create a user account.
4675 - designate it DES only
4677 - specify its UPN to be "afs@realm"
4678 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4680 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4681 is special when using long file names. (you must use "*.*" and not "*")
4682 Instead only enforce it when performing 8dot3 searches.
4684 * Fixed the DST problem with creation times being set one hour ahead
4686 * Fixed the problem when using \\afs\cell-alias. For example,
4687 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4688 for the alias name; instead simply expand the name. One of the
4689 symptoms of this problem was a loss of acquired tokens.
4691 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4692 of strings. (Only English strings provided.)
4694 * Fixed the installer to properly replace in use files.
4696 * Fixed the build system to cleanup generated component version files
4698 * The release build compiled with MSVC 6.0 compiler to avoid the
4699 afsd_service.exe shutdown crash. This does not solve the problem
4700 but simply avoids it for the time being.
4704 * fix afslogon.dll to not corrupt memory when High Security mode
4707 * fix afsd_service.exe to not attempt to restore the stack when
4708 an exception occurs. (not safe in multi-threaded programs)
4710 * fix uninstaller to properly remove the CRT and MFC DLLs
4712 * remove a Message Box from afscreds.exe when getcellconfig()
4713 fails on a kerberos realm which is not a cell
4715 The following is a list of changes to the OpenAFS for Window client
4718 * "fs setserverprefs" will leave afsd service deadlocked
4720 * "vos listaddrs" will core dump
4722 * installer sets the appropriate keys to support Integrated Logon
4724 * installer disables the "Find Lana by Name" functionality as it
4725 was causing headaches for many users
4727 * fix the intermittent crash of the power management thread when
4728 shutting down the AFS Client Service
4730 * optimizes the obtain drive mount list functionality which is
4731 executed every time the mount tab in afscreds.exe and afs_config.exe
4732 are refreshed. (this happens a lot)
4734 * fix the service shutdown logic. add the STOP_PENDING state
4735 and do not accept additional service events after we declare
4738 The following is a list of changes to the OpenAFS for Window client
4741 * flexelint was run against the source tree and hundreds (perhaps
4742 thousands) of corrections were applied to ensure prototypes
4743 were in use; types were used consistently; variables were
4744 initialized; unused variables were removed; etc.
4746 * A wide variety of instrumentation was added including the
4747 ability to produce a stack trace from within afsd_service.exe
4750 * Dynamic configuration of the RDRtimeout value based upon the
4751 LanMan Workstation Session Timeout
4753 * The mount root no longer needs to be called "/afs". This
4754 is now set by a registry value "MountRoot" within the key
4755 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4757 * The cell list is now only read out of afsdcell.ini when the
4758 file changes instead of each time a cell is resolved.
4760 * Thread synchronization was added to cm_server.c and ktc_nt.c
4762 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4763 calloc()/free(). The Global functions were needed on Windows 3.x
4764 but have caused a variety of problems on the Win32 platforms.
4765 Avoiding them is highly recommended by several Microsoft
4766 Knowledgebase articles
4768 * Support for Symbolic Links added to the AFS Shell Extension
4770 * Added a registry value "OverlayEnabled" to determine if
4771 Shell Extension Overlays should be enabled.
4772 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4774 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4775 separate trees for checked and free builds. Build system supports
4776 a custom directory src\WINNT\extra which can be used as a grafting
4777 location of organization specific additions to the build tree.
4779 * New installer built using NSIS 2.0.
4781 * Named all kernel objects in order to allow them to be monitored
4782 with tools such as SysInternals' ProcExp.exe.
4784 * Introduced new EventLog framework for AFSD
4786 * Introduced Power Management interface to AFSD for Standby and
4787 Hibernate modes to allow cache to be flushed prior to network
4790 * Utilize Win32 DNSQuery API instead of internal routines. This
4791 allows DNS SRV queries to be sent to all current domain name
4792 servers. Not just one specified in an INI file. DNS is now
4795 * "NetbiosName" registry value may be used to specify a fixed
4796 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4797 when the loopback adapter is in use. If you need to use the
4798 old notation with a loopback adapter installed specify a registry
4801 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4803 * Refactor all modules which depend on LAN Adapter and NetbiosName
4804 determination in a new library: lanahelper.lib. This allows for
4805 consistent behavior throughout the product.
4807 * Move the afsd.log and afsd_init.log files to the directory specified
4808 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4809 for services. Added the Date to the log entries.
4811 * New registry value "RxMaxMTU" used to limit the size of the RX
4812 packets sent by the AFS Client Service to the Server. In order
4813 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4814 size must be restricted to 1264 or smaller. The latest NSIS
4815 installer sets a value of 1260 by default.
4817 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4818 packets. This is not needed in order to work across the Cisco
4819 VPN but might be needed for other network environments. This
4820 value is not set by the NSIS installer.
4822 * New registry value "HideDotFiles" is used to apply the Hidden
4823 attribute to files whose names begin with a '.'. This value
4824 is set by the NSIS installer.
4826 * New registry value "MaxMpxRequests" allows the maximum number
4827 of multiplexed sessions to be configured at run time. This
4828 value is not set by the NSIS installer. The default value is
4831 * New registry value "MaxVCPerServer" allows the maxmimum number
4832 of VCs per server to be configured at run time. This value is
4833 not set by the NSIS installer. The default value is 100.
4835 * New registry value "AllSubmount" allows the "all" submount to
4836 be disabled by setting its value to 0x00.
4838 * Allow cells names to be valid mount points
4839 \\<netbiosName>\<cellname>
4841 * Store the active state of drive mappings in order for afscreds.exe
4842 to restore them upon startup
4844 * Add exception handling to generate a Stack Trace to the afsd_init.log
4845 file if one happens to occur.
4847 * Add lots of logging to help detect the cause of invalid SMB packets
4849 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4850 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4851 renewal of expiring tokens as long as afscreds.exe is running.
4853 * New afscreds.exe command line options:
4855 -M = renew drive maps
4856 -N = ip address change detection
4859 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4860 determines whether or not MIT Kerberos for Windows should be used
4861 to obtain tokens via Kerberos 5 tickets.
4863 * New registry value "AfscredsShortcutParams" in
4864 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4865 determines the command line parameters to be specified when "fixing"
4866 the AFS Shortcut in the user's startup folder.
4868 * The "ShowTrayIcon" registry value has been moved from
4869 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4870 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4872 * The <cell name> registry values used to store the token expiration
4873 reminders have been moved from
4874 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4875 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4877 * Obtain the Logon User Name from the Explorer key when available
4879 * new text document doc\txt\winnotes\registry.txt lists all registry
4880 values used by OpenAFS (excluding the AFS Server)
4882 * BUG: rx_securityClass objects were not properly reference
4883 counted and were never freed.
4885 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4886 would be generated. The existence of a server does not imply
4887 that it is not down. If all of the servers for a cell are down
4888 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4891 * BUG: the directory name lookup cache failed to free the entries
4892 in the cache when the name cache entries cycled. The entries
4893 in the cache would become dereferenced without being freed.
4895 * BUG: fs setserverprefs could be executed without Administrator
4898 * BUG: the number of allocated NCB objects (100) exceeded the number
4899 which could actually be waited upon by the kernel (64). Any objects
4900 which were utilized above the limit could never have event completions
4903 * BUG: smb_username_t objects were not being reference counted and
4904 were not properly freed.
4906 * BUG: smb_tid_t objects could under unusual circumstances be freed
4907 before they were no longer referenced.
4909 * BUG: smb_fid_t object pointer were frequently used even when
4910 their value could be NULL. They were not properly released and
4911 therefore they were never freed.
4913 * BUG: smb_packet_t data structures were not completely initialized
4916 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4917 Denied" because that causes the Explorer Shell to try again until
4918 access is obtained. Instead return "Remote Resources" which allows
4919 the shell to move on and treat the error as transient.
4921 * BUG: when initializing the NCBreturns structure, separate Event objects
4922 were created for each NCB although a single Event object was supposed
4923 to be shared by all.
4925 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4928 * BUG: smb_tran2Packet_t objects were not being properly referenced
4931 * BUG: directory path creation did not handle the case of multiple
4932 directories requiring creation in one attempt
4934 * BUG: SMB requests which required an Extended Response were ignored.
4935 This prevented some files from being written to AFS volumes.
4937 * BUG: character strings were being freed even after they were
4938 inserted into in use data structures
4940 * BUG: inconsistent usernames were used when High Security mode was
4941 enabled. (there is still much to do in this area)
4943 * BUG: pioctl() calls which require out of band RPC operations were
4944 susceptible to race conditions when performed by multiple processes
4946 * BUG: memory allocation and deallocation crossed instances of the
4947 C Runtime Library producing memory leakage and corruption in
4948 afscreds and the client configurator.