3 * Avoid a race when updating cell vldb server lists
4 that can result in a crash.
6 * Avoid a deadlock when managing CM_SCACHESYNC_STOREDATA
7 state operations for directory objects.
9 * Add new Windows Application Event log messages for
10 VBUSY, VRESTARTING, ALL_BUSY, ALL_OFFLINE, and ALL_DOWN.
11 Include message throttling to prevent the same message
12 from being logged repeatedly within a five second window.
14 * Reduce lock contention by waiting for cm_buf_t I/O
15 operations to complete before permitting cm_SetupStoreBIOD
16 to analyze a buffer for inclusion in a BIOD.
18 * Split the cm_buf_t flags field to separate the flags
19 that are protected by the cm_buf_t mutex from those
20 protected by the buf_globalLock. This eliminates the need
21 to hold both locks everytime the flags field is accessed.
22 Both locks were not held in the past resulting in race
23 conditions that could result in deadlocks.
25 * Add "vos setaddrs" command.
27 * Rx library lock contention avoidance between rx_NewCall and
30 * Rx library races due to inconsistent use of rx_connection
31 conn_data_lock to protect the flags field.
33 * Rx library inconsistent use of RX_CALL_TQ_WAIT which could
36 * Rx library must signal transmit queue waiters when flushing.
37 Otherwise, deadlocks can occur.
39 * In cm_UpdateVolumeLocation, avoid searching for a ".readonly"
40 volume on a numeric volume name.
42 * File buffer allocations whose offsets are beyond server EOF
43 should be locally allocated and zero filled. The file server
44 should not be issued a FetchData rpc which is guaranteed to
47 * Enable integrated logon to work with Windows 7/2008 when
48 user logons are performed with a non-Domain Kerberos principal.
50 * Add Protection Error messages to aklog output.
54 * Prevent the Explorer Shell extension from crashing if
55 symlink creation failed. (126406)
57 * A Rx level NAT ping has been implemented.
58 Add NatPingInterval registry value to
59 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
60 to permit Nat Ping to be enabled. The default value is 0 seconds.
62 * When a re-initialization is taking place, be sure to reset
63 cm_noLocalMountPoints to 0 in case someone deletes the "Freelance"
64 registry key out from underneath the service.
66 * Add krb5 error message translation to aklog, afscreds,
67 afslogon.dll, the network identity manager afs provider
70 * Mode bits aren't directly exposed by the Win32 API. We were leaving
71 them to default to 0777 when creating new files and directories.
72 This version introduces two configuration parameters;
73 'UnixModeFileDefault' and 'UnixModeDirDefault' which are DWORD
74 registry entries that are used to set the initial mode bits.
75 If the values are set to 0, then the behavior is identical to what we
78 * Minidump files are now produced with a timestamp appended
81 * An SMB request debugging monitor has been added. When activated
82 the monitor will automatically turn on trace logging if any SMB
83 request has required longer than 60 seconds to complete and will
84 then create a minidump every 60 seconds thereafter until the
89 * Restore use of DNS AFSDB and SRV records by kaserver clients.
93 * Avoid a potential Freelance deadlock during initial execution
94 of afsd_service.exe if the old ini file data has to be
97 * Three rx library corrections. (1) Idle data connection
98 processing could timeout if the send window filled and
99 took longer than the idle data timeout period for the
100 transmit window to re-open. (2) The transmit queue
101 could be emptied prematurely. A required check for the
102 queue being in use was forgotten. (3) The function that
103 is supposed to implement a wait for the transmit queue
104 to cease being busy failed to wait.
107 * Restore use of DNS AFSDB and SRV records which were
108 unintentionally disabled in 1.5.69
111 * Add a context menu to the NetIdMgr AFS Provider
114 * Prevent an empty directory Btree from being created and
115 marked as valid if cm_BPlusDirBuildTree fails.
117 * cm_BPlusEnumAlloc should not fail if the enumeration
118 contains zero entries.
120 * In cm_BPlusXXX functions, return ENOMEM if malloc() fails.
122 * Do not leak the cm_scache_t dirlock if cm_BPlusDirBuildTree
125 * fs examine should report owner and group ids as signed values.
127 * Protect buffers in smb_WriteData from simultaneous writes.
129 * Query network interface MTU values from the registry because
130 the IP Helper GetAdapterAddresses() fails to report a manually
133 * Improve performance by not dropping and reacquiring the rx_call
134 lock in rx_WriteProc* and rx_ReadProc*.
136 * When configuring the rx library, the network interface MTU
137 must be applied to both the maximum send size as well as the
138 maximum receive size as is done when rx_SetMaxMTU is called
141 There was no 1.5.67 release
144 * Convert afsauthent.dll to use the xdr_alloc and xdr_free
145 routines from afsrpc.dll in place of malloc and free.
146 This permits the two libraries to be built with different
147 C runtime libraries without crashing.
149 * Set the DOS Readonly attribute on a file/directory
150 whenever the unix mode combined with the mask 0200
151 is true. Previously there was a discrepency between
152 the mask used for testing for readonly behavior and
153 that used for setting the attribute.
155 * Disable AFSVolSync based .readonly "whole-volume callback"
156 support because the all file servers prior to 1.5.67
157 (and perhaps 1.4.12) do not properly assign a value to
158 the AFSVolSync structure in bulk status RPC responses.
160 * Improve the error output from aklog to output the value
161 from krb5 error_message() if the afs_com_err output
162 indicates an unknown value.
164 * Hold a lock on the cm_scache_t object when the smb
165 server is assigning a new clientModTime value.
167 * Permit custom build version numbers to be set
168 without modifying the src/config/NTMakefile.<platform>
171 * Add support for binary code signing with cross-signed
172 certificates and arbitrary versions of signtool.exe.
174 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and
175 do not permit them to be exposed to the smb redirector.
177 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT
178 to avoid confusion within the smb redirector.
180 * Fix the byte order assigned to port numbers associated
181 with AFSDB record lookups. They must be network byte
182 order not host byte order.
184 * Add dynamic server ranking based on RPC round trip
188 * Official support for Windows 7 and Server 2008 R2
190 * Improved Rx RTT computation following the lessons
193 * Prevent a file server bug (FetchData returning an
194 invalid length instead of zero) from causing an
195 "unexpected network error" when writing to files.
197 * Promote DNS SRV records as superior to DNS AFSDB
198 records. Support arbitrary port numbers for vldb
201 * Add "fs listacl -cmd" support.
203 * Add AFSVolSync based .readonly "whole-volume callback"
204 support. With this functionality, multiple objects from
205 a .readonly volume can have their status validated by
206 issuing a single RXAFS_FetchStatus RPC.
208 * Remove drive mapping functionality and service start/stop
211 * Remove drive mapping functionality from afs_config.exe.
213 * Adjust SMB error return codes to avoid returning
214 STATUS_TIMEOUT which results in the SMB redirector
217 * Network Identity Manager OpenAFS Provider now provides its
218 own "AFS lock" notification icon to report the status of
219 "have tokens, have no tokens, service not started, service
220 started but inaccessible". Hovering over the icon lists the
221 cells for which tokens exist (if any) and the OpenAFS version
222 number. Double-clicking executes the Network Identity Manager
225 * When merging the current status for an AFS object, ensure
226 that the volume status for that object is set to 'vl_online'.
228 * Permit cm_Analyze to request a retry even when the cm_req_t is
229 flagged CM_REQ_NORETRY if the RPC failed due to RX_CALL_DEAD.
230 The retry will force the use of a new connection.
232 * Add digital signatures to resource only dlls as required by
235 * Prevent pioctl calls from retrying indefinitely when a sharing
236 violation error occurs.
238 * Add {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab"
239 to restore access to drive mapping functionality in afscreds.exe
242 * Permit Rx minimum peer timeout, maximum receive window size and
243 maximum send window size to be configured via the registry.
245 * The MSI installer now sets the ARPINSTALLLOCATION as required by
248 * DNS lookups set the port numbers in network byte order within
251 * Replace use of the 16-bit compatibility API WinExec with
252 ShellExecuteEx in afs_config.exe and afscreds.exe. WinExec
253 is incompatible with Windows 7 / Server 2008 R2 UAC
257 * Short circuit background volume checks if a shutdown
258 or suspend request is received.
260 * Avoid contacting the file server in order to allocate
261 an empty buffer beyond the current length of the file.
263 * When a request for streams on a directory or mount
264 point object is received, do not offer a default stream.
265 This was affecting the saving of roaming profiles.
268 * Fix Windows 2000 compatibility. Do not call rand_s()
269 even if it is compatible with the C run time library
272 * Fix a data consistency error between the output of
273 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
274 workstation group name "AFS", and report server name as
275 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
277 * Enable executables to be run from \\AFS on Windows 7.
278 Return "Name not found" instead of "File not found" when
279 a directory or file name cannot be found.
281 * Prevent cache manager from marking file server "down" when
282 the data returned in response to either RXAFS_FetchData64
283 or RXAFS_StoreData64 is invalid.
285 * Add pioctl data validation to the AFS Explorer Shell extension.
288 * After 1.5.62 afslogon.dll would no longer crash but it
289 also would not load "domain" specific configurations.
292 * An uninitialized variable was present in the symlink
293 recursion detection loop.
295 * Reverse addition of unique per file GUID during
296 Create/Open responses. Inclusion of the GUID prevents
297 Cygwin applications (and others) from accessing //afs.
299 * Treat "filename::$DATA" as a synonym for "filename".
300 This addresses the recent report of roaming profiles
301 failing to save back to the file server.
303 * When given a choice between an error from rx_Write()
304 or RXAFS_EndStoreData(), use the rx_Write() error.
305 This ensures that over quota and access denied errors
308 * If an error occurs during RXAFS_StoreData operations,
309 apply that error to all of the buffers in the BIOD.
310 This ensures that failures will not be retried if they
311 will only fail again.
313 * Do not perform a DNS lookup for a cellname if the
314 cellname prefix is _._AFS_IOCTL_._. This will avoid
315 DNS lookups when the query is for
316 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
318 * Fix Freelance root.afs volume object callback processing.
319 Object synchronization and status merging bugs.
321 * Fix Explorer Shell "invalid parameter" bug introduced
324 * Log RXKAD errors to the trace log.
326 * Prevent simultaneous pioctl calls from multiple processes
327 from stepping on each other's data.
329 * Prevent simultaneous pioctl calls from crashing the
332 * Add data validation to pioctl output processing in
333 ktc_GetToken and ktc_ListTokens.
335 * Add data validation to pioctl output processing in
336 fs.exe and symlink.exe.
339 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
340 No longer will browsing \\AFS produce truncated share names.
342 * Registry specified server preferences did not result in
343 the CM_SERVERFLAG_PREF_SET flag being applied to the
346 * Avoid unnecessary DNS lookups of share names as cells
347 when it is known that the name cannot be a cell name.
348 Any name that does not contain a dot is skipped.
350 * When processing Dfs Referral requests, do not return an
351 error if the server is down or busy; if the volume is
352 offline; the cell vldb cannot be reached; etc.
353 These paths are still \\AFS paths and so the client
354 should not be sent to search elsewhere for resolution.
356 * CRITICAL: Prevent the cm_Daemon thread from terminating when
357 the machine enters the suspend state. IF the cm_Daemon thread
358 dies, there is nothing to execute down server checks.
360 * Prevent use of smb_StartedLock before initialization which
361 can result in a panic in the lock package.
363 * Improve error checking in afslogon.dll in order to prevent
364 NULL pointer dereferences within GetLogonDomainOptions()
365 if the Lsa operations fail.
367 * Add support to the build system to automatically update
368 a Microsoft Symbol Store.
370 * Do not access the cm_conn_t after executing the cm_Analyze
371 loops. Doing so can result in an access to memory that has
374 * CRITICAL: If a StoreData request offset is not aligned on the
375 buffer module blockSize, (offset % blockSize) bytes of dirty
376 data will not be written to the file server.
378 * CRITICAL: If a StoreData64 request is sent to a file server
379 that does not support large files (>2GB) and the client is
380 unaware that the server has no large file support, the first
381 chunksize worth of data will not be written to the file server.
382 This bug was introduced in 1.5.3. It affects all IBM AFS file
383 servers and all OpenAFS file servers older than 1.4.0.
386 * If a file server becomes inaccessible while the cache manager has
387 dirty buffers to write, the afsd_service buf_IncrSync thread can
388 attempt to use 100% of the cpu.
390 * Fix the locking that protects the cm_cell_t lists
391 (all, name, id, free).
393 * Remove cm_cell_t objects from the name and id hash tables before
394 placing them on the free list.
396 * Fix "fs newcell" which was broken in 1.5.60.
398 * Do not attempt to synchronize dirty buffers if the associated
399 volume is known to be unavailable.
401 * Log to the Windows Event Log dirty buffers that cannot be written
402 during the shutdown process.
404 * Modify behavior of a Freelance mountpoint target that does not
405 specify a cell. Instead of assuming the target volume is in the
406 Freelance.Local cell, use the workstation "Cell" specified in the
407 registry. A mountpoint target of "#root.cell." will now mean the
408 root.cell volume in the workstation cell for the current session.
409 If the workstation cell changes from "athena.mit.edu" to
410 "andrew.cmu.edu", the referenced volume will also change without
411 requiring that the mount point targets be altered.
413 * Avoid false infinite loop errors when validating the cm_cell_t
414 free list during startup.
416 * Fix two more locations where xdr_free() should be used instead of
417 free() to permit a checked built of afsd_service.exe to work with
418 release builds of afsrpc.dll.
420 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
421 to permit the server Uuid to be used to lookup the server object
422 and from that determine the cell. This permits callbacks that are
423 received from alternate addresses to be processed with a known server
424 object. Previously a request from an unknown server would clear all
425 callbacks from all cells.
427 * Fix a bug that prevented optimal performance when using a non-zero
428 value for 'daemonCheckVolCBInterval'. As a reminder, when
429 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
430 volume callbacks are automatically renewed 90 minutes before their
433 * Fix automatic ranking of vldb servers whose values are obtained from
436 * Add logging for RX CALL DEAD errors which are generated if the remote
439 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
440 or the call is to a vldb server.
442 * Add registry based cell search functionality to NetIdMgr,
443 afs_config.exe, and klog.exe.
445 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
446 lookups on the host names in the CellServDB instead of using the
447 specified IP addresses. This provides aklog, pts, vos, etc. the same
448 CellServDB behavior that the Windows Cache Manager uses.
450 * Completely a behind the scenes change. The snprintf() routines have
451 been completely rewritten.
453 * When updating the stat cache entry callback of a .readonly object
454 from the volume group object, update the file server reference to
455 ensure it matches the most update to date callback.
457 * Add proper support for processing callbacks from multi-homed file
458 servers. Instead of comparing servers by cm_server_t pointer,
459 compare them by UUID when the UUID is known.
461 * During a shutdown short circuit the offline volume check daemon
464 * Properly track link counts during cm_Link and cm_Unlink operations.
466 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
467 preference to an error code reported by rx_EndCall.
469 * Add "PerFileAccessCheck" registry value to permit testing against
470 experimental file servers that include per-file acl support. This
471 value is intentionally undocumented. It is not to be used by
472 production environment deployments.
474 * Log a warning to trace log if a volume group is marked as a DFS
475 File Set. The Windows OpenAFS client does not include DFS support.
477 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
478 provider from obtaining tokens when referrals are in play.
480 * Add "fs chown" and "fs chgrp" commands to permit the owner and
481 group of objects stored in AFS to be set from Windows.
483 * Avoid performing background daemon operations when the machine is
484 going into suspend mode.
486 * Perform offline volume checks in most recently used order.
488 * Prevent crash when a data version for a cache object goes backwards.
490 * fs setquota, fs setcachesize, vos setfields, and vos create now
491 accept human readable orders of magnitude. (K, M, G)
493 * fs listquota fixed to permit large quota sizes to be displayed.
495 * rx packet resend and data packets sent counts were incorrect.
497 * Multi-thread safe library versions are now being generated and used.
498 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
500 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
501 values are now available on XP through Windows 7. Add functionality
502 to autodetect if such support is present on the machine. If so,
503 configure it if necessary and dynamically adjust the AFS Rx timeout
506 * Correct another race condition in the Rx library that could result
507 in an unexpected panic while freeing the Rx call iovq.
510 * A fix to the pioctl library to support drive substitution
511 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
513 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
514 that corrects a deadlock in the smb redirector and also adds new
515 functionality that permits the AFS SMB server to be given a longer
516 timeout than is normally the case. New functionality has been added
517 to configure these additional LanmanWorkstation\Parameter values.
518 (This functionality has been backported to XP SP3 and is scheduled
519 to be released on June 5th.)
521 * The BackConnectionHostNames registry value configuration was broken
522 when dynamic re-establishment of Netbios Name registrations was added.
525 * Hidden vos commands are revealed.
527 * If the "DisableLoopbackCheck" registry value is set, do not unset
528 it during the same service session.
530 * Reorganize code that prevents multiple Store operations to the same
533 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
534 permit broken symlinks to be treated as being in AFS.
536 * Fix vos commands that output 64-bit integer values.
538 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
539 Permits building cygwin applications against OpenAFS libraries.
541 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
542 as Windows HTML Help (.CHM) files.
544 * NSIS installer does a much better job of cleaning up files left
545 over from previous installs.
547 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
548 or "fs flushall" and on-going directory operations that can result in
549 afsd_service.exe crashing.
551 * Add support for DNS SRV records in place of AFSDB records.Â
552 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
553 Weight is currently ignored.
555 * Add a method of specifying Client CellServDB information within the
556 registry that can be used to either override the CellServDB file or
557 force the use of DNS lookups for a given cell.
559 The registry schema is as follows:
561 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
562 "LinkedCell" REG_SZ "[cellname]"
563 "Description" REG_SZ "[comment]"
564 "ForceDNS" DWORD {0,1}
566 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
567 "HostName" REG_SZ "[hostname]" Default: [servername]
568 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
569 "IPv6Address" REG_SZ "[address]" <future>
570 "Comment" REG_SZ "[comment]"
571 "Rank" DWORD "0..65535" Default: 0
572 "Clone" DWORD "{0,1}" <future: server only>
573 "vlserver" DWORD "7003" <future>
574 "ptserver" DWORD ... <future>
576 ForceDNS is implied non-zero if there are no [servername]
577 keys under the [cellname] key. Otherwise, ForceDNS is zero.
578 If [servername] keys are specified and none of them evaluate
579 to a valid server configuration, the return code is success.
580 This prevents failover to the CellServDB file or DNS.
582 Only one of "HostName" or "IPv4Address" is required. "HostName"
583 is optional if [servername] is the host name.
585 * Extend registry based CellServDB functionality to
586 afsconf_GetCellInfo() interface used by aklog and many
587 other command line utilities.
589 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
590 as it is now used by both client and server components.
593 * PriorityClass of afsd_service.exe process raised to
594 "High" to match the priority of the system services
595 that are waiting for afsd_service.exe to respond.
596 Configurable using [TransarcAFSDaemon\Parameters]
597 DWORD "PriorityClass" value. (use values from Win32
600 * Change error reporting of CM_ERROR_BADSHARENAME to
601 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
602 The error we were sending would cause the smb
603 redirector to drop the connection.
605 * Change the error reporting of CM_ERROR_ALLOFFLINE
606 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
607 Previous we would return Bad Network Path which
608 would cause the smb redirector to cancel all file
609 handles on the same "share".
611 * A debug version of afsd_service.exe can now be used
612 with a release version of afsrpc.dll (or vice versa).
613 afsrpc.dll now implements and exports xdr_free().
615 * Removed all unprotected OutputDebugString() calls as
616 those calls can block for up to 10 seconds if there
617 is contention for the globally shared memory or if
618 the debugger fails to read data in a timely manner.
620 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
621 supported for builds. The production releases still
625 * Discovered that XP/2003 supports a variant on recvmsg
626 called WSARecvMsg. The input parameters are mostly
627 compatible with the Posix recvmsg. This will save
628 multiple memcpy() calls per received packet.
630 * Discovered that Vista/2008 supports a variant on
631 sendmsg called WSASendMsg. The input parameters
632 are mostly compatible with the Posix sendmsg. This
633 will save multiple memcpy() calls per sent packet.
635 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
636 and utilizing rx_ReadProc32 from within the cache manager
637 as part of the FetchData operations. rx_ReadProc32 and
638 rx_WriteProc32 are optimized versions of rx_ReadProc and
639 rx_WriteProc for use when reading 32-bit integer values.
641 * The Network Identity Manager AFS provider now reports
642 progress events. These provide a better user experience
643 with Network Identity Manager 2.0.
645 * [RT 124293] A race condition exists which permits the
646 scp field of the an smb_fid_t object to become invalid
647 after a request on the smb_fid_t has begun. Remove the
648 race by protecting all access to the scp field with the
649 smb_fid_t mx mutex and obtaining a local reference on
650 the cm_scache_t object for the length of the request.
652 * [RT 124276] If the vldb is out of sync with the contents
653 of the file servers, afsd_service will retry too many
654 times when a file server reports a volume as not being
655 present. Now if the list reported by the vldb is the
656 same as the previously seen list, then the retry is
659 * [RT 124276] Read-only volume failover was broken in
660 1.5.53 whenever accessing a volume results in VNOVOL or
663 * [RT 124276] Prior to 1.3.70 the volume server reference
664 list was not reference counted and would be prematurely
665 freed while in use. When reference counting was added
666 in 1.3.70 a bug was introduced that could result in
667 service reference list corruption.
669 * Add Windows Application Event Log warning messages for
670 "Client SMB MPX value too large" and
671 "Client SMB Buffer Size too small".
673 * Renaming of files across directory boundaries would result
674 in an invalid handle error when attempting to access the
675 files after the move.
677 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
678 when a smb file descriptor cannot be found. The whole point
679 of using a Path Info function is because an smb file
680 descriptor wasn't allocated.
682 * More edge cases in which dynamic addition of Freelance
683 root.afs entries would get the wrong FID or where the
684 root.afs directory would not be refreshed.
686 * Buffer overflow could occur if the workstation cell name was
687 longer than 64 characters. Crashes could occur in
688 afscreds.exe, afslogon.dll, and afsd_service.exe.
690 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
693 * Improve consistency of the bad data version value used for
696 * Log file server uuid values as part of the cm_server object
697 when available. Dump the cm_server object list in response
700 * Optimize the performance of resetting access control lists
701 when tokens are set or removed.
703 * Fix vos.exe so that it can work with very large volume
706 * Remove symlink recursion tests and increase max symlink count
710 * Undo the "UAC manifest fix" applied to afs_config.exe.
712 * Fixed some additional race conditions in the Rx library
714 * Ensure that Freelance allocation of vnodes follow
715 the AFS convention of odd vnodes are directories
716 and everything else is an even vnode.
718 * Add Freelance logic to mount point and symlink
719 evaluation functions.
721 * Enhance smb_ParseASCIIBlock() so that it can handle
722 all of the STRING formats defined by the CIFS Technical
725 * Validate the output of smb_ParseASCIIBlock() in all
726 callers. Return CM_ERROR_BADSMB if the STRING field
727 cannot be parsed. CM_ERROR_BADSMB will cause the
728 contents of the packet to be logged.
730 * If multiple SMB Raw Write operations were taking
731 place at the same time, there could be data corruption
732 because unique event objects were not generated for
733 each Netbios receive operation.
737 * Fix a race in rxi_WriteProc which could result
738 in a freed or in use packet being passed to
739 rxi_PrepareSendPacket if the call is currently
740 in a transmit wait state. Passing a freed packet
741 will result in a panic. Passing an in use packet
742 might result in a call failure or perhaps data
745 * Kerberos referrals broke the NIM AFS provider
746 logic for automated configuration of the workstation
747 cell with a new identity. Now if a referrals
748 response is returned, assume the cell belongs to the
751 * The interface to buf_CleanAsync() and
752 buf_CleanAsyncLocked() has been changed to permit
753 errors to be propagated to the caller.
755 * During an enumeration of a directory, if the btree
756 has not yet been generated, return an error.
758 * When mapping CM_ERROR values to NTSTATUS values
759 do not treat 0 as an error.
761 * Avoid a cm_scache_t object reference count underflow
762 when renaming via the SMB interface.
764 * smb_IsMask() can be called with a non-8.3 input buffer.
765 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
766 input buffers and smb_IsMask() for non-8.3 input buffers.
768 * Add cm_buf_t error state to the "fs memdump" output.
770 * pioctl drive mapping checks need to take into account
771 Global Drive Mappings and SUBST assignments.
773 * Rename operations should not be permitted if the
774 destination name collides with an existing name that
775 differs only by case UNLESS the collision is the object
776 that is being renamed.
778 * Increase the maximum number of background daemon threads
779 to 64. Leave the default at 4.
781 * Set the FILE_UNICODE_ON_DISK flag when reporting the
782 file system characteristics.
784 * Report the maximum component name length as 255 even
785 for Unicode as part of the file system characteristics.
787 * Add RX debugging information to "fs memdump" output
790 * Refcount the cm_server_t objects returned by
793 * Fix the UAC manifest applied to afs_config.exe
795 * rxdebug now indicates if any RX_CONN flag is set
796 not just some of them
798 * Change the mapping of CM_ERROR_WOULDBLOCK from
799 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
803 * Includes #GCO Public CellServDB 07 Nov 2008
805 * Prevent separate cell entries from being created for
806 "foobar.org" and "foobar.org." which can happen if
807 an AFSDB record is found instead of a CellServDB
810 * Fix a panic when checking the SMB virtual circuit list.
811 If a VC is destroyed while the list is being checked
812 it is possible for the next object to become invalid.
814 * Increased logging to afsd_init.log of netbios name
817 * Fix a DNLC panic when a UTF-16 string that cannot be
818 converted to a utf-8 string is provided as a filename.
820 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
821 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
822 location to refer directly to the new location in
823 %programfiles%\openafs\client\program
825 * Be more aggressive about responding to Netbios
826 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
827 the LANA within the listener thread without performing a
830 * Use the WNet Enumeration of all connected drives as a filter
831 in the pioctl GetIoctlHandle operation. This should remove
832 the last situations in which there was a delay when opening
833 an Explorer Shell properties menu with the AFS Shell
836 * Fix the parsing of CellServDB in Windows to handle linked
837 cells as is done on Unix.
839 * Add support for linked cells to the cache manager,
840 netidmgr afs provider, and aklog. When two cells are
841 linked, a volume lookup that fails in one cell is looked
842 up in the other. This can be used for construction of
843 test cells that override specific volumes; cell renaming;
846 * Increase the length of cell and realm names that can
847 be entered in the NIM AFS provider.
849 * Actively prevent a 32-bit afsd_service.exe from
850 executing in the WOW64 environment.
852 * Protect against invalid IP address values in CellServDB
857 * Fix an Rx panic caused by an attempt to double free
858 an rx_packet when an error occurs while writing the
861 * Fix an exception cause by a race condition when
862 processing multiple file change notifications registered
863 by the Explorer Shell or other applications.
865 * Modify the Rx debug interface to use var_args and
866 permit Rx debugging to be used on 64-bit Windows
867 which previously had it disabled.
871 * Automated lock hierarchy enforcement added to the
872 osi base lock package. This permits detection of lock
873 violations that might but often does not result in
874 deadlocks. Enabled by default in the checked builds.
875 May be activated in release builds via the registry
876 value "LockOrderValidation" in the TransarcAFSDaemon
879 When a violation is detection an assertion is thrown
880 generating a dump file and a message indicating which
881 locks were obtained out of order is logged to the
882 Windows Application Event Log.
884 * Attempt to ensure that a dump file will be generated
885 whenever the service panics.
887 * Removed unused lock cm_bufGetMutex
889 * Many out of order lock acquisitions were detected
890 using the lock order verifier and were corrected.
892 * A race condition within the function path cm_GetSCache() ->
893 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
894 cm_scache_t object to be issued simultaneously to two
895 threads. This would eventually result in a panic due
896 to the resulting undercount.
898 * Fix interpretation of the empty string as the ioctl path
899 to mean the current directory. "fs lsm", "symlink list",
900 etc. now return a "not a ..." error instead of "does not
903 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
904 to indicate that the input string is a nul terminated string.
905 Assign ti when input strings are nul terminated instead of
906 length counted. (from 1.5.50)
908 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
909 does not match the observed behavor. The 'nameLength' is specified
910 in bytes, not in characters. Fix the implementation to match
911 the observed behavior. (1.5.50)
913 * The cm_HaveAccessRights() attempt at deadlock avoidance by
914 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
915 does not avoid the deadlock. Avoid the deadlock by enforcing the
916 lock order of the lowest vnode first. Then remove the infinite
917 loop avoidance in cm_SyncOp() that was returning an unwarranted
920 * File Server lock synchronization was not propertly enforced in
921 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
923 * Another change to the short name generator. Plus and equal are
924 not valid characters in a short name according to the
925 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
926 with left and right brace in order to minimize the impact.
928 * Fix a race condition that can result in a heap overwrite error if
929 the number of known file servers or vldb servers increases during
930 the preparation phase of a cm_CheckServers() operation.
932 * If the SMB Authentication would fail, a LSA allocated memory
933 buffer would be leaked.
935 * Rewrite the cm_cell_t allocation code to permit cell vldb server
936 lookups to fail without leaking cm_cell_t objects.
938 * Add "fs setcrypt auth" to permit the use of rxkad authentication
939 only mode. Not documented.
941 * Initialize Rx mutex and lock objects once and only once as part of
942 the DLL load process.
944 * When reporting the Volume Name "AFS" do not include the terminating
945 NUL in the reported length.
947 * Refresh Volume Location information upon receipt of a VNOVOL error
948 from the file server and avoid a race condition that prevented the
949 updated volume location information from being destroyed immediately
950 after it was acquired.
952 * In response to VNOVOL and VMOVED, remove the volume id from the server's
953 volume list so that the volume's up and down status is not affected
954 when the machine is shutdown.
956 * Many fixes to the Rx stack related to packet allocation and deallocation;
957 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
958 by overwriting the rx_call currentPacket pointer when it was already
959 referring to a packet.
961 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
962 in the TransarcAFSDeamon/Parameters key.
964 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
965 constructed directory B+ trees for flushed directory objects.
967 * When cached pages for directory objects have been locally modified, do
968 not permit the locally modified pages to be mixed with pages subsequently
969 obtained from the file server. Doing so can result in directory
972 * If Ubix errors are received in response to VL_xxx operations record
973 the error on the server instance and permit the operation request to
974 be retried in case another server is able to answer the query. We have
975 seen at least one instance of a cell in which a VL server returned
976 UNOQUORUM (possible database truncation) when the other servers were
977 responding successfully.
980 * Optimize the assignment of write errors to dirty
981 buffers by applying the first error to all buffers
982 belonging to the same object instead of trying each
985 * Add resource version info to xstat_fs_test and xstat_cm_test
987 * Fix a delete vs delete[] usage problem in the WiX custom
990 * When creating a mount point, validate the volume name to
993 * When processing a mount point, pay attention to VL_BADNAME
996 * Remove a race condition from cm_volume_t recycling that could
997 have resulted in two threads making use of the same object at
1000 * Fix a reference count leak on smb_vc_t objects introduced in
1001 1.5.50. Permit the objects to be freed once again.
1003 * Fix the output of "fs memdump" for smb_vc_t objects that was
1006 * Add routines for debugging smb refcount errors.
1008 * When shutting down the cache manager, do not permit AFS
1009 callback requests to be processed. Doing so can result in
1010 invalid memory accesses.
1012 * Do not recycle a deleted cm_scache_t object unless the
1013 refCount is 0. Doing so can result in a deadlock and
1014 invalid memory access.
1016 * Minimize the number of RPCs sent to the file server for
1017 directory bulk status requests. Only request status for
1018 objects that we are actually reporting on and only if we
1019 do not already have status for the object and then make
1020 sure that we return as much as we can at once via the SMB
1021 interface but not so much that it exceeds the RDR timeout
1023 * Enable bos restricted operations
1025 * Fix the creation of submounts used by the afscreds and afs_config
1028 * Avoid repeated retries when attempting to obtain access rights.
1029 If the access rights were obtained successfully but do not
1030 satisfy the request, fail the request.
1032 * Fix error code reported when attempting delete a file on a
1033 readonly volume or one that is marked with the readonly DOS
1036 * Fix a short name truncation bug.
1038 * Fix a heap corruption error when reading the CellServDB file
1041 * Fix directory listings when the directory has "lookup" permission
1042 only and the directory has never been previously accessed.
1044 * Fix several tracking issues for cm_scache_t rw locks.
1046 * Ensure that all buffers used to store cell names are the
1049 * Restore Windows 2000 support (broken in 1.5.50).
1050 LOCALE_INVARIANT was introduced in XP.
1052 * Add RxUdpBufSize registry value. Default is 256K.
1054 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
1056 * Do not include trailing NULs when returning file names in
1057 a directory search response.
1059 * Prevent RX from crashing the afsd_service by attempting
1060 to allocate to many buffers at once when memory is fragmented.
1062 * Avoid a deadlock when freeing cm_server_t objects
1064 * Optimize search cell file. Do not search to end of file
1065 if the cell has already been found.
1067 * Remove explorer shell extension delay on local disks.
1069 * Automatically pre-allocate 64 Rx packets per thread
1071 * Add smb lock requests and stat cache lock allocations to
1072 the output from fs memdump.
1075 * Fix Large File Support (broken in 1.5.50)
1077 * Fix "fs setacl" (broken in 1.5.50)
1079 * Fix pioctl path evaluation of drive letter
1080 mapped paths. (broken in 1.5.50)
1082 * Fix installation of VC8 Runtime library in the EXE
1083 Retail installer. (broken in 1.5.50)
1085 * Avoid a potential deadlock during Volume Location
1086 updates if all the file servers are unreachable at
1087 the time the VL update succeeds.
1089 * Ensure that rx connection object is references
1092 * Fix propagation of over quota errors during cm_FSync()
1095 Since 1.5.39 [1.5.50 released 16 July 2008]
1096 * There were no releases between 1.5.39 and 1.5.50.
1098 * FIRST RELEASE WITH UNICODE SUPPORT
1100 * Pioctl interfaces to the cache manager refactored
1101 to provide layering between the SMB specific code
1102 and the general purpose ioctl operation.
1104 * Garbage collect dead SMB virtual circuits as soon
1105 as they are no longer being referenced. This avoids
1106 problems with outstanding locks not being dropped
1107 when the virtual circuit becomes invalid.
1109 * Remove the IBM Administration Reference documentation
1110 and replace it with the OpenAFS Command Reference
1113 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
1114 functions each time a connection is about to be used.
1115 Do not hold a lock on the rx connection object while
1116 it is being selected. This avoids a race between
1117 threads attempting to set the timeout values and
1118 removes a bottleneck that was hampering performance.
1120 * Ensure that the smb directory attribute is set
1121 for all directory objects.
1123 * Add vs2008 support to the NSIS installer scripts
1125 * Replace the VC Runtime EXE installer with the MSI
1126 installer in the NSIS installer scripts
1128 * Properly delete the VC7.1 C Runtime libraries.
1130 Since 1.5.36 [1.5.39 released 23 June 2008]
1131 * There were no 1.5.37 or 1.5.38 releases for Windows
1133 * Use Visual Studio version of vsprintf() instead of
1134 home grown version that resulted in linking conflicts
1136 * Prevent buserver, ptserver and fileserver from crashing
1137 due to non-portable strftime() format strings.
1139 * Prevent server crashes caused by rx_InitMorePackets()
1140 being called prior to allocation of all of the required
1141 synchronization objects.
1143 * Fixed a memory leak in the internal B+ tree directory.
1144 For objects with non-8.3 names, the 'longname' memory
1145 allocation would be leaked during object deletion or
1148 * Fixed a memory leak in the background daemon cm_CheckServer()
1149 routine. The memory allocated to store the server list
1152 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
1154 * Correct an abstraction layer violation. cm_scache_t objects
1155 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
1156 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
1157 performed in cm_ReleaseSCache() when the reference count hits zero.
1159 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
1161 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
1164 * Ensure that scache object 'nextp' pointers are set to NULL
1165 when the objects are removed from the hash table linked list.
1167 * If a dirty buffer cannot be written to the file server because
1168 of an access denied error, mark the error in the cm_buf_t
1169 structure so that it can be dealt with instead of retrying
1173 * Update CellServDB file to GCO Public 23 Apr 2008
1175 * Fix a cm_buf_t reference count leak when attempts to write
1176 dirty buffers to the file server from within cm_IncrSyncer()
1179 * Prevent udebug from crashing.
1181 * Activate RX Idle Timeouts. If the file server is busy for more
1182 then 30 seconds, attempt to failover to another server without
1183 marking the busy server down.
1185 * Another VNOVNODE issue fixed. When writing a dirty buffer
1186 to the file server, if VNOVNODE is received, mark all buffers
1187 as invalid without further attempts to contact the file server.
1189 * Improved performance on high latency links.
1192 * Fix a bug preventing the re-initialization of the Freelance
1193 root.cell contents when a change is made.
1195 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
1196 or mountpoint that does not exist.
1198 * aklog, afscreds, integrated logon, and the NIM AFS credential
1199 provider will all attempt to first acquire an afs service ticket
1200 using the service principal: afs/<cell>@<CLIENT_REALM>. This
1201 permits the proper detection of the cell realm when multiple
1202 realms are local realms for the cell.
1204 * Fix fs commands that re-use file identifiers instead of forcing
1205 a path evaluation each time. This was broken in 1.5.33 when
1206 hashes were added to the cm_fid_t structure.
1208 * Fix symlink and mountpoint removal and then recreation when
1209 applied to the fake Freelance root.afs volume.
1211 * Fix SMB change notifications for the Freelance root.afs volume
1212 when creating or removing symlinks and mount points.
1214 * Force DNS AFSDB cell name lookups to lower case if the DNS
1215 response is not all lower case.
1217 * 32-bit installers are now built with Visual Studio 2005
1219 * All vos.exe commands now support the "-noresolve" option which
1220 forces the display of IP addresses instead of DNS hostnames.
1222 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
1223 permits operations that do not require up to date volume
1224 location data to continue while waiting for the RPC to complete.
1225 Multiple threads waiting for a volume location update on the
1226 same volume set no longer trigger multiple calls VL_xxx RPCs.
1228 * Volume server references could be marked offline and never
1229 reset to the not_busy state.
1231 * Further optimizations to cm_Analyze() retry logic for
1232 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
1235 * Optimize the DNLC by applying Interlocked incrementing for
1236 statistics gathering and enabling greater use of read
1237 locks instead of write locks
1239 * Further optimize osi_Log operations when logging is disabled
1240 by removing the need for the function call overhead.
1242 * Add a new lock operation, Convert Read to Write, and apply
1243 it throughout the cache manager. This function provides a
1244 fast transition from a read lock to a write lock when the
1245 caller is the only reader.
1247 * Fix a timeout problem when opening files or request locks
1248 when the file in question is already write-locked by another
1251 * Fix a deadlock that can occur if "fs flushall" is executed
1252 while applications are writing to AFS.
1254 * The cache manager would re-read data from the file server
1255 after file truncation when obtaining buffers to write to the
1256 same file until the first StoreData RPC completes.
1258 * Fixes error handling in cm_NTCheckDelete()
1260 * Replaces cm_scache_t mutex with a read-write lock permitting
1261 additional parallel access.
1263 * Prevent an cm_scache_t reference undercount when evaluating
1264 symlinks containing @sys when none of the active sysnames
1265 match any of the available targets.
1267 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1270 * Set RxMaxMTU registry value to 0 which is the equivalent of
1271 it not being set at all.
1273 * Reduce contention between BkgDaemon threads
1275 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1278 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1279 objects to use the faster Interlocked operations which avoid
1280 the need to hold write locks.
1282 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1283 non-fatal error that forces a retry against another service
1286 * Fix a crash in smb_WriteData that would occur if an application
1287 opened the file as read-only and then sent a write request anyway.
1290 * Respond to VL_NOENT errors by removing cm_volume_t object and
1291 setting it to be the first object to be recycled. This is not
1292 only a more efficient use of resources but it also prevents
1293 repeated attempts to query the VLDB server for the volume from
1294 the CheckOfflineVolumes routine.
1296 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1299 * A first cut at a cache manager statistics monitor that can be used
1300 to determine the necessary cache parameters to support the
1301 application working set.
1303 Off by default, the performance package can be activated by setting
1304 "daemonPerformanceTuningInterval" in the service Parameters key.
1305 As with the other daemon interval values the unit is in seconds.
1307 At service start and each succeeding interval the cache manager
1308 will write statistics to %TEMP%\afsd_performance.txt showing the
1309 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1310 The FID statistics keep track of all FIDs seen by the cache manager
1311 during the service session whether or not they are backed by any
1312 live object in the cache.
1314 These statistics are not stored in the cache file.
1316 * Fix windows event logging broken in 1.5.33.
1318 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1319 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1320 as well as when computing the server preference list.
1322 * When merging status after a StoreData operation prevent data
1323 buffers from being discarded that we would prefer to keep.
1324 (Bug introduced in 1.5.33.)
1326 * Modify installers the set RxMaxMTU to 0.
1328 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1329 dialog, the Background Daemons and Service Threads fields were
1332 * Increase default number of Background Daemons to 4.
1334 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1335 cm_volume_t objects to be recycled.
1337 * Windows Error Reporting has shown a number of previously unseen
1338 stack corruption errors in the Explorer Shell extension. The
1339 minidump appeared to indicate the problem was in ParseAcl().
1340 Reviewing the function it appears that numerous things could go
1341 wrong if invalid input was provided. Added error checking and
1342 data validation. Hopefully this will address the problem.
1344 * Windows Error Reporting has shown crashes in aklog as a result
1345 of krb5_cc_get_default() failures not being handled. They are
1349 * The Rx library used a 32-bit type for sockets which was
1350 truncating the socket value on 64-bit Windows. This
1351 was not an issue in general but became an issue when
1352 Rx "hot thread" support was enabled.
1354 * Enable "Rx hot thread" support which ensures that while
1355 a received message is being processed that a new thread
1356 is allocated to receive the next message to arrive.
1358 * Increased parallel access to smb_rctLock protected
1359 resources. Reduce the number of times that the
1360 lock is dropped and re-obtained.
1362 * Increased parallel access to cm_scache_t and cm_buf_t
1363 resources by replacing the previous global write lock
1364 protected reference counting mechanism with one built
1365 upon the use of InterlockedIncrement and InterlockedDecrement.
1367 * Implement a hash for cm_fid_t to reduce comparison costs
1368 and inline the cm_FidCmp() function.
1370 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1371 hash which has a better distribution
1373 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1374 the most recent read request instead of the next chunkSize.
1375 cm_chunkSize can be dozens or hundreds of buffers. As a result
1376 too much time is spent performing the evaluation.
1378 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1379 lock is to prevent the creation of new buffers while a truncation
1380 is being performed. All references to bufCreateLock have been
1381 removed except in two places:
1382 i. a write-lock surrounding the function that calls buf_Truncate()
1383 ii. a read-lock within buf_GetNewLocked() that actually allocates
1386 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1387 CM_CONFIG_DATA_VERSION number which value be used to force the
1388 replacement of the cache file contents when incompatible changes
1389 are made between releases.
1391 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1392 cm_SyncOpCheckContinue
1394 * Avoid calls to multi_Rx if nconns == 0
1396 * Modify smb_WriteData to perform background writes based upon the
1397 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1398 boundaries. This will slow down writes from the SMB interface but
1399 will avoid the risk of the CIFS client disconnecting from the AFS
1400 client SMB server. The default value for 'smb_AsyncStoreSize'
1403 * SMB AsyncStore functionality is on by default but can be
1404 disabled using the registry value:
1406 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1408 smb_AsyncStoreSize can be adjusted with
1410 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1412 The value must be a multiple of the buffer block size
1413 and cannot be larger than the chunk size.
1415 * Do not leak a read lock on the bufCreateLock
1417 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1420 * In MergeStatus, only remove clean buffers from the hash table.
1422 * Avoid the need to update the data version number on each buffer
1423 associated with a scache when MergeStatus is called after a
1424 StoreData by maintaining a range of valid data versions as part
1425 of the cm_scache_t object.
1427 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1428 from cm_HaveAccess... We can safely avoid the call so do so.
1430 * Add "ForceAnonVLDB" service parameters registry value. Default
1431 is 0. Set to 1 to force the use of anonymous connections to
1432 the VLDB servers. This is useful when using krb5 tokens and the
1433 VLDB servers have not been updated.
1436 * Do not attempt to search for a cell name that matches one
1437 of the special share names "svrsvc", "wkssvc", "ipc$"
1439 * Issue permission denied warnings from afscreds if attempts
1440 to modify the startup shortcut or registry values fail.
1441 This happens under Vista UAC.
1443 * Do not return BAD_NETWORK_ERROR in response to attempts
1444 to create directories or files in the Freelance root.afs
1445 volume. Instead return ACCESS_DENIED.
1447 BAD_NETWORK_ERROR was being sent because the Freelance cell
1448 has no vldb servers. Short circuit the error by testing for
1449 freelance in CreateFile and MakeDir
1451 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1452 to be opened and closed without actually existing. For now
1453 we treat them as equivalent to ioctls but this should
1456 * Fix RAP processing to not reject names that exist in the
1457 root.afs volume for GetInfo queries.
1459 * In afscreds, use GetUserNameEx to obtain the username instead
1460 of searching the registry. Try UserPrincipalName first and
1461 fallback to SamCompatible.
1463 * Stop linking to rpcns4.dll which is no longer supported on
1464 Vista and 2008. TaAfsAdmSvr is going to have to be
1467 * When processing dfs referral requests, if the sharename is
1468 a partial match and therefore will not be added automatically
1469 to the share list, make sure that we return an error.
1471 * Avoid a crash when logging VNOVOL errors returned from
1472 cm_CheckOfflineVolume.
1474 * When trace logging is disabled, do not permit osi_LogSaveString
1475 to perform the string copying. By disabling the string
1476 copies, throughput of the service more than doubles.
1478 * when checking offline volumes, update the cell vlserver info
1479 if required and don't perform a check if there are no servers
1482 * Add volumes in the vl_unknown state to the list of objects to be
1483 checked as offline or retried when an allbusy error is received
1486 * Add support for configuring rx_ExtraPackets
1488 * All calls to buf_Get() must be protected by a read lock on the
1489 cm_scache_t bufCreateLock
1491 * When a MergeStatus results in the invalidation of buffers, do
1492 not remove buffers from the hash tables that have active references.
1494 * Upon receiving a VNOVNODE error from a file server, be sure to
1495 invalidate any file locks. This prevents on-going attempts to
1496 extend the locks and will cause subsequent attempts to access
1497 the file to receive a invalid handle error.
1499 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
1500 to directory entry names.
1502 * No longer discard tokens when RXKADBADTICKET error is received.
1503 Instead return STATUS_NO_KERB_KEY error.
1506 * Add "cmdebug <host> -cellservdb" option and client service
1507 implementation permits querying the vldb server for each
1508 cell known to the cache manager. The output is in a form
1509 that permits it to be used as a CellServDB file.
1511 * Add Symlinks->Show option to the AFS Explorer Shell Context
1514 * Prevent buffer overrun in the cmd parser used by command
1515 line tools such as fs, vos, pts, etc.
1517 * The path ioctl operations have several issues:
1519 (1) the specified path for "fs examine, whereis, whichcell,
1520 flush" and so always has the follow symlinks and mount
1521 points semantics. This makes it impossible to determine
1522 what the FID of a symlink or mount point is.
1524 (2) "fs examine" out is not the result of a single pioctl
1525 operation but is actually the combined output of half
1526 a dozen operations. Path evaluation is an expensive
1527 operation. It would be faster if the caller could
1528 evaluate the FID first and then perform all of the
1529 rest of the operations by specifying the FID instead
1532 (3) fs output reports all objects as files. By adding a
1533 GetFileType pioctl more informative output can be provided
1534 that indicates what type of object the path evaluates to.
1536 (4) the Windows fs command includes a number of commands that
1537 do nothing but exist only because the Unix cache manager
1540 A new extendible data structure cm_ioctl_query_opts_t has been
1541 added which can be optionally specified with pioctls that do not
1542 already require input data. The first two fields of this structure
1543 are 'literal' and 'fid'. The literal field is used to indicate
1544 whether the last component of the path should be evaluated following
1545 symlinks and mount points.
1547 The fid field permits a fid to be specified.
1549 * A new GetFileType pioctl has been added. The type of objects are
1552 * A new "-literal" option is available for "fs examine, flush, whereis,
1555 * Unimplemented fs commands have been removed.
1557 * There are circumstances where a volume object is being accessed and
1558 the volume is marked indicating that the volume location information
1559 is out of date but where it is also pointless and perhaps dangerous
1560 to block waiting for the rpc to complete. One example is when
1561 processing the cmdebug requests. If we know that we are not going
1562 to use the volume object to contact a server, then we can now set
1563 the CM_GETVOL_FLAG_NO_RESET flag.
1565 * when moving up in the directory tree we search the recorded fid list
1566 to find a matching fid that we have already crossed. we must also
1567 reset the fid count based upon what we discover.
1569 * Cell names are published as share names. As a result they are searched
1570 for as part of the Dfs Referral evaluation. If share "foo" can not be
1571 reached, the CIFS client will fallback to searching for "fo" as well.
1573 Since the freelance client automatically adds a symlink for prefixes
1574 of the cell name, this results in both "foo" and "fo" being added to
1575 the freelance root.afs volume. This patch prevents that negative side
1578 * when tracking Kerberos credential cache names be sure to include
1579 the cache type prefix.
1581 * Microsoft has assigned OpenAFS a network type value
1583 * The algorithms that were used to produce a Netbios name were broken.
1584 The name that was produced when the hostname was too long was in fact
1585 longer than the maximum permitted netbios name.
1587 Also, the "NetbiosName" value was not used as a suffix when the loopback
1588 adapter was not installed. The hardcoded string "AFS" was used instead.
1590 * avoid another deadlock during server probes initiated by the ipaddr
1591 change daemon thread.
1593 * prevent the afs client service from crashing when fs newcell is
1596 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
1597 to \Program Files\OpenAFS\Client\Program in order to get them out
1598 of the %windir% tree
1600 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
1601 end STATUS_OUT_OF_QUOTA can be returned to the application.
1603 * Modify the search order for determining the location of CellServDB
1604 and other client configuration files.
1606 1. AFSCONF environment variable
1608 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
1609 4. use Program Files\OpenAFS\Client
1611 * More Vista Power Management tweaks
1613 * Close a socket handle leak in the kauth routines
1615 * Don't hold a cm_volume_t mutex across RPCs
1617 * Convert downTimes to use clock time instead of relative times.
1618 This permits correct comparisons with expiration times.
1620 * Update the Wix installer to properly install the prerequisites
1621 for translate_et.exe
1623 * Do not translate pioctl paths as they are not converted by the
1626 * Implement multi_rx probes for checking whether servers are up
1627 or down. Turns probes into a constant time activity instead
1628 of one dependent upon the number of known servers.
1630 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
1631 update of the volume location data and retry. VNOVOL means
1632 that the volume could not be attached. It doesn't mean that
1633 the volume is not on this server.
1635 * Modify the Explorer Shell AFS Context menu Symlink->Add to
1636 not attempt to validate the target path. There is no reliable
1639 * Improve performance of \\afs\<cell><mount-type><volume>\
1640 access by eliminating extraneous attempts to search the
1641 CellServDB file and DNS AFSDB records. Share names
1642 containing '%' or '#' are not valid cell names.
1644 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
1645 to unlock a range that is not currently locked.
1647 * Enable the 32-bit tools to read the 64-bit registry keys
1648 when executing in the WOW64 environment.
1652 * Fix buffer refcount leak introduced in 1.5.29
1654 * Prevent the NIM AFS Provider from crashing when configured to use
1655 Kerberos v4 for token acquisition in situations where no Kerberos v4
1656 support is available. For example, 64-bit KFW or sites that delete
1657 the Kerberos v4 DLL.
1659 * In afsd_init.log, support logging cache sizes greater than 2GB.
1661 * The volume status output of "fs examine" was being determined based
1662 upon the wrong error value. Instead of using the pioctl() return
1663 code, the errno value must be used.
1665 * Prevent invalid pts auto-registration attempts by aklog and the
1666 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
1668 * Add synchronization protection to all of the SMB Listener State
1669 variables. This prevents race conditions when the addition or
1670 loss of a network adapter takes place.
1672 * On Vista, there is a race condition between the restart of the
1673 afs client service and the network adapters. Ensure that the
1674 afs client service can handle switching between loopback and
1675 non-loopback modes of operation.
1677 * Fix FollowBackupPath to work with mount points that are not
1678 explicitly read/write.
1680 * Fix a deadlock that has been observed on Vista while resuming
1681 after at least four hours of sleep.
1683 * Add new HKLM registry configuration options that can be used
1684 to provide configuration hints to the NIM AFS Provider when
1685 creating new identities. See release notes for details on
1686 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
1689 * Speed up write buffer operations by holding a global read-lock
1690 instead of a write-lock when updating the buffer data version
1691 number. The data version is protect by the associated stat
1694 * Speed up read buffer operations by removing out of date buffers
1695 from the buffer hash tables.
1697 * Protect against a null pointer dereference in afslogon.dll
1698 if QueryAdHomePathFromSid() fails.
1700 * Remove a race condition while accessing private freelance data
1703 * Update the buffer data versions when Freelance directory data
1706 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
1709 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
1710 instead of returning zero length.
1712 * Return ERROR_RETRY when a server sends VBUSY.
1714 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
1715 a volstat plug-in is in use.
1717 * Enhance the volstat interface to support Dfs path mappings.
1719 * Fix "symlink list" to do the right thing for Dfs links
1721 * Fix callback expiration that was broken in 1.5.27.
1723 * Fix a crash at shutdown caused by the lack of synchronization
1724 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
1726 * Hold the cm_server_t mutex while walking or modifying the
1727 cm_server_vols_t list in order to avoid a race condition.
1729 * Ensure that the cm_server_t refCount is safely modified.
1731 * The pthread library implementations of pthread_cond_wait
1732 and pthread_cond_timedwait were failing to return with
1733 the mutex held when error conditions occurred. This could
1734 result in race conditions wherever they are used. One case
1735 in which problems were seen was the Rx event_handler().
1737 * Prevent cm_GetBuffer() from locking all buffers associated with
1738 a stat cache object before deciding if the current buffer is up
1741 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
1742 object for each offset that we will attempt to reserve.
1744 * Reduce the default number of background daemon threads used to
1745 prefetch or store to one in order to prevent two or more threads
1746 from stepping on each other.
1748 * Reduce the default chunk size to 256KB in order to prevent read
1749 timeouts on low bandwidth connections.
1751 * Fix a race condition that could result in threads becoming stranded
1752 when synchronizing multiple requests on the same stat cache object.
1755 * The GiveUpAllCallBacks functionality that was added in 1.5.21
1756 has been disabled. It has been disabled because it turns out
1757 that the RPC is improperly implemented in all file servers
1758 previous to 1.4.6. The RPC is executed without the proper
1759 locks being held resulting in data corruption and eventual file
1760 server failure. The functionality will be re-enabled in a
1763 * a couple of .readonly Volume Callback optimizations were
1765 - Apply most recent volume callback to all stat cache objects
1766 in the volume. This avoids premature callback expirations
1767 and unnecessary FetchStatus calls.
1768 - Add option to permit attempts to renew .readonly Volume
1769 Callbacks on a periodic basis to avoid invalidation of stat
1770 cache data. [Requires daemonCheckVolCBInterval be set to
1771 non-zero number of seconds. 1800 (30 minutes) is suggested.]
1773 * Adds support for 64-bit data version values. Previous releases
1774 only supported 32-bit values. Larger values sent by the file server
1775 would result in rollover.
1777 * aklog and asetkey no longer generate an exception if Kerberos for
1778 Windows is not installed.
1780 * If the LSA authentication functions report a temporary out of memory
1781 condition, fail the current authentication and do not cause the AFS
1782 client service to terminate unexpectedly.
1784 * When the AFS client service is shutting down, prevent the daemon check
1785 thread from executing additional checks which might attempt to access
1786 discarded locks or unmapped memory.
1788 * Correct a double free condition in the NIM AFS credentials provider.
1790 * Add "FollowBackupPath" service parameter registry value.
1791 This provides equivalent functionality to the UNIX afsd -backuptree
1792 option. When set to a non-zero value, normal mount points originating
1793 in a .backup volume will prefer a .backup volume.
1796 * Fix NSIS uninstall from the .MSI installer.
1798 * Add Stack Frame logging for AMD64
1800 * Add support for EWOULDBLOCK to cm_Analyze. If the file
1801 server returns EWOULDBLOCK, retry the request every two
1802 seconds for up to the RDRtimeout.
1804 * Reorganize the order of the includes to ensure that EWOULDBLOCK
1805 is not assigned the same value as EIO.
1807 * Update CellServDB to GCO Public 25 Oct 2007
1809 * Add additional validation and error handling code after
1810 each call to getSlot(). If an invalid slot is returned,
1811 return NONODE. If the invalid slot is returned when
1812 extracting a data node, invalidate the tree.
1814 * Modify compareKeys() to always perform a case-insensitive
1815 comparison and only perform a case sensistive comparison
1816 if the case-insensitive one matches. This ensures the
1817 ordering is consistently reported.
1819 * Add lock assertions to ensure that all calls are being
1820 performed with the correct locks being held. There have been
1821 some crash reports that provide stack data that does not appear
1822 to be possible unless there is a race. However, there are no
1823 obvious locations where the race is taking place and the test
1824 suite indicates that all of the correct locks are being held.
1825 We shall see what happens in the field.
1827 * For consistency replace all calls to findKey in which the range i
1828 s (1,numentries) with calls to getSlot().
1830 * Optimize the depth search loop by testing the slot value in the
1831 for statement instead of forcing the loop to be broken later.
1833 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
1834 There are a number of locations where locks are obtained, dropped,
1835 and reobtained. This reorganization attempts to accomplish several
1838 (1) be optimistic for the most common case so it will be fast
1840 (2) add consistency checks after each location where locks are
1841 dropped and re-obtained. If we lose a race in cm_BeginDirOp
1842 and the bplus tree is out of date, retry until we get to a
1843 consistent state that we can use.
1845 (3) Ensure that all operations take place with the correct locks.
1847 * One of the issues that has become a serious problem since the
1848 addition of the local directory updates is that although cm_SyncOp
1849 synchronizes operations, it does not preserve the order of requests.
1850 This has always been a problem in that it has been possible for a
1851 request to fail to complete due to its worker thread's bad luck.
1852 When a request takes longer than the Windows SMB Redirector's timeout,
1853 the SMB Redirector tears down the SMB virtual circuit.
1855 When using the local directory updates it is really important that
1856 the directory update operations complete in the order that they were
1857 sent to the file server. If they don't, then the local directory state
1858 and the file server state will not match and the local directory state
1859 must be discarded which in turn forces a new read of the entire
1860 directory contents over the network.
1862 This patch adds a new cm_scache_waiter_t object that is used to store
1863 the current thread, buffer, and syncop flags within a waiters queue
1864 on each cm_scache_t object. If a thread is forced to sleep in
1865 cm_SyncOp, upon waking it will check to see if there are any other
1866 threads waiting that are attempting to perform a similar task ahead
1867 of it in the queue. If yes, the thread goes back to sleep. If not,
1868 it goes ahead and enters the cm_SyncOp conflict resolution block.
1870 This patch has the additional side effect of reducing the number of
1871 competing threads that must obtain the cm_scache_t mutex and process
1872 the cm_SyncOp conflict resolution block. As a result, the overall CPU
1873 utilization of the service and the clock time associated with processing
1874 requests will be reduced.
1876 * assert that the cm_scache_t mutex is held when calling
1877 buf_ForceDataVersion().
1879 * rename findNode to leafNode in bplus_Lookup
1881 * replace all OutputDebugString calls with osi_LogX calls
1883 * modify bestMatch to special case the return values for leaf nodes.
1884 If an entry is above or below the values available in the leaf node
1885 return BTLOWER or BTUPPER instead of BTERROR.
1887 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
1888 if true convert to either slot 0 or Max and perform the insertion.
1889 This produces easier to read code when performing lookups.
1891 * modify lock_AssertXXX macros to call osi_assertx() and provide a
1892 descriptive message.
1894 * the difference between osi_assert() and osi_assertx() is whether or not
1895 a description is specified as part of the call. When no message is
1896 specified afsd_notifier() is called with a NULL msgp parameter. This
1897 results in a null pointer dereference during the EventLog() call.
1899 As a result, none of the information describing the assertion is written
1902 This commit sets a default message for use when no other message is
1903 provided and it replaces all calls to osi_assert with osi_assertx and
1906 * Migrate B+ tree search key into thread local storage
1909 * Fix an uninitialized variable in lana_IsLoopback() which
1910 could result in a random determination that a loopback
1911 adapter is not a loopback adapter. This would result in
1912 a computed Netbios service name of %MACHINE%-AFS instead
1913 of AFS. This is a serious problem if the value computed
1914 by the AFS Client Service and the client process pioctl
1915 calls produce different results as it will result in a
1916 communication failure.
1918 * Fix a crash in the B+ Tree code that results when a file
1919 rename only results in a change of case.
1921 * Fix the pioctl path parsing. The previous algorithm
1922 broke the current directory of the client into two parts,
1923 the TIDpath which represented the device root and the
1924 relative path. The problem is that this is not sufficient
1925 from proper processing of symlinks and mount points.
1927 What is required is that the last component of the path
1928 must not have the symlink or mount point be evaluated
1929 whereas symlink and mount point evaluation must be performed
1930 for the rest of the relative path. Therefore, the path is
1931 now split into three parts: TIDpath, intermediate path,
1934 "fs examine" now reports the data of the symlink or mount
1935 point object instead of the target and "symlink list" and
1936 "symlink remove" now works. The "Explorer Shell Extension"
1937 menu now is displayed for symlink objects.
1939 * More informative symlink.exe error messages.
1941 * When the cmd.exe "del" command is used to remove a symlink,
1942 not only was the link object removed, but the stat cache
1943 object for the target was marked deleted. This resulted
1944 in cache corruption which would require the destruction
1945 of the cache to correct.
1947 * Add undocumented registry value, "OfflineReadOnlyIsValid"
1948 which results in cached contents from read only volumes
1949 to be treated as valid when there is no callback registered
1950 and all servers are down or all instances are offline.
1952 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
1954 * Add documented registry value, "AcceptDottedPrincipalNames"
1955 which is used to permit the acquisition of AFS tokens
1956 using Kerberos v5 principal names which include a dot in
1957 the first component.
1959 {HKLM\HKCU}\Software\OpenAFS\Client
1961 * "fs uuid" does not require "AFS Client Administrator" group
1962 membership to display the current UUID value.
1965 * Update NSIS installer to version 2.30
1967 * Update Wix installer to version 2.0.5325
1969 * 64-bit KFW support:
1970 - Network Identity Manager plug-in
1972 - 64-bit krb5 enable afscreds.exe, afs server manager,
1975 * Enable B+ tree Directory Search by default.
1977 * NIM plug-in is now compatible with FILE: ccaches
1979 * Install instloop.exe into Client/Programs/ even though
1980 it isn't required. Permit it to be used by organizations
1981 that clone Windows systems.
1983 * On a suspend, only drop callbacks for servers that
1984 are online. If the server is offline, maintain the callback
1985 until the server is once again available.
1987 * Prevent against acquisition of a directory write lock when
1988 there are no updates to be performed. This was resulting
1989 in a write-lock leak which later on would lead to a deadlock.
1991 * Add configurable file pre-fetch mode. For files that match
1992 the specified file extensions, pre-fetch the entire file into
1995 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
1996 (MULTI_SZ) PrefetchExecutableExtensions
1998 * Fix the client_osi large integer support so that return values
1999 can be properly assigned to large integer variables.
2001 * Do not hold cm_scache_t mutex across cm_GetCell() calls
2002 since cm_GetCell() can block in a RPC.
2004 * Add short file name support to B+ tree searches.
2006 * Resolve a conflict between the desire to notify file servers
2007 that the client workstation is suspending and the desire to
2008 permit objects with valid callbacks when the client suspends
2009 to continue to use those objects from the cache if the client
2010 resumes without network connectivity.
2012 * Fix stack variables to ensure that all buffers storing paths
2013 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
2014 would cause a stack overflow condition.
2016 * On Vista, force afs_config.exe to run with the highest privileges
2017 available to the end user. For a user in the Administrators
2018 group running under UAC, this will cause afs_config.exe to be
2019 be run with full Administrators privileges after the user
2020 approves the privilege escalation.
2022 * More VNOVNODE handling issues. If a dirty buffer cannot be
2023 written due to VNOVNODE, clear the buffer and ignore the error.
2024 If a cm_scache_t object has been marked deleted due to a VNOVNODE
2025 error, the next time the smb_fid_t object that references it is
2026 used return either NOSUCHFILE or NOSUCHPATH and close the
2027 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
2028 result in an invalid handle error.
2030 * For Vista, set the RequestExecutionLevel of the NSIS installer
2033 * For the build environment, support the generation of manifests
2034 when using the CL1310 (VS .NET 2003) compiler/linker.
2036 * There was a bug in OutputDebugHexString() which could overflow
2037 a stack buffer resulting in unexpected process termination
2038 of afsd_service.exe if the input Security Token contained
2039 many bytes equivalent to the '%' char.
2041 * Add directory lookup statistics to "fs memdump" output.
2045 * Windows uses case-insensitive file name pattern matching
2046 but AFS is a case sensitive file system. The AFS3 directory
2047 format is block based, uses network byte order and
2048 includes a hash table for fast case sensitive lookups.
2049 This causes several problems for the Windows AFS client.
2050 (1) Traversing the directory blocks is cpu expensive
2051 (2) A hash table miss does not indicate that the desired
2052 entry does not exist.
2053 (3) Determining whether a non-ambiguous inexact match or
2054 the entry does not exist requires a linear traversal
2055 of the entire directory.
2056 These issues often result in 100% CPU utilization.
2058 These issues are addressed by building a modified B+ tree for
2059 each directory and then using the B+ tree for searches.
2061 This feature is disabled by default. To enable it set
2062 the "BPlusTrees" registry value to 0x01 in the AFS Client
2063 Service Parameters key.
2065 * Correct a null pointer dereference in the symlink recursion
2066 detection code. (Introduced in 1.5.22)
2068 * Add the Network Identity Manager AFS Provider to AMD64
2073 * A new registry value, "BlockSize", can now be used to adjust
2074 the size of the internal buffers used to store file data.
2075 The default is 4KB. When set, the values 1..1024 specify
2076 multiples of 4KB. Values > 1024 specify actual block sizes
2077 and must be multiples of 4KB. The actual block size in KB
2078 can not be larger than the chunk size which defaults to 1MB.
2080 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
2083 * pts.exe and ptserver.exe built with -DSUPERGROUPS
2085 * Service Manager shutdown hints have been added. The hint
2086 specifies that it may take the AFS client service up to two
2087 minutes to shutdown.
2089 * Fix the recursion detection feature introduced in 1.5.22.
2090 The previous implementation broke the evaluation of paths
2091 using ../path or ./path. It also broke the evaluation of
2092 symlinks to absolute paths. For example, /afs/cell/path.
2094 * Fix an uninitialized variable in cm_MakeDir that could result
2095 in crash if the requested directory could not be created.
2097 * Fix a the VolumeStatusNotification log messages to save
2098 the input string in the circular log buffer so garbarge
2099 won't be output to afsd.log when the log is dumped.
2102 * Increase the SMB Maximum Path length to 260 characters
2103 from 256 as that is what CIFS supports.
2105 * On 64-bit Windows, use a unique description string for
2106 the 32-bit AFS Shell Extension.
2110 * In the AFS Admin Library, replaced all of the
2111 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
2112 allows parameter checking and proper type conversions to be
2113 performed. This is crucial when passing parameters that are
2114 64-bit wide on systems in which 'long' is a 32-bit value.
2115 Otherwise, the parameter list on the stack is interpretted
2116 incorrectly and the function not only doesn't succeed but it
2117 might corrupt memory and if lucky crash the application.
2119 The only program that we care about that uses this library
2120 is the AFS Server Manager. Now it doesn't crash.
2122 * Fixed smb directory searches for non-wildcard names. The
2123 optimized function should not have set the bulk in progress
2124 flag. This would adversely affect later directory searches
2125 that required a bulk search.
2127 * Detect infinite recursions (up to 512 stat cache objects
2128 in a chain.) If an infinite recursion is detected, return
2129 an error to stop the CIFS client from querying us forever.
2131 * Executables executed out of AFS name space can have their
2132 pages swapped out by Windows. Those pages will be reloaded
2133 when they are required. However, if the cache manager no
2134 longer has a callback and all of the file servers with
2135 the volume data are down, then the application will crash
2136 due to the inability to read the required page.
2138 We now track the down time of each server. If the callback
2139 expires after the server that issued it went down and all
2140 servers are down, we delay the expiration to permit the data
2141 in the cache to be continue to be used by Windows or an
2144 * The Windows AFS client has always had very poor performance
2145 with regards to directory management. In the case of
2146 directory updates caused by create, rename,and delete operations,
2147 the AFS cache manager will now perform local directory updates
2148 when possible in order to avoid re-reading the directory from
2149 the file server. This will be done in whenever both of the
2150 following conditions are true:
2152 - the data version on the directory increased by one as a
2153 result of the operation
2155 - all of the directory buffers necessary to make the change
2156 are present in the cache
2158 * The error message in the AFS Control Panel describing the
2159 need to be a Windows Administrator has been replaced with
2160 a message describing the need to be in the AFS Admins Group.
2162 * If the vldb server list is empty when attempting to evaluate
2163 a cell mountpoint it was possible for a crash to occur. This
2164 can only happen if there is an existing mount point and the
2165 is a failure attempting to load the server list from the
2166 CellServDB file or via AFSDB queries.
2168 * dirty buffers only have their dirty bytes stored to the file
2169 server instead of the entire chunksize
2171 * the default chunksize has been increased from 128KB to 1MB
2176 * changed the enum values for cm_serverRef_t state info to use a
2177 private name space to avoid collisions (srv_)
2179 * added a srv_deleted state for cm_serverRef_t objects. This
2180 state is set when cm_FreeServerList() is called with the
2181 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
2182 not always delete the cm_serverRef_t from the list if it is
2183 still in use by another thread. the srv_deleted state means
2184 the object's contents are no longer valid and it must be
2185 skipped. It will be deleted the next time the object is
2186 freed and the refcount hits zero.
2188 * the srv_deleted state is also used when a file server reports
2189 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
2190 as offline. This is done to prevent additional usage of the
2191 stale vldb data while waiting for the update volume request
2194 * added a state field to the cm_volume_t object (enum volstate
2195 vl_ name space) that maintains the state of the volume based
2196 upon the states of all of the cm_serverRef_t and cm_server_t
2199 * modified cm_UpdateVolume() to set the state of the cm_volume_t
2200 RW, RO, and BK to either vl_alldown or vl_online. There can't
2201 be any other states because cm_UpdateVolume() destroys any
2202 previous knowledge we might have had regarding busy or offline
2205 * modified cm_UpdateVolume() to update the volume name in the
2206 cm_volume_t to the volume base name if the previous value was
2209 * modified cm_FollowMountPoint() to check to see if the volume
2210 name is a volume ID and if so call cm_GetVolumeByID instead
2211 of cm_GetVolumeByName. This ensures that volume IDs are always
2212 looked up as numeric values. There is no longer a need to
2213 maintain a separate cm_volume_t containing the string representation
2216 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
2217 The first flag is a "CREATE" flag which is set by all existing
2218 calls. The flag is not set by calls to cm_GetVolumeByID() from
2219 the server probe code when volume status is being updated. We
2220 do not want the server probe operation to result in additional
2221 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
2222 which is set when the server probe code updates the volume status.
2223 This flag will be used to prevent the server probe operation from
2224 changing the order of the least recently used queue.
2226 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
2227 allocated for a given set of normal, readonly, and backup volumes
2228 regardless of whether or not the volume is accessed via name or
2229 ID number. The cm_volume_t namep field is always the base name
2232 * Added a new volume state, vl_unknown. This state is used as
2233 the initial state for all cm_volume_t when the cache manager starts,
2234 for each cm_volume_t at creation, and for each cm_volume_t when
2235 recycling. The cache manager does not know the state of all
2236 volumes in the world, only those that are in the cache and for
2237 which it has queried the VLDB and hosting file servers.
2239 * modified cm_GetVolumeByName() to initialize the state of a
2240 volume to vl_unknown. The actual state will be set when a
2241 cm_VolumeUpdate() call completes successfully.
2243 * changed name of scache hash table variables to avoid ambiguity
2244 when adding hash tables for volumes
2246 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
2248 * modified cm_UpdateVolume() to handle the case in which there is
2249 no RW volume but there is are RO volumes for a given base name.
2250 This is done by querying for the ".readonly" volume name if the
2251 base name does not exist in the VLDB. We never query for the
2252 .backup name because under the current usage model a .backup
2253 volume may only exist on the server that the read-write volume
2254 is located. If there is no RW volume, there can be no .backup.
2256 * Added four hash tables for cm_volume_t objects to improve the
2257 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2258 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2259 BKID. Three ID hash tables are necessary as long as it is
2260 desireable to maintain a single cm_volume_t containing all
2261 of the related RW, RO, and BK volume data. Having the RW and
2262 RO volume data in the same object is necessary for the
2263 implementation of cm_GetROVolumeID() which returns either the
2264 RO or RW ID depending upon the existence of RO volume instances.
2266 * Added a volume LRU queue so that volume reuse becomes fairer.
2267 This does not replace the all Volumes list which is used when
2268 it is desireable to walk a list of all the volumes whose order
2269 is not going to change out from underneath you which makes it
2270 safe to drop the cm_volumeLock.
2272 * handles volume hash table updates where volume name to
2273 volume ID number changes. The volume name remains
2274 constant in the cm_volume_t. if a vos rename is performed,
2275 the name of the volume will change and the volume IDs will be
2276 updated. Subsequent access to the old volume ID will create a
2277 new cm_volume_t with the new name.
2279 * Added a daemon thread operation to query the state of volumes
2280 listed as busy or offline. cm_CheckBusyVolumes() calls
2281 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2282 or vl_offline. If the volume is now online, the status on the
2283 volume is updated. The default period is 600 seconds. This can
2284 be configured with the BusyVolumeCheckInterval registry value.
2286 * Added prototype for smb_IoctlPrepareRead() which was missing a
2287 return type in the function definition.
2289 * Added volume id lists to the cm_server_t. These lists are
2290 allocated in blocks of ~32 IDs. When a cm_PingServer()
2291 detects a change in server state, the state of the cm_volume_t
2294 * Added volID to the cm_serverRef_t object. volID is used
2295 to identify the volume for which the object is a referral.
2296 cm_FreeServerList() uses the volID to remove the volume
2297 from the cm_server_t.
2299 * In cm_Analyze, when VNOVOL or VMOVED are received,
2300 call cm_ForceVolumeUpdate() to force a refresh of the volume
2303 * Added cm_VolumeStatusNotification() which is used at the moment
2304 to log volume status changes to the trace log. It will also
2305 be used as the access point to the File System Filter driver
2306 notification engine.
2308 * Added an all cm_scache_t list to cm_data. This replaces the use
2309 of the stat cache LRU queue when we need to enumerate all
2310 entries. The LRU list order is not static and when using it to
2311 enumerate all entries it can result in items being missed or
2312 items being processed more than once.
2314 * Modified cm_Analyze(). Instead of reseting the busy or offline
2315 state of a volume and forcing a retry of the operation
2316 cm_Analyze will defer to the background daemon thread that will
2317 update the state once every 600 seconds.
2319 * Added the automatic generation of a Freelance ".root" read-write
2320 mountpoint that refers to the root.afs volume of the workstation
2321 cellname at the time the mountpoint is created.
2323 * re-write cm_Analyze to make better use of the known volume
2324 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2327 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2328 busy volumes will be reset to srv_non_busy by the function
2329 but there is no mechanism for querying the busy state other
2330 than by attempting to access the resource.
2332 * cm_Analyze will query the state of an offline volume before
2333 deciding whether or not to retry when all volume instances
2336 * the algorithm used after an IP address list change detection
2337 has been modified. we now check the state of the VLDB servers,
2338 then reset the rx connections, and finally check the state of
2339 the file servers. Previously we flushed the rx connections
2340 and then checked the servers in whatever random order they
2341 were first discovered.
2343 * fix the volume status reported by "fs examine"
2345 * various improvements to the AFS Network Identity Manager Provider.
2346 - Don't access freed objects after the window has been destroyed.
2347 - Write back the AFS cell list in the identity configuration
2348 even when its empty.
2349 - Disable all use of the provider when AFS is disabled for a
2350 particular identity.
2352 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2353 alias instead of the full cellname, double check the full cell
2354 name and see if we have a matching cell before allocating a new
2357 * Dump more data with "fs memdump". All volumes and scaches and
2360 * Reverse the order of the all cells list. This ensures that
2361 the workstation cell is always first in the list.
2363 * If we recognize the server from which a callback revoke was
2364 received, then we can ensure that we only invalidate the status
2365 information for cm_scache_t objects belonging to that cell.
2367 * When building for amd64, use APPVER 5.02 as we do not support
2368 64-bit Windows 2000 and explicitly state the machine type in
2371 * no longer treat VMWare network adapters as equivalent to a
2372 loopback adapter. For some unknown reason, publishing the "AFS"
2373 netbios name on them frequently results in a name collision
2376 * Use read locks whenever possible instead of write locks when
2377 walking the allCellsp list. This permits better use of multiple
2380 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2381 is not present on the directory optimization. It would report
2382 mount points and directory objects for which there was no
2383 status information as a zero length file. Instead we can use
2384 the vnode number of the object to determine if it should be
2385 considered to be a directory or a file. If it is odd it is a
2386 directory; even it is a file.
2388 * Give Back Callbacks to the file servers in response to suspend
2389 and shutdown power management events. Only give up callbacks
2390 to UP servers and do so with a very short RX dead timeout.
2393 * Fix a race condition in the SMB FID cleanup code that could
2394 result in a stat cache mutex never being released.
2396 * Fix the graphics files that are installed as part of the
2397 IBM Administrators Guide.
2399 * Change the behavior of pioctl path parsing to not follow
2400 symlinks or mount points during evaluation. This results in
2401 the FID being the symlink or mount point and not the object
2402 they point to. This is consistent with the behavior of the
2405 * Enable AFSDB record lookup for command line tools such as
2408 * CellServDB update 2007-07-05
2410 * Added a Volume Status Plug-in Interface that permits third parties
2411 to receive notifications when the volume state changes occur.
2413 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2415 * Prevent the afsprocmgmt.dll from being unloaded while the
2416 background thread it starts is running.
2418 * Avoid a race condition when closing smb_fid_t objects.
2421 * fix install path of IBM Quick Start Guide when installed by
2424 * fix the client so that it can write into directories with
2425 only 'li' permissions.
2427 * fix installation of 32-bit Shell Extension for 64-bit systems.
2428 Install it under its own HKCR keys.
2430 * modify the wix installer to remove and upgrade install packages
2431 with the same version number.
2433 * when flushing volume location data every two hours or in
2434 response to "fs checkvolumes", also flush the mount point
2435 links in order to force their re-evaluation.
2437 * aklog was copying an uninitialized buffer as part of the
2438 foreign cell automatic pts registration
2440 * when obtaining directory listings, check for read permission
2441 on the directory. If read permission does not exist, fake the
2442 status information for the directory entries. This improves
2443 performance, prevents the client from pounding the file server,
2444 and prevents leakage of cached status information.
2446 * "fs getcacheparms" output was incorrect.
2448 * NetIdMgr AFS Provider should only automatically add token
2449 acquisition for cell foo.com if the realm of the cell matches
2450 the realm of the identity.
2452 * Fix a null pointer dereference in cm_MergeStatus()
2456 * network identity manager provider:
2458 - Correct for the fact that pr_CreateUser tampers with the
2459 contents of the username input buffer.
2461 - Take into account that the krb4 credential data structure
2462 does not have separate realm fields for the user and service.
2463 When krb524 is used with a cross-realm ticket, the realm
2464 describes the service not the user.
2466 * Many integrated logon changes:
2468 - Fix eventlog reporting. Do not attempt to log an event
2469 if the event source registration fails. Use DebugEvent0
2470 instead of DebugEvent when there are no parameters.
2472 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2475 - Fix the reading of Domain specific configuration for
2476 LogonScript and TheseCells. Previously the dwSize value was
2477 being overwritten so that subsequent RegQueryValueEx call
2480 - Fix a memory leak in the TheseCells reading code.
2482 - Add support for Domain specific "Realm" specification. The
2483 realm is the realm to be appended to the username. When
2484 logging in as a domain or to the local machine, the specified
2485 "Domain" name is not going to be a valid realm name.
2487 - Construct a proper principal name based upon the domain
2488 specified realm for use in obtaining tokens with KFW.
2490 - If the domain specified "TheseCells" list includes the
2491 default cell, do not obtain tokens twice.
2493 * An unprotected call to the Kerberos v5 function,
2494 krb5_free_context(), can result in an invalid memory access
2495 exception if the contents of the krb5 profile, krb5.ini, cannot
2496 be parsed correctly. The error is located within the afskfw
2497 library which is used by afscreds.exe, afssvrmgr.exe, and
2498 afslogon.dll. When the error is triggered within afslogon.dll
2499 a blue screen condition can occur when the winlogon.exe system
2500 process terminates unexpectedly.
2504 * Do not stop the afs client service with an assert
2505 if the cell name cannot be determined when responding
2506 to a "fs whereis" command.
2508 * Integrated Logon (afslogon.dll). No longer attempt
2509 to preserve Kerberos 5 tickets for the logon session
2510 as that job is now supported by KFW 3.1 and above.
2512 * The Network Identity Manager plug-in will attempt to
2513 remove expired service tickets from the cache.
2517 * CellServDB update - GCO Public CellServDB 09 Mar 2007
2519 * AFS Server Manager and AFS Account Manager data cache
2520 moved from HKCU\Software\TransarcCorporation to
2521 HKCU\Software\OpenAFS to prevent crashes due to data
2524 * Reduce the default UP server probe time from 10 minutes
2525 to 5 minutes to match UNIX. This will not solve the NAT
2526 port mapping timeout issue because most NATs default to
2527 a value less that 5 minutes and this value ensures that
2528 the afs client server will start probes 5 minutes after
2529 the last probe set ended. The probe period for any one
2530 server is therefore longer than 5 minutes. The more
2531 servers in the list the longer the time between probes.
2533 * Before marking a server as down in cm_Analyze, retry the
2534 request using a new rx connection to the server. If the
2535 cache manager is behind a NAT, it is possible that the
2536 response from the server is being sent to the original
2537 address:port from which the rx connection was established.
2538 This will appear to the cache manager as the server being
2539 down when in fact it is not. Forcing the use of a new rx
2540 connection will result in the server responding to the
2541 most recent address:port assigned to the client by the NAT.
2543 If the server is in fact down, it will simply take a little
2544 longer to notice it.
2548 * Removing a directory that contained entries was failing with
2549 the wrong error code being returned to the CIFS client.
2553 * rxdebug exponential back-off. don't take down a server that
2554 is already struggling to respond.
2556 * Network Identity Manager plug-in registers help on the Help
2557 menu if it is supported by NIM. (KFW 3.2 and above)
2559 * Plug-n-play networking did not always work after coming back
2560 from a Windows Suspend. The AFS client service failed to
2561 re-compute the LAN adapter to bind to and instead attempted
2562 to bind to all LAN adapters. This would polute the public
2563 network with the name "AFS" which could lead to conflicts
2566 * Build installers using Microsoft Hotfixes provided at
2567 http://tinyurl.com/243svk for Visual Studio C RunTime Library
2568 Daylight Savings Time issues.
2570 * "fs checkserver" does not require admin privileges unless
2571 the probe interval is being changed.
2575 * A reference leak was discovered in the mountpoint traversal
2576 functionality which resulted in cm_volume_t objects not becoming
2577 eligible for re-use. This will trigger a panic condition when
2578 more than the maximum number of volumes is accessed.
2580 * Fixed the success test for the "fs memdump" operation. Success
2581 is now reported as such. Added cm_volume_t object information
2584 * A data corruption error has been fixed. Write requests
2585 queued for background processing were not retried upon failure.
2586 This leads to inconsistencies between the contents of the file
2587 in the local cache and the contents of the file on the file
2590 The implemented fix is to re-queue any request that has failed
2591 due to timeout, volume busy, volume offline, server busy, or
2592 server offline. Requests are not processed while all servers
2593 for a volume are inaccessible. When servers become accessible
2594 the background request will be retried.
2596 In addition, if a background write fails, do not mark the buffers
2597 as clean. Failure to preserve the dirty flag prevents the buffers
2598 from being written to the file server in the future.
2600 This fix does not address the possibility that the credentials
2601 necessary to perform a write request might expire between the
2602 time the request is queued and when the request can be processed
2603 due to server accessibility.
2605 * The Freelance volume does not support per user ACLs. The
2606 cm_HaveAccessRights function did not handle this special case.
2607 Requests for privileges greater than read|lookup would result
2608 in afsd_service.exe entering an infinite loop. ISOBuster is
2609 known to cause this behavior.
2611 * When deleting files, the SMB Server would attempt to obtain
2612 the access permissions for the parent directory instead of
2613 for the file itself. This would result in an incorrect
2614 assessment of whether or not the file can be deleted.
2616 * The Network Identity Manager AFS plug-in has had its string
2617 tables updated for consistency with the language used by NIM.
2618 All credentials are referred to as credentials instead of tokens.
2619 Drop-down combo boxes now in fact drop down.
2621 * The AFS Salvager was not properly being built on Windows.
2622 This is relevant only if the AFS servers are in use.
2624 * The default signal handlers for the AFS servers incorrectly
2625 processed SIGQUIT. Instead of stopping the server, an
2626 exception was raised.
2628 * Fix another volume refcount leak. This one was triggered by
2629 an inability to contact the vlservers in order to update the
2632 * Fix checklist control used by the drive mappings tabs in
2633 afscreds.exe and afs_control.exe.
2635 * Improve output of "fs memdump" related to buffer management.
2637 * Fix refcount leak of cm_buf_t objects when prefetching buffers
2638 already in the process of being fetched.
2640 * Improve cache performance by computing the buffer hash table
2641 size based upon the number of buffers so that the average
2642 number of entries in a bucket is seven.
2644 * Improve cache performance of read/write data by locally
2645 updating the dataVersion of all buffers associated with the
2646 file on each StoreData or StoreStatus operation provided that
2647 the dataVersion was incremented by one and the dataVersion of
2648 the buffer matched the original dataVersion of the cached
2649 status information for the file.
2651 * When recycling cm_scache_t objects, make sure that we
2652 clear all of the previous fields. It is no longer the
2653 case that we can assume that the stat cache object is being
2654 obtained only after a successful FetchStatus call.
2655 InlineBulkStatus will allocate stat cache objects for which
2656 the no status information is available.
2658 * When a file is being opened with the OPEN_ALWAYS disposition,
2659 do not require write permissions when the file already exists
2660 unless of course the file is being opened requesting write
2663 * When a file has been opened for read-only access, if the
2664 application requests an exclusive lock, obtain a read-lock
2665 instead of a write-lock. The purpose of the exclusive lock
2666 in this situation is to prevent changes to the file and the
2667 full file read lock will provide this guarrantee.
2672 * Enforce Unix Mode Write bit to prevent file deletion of read only
2675 * The CIFS READ and LIST permissions use the same value 0x0001.
2676 When a directory is being accessed, check for PRSFS_LOOKUP
2677 instead of PRSFS_READ.
2681 * Fix computations of short names of mount points. They were
2682 broken as part of the directory search optimization. (1.5.13a)
2684 * When all buffers are in use, periodically sleep the threads
2685 waiting for a buffer to become free while the global buffer
2686 lock is not being held in order to permit those threads that
2687 can free buffers to do so.
2689 * Power Management improvements. Maintain a global flag that
2690 specifies whether or not the service is in a suspend state.
2691 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
2692 lana is no longer valid. Instead, stop the listener threads
2693 and if all listener threads are stopped, reset the lana_list.
2694 Allow the cm_Daemon() thread to periodically check the state
2695 of the smb listeners. If they are all stopped and the service
2696 is not suspended, attempt to restart them. If there are no valid
2697 lanas, return to the stopped state.
2699 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
2700 the user is permitted to obtain read or write locks. This function
2701 would obtain the lock and then drop it returning whether or not
2702 the lock could be obtained. If the lock was in fact required,
2703 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
2704 of course being that this pattern results in three RPCs to the
2705 file server (lock, unlock, lock). This is reduced to one RPC by
2706 implementing cm_CheckNTOpenDone() which frees the allocated byte
2707 range lock from cm_CheckNTOpen() after the cm_Lock() call is
2710 * Optimize the SMB FindFirst processing for the case in which there
2711 are no wildcards. If a cm_Lookup succeeds, generate a response
2712 from the stat cache entry we have a callback for. Its cheaper
2713 than walking each buffer of directory data associated with the
2714 directory stat cache entry when we have thousands of directory
2717 * Increase the SMB Ioctl MaxData size to support larger token sizes
2719 * Disable autorun of the AFS Server Wizard when the AFS Server is
2722 * Add AFS Server service dependencies on tcpip and PNP_TDI
2724 * Remove the AFS Server Configuration Tool shortcut
2726 * Add Freelance mode awareness to AFS Server Configuration Tool
2730 * Unlike previous versions of the OS, Vista performs a shutdown on
2731 the Microsoft Loopback adapter just like it would on a real adapter.
2732 This causes the smb_Listener threads to trigger a panic during a
2733 suspend/hibernate power event.
2735 The fix is to unbind from the network adapters in response to a
2736 suspend/hibernate power event and then rebind to the adapters
2737 when a resume power event is received. Note that the resume events
2738 are not reliably delivered so it is possible the afs service will
2739 not be accessible. However, this is the best we can do.
2741 * don't allow environment variables we set in afslogon.dll to be
2742 inherited by child processes
2744 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
2747 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
2748 thread walking the all buffer list periodically searching for dirty
2749 buffers to write to the file server. This had several negative
2750 results. The algorithm ate up ever increasing amounts of CPU time
2751 even when AFS is idle as the size of the cache increases. Also,
2752 buffers were written to the file server in an order based upon the
2753 original buffer allocation which has nothing to do with the order
2754 in which the buffers became dirty.
2756 The new algorithm maintains a dirty buffer list. Items are added
2757 when the buffer is originally marked dirty. A buffer is only
2758 removed from the list by the buf_IncrSyncer when the buffer is no
2759 longer dirty. If the list is empty the thread goes back to sleep
2760 immediately without additional processing requirements.
2764 * Enable renaming of directories from the Explorer Shell when
2765 Offline Folders is enabled
2767 * An error is reported in response to the SMB ReceiveCoreFlush
2768 request when an error has not occurred.
2772 * optimize dirty buffer flushes in response to suspend operations
2773 do not flush if only loopback adapter is installed
2774 do no retry dirty buffer writes
2776 * do not reset netbios when performing loopback checks. doing
2777 so will break the existing lana bind
2779 * remove a deadlock in smb3.c QPathInfo
2783 * Add NetIDMgr AFS Plug-in
2785 * "Show Tray Icon" checkbox removed from afs_config.exe
2787 * Disable DNS Registrations for the Loopback Adapter and make sure
2788 that Netbios is turned on.
2790 * A reference count leak on stat cache objects was fixed
2792 * For debugging, tread ids of writers are now tracked on locks
2794 * Corrected an incorrect lock that was being held but not
2795 freed in a rare error case
2797 * Changed trace log end of line to CR-LF so it can be read
2800 * modify the buffer synchronization thread to write all dirty
2801 buffers and sleep only when there is no work to be done
2803 * improve behavior of cache manager when large numbers of
2804 temporary files are being created by multiple clients within
2807 * do not permit dirty buffers that belong to deleted files
2808 to be written to the file server. its a waste of bandwidth
2810 * do not leak stat cache references when the file server
2813 * afscreds: do not permit an attempt to obtain tokens if the
2814 afs service could not be started. this was causing stack
2815 overflows on Vista after resuming from a suspend
2819 * better handling of dot directory
2821 * fs sq and fs mkm now behave as on UNIX
2823 * add license to installers
2825 * Return Invalid Parameter error for unknown InfoLevels in
2826 SMB Tran2 Search Directory.
2828 * Correctly use WSAStartup and WSACleanup within the RX library
2830 * Check that afsd_service.exe is running before performing
2833 * Force AFS Shell Extension to load resource DLL using
2834 afs_shl_ext.dll as a base name.
2836 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
2837 OpenAFS\Client\Program
2839 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
2841 * Do not panic if the maximum number of volume entries are in use
2842 and one of them can be recycled.
2846 * Fix the code that reads the backconnnectionhostnames value
2847 so that there is not an uninitialized variable
2851 * integrated logon could cause winlogin.exe to blue screen windows
2852 if it was called with a domain that does not exist on the machine.
2853 this is possible during a sysprep.
2855 * corrections to cm_buf_t handling to ensure that buffers are
2856 removed from queues before they are freed and that holds are
2857 obtained on the next buffer in a chain before releasing the
2860 * do not generate MessageBox dialogs on network errors. they
2861 cause more trouble then they are worth.
2863 * move GlobalAutoMapper into a separate thread so that it does
2864 not prevent service startup if the WNetAddConnection2 call
2867 * restrict hard dead and connection timeouts to the CIFS Session
2868 Timeout value. This is being done to try to prevent virtual
2869 circuits from being destroyed prematurely.
2871 * Fix the code that writes the backconnectionhostnames value
2872 to ensure that the data buffer is written with the correct
2875 (there were no windows specific changes between 1.5.4 and 1.5.5)
2879 * Corrections to the Integrated Logon functionality that permits
2880 Kerberos 5 TGTs to be communicated into the user session. The
2881 transfer is now more secure and succeeds when the user does not
2882 have read permission to the %SystemRoot%\TEMP directory.
2884 * NSIS installer updated to version 2.18
2886 * WiX installer updated to version 2.0.4221.0
2888 * Updates to the IFS functionality for 64-bit correctness and
2889 permit compilation. The IFS functionality works for pioctl()
2890 operations but not much else at this point.
2892 * A major re-write of the OpenAFS implementation of the CIFS Tran2
2893 query and set functions.
2895 * Ignore the Windows CIFS client session timeout entirely. Timeout
2896 only after the AFS Hard Dead Timeout period expires.
2900 * move headers, libraries, and samples from OpenAFS\Client to
2903 * remove the auto-registration of AFS IDs for foreign cells
2904 from integrated logon. this permits afslogon.dll to avoid
2905 initializing the RX library. afslogon.dll is a network
2906 provider library and as such it is loaded into all applications
2907 which perform a CreateProcessAsUser call. Initializing the
2908 RX library spawns background threads to listen for inbound
2909 packets. When the network provider is unloaded from the
2910 process, the RX library does not cleanup the threads. If they
2911 continue to execute after the library is unloaded, the process
2914 * implement inline bulk status calls. this will significantly
2915 improve the performance of directory listings.
2917 * prevent a crash when evaluating mountpoints to volumes that do not
2920 * new command: fs uuid [-generate]
2921 displays the current uuid and allows a new one to be generated
2922 on the fly without restarting the afs client service.
2924 * fix the 64-bit installers to properly document the install
2927 * move help files to the proper location so they can be located by
2928 the system tray tool and the control panel.
2930 * "over quota" and "disk full" errors were not being reported to the
2933 * largefile support has been added. Files greater than 31-bits can no
2934 longer be written to file servers which do not support large files.
2935 When file servers support large files, files up to 64-bits can be
2938 * cm_SyncOp could return with the input buffer unlocked upon error
2939 which would result in a panic condition. This bug was introduced
2942 * vlserver failover when freelance mode did not work. This prevented
2943 the root.afs volume from being loaded if the first vlserver could
2948 * A bug in the CIFS Standard Info response packet caused directories
2949 to appear as standard files. This prevented Microsoft Vista from
2950 being able to walk the AFS directory tree.
2952 * Mount points must be reported as directories in response to
2953 Query Path Info operations.
2955 * A deadlock was detected when performing "fs flushall" if the
2956 file server reports VNOVNODE. The scp->createBufferLock is
2957 already held by the current thread and there is no mechanism
2958 to propagate the knowledge. Therefore, an alternate mechanism
2959 for clearing the cache must be developed. A new function
2960 cm_RecycleSCache(scp,flags) has been extracted from
2961 cm_GetNewSCache(). This function performs the task of recycling
2962 an cm_scache_t object. When called from cm_GetNewSCache() with
2963 no flags the expectation is that there are no associated buffers
2964 that are queued to be read or written. When called from
2965 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
2966 flag, any queued buffers will be de-queued and marked as if
2967 the operations were performed so the data can be discarded.
2969 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
2971 * In response to VNOVNODE, the parent is only discarded if the current
2972 cm_scache_t is not a directory.
2974 * In the Ioctl FlushFile and FlushVolume functions, there were no
2975 checks to protect against flushing the Freelance SCache entries.
2977 * In FlushFile, the wrong cm_scache_t object was being released.
2979 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
2980 recycled. Choose a new entry if cm_RecycleSCache fails.
2982 * Speed up the performance of the cache manager by not
2983 holding the smb_fid_t mutex across calls to cm_SyncOp
2986 * Ensure that all smb_fid_t flag references are protected.
2988 * Remove a deadlock between smb_fid_t mutex and
2989 smb_rctLock introduced in 1.4.1.
2991 * Fix aklog AFSID lookup to use not send the realm
2992 name when the realm is the local realm for the
2993 cell. The PT Server doesn't strip the realm
2994 name when it is the local realm.
2996 * Treat "\\afs\*." as an alias for "\\afs\all"
2998 * Dynamically adjust the priority of server threads
2999 based upon the age of the CIFS request that is
3000 being processed. The priority is bumped one level
3001 for every 15 seconds of processing time.
3003 * The original openafs contribution mis-used the
3004 cm_GetCallback function. This function should only
3005 be called as a side effect of a call to
3006 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
3007 was being called results in a FetchStatus RPC being
3008 performed even when the client already has a valid
3009 callback and does so in a manner that destroys the
3010 synchronization of the threads that are calling
3011 cm_SyncOp. If nothing else this results in a
3012 significant performance penalty.
3014 * More changes to cleanup of smb_vc_t objects in
3015 order to prevent race conditions.
3017 * Reduce the amount of Power Management event logging
3020 * If the client detects that the IP address has
3021 changed, force the use of new RX connections. This
3022 enables the file server to recognize that the client
3025 * Remove a race condition and deadlock associated
3026 with cleanup of smb_vc_t objects.
3028 * Fix the ANSI filename support. One function that
3029 required it was missing the translations.
3031 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
3034 * Add a DllMain() function to afspthread.dll. Perform
3035 cleanup of thread local storage, rx queues, critical
3036 sections, event handles, etc. This was implemented
3037 to avoid crashes in afspthread.dll when applications
3038 which load/unload the dll attempt to terminate.
3040 * Lock request behavior has been altered. As of this
3041 release a locally allocated lock will be used whenever
3042 the volume is RO or the requesting user's permissions
3043 are 'rl'. This will allow end users to execute programs
3044 from RW volumes when they do not have the 'k' privilege.
3046 Note that a recent discovery indicates that the 'k' bit
3047 is only supposed to control the ability to obtain read
3048 locks. Write locks are to be controlled by the 'w' and
3051 * Code has been added to enable the cache manager to be
3052 able to obtain the AFSID of the authenticated user.
3053 This is required to be able to determine when "Insert"
3054 privileges should apply. AFS considers a file to be
3055 a new file whenever the owner's AFSID matches the
3056 AFSID of the user. The current implementation relies
3057 on the name specified during the SetToken call to be
3058 correct. This will be replaced in the future with a
3059 new RPC that allows the AFSID of the authenticated
3060 user to be obtained.
3062 * A 32-bit-tools installer for use on 64-bit systems
3063 is now available. These tools are for use with 32-bit
3064 versions of Kerberos for Windows or JPSoftware's 4NT.
3066 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
3068 * Remove the use of the "active_vcp" variable. This variable
3069 no longer serves any useful purpose and was the cause of
3070 race conditions that produced under counts of the smb_vc_t
3073 * Change default minidump output to include data segments
3075 * restore access to cm_IoctlGetVolumeStatus and
3076 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
3077 Still need to figure out why they are being called
3078 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
3080 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
3082 * Update build system to include the "sys_name" in the
3085 * optimize the number of calls to obtain the smb_user_t for a
3086 smb virtual circuit.
3088 * add logging of server addresses, status and capabilities when
3091 * prevent multiple threads from pinging the same server simultaneously.
3092 instead, threads block and then use the result of the ping that was
3093 already taking place.
3095 * Add registry values to allow the defaults for daemon checks to be altered
3096 at startup. Log values to afsd_init.log.
3098 daemonCheckDownInterval
3099 daemonCheckUpInterval
3100 daemonCheckVolInterval
3101 daemonCheckCBInterval
3102 daemonCheckLockInterval
3103 daemonCheckTokenInterval
3105 * Add EventLog messages for transitions in fs crypt state.
3107 * Byte range locking (when locks will be obtained from the
3108 is still under discussion)
3110 * 64-bit Windows operating systems now supported
3111 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
3112 This is a native 64-bit implementation of OpenAFS for Windows.
3113 There is currently no 64-bit version of KFW and therefore
3114 there is no (easy) method for obtaining AFS credentials.
3115 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
3116 to obtain AFS tokens but there is not for the feint of
3119 * A faster implementation of the fcrypt algorithms improves
3122 * The afsd_service.exe now supplies Event Log Message Text.
3123 The event log no longer generates warning text due to lack
3124 of message configuration.
3126 * When afs servers are unreachable and network stack returns
3127 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
3128 mark the attempted server as being down. This allows for
3129 very fast response to the calling application. This change
3130 has a pronounced improvement when the Windows Explorer is
3131 used on disconnected mobile computers.
3133 * Added DebugOutputString debugging of the RX library
3135 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
3137 * Fix the rx-lwp implementation to randomly generate rx call
3138 identifiers which are supposed to be unique at a given point
3139 in time. This affects some of the command line tools used
3140 for debugging but not the AFS Client Service itself which
3141 uses the rx-pthreads implementation.
3143 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
3144 Windows although they are not distributed
3146 * don't generate an exception when logging unexpected SMB packets
3148 * the data structures used by cm_IoctlGetVolumeStatus and
3149 cm_IoctlSetVolumeStatus are not compatible with the CIFS
3150 Get/Set File Information calls so don't bother executing them.
3154 * Fix data written to registry as part of BackConnectionHostnames
3157 * Fix initialization bug when allocating cm_server_t objects that
3158 would cause servers to appear to be down after each allocation
3159 or renewal. Server would appear up after the next check servers
3164 * Windows will return WSAECONNRESET to the next receive packet
3165 operation on a socket if an ICMP packet is received in response
3166 to a packet that was sent and could not be delivered because
3167 the port was unreachable. We have to treat this as
3168 WSAEWOULDBLOCK so that we can retry the select() check and
3169 not mark the connection as bad.
3171 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
3173 * The client manager maintains a least recently used list of
3174 all the ACLs. If an ACL is already the most recently used
3175 do not bother to remove it from the list and place it on the
3178 * As part of the process of adding a new server, set the
3179 server is down flag and then perform a ping. If the server
3180 is up, the flag will be reset. If the server is down, the
3181 client will not timeout unnecessarily while attempting to
3184 * If a Mountpoint String is the empty string, return PATH NOT
3187 * Modify behavior of the Network Provider to avoid querying the
3188 profile location if integrated logon is known to be disabled.
3190 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
3191 will always be preserved regardless of whether or not integrated
3192 logon was used to obtain tokens. When unset, tokens will be
3193 destroyed regardless of whether integrated logon was used to
3194 obtain the tokens unless the profile is stored in AFS.
3196 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
3197 message is being delivered even when the user is not logging
3198 off of the computer.
3200 * Fix a deadlock condition introduced in rc6.
3202 * Cell names are case-insensitive. Since the founding of OpenAFS
3203 there has been a bug that treated "cell", "Cell", and "CELL" as
3204 different cells. This would result in what appeared to be
3205 random disappearance of user tokens or a failure to authenticate
3206 even when tokens were listed by the "tokens" command.
3208 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
3210 * Additional auditing of the smb code revealed that smb ioctl
3211 operations were failing to free smb_fid_t objects
3213 * The File Identifiers assigned to smb_fid_t objects could be
3214 given the value 65535 which is 0xFFFF which is also the
3215 INVALID_HANDLE value. Both 0 and 65535 are now treated as
3218 * Audited all assignments to fields and ensured that they are
3219 protected by obtaining the relevant mutex object.
3221 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
3222 if both the current cm_scache_t object and its parent are
3223 both returning VNOVNODE.
3225 * Removed all the logic associated with the 'dead_vcp' variable
3226 which must have been added for debugging because its used
3227 was never thread safe and it truly served no purpose.
3229 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
3231 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
3232 registry settings. Default is ON and 120 seconds. If set to
3233 OFF, tokens will not be destroyed.
3235 * Fixed a number of race conditions and reference count problems
3236 associated with global queues.
3238 * Added a method to detect abandoned smb virtual circuits so their
3239 associated file handles, locks, user tokens, and smb session
3240 resources can be released.
3242 * Added a mechanism to allow user tokens to be preserved
3243 during integrated login and be freed after logoff.
3245 * Check access permissions to the file before the file is
3248 * More improvements to the handling of SMB Virtual Circuits
3249 when SMB sessions are closed.
3251 * Correct a race condition introduced in rc4 that could result
3252 in a crash of afsd_service.exe.
3254 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3256 * Properly handle pre-mature SMB session closures.
3257 File handles, tree connections, and user sessions are now
3258 freed when the virtual circuit is closed.
3260 * smb_username_t and smb_user_t objects are now properly reference
3261 counted and freed as appropriate.
3263 * disable to use of AFS file locks. byte range locking is still
3264 implemented but all locks are managed locally. This prevents
3265 multiple processes from editing the same file on the same
3266 machine but it does prevent multiple processes on separate
3267 machines from modifying the file.
3269 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3271 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3272 the same server parameter would result in a panic condition
3273 if the server had not already been contacted by the AFS
3276 * Better warnings are written to the afsd.log file when
3277 the Windows RPC Protocols are not properly configured as
3278 is often the case on Windows 2000.
3280 * fix the procmgmt library to restore C RTL signal handlers when
3281 it is being detached from the process prior to process termination.
3283 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3285 * close a lock leak in the rx library that could be triggered while
3288 * prevent removal of drive mappings in response to afscreds.exe -M
3290 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3291 strings and flush the stat cache entry.
3293 * prevent server objects from being freed if user preferences are set.
3295 * Update WIX support to 2.0.2419.0
3297 * Release all locks on the server at service shutdown.
3299 * A down server should be marked up as long as it responds to the client.
3300 It doesn't matter if the server returned an error as long as the error
3301 is something other than VRESTARTING. If there is such an error, all
3302 connections on that server should be marked for replacement.
3304 * Change the check "up" servers period from once per hour to once every
3305 ten minutes. This matches the behavior of the Unix cache manager.
3307 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3308 * Byte Range Locking support has been added. Architectural documentation
3309 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3310 made to enforce Windows mandatory locking model using locally managed
3311 byte range locks with full file locks obtained from the file server.
3312 Read-only volumes provide the user with an implied locking privilege.
3313 Read-write volumes require that a user be granted the locking privilege
3314 in order for a lock to be obtained. As most Windows applications use
3315 Shared file access modes, they require the ability to obtain locks in
3316 order to access files.
3318 * Implement error translation functions to convert Universal Error Codes
3319 output by modern servers to local errno and winsock error values.
3320 This fixes the error message problems using pioctl() calls from fs.exe
3321 and the AFS Shell Extension.
3323 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3324 * Allow the AFS Client Service to set the firewall configuration by
3325 retrying until after Windows determines the boot period is over.
3327 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3330 * Detect IP address changes from within the AFS Client Service for the
3331 purpose of checking down servers sooner
3333 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3334 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3335 user that there is a Kerberos key problem. Previously the user would
3336 receive no feedback when there was something wrong.
3338 * Fix the large integer library support to properly handle values greater
3341 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3342 * Updated CellServDB
3344 * Fixed file name pattern matching. Question marks can match nothing
3345 if found at the end of a component.
3347 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3348 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3350 * The MSI installer now has a new Product ID for each build. This
3351 is to allow for automated upgrades. The Upgrade ID is remaining
3354 * Volumes that were cached and then subsequently were salvaged,
3355 moved or otherwise taken temporarily offline (as opposed to being
3356 busy) would become unreachable. A distinction is now made between
3357 a volume being offline and a server being down.
3359 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3361 * Fixed several AFS Shell Extension issues:
3362 - Checking Server Status would only work if a specific cell was
3364 - If servers were down, the Server Down dialog would write to
3365 invalid memory and cause explorer.exe to crash
3366 - Default Push Buttons were not specified for many dialogs
3367 - The extension would be unloaded by the Explorer Shell and
3368 never be reloaded under some circumstances.
3370 * Fixed the AFS System Tray menu to disappear if another Window
3371 was selected before a menu item was selected
3373 * Added volume owner and group information to "fs examine" output
3375 * Added fs and registry support for enabling/disabling/clearing
3376 rx statistics gathering.
3378 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3380 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3381 GUIDs for each platform.
3383 * Documentation directory structure and default installation rules
3386 * New OpenAFS for Windows HTML Release Notes installed
3387 as part of the documentation set.
3389 * Add thread id to rx debugging messages
3391 * When all servers are offline, return a bad network path error
3392 immediately. This is necessary to allow Explorer to do the
3393 right thing and not hang. The background daemon thread will
3394 check the down servers every 30 seconds to see if they are
3397 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3399 * The AFSCache file has been modified to store both the
3400 serial number of the volume on which the file resides
3401 and the SID of the local machine. If either change,
3402 a new UUID is generated.
3404 * add "fs flushall" command that forces all data buffers
3407 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3409 * add support for '/' instead of '\' in pioctl() calls
3411 * Apply AFS Client Admins group protect to AFS Shell Extension
3413 * Add support for \\afs\<name> to most AFS Shell Extension
3414 commands except the symlink methods.
3416 * when installing the Microsoft Loopback Adapter, enable
3417 MS Client for Networks and prevent an install failure
3418 by not calling CoInitialize twice in the same thread.
3420 * reload cell vldb values from the CellServDB every two
3421 hours in case it changes
3423 * When updating cell information from DNS, be sure to set
3426 * Add support to allow use of \\AFS\<foo> where <foo> is
3427 either a mount point or symlink. As <foo> is normally
3428 treated as a share name, we transform it into \\AFS\all\<foo>
3431 * Init 'code' to prevent false errors when integrated logon
3432 is disabled and the service is not running
3434 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3436 * Removed trace log messages that were hampering performance.
3438 * Fixed a deadlock that was being triggered by editing Word
3439 documents stored within AFS with WinWord 2003.
3441 * Bit 3 of the TraceOption registry value is now used to set the
3442 default for "fs trace" in the non-Debug builds. The new default
3443 is off for release builds and on for debug builds.
3445 * Bit 2 of the TraceOption registry value can be used to turn on
3446 real-time output of debug log entries to the Windows Debug
3447 Output monitor interface. This data can be viewed with tools
3448 such as Sysinternal's DbgView.
3450 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3451 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3452 a resource we should not be returning status codes such
3453 as Network Busy or Remote Host Down because these errors
3454 will imply that the AFS Client Service is Busy or Down
3455 and that is simply not the case. Instead we will return
3456 Bad Network Path as the path is temporarily not available.
3458 Instead of returning Sharing Paused when there is a
3459 timeout error, return Timeout. Once again, the AFS Client
3460 Service is not paused.
3462 * afscreds.exe would display an Obtain Creds dialog when
3463 the expired credentials reminder was triggered even
3464 if there was no network path to the KDC. This is prevented
3465 by adding KDC probe logic to the reminder thread.
3467 * afscreds.exe would display expired tokens no differently
3468 than unexpired ones. This would make it difficult for the
3469 user to distiguish when the tokens were expired. For the
3470 English build added a new resource string "(expired) that
3471 is displayed instead of the expiration time.
3474 * "fs wscell" when executed in freelance mode will return the
3475 name of the cell configured in the registry. The root.afs
3476 volume is not loaded from this cell, but it is used for the
3477 default for aklog and integrated login.
3479 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3480 no longer work on \\AFS\all when freelance is being used
3481 unless the user is a member of the "AFS Client Admins" Windows
3484 * some more checks performed during persistent cache validation.
3487 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3488 down servers for the current cell not all cells.
3490 * If the scache LRU list becomes corrupted, the AFS Client Service
3491 might enter an infinite loop at startup. Detect the corruption
3492 and discard the cache contents in this case.
3494 * Fixed the Explorer Shell's remove mountpoint functionality.
3496 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
3497 obtain a mutex lock on a cm_scache_t object whose lock is held by
3498 a thread attempting to obtain the cm_aclLock.
3501 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
3502 misinterpretted as "csail.mit.edu" when the resolver
3503 is configured to append ".mit.edu" to failed DNS queries.
3505 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
3506 that can be used to force the preservation of user tokens upon logout.
3508 * Update the NSIS install scripts to use NSIS 2.07. This release adds
3509 recommended installation categories: AFS Client, AFS Administrator,
3510 AFS Server, AFS Development Kit. Each category includes a different
3511 default subset of the OpenAFS components.
3513 The OpenAFS logo is now associated with the NSIS uninstall entry
3514 in the Add/Remove Programs control panel.
3516 * The user name associated with AFS tokens when obtained with
3517 integrated login, the afs systray tool, or aklog will always
3518 include the full Kerberos 5 user principal name regardless of
3519 whether or not the cell is local to the realm.
3521 * Modify integrated login so that it does not enter an infinite
3522 loop if the service is not set to auto start.
3524 * Added asetkey.exe used to set a Kerberos 5 key for use by the
3527 * Added uninstall.exe to Wix installation
3529 * More modifications to algorithms used to wake sleeping threads.
3533 * Changes to the algorithms used to wake threads when they are
3534 put to sleep because of conflicts over buffer operations.
3537 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
3538 the process of producing AFS tokens. The tokens are stored within
3539 the AFS Client Service but the Kerberos 5 tickets have been discarded.
3540 New functionality has been added to temporarily copy the tickets to
3541 a file ccache during the short window between the NPLogonNotify and
3542 WinLogon Logon Event. The Logon Event handler starts a process as
3543 user, afscpcc.exe, that copies the tickets into the default ccache.
3544 The file ccache is then destroyed by the event handler.
3546 * During synchronization operations on stat cache entries, if threads
3547 are forced to sleep it was possible that the threads would not be
3548 woken under heavy load.
3550 * Enforce the delete on close flag specified by the NTCreateX()
3553 * A race condition in the rx library was fixed that resulted in
3554 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
3555 was cleared when threads were still waiting to use the
3556 connection. Also, fixed a potential case in which the
3557 connection could be destroyed while threads were waiting to use it.
3559 * The DNLC freelist has been seen to become corrupted with still
3560 active entries being stored on the freelist. Changes were made
3561 to perform a better job of cleaning entries before placing them
3562 onto the freelist; marking them with the correct magic value when
3563 purging the list; and allowing the list to be purged on startup
3564 if the cache validation fails.
3566 * Windows Crash Reporting does not get triggered for afsd_service.exe
3567 because it provides its own top level exception handler. This patch
3568 places an explicit call to ReportCrash() on platforms which support
3571 If you configure Crash Reporting via AD Policy to capture crash
3572 reports within the domain, then you will receive this output.
3574 * Generate MiniDumps independent of the Windows Crash Reporting.
3576 (a) If an exception occurs,a minidump will be generated at the path
3577 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
3578 can be set to any other type via use of the
3581 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
3583 (b) MiniDumps can be generated at any time using the "fs minidump"
3584 command. This allows you to generate a minidump without
3585 stopping the process or even requiring the use of a debugger
3588 * an RX connection which reports bad ticket now treats it the same way
3589 the expired ticket is treated. The ticket is removed from the
3590 connection and a new anonymous connection is established.
3592 * fs memdump now requires that the caller be in the AFS Client Admins
3595 * additional debugging information is provided in cm_Analyze when the
3596 error code is ignored.
3598 * Fix a regression introduced into the pioctl() client code which
3599 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
3600 MIT KFW is not installed. (1.3.82a)
3602 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
3605 * Fix a race condition in cm_EndCallbackGrantingCall() that
3606 could leave 'nrevp' pointing to freed memory. In the process,
3607 optimize the processing to ensure that the applications monitoring
3608 the status of the stat cache entry are only notified at most
3611 * Conditionalize access to the Explorer Shell AFS->Mount Point
3612 ->Remove menu item. Disable the item if the selected item(s)
3613 are not mount points.
3615 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
3616 Local Storage. This eliminates thread contention on the
3617 global RX free packet queue. This should improve performance
3618 on multi-CPU systems.
3620 * Fix the IP address filtering code to properly load/unload
3621 the IPHelper DLL so that Windows XP/2003 and beyond systems
3622 do not have to manually probe the registry.
3624 * CleanupACLEnt() was not being called consistently with
3625 the cm_scache_t object referred to by the back pointer
3626 mutex locked. This could in very rare conditions lead to
3627 an invalid memory access.
3629 * Added a script command to the msi installer "afs_replace"
3630 which can be used to replace one version of OpenAFS with
3631 another without requiring an uninstall and reinstall if
3632 the installed msi is no longer in its original location.
3634 * Apparently the problem with multi-domain forests with cross-
3635 realm trusts to non-Windows realms was not entirely solved.
3636 The authentication to the AFS SMB service failed because
3637 the wrong name was being used. Using ASU as an example,
3638 the authentication was being performed with the name
3639 "QAAD\user" (an account in the forest root) and not
3640 "user@ASU.EDU (the MIT Kerberos principal used to login with)
3642 The solution was to add an additional dependency on KFW
3643 in order or to be able to easily obtain the client principal
3644 name stored in the MSLSA ccache TGT. This information is
3645 used in two locations:
3647 - the pioctl() function
3649 - a new WinLogon Event Handler for the "logon" event.
3651 The pioctl function will now be able to use the correct
3652 name when calling WNetAddConnection2() and the "logon"
3653 event handler will now be able to call WNetAddConnection2().
3654 The hope is that the "logon" event handler will be called
3655 before the profile is loaded but I have not guarrantee
3659 * Fixed a locking error in cm_TryBulkStat() which had the
3660 potential of deadlocking the system for the length of time
3661 it takes to perform a bulk status fetch operation.
3663 * Replaced time conversion code (UnixTime <-> FILETIME) to
3664 be completely arithmetic instead of relying on a bizarre
3665 algorithm involving a variety of C RTL time functions.
3666 This has the side effect that UnixTime and FILETIME which
3667 are both stored in UTC are interpretted as UTC throughout
3668 the year. Windows will apply the same localization to AFS
3669 as it does to NTFS. Applications which rely on the ability
3670 to sync files between the two file systems will no longer
3671 see the timestamps of files in AFS change an hour relative
3672 to the files stored in NTFS or Windows based backup devices.
3674 * Fixed a invalid memory access under a bizarre circumstance.
3675 Windows will allow a physical mass media device to be
3676 installed via Plug N Play to the system and will assign it
3677 the lowest available drive letter. This is true even if the
3678 drive letter is currently assigned in the user session to
3679 a network device via NET USE (or its equivalent.) When this
3680 happens, queries sent to the CIFS server will contain invalid
3681 data. This invalid data was not being caught by the AFS
3682 Client Service and was resulting in a crash when Freelance
3685 * Fixed a reference count error when registering callbacks
3686 on a stat cache object if the callback was already registered
3688 * Add a case to cm_Analyze() in order to handle the case where
3689 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
3690 the server list for the volume is empty. In this case, force
3691 an update of the volume info and retry.
3693 * Insert a missing cm_EndCallbackGrantingCall() which could
3694 result in threads waiting for a callback to complete to never
3697 * In the persistent cache, there is a maximum number of volume
3698 entries. Allow volume entries to be reused if the maximum
3699 number have been allocated and their reference count is zero.
3701 * If we already have a dead virtual connection object, don't
3702 ignore the fact that additional dead vc objects must be taken
3705 * Removed a deadlock condition introduced in the summer of 2004.
3706 Do not hold mx locks around calls to RX functions including
3707 rx_NewCall(). That is what reference counts on the rx_connection
3710 * Fixed an initialization error in afslogon.dll which could
3711 result in random behavior including a failure to terminate
3712 the NPLogon function. (1.3.80b)
3714 * Fixed an error preventing the use of SMB authentication
3715 on Windows 2000. (1.3.80a)
3718 * Updated CellServDB to Public CellServDB 16 Mar 2005
3720 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
3721 misinterpretted as "csail.mit.edu" when the resolver
3722 is configured to append ".mit.edu" to failed DNS queries.
3724 * Fixed another case in which the client would replace connections
3725 to the host when it was not necessary. In this case the
3726 situation would occur if cryptall was on and the connections
3727 were unauthenticated due to lack of an appropriate token.
3729 * OpenAFS for Windows has failed to work at sites which are
3730 utilizing a cross-realm trust between an MIT/Heimdal realm
3731 and a multi-domain Windows forest when the workstation being
3732 accessed is not located in the root domain. This is caused
3733 by a bug in the workstation which was triggered after the
3734 introduction of Windows 2003 Server. When the bug is triggered,
3735 the workstation attempts to authenticate users to afsd_service.exe
3736 by contacting the Domain Controller instead of using the
3737 LSA loopback authentication mechanism.
3739 One of the reasons this bug occurs is because the workstation
3740 does not have a reliable way of knowing that the service whose
3741 netbios name is "AFS" is located on the workstation. This will
3742 be fixed starting in Longhorn Beta 1 by Microsoft. The
3743 "BackConnectionHostNames" registry value will be used to
3744 indicate that the authentications to that service name should
3745 be performed using the loopback authentication mechanism.
3747 In the meantime, when Logon Caching is enabled, we can force
3748 afsd_service.exe to authenticate using the logon cache before
3749 contacting the Domain Controller. This will work with both
3750 password and smart card based logons.
3752 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
3753 flag should not be cleared if a volume's server reference is
3754 marked as "offline". In the case where all of the volume's
3755 servers are either "down" or the volumes are "offline", we want
3756 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
3757 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
3758 should never occur in practice.
3760 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
3761 5 seconds, clear the server down and volume busy flags, and
3762 then force an update from the VLDB. This allows the client to
3763 update the location of a volume if the reason for it being
3764 marked offline is because it is being moved. Calls to
3765 cm_ConnByMServers() will be retried either until success or
3766 the RDRtimeout period is reached.
3768 * Correct the Power Management code behavior in response to
3769 Standby, Suspend, and Shutdown events. Instead of flushing the
3770 buffers associated with the mounted SMB submounts, simply write
3771 all buffers which are marked dirty.
3773 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
3774 provides a list of cells other than the default cell for which
3775 tokens should be obtained using the default Kerberos principal.
3776 This functionality is logon domain specific and is only available
3777 when using KFW for authentication.
3779 * Fixed FindNext Invalid Handle error caused by over agressive
3780 attempts at garbage collecting dirSearch entries when the
3781 dirSearch ID wraps from the maximum value to 1.
3783 * Add support for registry defined server preferences for VLDB and
3784 File servers. See registry.txt for details.
3786 * Increased default cache size to 96MB and default number of cache
3789 * Fixed refCount leaks related to directory lookups and pioctl calls.
3791 * Callbacks revoked during a race condition with an attempt to
3792 obtain the same callback no longer result in an inaccessible
3795 * New command line tool:
3797 afsdacl : Set or reset the DACL to allow starting or stopping
3798 the afsd service by any ordinary user.
3800 Usage : afsdacl [-set | -reset] [-show]
3801 -set : Sets the DACL
3802 -reset : Reset the DACL
3803 -show : Show current DACL (SDSF)
3805 * IP addresses are no longer queried once at startup. Instead IP
3806 addresses are obtained as needed. Loopback adapter addresses are
3807 no longer published to the server.
3809 * Pay attention to the MIT KFW registry configuration for automatic
3810 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
3812 * Fix (once again) case-sensitive comparisons which was apparently
3815 * Activate support for vos listvol -format
3817 * Population of the "AFS Client Admins" group failed on non-English
3818 versions of Windows because the name of the "Administrators" group
3819 is localized. Now we lookup the name of the group by using the
3820 Administrators Group Alias Relative ID.
3822 * Multiple cell token acquisition within afscreds.exe was broken in the
3823 case of Kerberos 5 cross realm authentication. Instead of contacting
3824 the KDC associated with the cell's realm, afscreds would obtain a
3825 token from the KDC of the user's realm. This would result in
3826 "invalid kvno errors" while authenticating to the AFS servers.
3827 Unauthenticated access would work. The symptoms would vary based upon
3828 whether or not the VLDB servers had been contacted using unauthenticated
3829 connections prior to the user obtaining tokens.
3831 * The list of ACL entries was becoming corrupted. It appears as if
3832 an ACL when it expires was not being appropriately cleaned up. In
3833 fact, it was left in the list of ACLs associated with the scache
3836 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
3837 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
3838 and "amd64_win64" for amd 64-bit processors.
3840 * When executing executables, Windows provides a mechanisms by which
3841 the normal search for DLLs can be bypassed. If a file foo.exe is
3842 being executed and there exists a file or directory "foo.exe.local"
3843 (the contents are ignored if a file), then Windows will search for
3844 DLLs first in the "foo.exe.local" directory and second in the
3845 directory in which the "foo.exe" file is located.
3847 Previous releases of OAFW would improperly return
3848 STATUS_NOT_A_DIRECTORY instead of the expected
3849 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
3850 to terminate the search for the DLL causing the execution of the
3851 application to fail.
3853 * Rework the reference counting of smb_vc_t objects. The references
3854 stored in the waiting locks were not counted. This could result in
3855 an assertion if the reference count drops to zero.
3857 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
3858 when running in freelance mode.
3860 * Added support for persistent caching of file contents, stat entries,
3861 volume data, cell data, and name lookup data. When the registry
3862 value, "NonPersistentCaching", is not defined or set to 0 the file
3863 specified by "CachePath" becomes a persistent cache file. The
3864 size of the cache file is computed at run time. It has a maximum
3865 size of approximately 1.2GB. The contents of the file will be
3866 validated according to the rules specified by the "ValidateCache"
3869 * Added support for UUIDs. UUIDs are kept for as long as the
3870 cache file is intact.
3872 * Added cmdebug.exe and support cache manager debugging callback
3873 interfaces in afsd_service.exe.
3875 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
3876 [-addrs] [-cache] [-help]
3877 Where: -long print all info
3878 -addrs print only host interfaces
3879 -cache print only cache configuration
3881 * Symlinks to \\AFS\[all\]... will now be treated the same as
3882 symlinks to /afs/.... However, please use /afs/... as otherwise
3883 the symlinks won't work on Unix.
3885 * Correct a problem with local CellID allocation for cells whose
3886 server list is obtained via DNS instead of a CellServDB file. If
3887 the DNS information expires the CellID assigned to the entry will
3888 be changed. This causes all of the cm_scache_t objects which refer
3889 to the old cellID number to become useless. Attempts to access
3890 files or directories with cm_scache_t objects using the old cellID
3891 will fail since the server list cannot be obtained.
3893 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
3894 be triggered if the volume referenced by the cm_scache_t object is
3897 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
3898 in addition to SERVICE_ACCEPT_STOP.
3900 * Move RPC shutdown until after the SMB and RX shutdown procedures
3901 complete. Block until RPC shutdown is complete.
3903 * Modify afslogon.dll (integrated logon) to wait for service if its
3904 state is START_PENDING. If the timeout period occurs, reset to
3905 the retry interval and not the sleep interval.
3907 * When renewing the server list for a cell obtained via DNS AFSDB
3908 records, the cm_cell_t entry must be removed from the list of all
3909 cells. Otherwise, the list of cells will be corrupted.
3911 * In the dcache and scache modules, use the cm_scache_t dataVersion
3912 instead of the cm_fakeDirVersion.
3914 * Synchronize fs.c with the unix version.
3916 * The variable used to determine whether a file or virtual memory
3917 mapped cache is used was not properly initialized to a default
3918 value. If the registry setting "NonPersistentCaching" was not
3919 set, the choice would be random. Properly initialized to be
3922 * The memory mapped view was never unmapped before closing the file
3923 at service shutdown. This is now properly cleaned up.
3925 * Default location of Cache file is now %TEMP%\AFSCache
3929 * A bug affecting new installations of 1.3.75/76 would result in
3930 the creating of incorrect mountpoints in the freelance root.afs
3931 volume for the default cell.
3933 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
3934 of "root.cell", you have been affected by the bug. To correct the
3935 problem, execute the following commands:
3937 fs rmmount \\afs\all\<cellname>
3938 fs rmmount \\afs\all\.<cellname>
3940 fs mkmount \\afs\all\<cellname> root.cell <cellname>
3941 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
3945 * A bug has been fixed in the auto-generation of Freelance root.afs
3946 symlinks which produced random entries in the \\AFS\all directory.
3948 * Support has been added for multi-homed servers
3951 * Added a new registry value, "StoreAnsiFilenames", which can be used
3952 to force the use of ANSI character sets instead of OEM Code Pages.
3953 This feature is useful when users require the ability to create
3954 filenames with 8-bit characters and need to access the files from
3955 both Latin-1 based Unix systems as well as from Windows.
3957 Activation of this feature will prevent access to files stored with
3958 8-bit OEM characters.
3960 * Shutdown all SMB threads in a synchronized manner when stopping the
3963 * There is currently a maximum cache size of 1.3GB. The limit is imposed
3964 by the largest contiguous block of unused memory within the 2GB process
3965 space which can be assigned to the memory mapped file. Unfortunately,
3966 when the executable digital signature verification code is activated
3967 Windows sees fit to further segment the process memory which in turn
3968 reduces the size of the maximum cache file to less then 800MB. If
3969 larger cache sizes are desired, a new registry value should be set:
3971 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
3973 Setting this value will disable the runtime verification of digital
3974 signatures on afsd_service.exe and the afs dlls which it loads. It
3975 will not disable the the version number check on those same files.
3976 The signature verification is not a security messure and is only meant
3977 to enhance the ability to afsd_service.exe to detect potential
3978 destablizing mixtures of DLLs from incompatible distributions.
3980 Added code to auto-disable the signature verification check if
3981 the desired cache size is greater then 700MB.
3983 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
3984 used to verify the validity of the afsd_service.exe binary
3985 as well as each of the AFS DLLs loaded by the service. Not only
3986 must the digital signature be valid but the signatures of the
3987 DLL must be signed by the same entity as the service.
3989 * Implement new functions: cm_freelanceMountPointExists and
3990 cm_freelanceSymlinkExists. Use them along with other validity
3991 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
3992 ensure that name collisions do not occur and that empty strings
3993 are not valid file names.
3995 A symlink may not have a name which would resolve to a valid
3996 cell name. Doing so would prevent access to the cell.
3998 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
3999 re-initialization code. The reference counts of the fake root.afs
4000 volume scache object(s) would become invalid when the mount point
4001 or symlink lists were altered.
4003 * Add registry entries to provide mappings from the afsdsbmt.ini
4004 to the new locations for applications which count on the use
4005 of the old Profile file APIs. These apps are likely to fail
4006 if the user does not have administrator privileges and the
4007 registry is locked down.
4009 * The afs_config.exe submounts dialog had two errors.
4010 First, attempts to remove entries failed because the registry
4011 key was being opened without KEY_WRITE privileges.
4012 Second, when editing a submount entry, changing the name
4013 would add a new key and leave the original one in place.
4014 Now the original submount will be removed if its name is
4017 * In recent months there have been several incidents in which
4018 users have experienced problems starting or accessing
4019 afsd_service.exe and after significant effort has been spent
4020 it has turned out that they have two versions of AFS on the
4021 machine or an inconsistent set of DLLs.
4023 Code has now been added to afsd_service.exe which will walk
4024 the list of modules loaded by afsd_service.exe and validate
4025 that the version of the AFS DLLs matches the version of the
4026 afsd_service.exe executable. If they do not match the service
4029 * When Freelance mode is enabled and there is no registry
4030 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
4031 will attempt to import the afs_freelance.ini file contents.
4032 If the file does not exist, it was creating a dummy file
4033 with a r/o and r/w entry for the default cell and then
4034 importing those values.
4036 This process has been changed. The temporary file is no
4037 longer created. Also, both the OpenAFS Client install
4038 directory as well as %WINDIR% are checked for previous
4039 afs_freelance.ini files.
4041 * Added support for VL_GetEntryByNameN(). Still need to add
4042 support for VL_GetEntryByNameU() for multi-homed support.
4044 * Fix a deadlock situation in afscreds.exe when canceling an
4045 auto-generated Obtain Tokens dialog
4048 * Fix the locking of objects during Directory Searches in the
4049 SMB/CIFS server. The failure to properly lock the reference
4050 counts was resulting in the premature freeing of smb_dirSearch_t
4051 objects while they were still in use by the SMB/CIFS client.
4052 This does not solve the "Invalid Handle" problem.
4054 * Fix Find Cell By Name pioctl call to return a valid cell
4055 name for the Freelance fake root.afs volume.
4056 "Freelance.Local.Root".
4058 * Fix the Explorer Shell Extension Symlinks->Add operation.
4059 The dialog template was missing and the link destination
4060 string was too short.
4062 * Add support for symlinks to Freelance root.afs volume
4063 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
4064 <number> = "<linkname>:<relative-path>."
4065 Use symlink.exe to create, list, or remove
4067 * Remove the fallback to the use of KFW's KRB4 library when obtaining
4068 tokens. We never obtain KRB4 tickets.
4070 * Fix AFS Client Configuration Control Panel to support new SysName
4073 * Fix a bug in afsd_service.exe which could result in the SysName
4074 not being read from the registry.
4077 * Add code to block the issuance of AFS tokens by aklog.exe or
4078 afscreds.exe when the Kerberos 5 principal name contains a dot.
4080 * Modify the IsAdmin() function to always treat the local SYSTEM
4081 account as an AFS client administrator. Affects fs.exe and
4084 * Modify the internal handling of Quota Exceeded errors
4086 * Upgrade all reference count fields in the Windows cache manager
4087 and the osi library to use unsigned long instead of signed short.
4088 A similar fix has been applied to the afs rpc (rx) library.
4090 * fix the Windows cache manager to prevent it from replacing the
4091 rx_connection object associated with the cm_conn_t object on each
4092 and every operation if "fs crypt" was set. This explains the
4093 dramatic performance difference when crypt is used vs clear.
4094 The problem: 'cryptall', a boolean flag indicating whether or not
4095 "fs crypt" is set, was being compared to the rx_connection
4096 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
4097 1 != 2 and therefore the rx_connection was always destroyed
4098 and replaced on each and every operation.
4100 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
4101 It is not safe for the cm_conn_t object to not be locked because
4102 rx_DestroyConnection might be called from another thread if:
4103 - the user's tokens have changed or expired
4104 - the crypt mode has changed
4106 This fix appears to have also taken care of the problems associated
4107 with Overlapped Writes resulting in Delayed Write errors.
4109 * fix NSIS installer's AdminGroup.exe to properly create and
4110 remove groups when given -create or -remove. The string comparison
4113 * fs sysname now accepts a list of sysname values
4115 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
4116 DWORD which when set to a non-zero value will cause error message
4117 text to be output to stderr from the pioctl() routine. Useful in
4118 debugging failures of fs.exe, tokens.exe, etc.
4120 * added a test to the power management code to only perform a
4121 flush operation if there is at least one network adapter which
4122 is not a loopback adapter.
4124 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
4125 "EnableKFW". This value will not be read if the key
4126 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
4127 value under that key does not.
4129 * provide mechanisms to force the use of krb524d for Kerberos 5
4130 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
4131 there is a new registry value "Use524" and for aklog.exe a new
4132 command line parameter "-m".
4134 * Fix the pattern matching algorithm to properly match patterns
4137 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
4138 which is used by both the original function and the new
4139 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
4140 creation of HardLinks in addition to Renaming. smb_Link()
4141 is a new function which creates HardLinks via cm_Link().
4142 cm_Link() is a new vnodeops function which creates links
4145 smb_ReceiveNTRename() does not support the File Copy and
4146 Move Cluster Information operations described in its interface.
4147 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
4149 * When opening files via symlinks, we should follow the symlinks
4150 until we reach the actual file stat cache entry. The stat cache
4151 entry of the file should then be stored in the FID instead of
4152 stat scache entry of the symlink.
4154 * return bad operation errors for all unimplemented functions
4155 even if we do not know the functions exist.
4157 * Log bad packets and unknown operation packets to the trace log
4159 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
4162 * Update list of known CIFS operations to include all those listed
4163 in CIFS-TR-1p00_FINAL.pdf.
4165 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
4166 to support the REG_EXPAND_SZ type.
4169 * A new Windows authorization group "AFS Client Admins" is now
4170 created and populated with the members of the "Administrators"
4171 group. The group is used to determine which accounts on the
4172 machine may be used to modify the AFS Client Configuration via
4173 the UI and command line tools. afs_config.exe, fs.exe,
4175 * Modify the WinLogon Logoff Event Handler to query NT4 domain
4176 controllers for the remote profile path if Active Directory
4177 services are not available.
4179 * Fix aklog.exe to not add the AFS ID to the username
4181 * PTS registration of new users to foreign cells has been added to
4184 * The cm_Daemon thread is used to perform checks for
4185 down servers, up servers, volumes, callback expirations,
4186 lock maintenance and token expiration. Due to a gaff in
4187 larger integer division the thread never performed any
4188 work. Instead the current time computation would always
4189 be less then the trigger times. This had an adverse affect
4190 on the client's ability to maintain communication with servers,
4191 keep volumes up to date, and flush user tokens and acls
4192 when they have expired. This was broken when the 1.3 branch
4193 was modified to support VC7 which no longer included
4196 * An initialization problem with the Freelance code was
4197 detected while fixing the callbackRequest. The cm_rootSCachep
4198 object is obtained during afsd_InitDaemons() but the callback
4199 information is incomplete. The callback information will not
4200 be obtained until cm_MergeStatus is called from within
4201 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
4202 test for the conditions under which the callback information
4205 * Reports have been filed indicating that callbacks were
4206 being lost. An examination of the code indicated that the
4207 cm_server_t objects were not being properly reference
4208 counted by the cm_scache_t and cm_callbackRequest_t objects.
4209 In particular, the cm_server_t objects may have been freed
4210 from beneath the cm_conn_t objects.
4212 All of the reference counting is now done via the functions:
4217 this improves the ability to track the referrals.
4219 Each cm_BeginCallbackGranting Call now allocates a reference
4220 to the cm_server_t. The cm_EndCallbackGrantingCall either
4221 frees the reference or transfers it to the cm_scache_t
4222 cbServerp field. These are then appropriately tracked
4223 through the cm_Analyze call.
4225 * Ensure that the dnlc hash table is the same size as the
4226 dir name hash table (as per original author's note).
4227 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
4228 for compatibility with 64-bit systems.
4230 * fix smb_ApplyV3DirListPatches to properly apply the hidden
4231 attribute to dotfiles when the infoLevel < 0x101 and
4232 cm_SyncOp has failed.
4234 * Fix the Freelance registry initialization code. There
4235 was a possibility that some systems could end up with
4236 garbage in the registry during a clean install.
4239 * file and directory names beginning with "." will now be given the
4240 hidden attribute when the volume access is anonymous. this matches
4241 the behavior when the volume access is via an authenticated user.
4243 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
4244 key. When a change occurs mark the root.afs data as invalid and
4245 for it to be reloaded on the next access. This allows administrators
4246 to modify the mount point list without restarting the service.
4248 The freelance client used to provide a fake modification time for
4249 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
4250 Added code to extract the last modification time of the Freelance
4251 registry key and use that instead. The time now represents the
4252 most recent mount point change.
4254 * PTS registration of new users to foreign cells has been added to
4257 * Additional Cache Control and Credential Manager options have been
4258 added to the WiX installer. See deployment guide for details.
4260 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4262 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4265 * The loopback installation code may have had a problem updating the
4266 %ETC%\HOSTS file which could have resulted in a premature failure.
4267 Work around code has been added for the case where the file cannot
4270 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4271 because Windows sends 64K blocks when using overlapped writes.
4273 * The default number of server threads was increased from 4 to 25 to
4274 better handle overlapped writes.
4276 * The "AfscredsShortcutParams" registry value was not being properly
4277 loaded by afscreds.exe. Therefore, the default value was always being
4278 used instead of the value set by the installer.
4280 * Windows XP provides downgrade attack detection to prevent an attacker
4281 from being able to force the use of NTLM simply by disrupting
4282 communication with the KDC. This attack cannot exist between the
4283 Windows CIFS client and the AFS Client Service. Therefore, when a
4284 downgrade has been detected the afs pioctl library will force the
4285 establishment of a new CIFS connection using NTLM.
4287 * A locking error was discovered surrounding all references to volume
4288 server lists within the cm_cell.c source file.
4290 * The logged into Windows username was incorrect on Terminal Server
4293 * A new registry value "NonPersistentCaching" was added to the service
4294 parameters key. When set to a non-zero value, the afs cache is stored
4295 in the Windows paging file. There are two limitations to choosing
4297 1. when persistent caching is implemented it won't work with
4298 this flag set since there will be nothing to persist.
4299 2. with this flag set the initial paging allocation cannot be
4300 changed while the service is running
4302 * An initialization bug was discovered in aklog.exe which affected users
4303 who have a domain name for their afs servers which could not be mapped
4307 * afs_config.exe now validates cell names against DNS in addition
4308 to the CellServDB file.
4310 * In order to allow the freelance client to connect to a volume with ID
4311 equal to 1 on the default cell we changed the fake root.afs volume ID
4312 once again. This time we choose 0xFFFFFFFF. In addition, we change
4313 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4314 It will now be impossible for a volume ID to match that of another
4315 cell unless the client is connected to 0xFFFFFFFD cells. That should
4316 be enough room for growth.
4318 * Fix "fs mkmount" command to work with UNC paths and when
4319 started from non-AFS drives. It is now possible to create a mount
4320 point in the freelance fake root.afs volume with the command
4322 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4326 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4327 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4329 * The algorithm used to re-attempt access to the servers associated with
4330 a volume has been altered to properly address the case in which all
4331 servers have been marked down. The previous algorithm did not reset
4332 the server's down flags so the servers were never actually retried.
4333 This caused a problem with active volumes if the network connectivity
4334 was lost as could be the case with a network cable removal, wireless
4335 drop, or laptop hibernation. With the fix volume access is restored
4336 almost instantenously when network connectivity becomes available.
4338 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4339 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4340 listing of AFS submounts and freelance mount points. Support for
4341 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4342 significantly enhances the behavior of AFS volumes within the Explorer
4343 Shell. For instance, "AFS" now shows up as server in the Explorer
4344 with each submount or freelance mount point visible as a share.
4345 The right click menu in each folder now works with full functionality
4346 on a consistent basis.
4348 * The network provider can be configured to have different behavior
4349 depending on the domain that the user logs into. These settings are
4350 only relevant when using integrated login. A domain refers to an
4351 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4352 local machine (i.e. local account logins). The domain name that is
4353 used for selecting the domain would be the domain that is passed into
4354 the NPLogonNotify function of the network provider. (see registry.txt
4357 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4358 "Authentication Cell" which may be used to specify a default
4359 authentication cell for afscreds.exe which is different from
4360 the default cell for the AFS Client Service daemon.
4362 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4363 afslogon.dll moved to %WINDIR%\System32\.
4364 New registry entries added to register the dll for Winlogon events.
4366 The logoff event will now force a call to ktc_ForgetAllTokens()
4367 using the context of the user being logged off as long as the
4368 user's profile is not loaded from within AFS. If the profile
4369 was loaded from AFS we can't release the tokens since the Logoff
4370 event is triggered prior to the profile being written back to
4371 the its source location. This is now performed in an XP SP2
4374 * Windows XP SP2 Internet Connection Firewall interoperability
4377 * The %WINDIR%\afsdsbmt.ini contains four sections:
4378 Submounts, Drive Mappings, Active Maps and CSC Policies.
4379 The Submounts and CSC policies are now stored in the registry under
4380 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4381 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4382 The Drive Mappings and Active Maps are stored in the registry under
4383 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4384 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4386 There is no automatic migration of this data as it would be impossible
4387 to consistently migrate data to user profiles which may not be active
4388 when the machine is updated.
4390 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4391 fake root.afs volume. For the same reasons as for the cellservdb file,
4392 this information should not be in %WINDIR%. This information is now
4393 kept under the registry key
4394 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4396 The data from the afs_freelance.ini file will be automatically
4397 migrated to the registry on first execution of afsd_service.exe
4399 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4400 troublesome for several reasons. One, it is confusing for those who
4401 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4402 Two, this file is not a Windows Profile formatted file. Three,
4403 applications should not be reading or writing to %WINDIR%. It causes
4404 problems for Windows Terminal Server.
4406 The new location of CellServDB will be the OpenAFS Client install
4407 directory which is by default C:\Program Files\OpenAFS\Client and can
4408 be determined by querying the registry for
4409 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4411 The existing afsdcell.ini will be migrated by the NSIS installer.
4412 The Wix installer must still be updated to do the same.
4414 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4415 High Security mode; and to add Terminal Services compatibility registry
4416 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4417 configuration files in %WINDIR%.
4419 * Add support for authenticated SMB connections. This will remove
4420 the need for high security mode in most situations. Both NTLM
4421 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4422 only NTLM can be used even though Kerberos is now supported. The
4423 reason is that it is not possible to construct a service principal
4424 which is unique to each individual machine.
4426 SMB Extended Auth does not work on XP SP2 unless one of two registry
4427 modifications are made:
4429 (1) To disable the check for matching host names on loopback connections
4430 set this key. This does not require a reboot:
4432 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4433 "DisableLoopbackCheck"=dword:00000001
4435 (2) To add the AFS SMB/CIFS service name to an approved list. This
4436 does require a reboot:
4438 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4439 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4441 afsd_service.exe will automatically add the current Netbios Name
4442 to the BackConnectionHostNames list and then temporarily disable
4443 the loopback check for one cycle of startup/shutdown of the service.
4444 We assume most folks do not start/stop without a reboot so this
4445 will be adequate in most cases.
4447 * Fix security hole in afslogon.dll which allowed passwords to be
4448 sent in clear text to the KDC in a misformed principal name.
4450 * Fix cm_GetCell() to properly handle expired dns entries
4453 * If Freelance mode is active and the afs_freelance.ini
4454 file does not exist, do not create an empty file.
4455 Instead create a file containing ro and rw mountpoints
4456 to the default cell using the standard conventions.
4458 * Modify the Freelance support to handle the ability
4459 to create rw mount points in the fake root.afs volume.
4461 * Changed the RPC mechanism used for token setting from
4462 named pipes to local. Use of named pipes can be restored
4463 by setting the environment variable AFS_RPC_PROTSEQ to
4466 Named pipes were required when a Windows 9x system was
4467 using a NT system in gateway mode which is incompatible
4468 with our use of local loopback adapters.
4470 * In afscreds.exe, if a username of the form user@REALM is
4471 specified and no password is specified, do not perform a
4472 kinit operation. Only perform the aklog functionality.
4474 * Add a new registry value which allows the number of processors
4475 on which afsd_service.exe executes to be restricted. Valid
4476 values are 1..numOfProcessors
4478 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4482 * A second MSI based installer option is now available.
4484 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4485 request tickets for user/instance@REALM instead of just user@REALM
4487 * Modify the Power Management Notify routine to wait for the Hard Dead
4488 timeout period instead of a fixed 19 seconds. With the longer timeout
4489 periods Hibernation and Standby could never succeed when network
4490 connectivity is not available.
4492 * The following fs.exe commands are now restricted to Administrator:
4493 - checkservers with a non-zero timer value
4496 - sysname with a new sysname list
4505 setting the default sysname for a machine should be done via the
4506 registry and not via "fs sysname".
4508 * NSIS installer adds options to install Debugging Symbols
4509 and the Microsoft Loopback Adapter; the user is now also
4510 given the ability to select the afscreds.exe startup options.
4512 * Build system modified to generate symbols for FREE (aka RELEASE)
4513 builds as well as CHECKED (aka DEBUG) builds
4515 * Sites which have a volume ID of 0x20000001 assigned to their
4516 root.cell volumes have been experiencing problems with accessing
4517 the root.cell volume of their cell when Freelance mode has been
4518 active. This was because 0x20000001 was assigned to the fake
4519 root.afs volume created by freelance. The fake volume id is
4520 now set to 0x00000001 to prevent conflicts.
4522 * The timeout logic in the AFS Client Service has been wrong
4523 for sometime. It is based on two different assumptions.
4524 First, the SMB client timeout is a fix value as was the case
4525 with OS/2 Lan Manager. This assumption is incorrect. The
4526 SMB timeout in Windows is a dynamic value computed based upon
4527 a fixed minimum timeout to which is added time based upon the
4528 size of the request and the performance characteristics of
4529 the connection. Second, it is the responsibility of the
4530 SMB Server to enforce the timeout requirements of the client.
4531 This is untrue. The SMB Server cannot be expected to know
4532 the requirements of the client. More importantly, if the
4533 SMB server uses the SMB client timeout as a value to restrict
4534 its behavior as an RX client, the performance characteristics
4535 of the local SMB session would be used to prematurely terminate
4536 WAN connections with significantly different performance
4539 The timeout logic has therefore been modified in the following
4541 . the Lan Manager Workstation (SMB) Session Timeout is used only
4542 as a basis for configuring the Connection Dead Timeout
4543 and Hard Dead Timeout values. The Connection Dead Timeout
4544 must be at least 15 seconds longer than the SMB Timeout
4545 and the Hard Dead Timeout must be at least double the
4546 Connection Dead Timeout.
4547 . New registry entries have been added to allow the Connection
4548 Dead Timeout and Hard Dead Timeout values independent of the
4549 Lan Manager Workstation Session Timeout
4550 . The test to enforce the SMB Client Timeout has been removed.
4552 One of the side-effects of removing the enforcement of the SMB
4553 Client Timeout is that regardless of whether or not the SMB client
4554 is available to receive the response (and how would the SMB server
4555 know) the RX protocol response can be used to update the AFS
4556 Client Service state for ready access by future SMB client
4559 This should be the end of the "Server paused or restarting messages"
4561 * Add support for arbitrary UNC paths to the pioctl() support.
4562 This enables the fs commands as well as the AFS Shell Extension
4563 to work correctly when UNC paths are being used.
4565 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
4566 DNS if the cell configuration cannot be determined via CellServDB
4568 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
4569 are ever set more than once at a time
4571 * Fix the management of lists of cm_cell_t structures when using
4572 DNS to lookup cell information. The previous code would fail to
4573 reuse the same cellID for a cell if DNS was used more than once
4574 for a given cell name. When the ttl expired, a single cm_cell_t
4575 could be inserted into the cm_allCellsp list more than once
4576 producing a loop. In addition, the vlServerp list belonging to
4577 the cell was not freed resulting in improper refCounting of the
4580 * Add DNS support to cm_IoctlNewCell() which previous only examined
4583 * Add cm_FreeServer() function and call it from cm_FreeServerList()
4584 to properly garbage collect cm_server_t objects
4586 * Add numVCs variable to smb.c to track the number of smb_vc_t
4587 objects created and use it to initialize the vcID field which
4588 previously was set to 0 in all objects resulting in FindByID
4591 * Fixed DNS lookups to work consistently throughout the OpenAFS
4592 product instead of just from within the afsd_service.exe
4594 * Added a runtime check to ensure that AFS Client Service SMB
4595 Server is accepting connections before attempting to mount
4598 * Read IP addresses for volume servers out of the CellServDB
4599 file if gethostbyname() on the hostname fails.
4601 * Fix getcellconfig() to populate both the Hostnames as well
4602 as IP addresses when loading cell data via DNS
4604 * Increase the Connection Dead Time to 50 from 20 seconds
4605 Increase the Hard Dead Time to 120 from 40 seconds
4606 (matches the Unix values)
4608 * Fixed an assertion validating the number of allocated NCBs
4610 * Fixed the build environment to consistently build for
4611 Windows 2000 and above. (APPVER = 5.0)
4613 * Fixed rx_debug to properly validate the receipt in incoming
4614 data with select() and recvfrom(). Do not copy data out of
4615 the socket buffer unless success is indicated.
4618 * afsd_service.exe will now display a message box to the
4619 desktop when it terminates due to an IP Address Change.
4621 * installer no longer deletes AFS Server configuration data
4624 * installer generates a warning dialog if the RPC service
4625 is not properly configured
4627 * installer compressed with lzma instead of bzip2
4629 * afsd_service.exe shutdown crash solved once and for all
4631 * reference counting of smb_vc_t data structures improved
4633 * name space collision of smb_fid_t event objects corrected
4635 * the output of "fs memdump" is now written to
4636 %WINDIR%\TEMP\afsd_alloc.log
4638 * the file TaAfsApp_1033.dll is now properly installed allowing
4639 the User Manager to start
4641 * a new algorithm is used for computing filename pattern matches
4643 * afscreds.exe now accepts user names containing instance
4646 * Fix the Directory Name Lookup Cache to be case-sensitive.
4647 This is crucial in environments in which a Windows client
4648 is accessing a directory with more than one filename that
4649 differs only by case. If the directory contains "FOO"
4650 and "Foo". You want "DEL Foo" to delete the correct one.
4651 We still have a problem in that "DEL foo" will delete a
4652 random filename. This will be addressed in a future release.
4654 * Fix afscreds.exe -M option (renewMaps) to work when High
4655 Security mode is off. Also, remember to disable the ActiveMap
4656 flag in afsdsbmt.ini when a drive mapping is removed.
4658 * Updates to NSIS installer script. AFS Server configuration
4659 data will not be destroyed on un-install or re-install.
4660 Use a better compression algorithm.
4662 * afslogon.dll now uses KFW to obtain tokens when available
4664 * afslogon.dll when given an all uppercase username will
4665 attempt to authenticate with both the uppercase name
4666 and an all lowercase variation
4668 * DST modification removed. The fix appears to make things
4669 worse after a reboot of the machine.
4671 * fs.exe: added "cscpolicy" which is used to
4672 change client side caching policy for AFS shares
4674 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
4675 [-documents] [-disable] [-help]
4677 * Several uninitialized variables have been initialized
4679 * It is now possible to obtain tokens using cross realm
4680 Kerberos within afscreds.exe:
4682 user: jaltman@ATHENA.MIT.EDU
4684 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
4685 will will in turn be used to obtain afs@DEMENTIA.ORG.
4686 The resulting token will be stored with the display name
4687 jaltman@ATHENA.MIT.EDU@dementia.org
4689 * aklog.exe has been added to the client
4691 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
4692 [[-p | -path] pathname]
4696 -d gives debugging information.
4697 krb_realm is the kerberos realm of a cell.
4698 pathname is the name of a directory to which you wish to authenticate.
4699 -noprdb means don't try to determine AFS ID.
4700 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
4701 (default is Kerberos V)
4702 No commandline arguments means authenticate to the local cell.
4705 * All of the resource files have been restructured to adhere to
4706 a set of rules IBM implemented for loading string resources.
4707 These rules had either been forgotten or were not discovered
4708 by folks working on the OpenAFS sources. The end result was
4709 memory corruption. This is primary item which was preventing
4710 the AFS Server from working.
4712 * Increased the size of the maximum ticket size stored in a token
4713 from 344 bytes to 12,000. Increased the buffers used to convey
4714 messages between the pioctl() caller and the SMB Server from
4715 1000 bytes to 12,512. The code appeared to have been writing
4716 above the top of the stack by quite a few number of bytes.
4717 (The increased ticket size is necessary for the next item.)
4719 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
4720 Instead the raw Kerberos 5 ticket is used in its entirety. This
4721 is extremely important as it allows us to use pure Kerberos 5 KDCs
4722 as the source of the AFS authentication. The use of up to 12,000 byte
4723 tickets will allow tickets produced by all versions of Microsoft
4724 Active Directory to be used.
4725 - create a user account.
4726 - designate it DES only
4728 - specify its UPN to be "afs@realm"
4729 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
4731 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
4732 is special when using long file names. (you must use "*.*" and not "*")
4733 Instead only enforce it when performing 8dot3 searches.
4735 * Fixed the DST problem with creation times being set one hour ahead
4737 * Fixed the problem when using \\afs\cell-alias. For example,
4738 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
4739 for the alias name; instead simply expand the name. One of the
4740 symptoms of this problem was a loss of acquired tokens.
4742 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
4743 of strings. (Only English strings provided.)
4745 * Fixed the installer to properly replace in use files.
4747 * Fixed the build system to cleanup generated component version files
4749 * The release build compiled with MSVC 6.0 compiler to avoid the
4750 afsd_service.exe shutdown crash. This does not solve the problem
4751 but simply avoids it for the time being.
4755 * fix afslogon.dll to not corrupt memory when High Security mode
4758 * fix afsd_service.exe to not attempt to restore the stack when
4759 an exception occurs. (not safe in multi-threaded programs)
4761 * fix uninstaller to properly remove the CRT and MFC DLLs
4763 * remove a Message Box from afscreds.exe when getcellconfig()
4764 fails on a kerberos realm which is not a cell
4766 The following is a list of changes to the OpenAFS for Window client
4769 * "fs setserverprefs" will leave afsd service deadlocked
4771 * "vos listaddrs" will core dump
4773 * installer sets the appropriate keys to support Integrated Logon
4775 * installer disables the "Find Lana by Name" functionality as it
4776 was causing headaches for many users
4778 * fix the intermittent crash of the power management thread when
4779 shutting down the AFS Client Service
4781 * optimizes the obtain drive mount list functionality which is
4782 executed every time the mount tab in afscreds.exe and afs_config.exe
4783 are refreshed. (this happens a lot)
4785 * fix the service shutdown logic. add the STOP_PENDING state
4786 and do not accept additional service events after we declare
4789 The following is a list of changes to the OpenAFS for Window client
4792 * flexelint was run against the source tree and hundreds (perhaps
4793 thousands) of corrections were applied to ensure prototypes
4794 were in use; types were used consistently; variables were
4795 initialized; unused variables were removed; etc.
4797 * A wide variety of instrumentation was added including the
4798 ability to produce a stack trace from within afsd_service.exe
4801 * Dynamic configuration of the RDRtimeout value based upon the
4802 LanMan Workstation Session Timeout
4804 * The mount root no longer needs to be called "/afs". This
4805 is now set by a registry value "MountRoot" within the key
4806 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4808 * The cell list is now only read out of afsdcell.ini when the
4809 file changes instead of each time a cell is resolved.
4811 * Thread synchronization was added to cm_server.c and ktc_nt.c
4813 * All calls to GlobalAlloc()/GlobalFree() were replaced with
4814 calloc()/free(). The Global functions were needed on Windows 3.x
4815 but have caused a variety of problems on the Win32 platforms.
4816 Avoiding them is highly recommended by several Microsoft
4817 Knowledgebase articles
4819 * Support for Symbolic Links added to the AFS Shell Extension
4821 * Added a registry value "OverlayEnabled" to determine if
4822 Shell Extension Overlays should be enabled.
4823 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
4825 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
4826 separate trees for checked and free builds. Build system supports
4827 a custom directory src\WINNT\extra which can be used as a grafting
4828 location of organization specific additions to the build tree.
4830 * New installer built using NSIS 2.0.
4832 * Named all kernel objects in order to allow them to be monitored
4833 with tools such as SysInternals' ProcExp.exe.
4835 * Introduced new EventLog framework for AFSD
4837 * Introduced Power Management interface to AFSD for Standby and
4838 Hibernate modes to allow cache to be flushed prior to network
4841 * Utilize Win32 DNSQuery API instead of internal routines. This
4842 allows DNS SRV queries to be sent to all current domain name
4843 servers. Not just one specified in an INI file. DNS is now
4846 * "NetbiosName" registry value may be used to specify a fixed
4847 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
4848 when the loopback adapter is in use. If you need to use the
4849 old notation with a loopback adapter installed specify a registry
4852 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
4854 * Refactor all modules which depend on LAN Adapter and NetbiosName
4855 determination in a new library: lanahelper.lib. This allows for
4856 consistent behavior throughout the product.
4858 * Move the afsd.log and afsd_init.log files to the directory specified
4859 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
4860 for services. Added the Date to the log entries.
4862 * New registry value "RxMaxMTU" used to limit the size of the RX
4863 packets sent by the AFS Client Service to the Server. In order
4864 to enable OpenAFS to work across the Cisco IPSec VPN the packet
4865 size must be restricted to 1264 or smaller. The latest NSIS
4866 installer sets a value of 1260 by default.
4868 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
4869 packets. This is not needed in order to work across the Cisco
4870 VPN but might be needed for other network environments. This
4871 value is not set by the NSIS installer.
4873 * New registry value "HideDotFiles" is used to apply the Hidden
4874 attribute to files whose names begin with a '.'. This value
4875 is set by the NSIS installer.
4877 * New registry value "MaxMpxRequests" allows the maximum number
4878 of multiplexed sessions to be configured at run time. This
4879 value is not set by the NSIS installer. The default value is
4882 * New registry value "MaxVCPerServer" allows the maxmimum number
4883 of VCs per server to be configured at run time. This value is
4884 not set by the NSIS installer. The default value is 100.
4886 * New registry value "AllSubmount" allows the "all" submount to
4887 be disabled by setting its value to 0x00.
4889 * Allow cells names to be valid mount points
4890 \\<netbiosName>\<cellname>
4892 * Store the active state of drive mappings in order for afscreds.exe
4893 to restore them upon startup
4895 * Add exception handling to generate a Stack Trace to the afsd_init.log
4896 file if one happens to occur.
4898 * Add lots of logging to help detect the cause of invalid SMB packets
4900 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
4901 conversion of Kerberos 5 "afs" service tickets. Supports auto-
4902 renewal of expiring tokens as long as afscreds.exe is running.
4904 * New afscreds.exe command line options:
4906 -M = renew drive maps
4907 -N = ip address change detection
4910 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4911 determines whether or not MIT Kerberos for Windows should be used
4912 to obtain tokens via Kerberos 5 tickets.
4914 * New registry value "AfscredsShortcutParams" in
4915 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4916 determines the command line parameters to be specified when "fixing"
4917 the AFS Shortcut in the user's startup folder.
4919 * The "ShowTrayIcon" registry value has been moved from
4920 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4921 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
4923 * The <cell name> registry values used to store the token expiration
4924 reminders have been moved from
4925 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
4926 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
4928 * Obtain the Logon User Name from the Explorer key when available
4930 * new text document doc\txt\winnotes\registry.txt lists all registry
4931 values used by OpenAFS (excluding the AFS Server)
4933 * BUG: rx_securityClass objects were not properly reference
4934 counted and were never freed.
4936 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
4937 would be generated. The existence of a server does not imply
4938 that it is not down. If all of the servers for a cell are down
4939 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
4942 * BUG: the directory name lookup cache failed to free the entries
4943 in the cache when the name cache entries cycled. The entries
4944 in the cache would become dereferenced without being freed.
4946 * BUG: fs setserverprefs could be executed without Administrator
4949 * BUG: the number of allocated NCB objects (100) exceeded the number
4950 which could actually be waited upon by the kernel (64). Any objects
4951 which were utilized above the limit could never have event completions
4954 * BUG: smb_username_t objects were not being reference counted and
4955 were not properly freed.
4957 * BUG: smb_tid_t objects could under unusual circumstances be freed
4958 before they were no longer referenced.
4960 * BUG: smb_fid_t object pointer were frequently used even when
4961 their value could be NULL. They were not properly released and
4962 therefore they were never freed.
4964 * BUG: smb_packet_t data structures were not completely initialized
4967 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
4968 Denied" because that causes the Explorer Shell to try again until
4969 access is obtained. Instead return "Remote Resources" which allows
4970 the shell to move on and treat the error as transient.
4972 * BUG: when initializing the NCBreturns structure, separate Event objects
4973 were created for each NCB although a single Event object was supposed
4974 to be shared by all.
4976 * BUG: smb_dirSearch_t objects were not being properly referenced counted
4979 * BUG: smb_tran2Packet_t objects were not being properly referenced
4982 * BUG: directory path creation did not handle the case of multiple
4983 directories requiring creation in one attempt
4985 * BUG: SMB requests which required an Extended Response were ignored.
4986 This prevented some files from being written to AFS volumes.
4988 * BUG: character strings were being freed even after they were
4989 inserted into in use data structures
4991 * BUG: inconsistent usernames were used when High Security mode was
4992 enabled. (there is still much to do in this area)
4994 * BUG: pioctl() calls which require out of band RPC operations were
4995 susceptible to race conditions when performed by multiple processes
4997 * BUG: memory allocation and deallocation crossed instances of the
4998 C Runtime Library producing memory leakage and corruption in
4999 afscreds and the client configurator.