1 Since 1.5.9904 (1.6pre4)
2 * Fix caching of non-existent volumes. The test to
3 trigger an immediate CM_ERROR_NOSUCHVOLUME in
4 cm_UpdateVolumeLocation() was backwards.
6 * Prevent the background daemon from checking the
7 status of non-existent volumes. cm_CheckOfflineVolumes()
8 should skip volume groups with the CM_VOLUMEFLAG_NOEXIST
11 * The afskfw library should return an error immediately
12 if the krb5_32.dll library cannot be loaded. Affects
13 afslogon.dll and afscreds.exe.
15 * No longer depend on leashw32.dll in afskfw library.
17 * NPLogonNotify must provide the user password in all
18 calls to KFW_AFS_get_cred(). It cannot count on a
19 credential cache being preserved between calls. Permits
20 tokens to be acquired for all cells listed in the
21 TheseCells registry value for a domain.
23 * Improve the trace logging from NPLogonNotify().
25 * Avoid a race when writing the cm_scache_t mountPointString
26 when acquiring mount point or symlink target data via
27 cm_GetData(). The race could result in bogus target
30 * Permit the use of des-cbc-md5 and des-cbc-md4 enctypes
31 as DES keys in asetkey.exe.
33 Since 1.5.9903 (1.6pre3)
34 * Fix automatic addition of Freelance cell mount points
37 * Avoid recursive offline volume checks which could cause
38 a crash due to stack exhaustion.
40 Since 1.5.9902 (1.6pre2)
41 * rx calls issued on a busy call channel are now specially
42 handled and logged. All calls are automatically retried.
44 * The vos support for multi-homed servers was incorrectly
45 implemented. The changes have been reverted.
47 * Avoid wasting cm_buf_t objects and associated data buffers
48 for obtaining mountpoint and symlink target strings. Instead
49 read them directly into the cm_scache_t mountPointTarget
52 * Optimize the evaluation of mountpoint and symlink targets
53 by skipping the FetchStatus and a full round-trip if
54 the cache manager does not have valid status information
57 * Fix the abstraction of cm_FreelanceAddMount() by passing
58 in volume names that do not include the trailing dot.
61 [there was no 1.6pre1 for Windows]
63 * vos commands now manipulate servers by UUID and can
64 recognize multi-homed servers.
66 * afs_config will not longer set the Tray Icon State
67 in the registry if the checkbox is not present in
68 the dialog. [RT 128591]
70 * Heimdal's roken utility library has been added
71 as \Program Files\Common\afsroken.dll
73 * When probing servers to determine if they are up or down
74 no longer issue an RXAFS_GetTime RPC to servers that indicated
75 that they do not understand the RXAFS_GetCapabilities RPC.
76 Since they responded it is known that they are up.
78 * AFS Explorer Shell Extension now works from folder
79 backgrounds. Overlays for mount points and symlinks
80 are present in the dll, but are not registered at present
83 * Do not use RankServerInterval registry value as the value for
84 PerformanceTuningInterval.
86 * CellServDB updated to 13 Dec 2010 release from grand.central.org
88 * Add "fs chmod" command and display current mode as part
89 of "fs examine" output.
91 * When the data version of a mountpoint or symlink changes,
92 the target string in the cm_scache_t object must be cleared.
93 Otherwise, the new target will not be queried.
95 * "fs checkservers" now includes vldb servers in the output
96 and only lists multi-homed servers once. A multi-homed
97 server that has at least one up interface is no longer
98 considered to be down.
100 * When asynchronously storing dirty data buffers to the
101 file server ensure that (a) the cm_scache_t object and
102 the cm_buf_t object are for the same File ID so that
103 locking and signalling work properly; and (b) if the
104 FID no longer exists on the file server, do not panic,
105 just discard the buffer.
107 * When processing VNOVOL, VMOVED and VOFFLINE errors perform
108 server comparisons by UUID or address and not simply by
109 cm_server_t pointer. Otherwise, server failover may not
112 * Do not preserve status information for cm_scache_t objects
113 when the issuing server is multi-homed.
115 * Giving up all callbacks when shutting down or suspending
116 the machine is now significantly faster due to the use
117 of an rx_multi implementation. (This functionality is
118 still off by default and must be activated by a registry
121 * Race conditions were possible when updating the state
122 of the cm_volume_t flags and when moving the volumes
123 within the least recently used list.
125 * Ensure that the lanahelper library does not perform a
126 NCBRESET of each lan adapter when enumerating the
127 current network bindings. Correcting this permits OpenAFS
128 to work on Windows 7 when the network adapter settings
131 * Fix creation of mount points and symlinks as \\AFS\xxxx
135 * Reimplement FetchData and StoreData operations to make
136 use of rx_Writev and rx_Readv which is faster than than
137 multiple calls to rx_Write and rx_Read. Throughput
138 improvements of up to 20% are possible.
140 * Many improvements to the rx rpc stack:
144 - Fast recovery is not triggered by packet resend timeouts
146 - Socket errors are now detected and reported
148 - avoid repetitive memory allocations
150 - avoid calling gettimeofday() for every packet
152 - reduced dependency on the call->lock
154 - When a call receive is completed, ack all of the packets
155 immediately. Do not wait for the first response packet
158 - Do not hold the call->lock during rx_Write* and rx_Read*
159 operations when the call fields in use can only be accessed
160 by the application thread.
162 - Reduce the default window size from 128 back to 32 because
163 the current overhead from walking packet queues in the rx
164 listener thread significantly hampers the ability to
165 process incoming packets from the network.
167 - Only backoff the peer timeout once for a series of naks.
169 - Release builds are lean and mean. All debugging and
170 packet tracking code is disabled.
172 * Version number is logged to the Windows Application Event
173 Log as part of the start pending message.
175 * MSI installers now backup and restore configuration
176 data during upgrades.
178 * Rx pmtu discovery has been disabled by default. Use
179 HKLM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
180 RxPMTUDiscovery "DWORD"
181 value to re-activate.
183 * cm_BkgDaemon event evaluation has been optimized to avoid
184 unnecessary overhead.
186 * If a volume is known to be readonly, then locally fail any
187 change request without sending it to the file server. This
188 avoids triggering the file server's abort threshold.
190 * Do not leak cm_volume_t objects from the volume LRU queue.
192 * Ensure that cm_NameI errors are acted upon promptly.
193 There are many cases in the SMB server where an error from cm_NameI()
194 was either ignored or not acted upon until several other operations
195 are performed that could result in the same error being repeated.
196 This is a mistake which did not have negative side effects until
197 additional checks for callback status were added recently.
199 At present, if a CM_ERROR_ACCESS error is returned and ignored,
200 subsequent attempts to operate on the same cm_scache_t will result
201 in additional queries to the file server that will also end in an
202 abort response. This can trigger the file server to delay responses
205 * Fix fs mount point and symlink create operations which were
206 failing due to improper directory separator termination.
208 * RXAFS_GetVolumeStatus can return VNOVOL, VMOVED, etc. In
209 order to process them and update volume state a fid must
210 be passed to cm_Analyze(). Use the volume root fid.
212 * Negative caching for volume lookups that fail with either
213 VL_NOENT or VL_BADNAME. The expiration time is five minutes.
214 This prevents volume lookup storms.
216 * Provide each volume lookup its own two hour lifetime instead
217 of reseting all volume location information every two hours.
218 This prevents the first smb path evaluation after the volume
219 location data has been discarded from potentially causing an
223 * When the SYSTEM account is in use, the SMB connection is
224 not always established as S-1-5-18. Sometimes it uses an
225 anonymous, S-1-5-7, connection. In both cases the SMB
226 authentication name is the nul string. To better identify
227 when an SMB connection is local system, the smb_username_t
228 is now stored as a SID instead of a user name. When a SID
229 is used, a new flag bit is set SMB_USERNAMEFLAG_SID.
231 In smb_SetToken if AFS_PIOCTL_AFSLOGON is set and the user
232 connection is not SYSTEM, the RPC SID is checked. If RPC
233 SID is SYSTEM, that supercedes the SMB authentication.
235 * Replace STATUS_IO_TIMEOUT with RPC_NT_SERVER_TOO_BUSY for
238 * Replace STATUS_PATH_NOT_FOUND with RPC_NT_SERVER_UNAVAILABLE
239 for CM_ERROR_ALLOFFLINE and CM_ERROR_ALLBUSY.
241 * Optimize behavior of buf_CleanAsyncLocked(). Avoid searching
242 for the cm_scache_t object by FID if it is already known.
244 * If the readonly file attribute is set (stored as a unix mode)
245 then a CreateFile operation should fail if the file is opened
246 for DELETE in combination with any other privilege.
248 * If the directory buffer contents are garbage we can crash
249 the service. Add some simple validation checks to ensure
250 that cm_dirEntry_t objects have the correct flag value and
251 that the name strings are not too long.
253 * If the bulkStat errorCode indicates that a particular object
254 is inaccessible due to a VIO error, we must update the server
255 status appropriately in order to permit failover.
257 * An RX_MSGSIZE error is returned by the new PMTU detection
258 code. It is critical that such an error result in a retry of
259 the operation that failed. Otherwise, the PMTU detection can't
260 work and the server will be marked down.
261 Secondly, it is important that such errors not leak to the
262 application layer. Map them to CM_ERROR_RETRY in all cases.
264 * Translate RX_RESTARTING to the same cache manager error code
267 * When logging server volume instance errors to the windows
268 application event log, be sure to log the cell as well.
269 Translating from server ip address is non-trivial. Make it
270 easier for administrators triaging issues to plug the volume
271 and cell info into vos commands.
273 * Since unix mode bits are represented in octal in most cases
274 make sure we log them that way.
276 * The cm_enforceTrailingDot() function failed if there was a
277 trailing dot followed by white space.
279 * Rx: Do not send RX_MSGSIZE if the PMTU size did not change.
282 * A crash could occur if a single letter share name was
283 dynamically evaluated as matching a cell name. This was
284 due to a failure to treat the comparison of two nul
285 strings as identical.
287 * Docbook validation is performed by xmllint during builds.
289 * Volume package bugs in the file server can result in VOFFLINE
290 being returned to the client instead of VNOVOL or VMOVED. As
291 a result the Unix CM treats VOFFLINE the same as VMOVED and VNOVOL.
292 The Windows client has not. As a result, bugs in the file server
293 can cause the Windows client to lose if the volume has in fact
294 been moved to another server.
296 As part of this change, the volume location list is updated prior
297 to the volume status being applied to the server from which the
300 * If the pages to be flushed are from a readonly or backup volume
301 they can't be dirty. In this case there is no need to stabilize
302 the pages before they are locked and recycled.
303 Stablilization is performed on the cm_scache_t object so do not
304 stabilize and unstablize for each cm_buf_t object.
306 * Over the years the processing of the Freelance callbacks have
307 added functionality that behaves much more like FetchStatus checks
308 to a file server. If the data version of the object has changed,
309 get the new data. Given that is the case, we can remove much of
310 the original refresh logic that is rather race prone. Say goodbye
311 to cm_fakeGettingCallback and cm_fakeDirCallback.
313 * When processing a pioctl path with either smb_ParseIoctlPath or
314 smb_ParseIoctlParent, cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK|GETSTATUS)
315 must be called on the cm_scache_t object to ensure that it is up
316 to date before we permit cm_Lookup or other operations to be performed
317 on it. Add the cm_SyncOp() call to smb_ParseIoctlPath and
318 smb_ParseIoctlParent to ensure it is done for all pioctl operations.
320 * The 32-bit tools installer was displaying client configuration wizard
321 pages. Since, the 32-bit tools share configuration with the AFS
322 service, the 32-bit tools installer should not be attempting to change
323 client configuration. Remove said wizard pages from the 32-bit tools
326 * cm_data.rootSCachep is a global pointer to the cm_scache_t that
327 represents the root.afs volume root directory. Throughout the
328 code this cm_scache_t was being used without ensuring that a
329 callback to the volume is in fact valid.
331 Under most circumstances this would not be a problem. However,
332 it is possible for a request to fail due to the lack of a callback
333 at a critical moment. Add a new function cm_rootScachep() that
334 attempts to ensure that a callback is present (if possible) prior
335 to use the cm_scache_t object.
337 * When the contents of the Freelance root directory changes the fake
338 directory buffers are updated and a fakeDirVersion is incremented.
339 The dataVersion of the cm_scache_t object is supposed to be updated
340 on the next access by performing a fake get callback request.
341 Unfortunately, this did not always occur because of a race. If another
342 Freelance object is updated first, the root directory object would
343 never successfully get a fake callback.
345 This patchset ensures that the generation of the fake directory
346 buffer content and the callback are obtained under the same set
347 of locks thereby removing the race.
349 * Enforce in afsd_service that tokens can be set for alternate
350 userids only if the SetToken pioctl is called from the
351 Local System account.
354 * Revise SMB QuerySecurityInfo Response for MS10-020
356 MS10-020 (http://support.microsoft.com/kb/980232) has caused
357 many problems for implementors of SMB 1.0 servers and applications
358 that call GetFileSecurity() without checking the return code to
359 determine if the call succeeded. The gist of the vulnerability
360 was that the SMB redirector would pass any buffer it received
361 to the application regardless of whether or not it was valid.
362 MS10-020 protects the applications by strictly validating the
363 SMB response data structure and the data in the security descriptor
366 The problem for SMB 1.0 server implementors is that there have
367 been at least three different protocol descriptions for
368 NT_TRANSACT_QUERY_SECURITY_DESC published over the last decade
369 and all of them are incomplete. Therefore, just about no one but
370 Microsoft has an SMB 1.0 server implementation that produces the
371 exact out that they are expecting to validate.
373 The end result is that in an attempt to protect applications from
374 crashing due to invalid input being passed in directly caused
375 dozens of applications to crash by not returning any security
376 descriptor data at all. Even when the applications didn't crash
377 they might not have been able to save their data. Cisco WAAS
378 and NetApp DataOnTap systems were most adversely affected and
379 they have had CIFS protocol licenses for many many years.
381 To fix OpenAFS here is what needed to be done:
383 1. Instead of returning a security descriptor that gives ownership
384 to the NUL SID, give it to the Everyone SID and set the flag
385 that states that everyone has full access.
387 2. Validate the input parameters. In particular, check to ensure
388 that the SMB file descriptor is valid and the file has not
391 3. Enforce the maximum output data and parameter counts.
393 4. Handle buffer overflow and buffertoosmall conditions
394 in the manner that Microsoft expects them to be handled.
395 In particular, note that the parameter data which is returned
396 in the SMB Data Region is not counted in the Data Count.
397 Even if MaxData is 0, we can still return parameters values
398 as long as MaxParm is large enough.
400 * Prevent use of AFSCache file contents if mapped to
403 * The Windows version of "fs newcell" did not accept any parameters
404 and behaved quite differently from the Unix version. Instead of
405 permitting new cell information to be added, the Windows version
406 simply forced the existing cell information to be reacquired.
408 This update adds a new pioctl, VIOCNEWCELL2, to support the
409 implementation of a Unix-style "fs newcell". The functionality
410 added here differs from the Unix version in the following ways:
412 1. "fs newcell" with no arguments is still accepted
413 in order to maintain compatibility with prior Windows
416 2. "fs newcell -cell <cell> -dns" instructs the cache manager
417 to add the new cell but obtain the vldb server info from
420 3. "fs newcell -cell <cell> ... -registry" instructs the cache
421 manager to add the new cell and also save the cell configuration
422 data in the registry for use the next time the service restarts.
424 4. The -vlport and -fsport options are accepted although the
425 -fsport value is currently unsupported by the cache manager.
427 * New registry value "FreelanceImportCellServDB" instructs Freelance
428 to create a mount point for every cell name listed within the
431 * Path MTU discovery for Rx is activated.
433 * Rx socket input buffer is converted to a circular buffer.
435 * Fix usage of cm_FreeServerList(). Do not set the server list
436 pointer to NULL after calling cm_FreeServerList(). Doing so
437 can result in a memory leak.
439 * Only enable Rx NAT pings on a single anonymous connection at a
442 * Fix cm_IoctlSkipQueryOptions() buffer management. Prevents a
443 potential read beyond end of memory buffer.
445 * Reduce requested privileges when reading registry CellServDB
446 to the minimum required.
448 * Add support for RPC Pipe Service NetWkstaGetInfo levels
449 101 and 102 which are called on Windows 7 and 2008-R2.
451 * Prevent integer overflow during quota percent used calculation
452 in Explorer Shell Extension (RT 126846)
454 * Generate a meaningful error if "fs listacls" or "fs setacls"
455 are executed on the Freelance root.afs volume.
457 * RXAFS_InlineBulkStat errors must be processed via cm_Analyze.
458 RXAFS_InlineBulkStatus does not return errors such as EACCES,
459 VNOVOL, VNOVNODE, VOFFLINE, VBUSY, VIO, VMOVED, etc. as an RPC return
460 code. Instead they are returned in the status info errorCode field
463 Traditionally, the error associated with the first FID in the query
464 list has been returned to the caller of cm_TryBulkStatRPC().
465 However, the error has never been processed through cm_Analyze()
466 which means that the per-vnode processing for VNOVNODE and the volume
467 global processing for VMOVED, VNOVOL, etc. has never been performed.
468 As a result, failover to other .readonly volume instances cannot occur,
469 volume moves will not be handled, and files that have been deleted
472 This patchset makes the following changes:
474 1. If an inline bulk operation has been performed and the inline
475 errorCode is a volume global error, then that error replaces
476 the RPC return code within the cm_Analyze() processing for
477 the RPC. This will affect whether or not a retry operation
480 2. The variable 'inlinebulk' is reset to 0 at the top of the
481 cm_Analyze() loop in case failover from an inlinebulk capable
482 file to an inlinebulk incapable file server takes place.
484 3. The FID that is passed into cm_Analyze() is not a real fid.
485 Instead it consists of the cell and volume but vnode = 0.
486 This ensures that the error (if any) is not applied to the
489 4. If an inline bulk operation was performed, prior to performing
490 the cm_MergeStatus() operation a vnode a check is made to
491 determine if an error was returned for that vnode. If so,
492 cm_Analyze() is called with no connection, a fake cm_req_t,
493 the fid, and the error. This permits cm_Analyze() processing
494 to be performed on the file.
496 * Show configuration pages for all types of MSI installations
498 The OpenAFS MSI installer wizard used to not show any configuration
499 pages for "Typical" and "Complete" installations. Setting the
500 workstation cell and logon options during installation required
501 selecting the "Custom" option. Many users choose the "Typical" option
502 during installation, and thus would never see the configuration pages.
503 Therefore, for these users, the workstation cell was being set to the
506 This patch makes the workstation cell and logon option configuration
507 pages visible to all types of installations (except silent
508 installations which show no UI).
510 * cm_LookupInternal creates Freelance mount points and symlinks
511 when queries cannot be found in the Freelance root.afs directory.
512 If the search name is a full cell name for which vldb information
513 can be obtained, then a mount point is added. If the search name
514 is a left-most substring or the full cell name with a dot appended
515 to it, then a symlink was created. This approach created a very
516 poluted Freelance name space.
518 This patchset makes the following changes:
520 1. Do not create symlinks with a dot appended to the cellname
522 2. Do not create symlinks where the left-most substring is not
523 a full dot separated component of the cellname.
525 3. Permit lookups to succeed when we would have created a
526 symlink in the past without creating the symlink.
528 * BPlus tree lookups are much faster than searching through
529 the native directory format on Windows because the case sensitive
530 hash tables cannot be used successfully. Permit BPlus trees
531 to be used except when called with cm_BPlusDirFoo as the action
532 function because cm_BPlusDirFoo is used to build the BPlus trees
533 from the native directory format.
535 * Symlinks are ending up in the Freelance root.afs directory that
536 end with a dot. Make sure it cannot happen.
538 * cm_FreelanceAddMount and cm_FreelanceAddSymlink is supposed to
539 return the allocated FID of the entry that was added. However,
540 cm_NameI is called to perform the lookup without forcing an update
541 of the Freelance fake directory. As a result the entry may not be
544 Force an update prior to calling cm_NameI() by using
545 cm_clearLocalMountPointChange() and cm_reInitLocalMountPoints()
548 * The Freelance fake root directory buffers were not zero-filled.
549 This results in random behavior that can cause the service to
550 terminate unexpectedly.
552 * The validation check for the response from the GetVolumeStatus
553 pioctl is incorrect. The response is not simply a VolumeStatus
554 structure but also several C strings appended to it.
556 * When flushing a file, we need to commit the file length changes
557 as well as the dirty buffers. Call cm_FSync instead of buf_CleanVnode
558 which is called by cm_FSync.
560 * Prevent rx_rpc_stats global lock from being a bottleneck in the
565 * Avoid a race when updating cell vldb server lists
566 that can result in a crash.
568 * Avoid a deadlock when managing CM_SCACHESYNC_STOREDATA
569 state operations for directory objects.
571 * Add new Windows Application Event log messages for
572 VBUSY, VRESTARTING, ALL_BUSY, ALL_OFFLINE, and ALL_DOWN.
573 Include message throttling to prevent the same message
574 from being logged repeatedly within a five second window.
576 * Reduce lock contention by waiting for cm_buf_t I/O
577 operations to complete before permitting cm_SetupStoreBIOD
578 to analyze a buffer for inclusion in a BIOD.
580 * Split the cm_buf_t flags field to separate the flags
581 that are protected by the cm_buf_t mutex from those
582 protected by the buf_globalLock. This eliminates the need
583 to hold both locks everytime the flags field is accessed.
584 Both locks were not held in the past resulting in race
585 conditions that could result in deadlocks.
587 * Add "vos setaddrs" command.
589 * Rx library lock contention avoidance between rx_NewCall and
592 * Rx library races due to inconsistent use of rx_connection
593 conn_data_lock to protect the flags field.
595 * Rx library inconsistent use of RX_CALL_TQ_WAIT which could
598 * Rx library must signal transmit queue waiters when flushing.
599 Otherwise, deadlocks can occur.
601 * In cm_UpdateVolumeLocation, avoid searching for a ".readonly"
602 volume on a numeric volume name.
604 * File buffer allocations whose offsets are beyond server EOF
605 should be locally allocated and zero filled. The file server
606 should not be issued a FetchData rpc which is guaranteed to
609 * Enable integrated logon to work with Windows 7/2008 when
610 user logons are performed with a non-Domain Kerberos principal.
612 * Add Protection Error messages to aklog output.
616 * Prevent the Explorer Shell extension from crashing if
617 symlink creation failed. (126406)
619 * A Rx level NAT ping has been implemented.
620 Add NatPingInterval registry value to
621 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
622 to permit Nat Ping to be enabled. The default value is 0 seconds.
624 * When a re-initialization is taking place, be sure to reset
625 cm_noLocalMountPoints to 0 in case someone deletes the "Freelance"
626 registry key out from underneath the service.
628 * Add krb5 error message translation to aklog, afscreds,
629 afslogon.dll, the network identity manager afs provider
632 * Mode bits aren't directly exposed by the Win32 API. We were leaving
633 them to default to 0777 when creating new files and directories.
634 This version introduces two configuration parameters;
635 'UnixModeFileDefault' and 'UnixModeDirDefault' which are DWORD
636 registry entries that are used to set the initial mode bits.
637 If the values are set to 0, then the behavior is identical to what we
640 * Minidump files are now produced with a timestamp appended
643 * An SMB request debugging monitor has been added. When activated
644 the monitor will automatically turn on trace logging if any SMB
645 request has required longer than 60 seconds to complete and will
646 then create a minidump every 60 seconds thereafter until the
651 * Restore use of DNS AFSDB and SRV records by kaserver clients.
655 * Avoid a potential Freelance deadlock during initial execution
656 of afsd_service.exe if the old ini file data has to be
659 * Three rx library corrections. (1) Idle data connection
660 processing could timeout if the send window filled and
661 took longer than the idle data timeout period for the
662 transmit window to re-open. (2) The transmit queue
663 could be emptied prematurely. A required check for the
664 queue being in use was forgotten. (3) The function that
665 is supposed to implement a wait for the transmit queue
666 to cease being busy failed to wait.
669 * Restore use of DNS AFSDB and SRV records which were
670 unintentionally disabled in 1.5.69
673 * Add a context menu to the NetIdMgr AFS Provider
676 * Prevent an empty directory Btree from being created and
677 marked as valid if cm_BPlusDirBuildTree fails.
679 * cm_BPlusEnumAlloc should not fail if the enumeration
680 contains zero entries.
682 * In cm_BPlusXXX functions, return ENOMEM if malloc() fails.
684 * Do not leak the cm_scache_t dirlock if cm_BPlusDirBuildTree
687 * fs examine should report owner and group ids as signed values.
689 * Protect buffers in smb_WriteData from simultaneous writes.
691 * Query network interface MTU values from the registry because
692 the IP Helper GetAdapterAddresses() fails to report a manually
695 * Improve performance by not dropping and reacquiring the rx_call
696 lock in rx_WriteProc* and rx_ReadProc*.
698 * When configuring the rx library, the network interface MTU
699 must be applied to both the maximum send size as well as the
700 maximum receive size as is done when rx_SetMaxMTU is called
703 There was no 1.5.67 release
706 * Convert afsauthent.dll to use the xdr_alloc and xdr_free
707 routines from afsrpc.dll in place of malloc and free.
708 This permits the two libraries to be built with different
709 C runtime libraries without crashing.
711 * Set the DOS Readonly attribute on a file/directory
712 whenever the unix mode combined with the mask 0200
713 is true. Previously there was a discrepency between
714 the mask used for testing for readonly behavior and
715 that used for setting the attribute.
717 * Disable AFSVolSync based .readonly "whole-volume callback"
718 support because the all file servers prior to 1.5.67
719 (and perhaps 1.4.12) do not properly assign a value to
720 the AFSVolSync structure in bulk status RPC responses.
722 * Improve the error output from aklog to output the value
723 from krb5 error_message() if the afs_com_err output
724 indicates an unknown value.
726 * Hold a lock on the cm_scache_t object when the smb
727 server is assigning a new clientModTime value.
729 * Permit custom build version numbers to be set
730 without modifying the src/config/NTMakefile.<platform>
733 * Add support for binary code signing with cross-signed
734 certificates and arbitrary versions of signtool.exe.
736 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and
737 do not permit them to be exposed to the smb redirector.
739 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT
740 to avoid confusion within the smb redirector.
742 * Fix the byte order assigned to port numbers associated
743 with AFSDB record lookups. They must be network byte
744 order not host byte order.
746 * Add dynamic server ranking based on RPC round trip
750 * Official support for Windows 7 and Server 2008 R2
752 * Improved Rx RTT computation following the lessons
755 * Prevent a file server bug (FetchData returning an
756 invalid length instead of zero) from causing an
757 "unexpected network error" when writing to files.
759 * Promote DNS SRV records as superior to DNS AFSDB
760 records. Support arbitrary port numbers for vldb
763 * Add "fs listacl -cmd" support.
765 * Add AFSVolSync based .readonly "whole-volume callback"
766 support. With this functionality, multiple objects from
767 a .readonly volume can have their status validated by
768 issuing a single RXAFS_FetchStatus RPC.
770 * Remove drive mapping functionality and service start/stop
773 * Remove drive mapping functionality from afs_config.exe.
775 * Adjust SMB error return codes to avoid returning
776 STATUS_TIMEOUT which results in the SMB redirector
779 * Network Identity Manager OpenAFS Provider now provides its
780 own "AFS lock" notification icon to report the status of
781 "have tokens, have no tokens, service not started, service
782 started but inaccessible". Hovering over the icon lists the
783 cells for which tokens exist (if any) and the OpenAFS version
784 number. Double-clicking executes the Network Identity Manager
787 * When merging the current status for an AFS object, ensure
788 that the volume status for that object is set to 'vl_online'.
790 * Permit cm_Analyze to request a retry even when the cm_req_t is
791 flagged CM_REQ_NORETRY if the RPC failed due to RX_CALL_DEAD.
792 The retry will force the use of a new connection.
794 * Add digital signatures to resource only dlls as required by
797 * Prevent pioctl calls from retrying indefinitely when a sharing
798 violation error occurs.
800 * Add {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab"
801 to restore access to drive mapping functionality in afscreds.exe
804 * Permit Rx minimum peer timeout, maximum receive window size and
805 maximum send window size to be configured via the registry.
807 * The MSI installer now sets the ARPINSTALLLOCATION as required by
810 * DNS lookups set the port numbers in network byte order within
813 * Replace use of the 16-bit compatibility API WinExec with
814 ShellExecuteEx in afs_config.exe and afscreds.exe. WinExec
815 is incompatible with Windows 7 / Server 2008 R2 UAC
819 * Short circuit background volume checks if a shutdown
820 or suspend request is received.
822 * Avoid contacting the file server in order to allocate
823 an empty buffer beyond the current length of the file.
825 * When a request for streams on a directory or mount
826 point object is received, do not offer a default stream.
827 This was affecting the saving of roaming profiles.
830 * Fix Windows 2000 compatibility. Do not call rand_s()
831 even if it is compatible with the C run time library
834 * Fix a data consistency error between the output of
835 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
836 workstation group name "AFS", and report server name as
837 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
839 * Enable executables to be run from \\AFS on Windows 7.
840 Return "Name not found" instead of "File not found" when
841 a directory or file name cannot be found.
843 * Prevent cache manager from marking file server "down" when
844 the data returned in response to either RXAFS_FetchData64
845 or RXAFS_StoreData64 is invalid.
847 * Add pioctl data validation to the AFS Explorer Shell extension.
850 * After 1.5.62 afslogon.dll would no longer crash but it
851 also would not load "domain" specific configurations.
854 * An uninitialized variable was present in the symlink
855 recursion detection loop.
857 * Reverse addition of unique per file GUID during
858 Create/Open responses. Inclusion of the GUID prevents
859 Cygwin applications (and others) from accessing //afs.
861 * Treat "filename::$DATA" as a synonym for "filename".
862 This addresses the recent report of roaming profiles
863 failing to save back to the file server.
865 * When given a choice between an error from rx_Write()
866 or RXAFS_EndStoreData(), use the rx_Write() error.
867 This ensures that over quota and access denied errors
870 * If an error occurs during RXAFS_StoreData operations,
871 apply that error to all of the buffers in the BIOD.
872 This ensures that failures will not be retried if they
873 will only fail again.
875 * Do not perform a DNS lookup for a cellname if the
876 cellname prefix is _._AFS_IOCTL_._. This will avoid
877 DNS lookups when the query is for
878 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
880 * Fix Freelance root.afs volume object callback processing.
881 Object synchronization and status merging bugs.
883 * Fix Explorer Shell "invalid parameter" bug introduced
886 * Log RXKAD errors to the trace log.
888 * Prevent simultaneous pioctl calls from multiple processes
889 from stepping on each other's data.
891 * Prevent simultaneous pioctl calls from crashing the
894 * Add data validation to pioctl output processing in
895 ktc_GetToken and ktc_ListTokens.
897 * Add data validation to pioctl output processing in
898 fs.exe and symlink.exe.
901 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
902 No longer will browsing \\AFS produce truncated share names.
904 * Registry specified server preferences did not result in
905 the CM_SERVERFLAG_PREF_SET flag being applied to the
908 * Avoid unnecessary DNS lookups of share names as cells
909 when it is known that the name cannot be a cell name.
910 Any name that does not contain a dot is skipped.
912 * When processing Dfs Referral requests, do not return an
913 error if the server is down or busy; if the volume is
914 offline; the cell vldb cannot be reached; etc.
915 These paths are still \\AFS paths and so the client
916 should not be sent to search elsewhere for resolution.
918 * CRITICAL: Prevent the cm_Daemon thread from terminating when
919 the machine enters the suspend state. IF the cm_Daemon thread
920 dies, there is nothing to execute down server checks.
922 * Prevent use of smb_StartedLock before initialization which
923 can result in a panic in the lock package.
925 * Improve error checking in afslogon.dll in order to prevent
926 NULL pointer dereferences within GetLogonDomainOptions()
927 if the Lsa operations fail.
929 * Add support to the build system to automatically update
930 a Microsoft Symbol Store.
932 * Do not access the cm_conn_t after executing the cm_Analyze
933 loops. Doing so can result in an access to memory that has
936 * CRITICAL: If a StoreData request offset is not aligned on the
937 buffer module blockSize, (offset % blockSize) bytes of dirty
938 data will not be written to the file server.
940 * CRITICAL: If a StoreData64 request is sent to a file server
941 that does not support large files (>2GB) and the client is
942 unaware that the server has no large file support, the first
943 chunksize worth of data will not be written to the file server.
944 This bug was introduced in 1.5.3. It affects all IBM AFS file
945 servers and all OpenAFS file servers older than 1.4.0.
948 * If a file server becomes inaccessible while the cache manager has
949 dirty buffers to write, the afsd_service buf_IncrSync thread can
950 attempt to use 100% of the cpu.
952 * Fix the locking that protects the cm_cell_t lists
953 (all, name, id, free).
955 * Remove cm_cell_t objects from the name and id hash tables before
956 placing them on the free list.
958 * Fix "fs newcell" which was broken in 1.5.60.
960 * Do not attempt to synchronize dirty buffers if the associated
961 volume is known to be unavailable.
963 * Log to the Windows Event Log dirty buffers that cannot be written
964 during the shutdown process.
966 * Modify behavior of a Freelance mountpoint target that does not
967 specify a cell. Instead of assuming the target volume is in the
968 Freelance.Local cell, use the workstation "Cell" specified in the
969 registry. A mountpoint target of "#root.cell." will now mean the
970 root.cell volume in the workstation cell for the current session.
971 If the workstation cell changes from "athena.mit.edu" to
972 "andrew.cmu.edu", the referenced volume will also change without
973 requiring that the mount point targets be altered.
975 * Avoid false infinite loop errors when validating the cm_cell_t
976 free list during startup.
978 * Fix two more locations where xdr_free() should be used instead of
979 free() to permit a checked built of afsd_service.exe to work with
980 release builds of afsrpc.dll.
982 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
983 to permit the server Uuid to be used to lookup the server object
984 and from that determine the cell. This permits callbacks that are
985 received from alternate addresses to be processed with a known server
986 object. Previously a request from an unknown server would clear all
987 callbacks from all cells.
989 * Fix a bug that prevented optimal performance when using a non-zero
990 value for 'daemonCheckVolCBInterval'. As a reminder, when
991 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
992 volume callbacks are automatically renewed 90 minutes before their
995 * Fix automatic ranking of vldb servers whose values are obtained from
998 * Add logging for RX CALL DEAD errors which are generated if the remote
1001 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
1002 or the call is to a vldb server.
1004 * Add registry based cell search functionality to NetIdMgr,
1005 afs_config.exe, and klog.exe.
1007 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
1008 lookups on the host names in the CellServDB instead of using the
1009 specified IP addresses. This provides aklog, pts, vos, etc. the same
1010 CellServDB behavior that the Windows Cache Manager uses.
1012 * Completely a behind the scenes change. The snprintf() routines have
1013 been completely rewritten.
1015 * When updating the stat cache entry callback of a .readonly object
1016 from the volume group object, update the file server reference to
1017 ensure it matches the most update to date callback.
1019 * Add proper support for processing callbacks from multi-homed file
1020 servers. Instead of comparing servers by cm_server_t pointer,
1021 compare them by UUID when the UUID is known.
1023 * During a shutdown short circuit the offline volume check daemon
1026 * Properly track link counts during cm_Link and cm_Unlink operations.
1028 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
1029 preference to an error code reported by rx_EndCall.
1031 * Add "PerFileAccessCheck" registry value to permit testing against
1032 experimental file servers that include per-file acl support. This
1033 value is intentionally undocumented. It is not to be used by
1034 production environment deployments.
1036 * Log a warning to trace log if a volume group is marked as a DFS
1037 File Set. The Windows OpenAFS client does not include DFS support.
1039 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
1040 provider from obtaining tokens when referrals are in play.
1042 * Add "fs chown" and "fs chgrp" commands to permit the owner and
1043 group of objects stored in AFS to be set from Windows.
1045 * Avoid performing background daemon operations when the machine is
1046 going into suspend mode.
1048 * Perform offline volume checks in most recently used order.
1050 * Prevent crash when a data version for a cache object goes backwards.
1052 * fs setquota, fs setcachesize, vos setfields, and vos create now
1053 accept human readable orders of magnitude. (K, M, G)
1055 * fs listquota fixed to permit large quota sizes to be displayed.
1057 * rx packet resend and data packets sent counts were incorrect.
1059 * Multi-thread safe library versions are now being generated and used.
1060 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
1062 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
1063 values are now available on XP through Windows 7. Add functionality
1064 to autodetect if such support is present on the machine. If so,
1065 configure it if necessary and dynamically adjust the AFS Rx timeout
1068 * Correct another race condition in the Rx library that could result
1069 in an unexpected panic while freeing the Rx call iovq.
1072 * A fix to the pioctl library to support drive substitution
1073 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
1075 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
1076 that corrects a deadlock in the smb redirector and also adds new
1077 functionality that permits the AFS SMB server to be given a longer
1078 timeout than is normally the case. New functionality has been added
1079 to configure these additional LanmanWorkstation\Parameter values.
1080 (This functionality has been backported to XP SP3 and is scheduled
1081 to be released on June 5th.)
1083 * The BackConnectionHostNames registry value configuration was broken
1084 when dynamic re-establishment of Netbios Name registrations was added.
1087 * Hidden vos commands are revealed.
1089 * If the "DisableLoopbackCheck" registry value is set, do not unset
1090 it during the same service session.
1092 * Reorganize code that prevents multiple Store operations to the same
1095 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
1096 permit broken symlinks to be treated as being in AFS.
1098 * Fix vos commands that output 64-bit integer values.
1100 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
1101 Permits building cygwin applications against OpenAFS libraries.
1103 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
1104 as Windows HTML Help (.CHM) files.
1106 * NSIS installer does a much better job of cleaning up files left
1107 over from previous installs.
1109 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
1110 or "fs flushall" and on-going directory operations that can result in
1111 afsd_service.exe crashing.
1113 * Add support for DNS SRV records in place of AFSDB records.Â
1114 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
1115 Weight is currently ignored.
1117 * Add a method of specifying Client CellServDB information within the
1118 registry that can be used to either override the CellServDB file or
1119 force the use of DNS lookups for a given cell.
1121 The registry schema is as follows:
1123 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
1124 "LinkedCell" REG_SZ "[cellname]"
1125 "Description" REG_SZ "[comment]"
1126 "ForceDNS" DWORD {0,1}
1128 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
1129 "HostName" REG_SZ "[hostname]" Default: [servername]
1130 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
1131 "IPv6Address" REG_SZ "[address]" <future>
1132 "Comment" REG_SZ "[comment]"
1133 "Rank" DWORD "0..65535" Default: 0
1134 "Clone" DWORD "{0,1}" <future: server only>
1135 "vlserver" DWORD "7003" <future>
1136 "ptserver" DWORD ... <future>
1138 ForceDNS is implied non-zero if there are no [servername]
1139 keys under the [cellname] key. Otherwise, ForceDNS is zero.
1140 If [servername] keys are specified and none of them evaluate
1141 to a valid server configuration, the return code is success.
1142 This prevents failover to the CellServDB file or DNS.
1144 Only one of "HostName" or "IPv4Address" is required. "HostName"
1145 is optional if [servername] is the host name.
1147 * Extend registry based CellServDB functionality to
1148 afsconf_GetCellInfo() interface used by aklog and many
1149 other command line utilities.
1151 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
1152 as it is now used by both client and server components.
1155 * PriorityClass of afsd_service.exe process raised to
1156 "High" to match the priority of the system services
1157 that are waiting for afsd_service.exe to respond.
1158 Configurable using [TransarcAFSDaemon\Parameters]
1159 DWORD "PriorityClass" value. (use values from Win32
1162 * Change error reporting of CM_ERROR_BADSHARENAME to
1163 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
1164 The error we were sending would cause the smb
1165 redirector to drop the connection.
1167 * Change the error reporting of CM_ERROR_ALLOFFLINE
1168 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
1169 Previous we would return Bad Network Path which
1170 would cause the smb redirector to cancel all file
1171 handles on the same "share".
1173 * A debug version of afsd_service.exe can now be used
1174 with a release version of afsrpc.dll (or vice versa).
1175 afsrpc.dll now implements and exports xdr_free().
1177 * Removed all unprotected OutputDebugString() calls as
1178 those calls can block for up to 10 seconds if there
1179 is contention for the globally shared memory or if
1180 the debugger fails to read data in a timely manner.
1182 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
1183 supported for builds. The production releases still
1187 * Discovered that XP/2003 supports a variant on recvmsg
1188 called WSARecvMsg. The input parameters are mostly
1189 compatible with the Posix recvmsg. This will save
1190 multiple memcpy() calls per received packet.
1192 * Discovered that Vista/2008 supports a variant on
1193 sendmsg called WSASendMsg. The input parameters
1194 are mostly compatible with the Posix sendmsg. This
1195 will save multiple memcpy() calls per sent packet.
1197 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
1198 and utilizing rx_ReadProc32 from within the cache manager
1199 as part of the FetchData operations. rx_ReadProc32 and
1200 rx_WriteProc32 are optimized versions of rx_ReadProc and
1201 rx_WriteProc for use when reading 32-bit integer values.
1203 * The Network Identity Manager AFS provider now reports
1204 progress events. These provide a better user experience
1205 with Network Identity Manager 2.0.
1207 * [RT 124293] A race condition exists which permits the
1208 scp field of the an smb_fid_t object to become invalid
1209 after a request on the smb_fid_t has begun. Remove the
1210 race by protecting all access to the scp field with the
1211 smb_fid_t mx mutex and obtaining a local reference on
1212 the cm_scache_t object for the length of the request.
1214 * [RT 124276] If the vldb is out of sync with the contents
1215 of the file servers, afsd_service will retry too many
1216 times when a file server reports a volume as not being
1217 present. Now if the list reported by the vldb is the
1218 same as the previously seen list, then the retry is
1221 * [RT 124276] Read-only volume failover was broken in
1222 1.5.53 whenever accessing a volume results in VNOVOL or
1225 * [RT 124276] Prior to 1.3.70 the volume server reference
1226 list was not reference counted and would be prematurely
1227 freed while in use. When reference counting was added
1228 in 1.3.70 a bug was introduced that could result in
1229 service reference list corruption.
1231 * Add Windows Application Event Log warning messages for
1232 "Client SMB MPX value too large" and
1233 "Client SMB Buffer Size too small".
1235 * Renaming of files across directory boundaries would result
1236 in an invalid handle error when attempting to access the
1237 files after the move.
1239 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
1240 when a smb file descriptor cannot be found. The whole point
1241 of using a Path Info function is because an smb file
1242 descriptor wasn't allocated.
1244 * More edge cases in which dynamic addition of Freelance
1245 root.afs entries would get the wrong FID or where the
1246 root.afs directory would not be refreshed.
1248 * Buffer overflow could occur if the workstation cell name was
1249 longer than 64 characters. Crashes could occur in
1250 afscreds.exe, afslogon.dll, and afsd_service.exe.
1252 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
1255 * Improve consistency of the bad data version value used for
1258 * Log file server uuid values as part of the cm_server object
1259 when available. Dump the cm_server object list in response
1262 * Optimize the performance of resetting access control lists
1263 when tokens are set or removed.
1265 * Fix vos.exe so that it can work with very large volume
1268 * Remove symlink recursion tests and increase max symlink count
1272 * Undo the "UAC manifest fix" applied to afs_config.exe.
1274 * Fixed some additional race conditions in the Rx library
1276 * Ensure that Freelance allocation of vnodes follow
1277 the AFS convention of odd vnodes are directories
1278 and everything else is an even vnode.
1280 * Add Freelance logic to mount point and symlink
1281 evaluation functions.
1283 * Enhance smb_ParseASCIIBlock() so that it can handle
1284 all of the STRING formats defined by the CIFS Technical
1287 * Validate the output of smb_ParseASCIIBlock() in all
1288 callers. Return CM_ERROR_BADSMB if the STRING field
1289 cannot be parsed. CM_ERROR_BADSMB will cause the
1290 contents of the packet to be logged.
1292 * If multiple SMB Raw Write operations were taking
1293 place at the same time, there could be data corruption
1294 because unique event objects were not generated for
1295 each Netbios receive operation.
1299 * Fix a race in rxi_WriteProc which could result
1300 in a freed or in use packet being passed to
1301 rxi_PrepareSendPacket if the call is currently
1302 in a transmit wait state. Passing a freed packet
1303 will result in a panic. Passing an in use packet
1304 might result in a call failure or perhaps data
1307 * Kerberos referrals broke the NIM AFS provider
1308 logic for automated configuration of the workstation
1309 cell with a new identity. Now if a referrals
1310 response is returned, assume the cell belongs to the
1313 * The interface to buf_CleanAsync() and
1314 buf_CleanAsyncLocked() has been changed to permit
1315 errors to be propagated to the caller.
1317 * During an enumeration of a directory, if the btree
1318 has not yet been generated, return an error.
1320 * When mapping CM_ERROR values to NTSTATUS values
1321 do not treat 0 as an error.
1323 * Avoid a cm_scache_t object reference count underflow
1324 when renaming via the SMB interface.
1326 * smb_IsMask() can be called with a non-8.3 input buffer.
1327 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
1328 input buffers and smb_IsMask() for non-8.3 input buffers.
1330 * Add cm_buf_t error state to the "fs memdump" output.
1332 * pioctl drive mapping checks need to take into account
1333 Global Drive Mappings and SUBST assignments.
1335 * Rename operations should not be permitted if the
1336 destination name collides with an existing name that
1337 differs only by case UNLESS the collision is the object
1338 that is being renamed.
1340 * Increase the maximum number of background daemon threads
1341 to 64. Leave the default at 4.
1343 * Set the FILE_UNICODE_ON_DISK flag when reporting the
1344 file system characteristics.
1346 * Report the maximum component name length as 255 even
1347 for Unicode as part of the file system characteristics.
1349 * Add RX debugging information to "fs memdump" output
1352 * Refcount the cm_server_t objects returned by
1355 * Fix the UAC manifest applied to afs_config.exe
1357 * rxdebug now indicates if any RX_CONN flag is set
1358 not just some of them
1360 * Change the mapping of CM_ERROR_WOULDBLOCK from
1361 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
1365 * Includes #GCO Public CellServDB 07 Nov 2008
1367 * Prevent separate cell entries from being created for
1368 "foobar.org" and "foobar.org." which can happen if
1369 an AFSDB record is found instead of a CellServDB
1372 * Fix a panic when checking the SMB virtual circuit list.
1373 If a VC is destroyed while the list is being checked
1374 it is possible for the next object to become invalid.
1376 * Increased logging to afsd_init.log of netbios name
1379 * Fix a DNLC panic when a UTF-16 string that cannot be
1380 converted to a utf-8 string is provided as a filename.
1382 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
1383 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
1384 location to refer directly to the new location in
1385 %programfiles%\openafs\client\program
1387 * Be more aggressive about responding to Netbios
1388 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
1389 the LANA within the listener thread without performing a
1392 * Use the WNet Enumeration of all connected drives as a filter
1393 in the pioctl GetIoctlHandle operation. This should remove
1394 the last situations in which there was a delay when opening
1395 an Explorer Shell properties menu with the AFS Shell
1398 * Fix the parsing of CellServDB in Windows to handle linked
1399 cells as is done on Unix.
1401 * Add support for linked cells to the cache manager,
1402 netidmgr afs provider, and aklog. When two cells are
1403 linked, a volume lookup that fails in one cell is looked
1404 up in the other. This can be used for construction of
1405 test cells that override specific volumes; cell renaming;
1408 * Increase the length of cell and realm names that can
1409 be entered in the NIM AFS provider.
1411 * Actively prevent a 32-bit afsd_service.exe from
1412 executing in the WOW64 environment.
1414 * Protect against invalid IP address values in CellServDB
1419 * Fix an Rx panic caused by an attempt to double free
1420 an rx_packet when an error occurs while writing the
1423 * Fix an exception cause by a race condition when
1424 processing multiple file change notifications registered
1425 by the Explorer Shell or other applications.
1427 * Modify the Rx debug interface to use var_args and
1428 permit Rx debugging to be used on 64-bit Windows
1429 which previously had it disabled.
1433 * Automated lock hierarchy enforcement added to the
1434 osi base lock package. This permits detection of lock
1435 violations that might but often does not result in
1436 deadlocks. Enabled by default in the checked builds.
1437 May be activated in release builds via the registry
1438 value "LockOrderValidation" in the TransarcAFSDaemon
1441 When a violation is detection an assertion is thrown
1442 generating a dump file and a message indicating which
1443 locks were obtained out of order is logged to the
1444 Windows Application Event Log.
1446 * Attempt to ensure that a dump file will be generated
1447 whenever the service panics.
1449 * Removed unused lock cm_bufGetMutex
1451 * Many out of order lock acquisitions were detected
1452 using the lock order verifier and were corrected.
1454 * A race condition within the function path cm_GetSCache() ->
1455 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
1456 cm_scache_t object to be issued simultaneously to two
1457 threads. This would eventually result in a panic due
1458 to the resulting undercount.
1460 * Fix interpretation of the empty string as the ioctl path
1461 to mean the current directory. "fs lsm", "symlink list",
1462 etc. now return a "not a ..." error instead of "does not
1465 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
1466 to indicate that the input string is a nul terminated string.
1467 Assign ti when input strings are nul terminated instead of
1468 length counted. (from 1.5.50)
1470 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
1471 does not match the observed behavor. The 'nameLength' is specified
1472 in bytes, not in characters. Fix the implementation to match
1473 the observed behavior. (1.5.50)
1475 * The cm_HaveAccessRights() attempt at deadlock avoidance by
1476 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
1477 does not avoid the deadlock. Avoid the deadlock by enforcing the
1478 lock order of the lowest vnode first. Then remove the infinite
1479 loop avoidance in cm_SyncOp() that was returning an unwarranted
1480 access denied error.
1482 * File Server lock synchronization was not propertly enforced in
1483 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
1485 * Another change to the short name generator. Plus and equal are
1486 not valid characters in a short name according to the
1487 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
1488 with left and right brace in order to minimize the impact.
1490 * Fix a race condition that can result in a heap overwrite error if
1491 the number of known file servers or vldb servers increases during
1492 the preparation phase of a cm_CheckServers() operation.
1494 * If the SMB Authentication would fail, a LSA allocated memory
1495 buffer would be leaked.
1497 * Rewrite the cm_cell_t allocation code to permit cell vldb server
1498 lookups to fail without leaking cm_cell_t objects.
1500 * Add "fs setcrypt auth" to permit the use of rxkad authentication
1501 only mode. Not documented.
1503 * Initialize Rx mutex and lock objects once and only once as part of
1504 the DLL load process.
1506 * When reporting the Volume Name "AFS" do not include the terminating
1507 NUL in the reported length.
1509 * Refresh Volume Location information upon receipt of a VNOVOL error
1510 from the file server and avoid a race condition that prevented the
1511 updated volume location information from being destroyed immediately
1512 after it was acquired.
1514 * In response to VNOVOL and VMOVED, remove the volume id from the server's
1515 volume list so that the volume's up and down status is not affected
1516 when the machine is shutdown.
1518 * Many fixes to the Rx stack related to packet allocation and deallocation;
1519 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
1520 by overwriting the rx_call currentPacket pointer when it was already
1521 referring to a packet.
1523 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
1524 in the TransarcAFSDeamon/Parameters key.
1526 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
1527 constructed directory B+ trees for flushed directory objects.
1529 * When cached pages for directory objects have been locally modified, do
1530 not permit the locally modified pages to be mixed with pages subsequently
1531 obtained from the file server. Doing so can result in directory
1534 * If Ubix errors are received in response to VL_xxx operations record
1535 the error on the server instance and permit the operation request to
1536 be retried in case another server is able to answer the query. We have
1537 seen at least one instance of a cell in which a VL server returned
1538 UNOQUORUM (possible database truncation) when the other servers were
1539 responding successfully.
1542 * Optimize the assignment of write errors to dirty
1543 buffers by applying the first error to all buffers
1544 belonging to the same object instead of trying each
1547 * Add resource version info to xstat_fs_test and xstat_cm_test
1549 * Fix a delete vs delete[] usage problem in the WiX custom
1552 * When creating a mount point, validate the volume name to
1553 make sure it exists.
1555 * When processing a mount point, pay attention to VL_BADNAME
1558 * Remove a race condition from cm_volume_t recycling that could
1559 have resulted in two threads making use of the same object at
1562 * Fix a reference count leak on smb_vc_t objects introduced in
1563 1.5.50. Permit the objects to be freed once again.
1565 * Fix the output of "fs memdump" for smb_vc_t objects that was
1568 * Add routines for debugging smb refcount errors.
1570 * When shutting down the cache manager, do not permit AFS
1571 callback requests to be processed. Doing so can result in
1572 invalid memory accesses.
1574 * Do not recycle a deleted cm_scache_t object unless the
1575 refCount is 0. Doing so can result in a deadlock and
1576 invalid memory access.
1578 * Minimize the number of RPCs sent to the file server for
1579 directory bulk status requests. Only request status for
1580 objects that we are actually reporting on and only if we
1581 do not already have status for the object and then make
1582 sure that we return as much as we can at once via the SMB
1583 interface but not so much that it exceeds the RDR timeout
1585 * Enable bos restricted operations
1587 * Fix the creation of submounts used by the afscreds and afs_config
1590 * Avoid repeated retries when attempting to obtain access rights.
1591 If the access rights were obtained successfully but do not
1592 satisfy the request, fail the request.
1594 * Fix error code reported when attempting delete a file on a
1595 readonly volume or one that is marked with the readonly DOS
1598 * Fix a short name truncation bug.
1600 * Fix a heap corruption error when reading the CellServDB file
1603 * Fix directory listings when the directory has "lookup" permission
1604 only and the directory has never been previously accessed.
1606 * Fix several tracking issues for cm_scache_t rw locks.
1608 * Ensure that all buffers used to store cell names are the
1611 * Restore Windows 2000 support (broken in 1.5.50).
1612 LOCALE_INVARIANT was introduced in XP.
1614 * Add RxUdpBufSize registry value. Default is 256K.
1616 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
1618 * Do not include trailing NULs when returning file names in
1619 a directory search response.
1621 * Prevent RX from crashing the afsd_service by attempting
1622 to allocate to many buffers at once when memory is fragmented.
1624 * Avoid a deadlock when freeing cm_server_t objects
1626 * Optimize search cell file. Do not search to end of file
1627 if the cell has already been found.
1629 * Remove explorer shell extension delay on local disks.
1631 * Automatically pre-allocate 64 Rx packets per thread
1633 * Add smb lock requests and stat cache lock allocations to
1634 the output from fs memdump.
1637 * Fix Large File Support (broken in 1.5.50)
1639 * Fix "fs setacl" (broken in 1.5.50)
1641 * Fix pioctl path evaluation of drive letter
1642 mapped paths. (broken in 1.5.50)
1644 * Fix installation of VC8 Runtime library in the EXE
1645 Retail installer. (broken in 1.5.50)
1647 * Avoid a potential deadlock during Volume Location
1648 updates if all the file servers are unreachable at
1649 the time the VL update succeeds.
1651 * Ensure that rx connection object is references
1654 * Fix propagation of over quota errors during cm_FSync()
1657 Since 1.5.39 [1.5.50 released 16 July 2008]
1658 * There were no releases between 1.5.39 and 1.5.50.
1660 * FIRST RELEASE WITH UNICODE SUPPORT
1662 * Pioctl interfaces to the cache manager refactored
1663 to provide layering between the SMB specific code
1664 and the general purpose ioctl operation.
1666 * Garbage collect dead SMB virtual circuits as soon
1667 as they are no longer being referenced. This avoids
1668 problems with outstanding locks not being dropped
1669 when the virtual circuit becomes invalid.
1671 * Remove the IBM Administration Reference documentation
1672 and replace it with the OpenAFS Command Reference
1675 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
1676 functions each time a connection is about to be used.
1677 Do not hold a lock on the rx connection object while
1678 it is being selected. This avoids a race between
1679 threads attempting to set the timeout values and
1680 removes a bottleneck that was hampering performance.
1682 * Ensure that the smb directory attribute is set
1683 for all directory objects.
1685 * Add vs2008 support to the NSIS installer scripts
1687 * Replace the VC Runtime EXE installer with the MSI
1688 installer in the NSIS installer scripts
1690 * Properly delete the VC7.1 C Runtime libraries.
1692 Since 1.5.36 [1.5.39 released 23 June 2008]
1693 * There were no 1.5.37 or 1.5.38 releases for Windows
1695 * Use Visual Studio version of vsprintf() instead of
1696 home grown version that resulted in linking conflicts
1698 * Prevent buserver, ptserver and fileserver from crashing
1699 due to non-portable strftime() format strings.
1701 * Prevent server crashes caused by rx_InitMorePackets()
1702 being called prior to allocation of all of the required
1703 synchronization objects.
1705 * Fixed a memory leak in the internal B+ tree directory.
1706 For objects with non-8.3 names, the 'longname' memory
1707 allocation would be leaked during object deletion or
1710 * Fixed a memory leak in the background daemon cm_CheckServer()
1711 routine. The memory allocated to store the server list
1714 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
1716 * Correct an abstraction layer violation. cm_scache_t objects
1717 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
1718 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
1719 performed in cm_ReleaseSCache() when the reference count hits zero.
1721 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
1723 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
1726 * Ensure that scache object 'nextp' pointers are set to NULL
1727 when the objects are removed from the hash table linked list.
1729 * If a dirty buffer cannot be written to the file server because
1730 of an access denied error, mark the error in the cm_buf_t
1731 structure so that it can be dealt with instead of retrying
1735 * Update CellServDB file to GCO Public 23 Apr 2008
1737 * Fix a cm_buf_t reference count leak when attempts to write
1738 dirty buffers to the file server from within cm_IncrSyncer()
1741 * Prevent udebug from crashing.
1743 * Activate RX Idle Timeouts. If the file server is busy for more
1744 then 30 seconds, attempt to failover to another server without
1745 marking the busy server down.
1747 * Another VNOVNODE issue fixed. When writing a dirty buffer
1748 to the file server, if VNOVNODE is received, mark all buffers
1749 as invalid without further attempts to contact the file server.
1751 * Improved performance on high latency links.
1754 * Fix a bug preventing the re-initialization of the Freelance
1755 root.cell contents when a change is made.
1757 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
1758 or mountpoint that does not exist.
1760 * aklog, afscreds, integrated logon, and the NIM AFS credential
1761 provider will all attempt to first acquire an afs service ticket
1762 using the service principal: afs/<cell>@<CLIENT_REALM>. This
1763 permits the proper detection of the cell realm when multiple
1764 realms are local realms for the cell.
1766 * Fix fs commands that re-use file identifiers instead of forcing
1767 a path evaluation each time. This was broken in 1.5.33 when
1768 hashes were added to the cm_fid_t structure.
1770 * Fix symlink and mountpoint removal and then recreation when
1771 applied to the fake Freelance root.afs volume.
1773 * Fix SMB change notifications for the Freelance root.afs volume
1774 when creating or removing symlinks and mount points.
1776 * Force DNS AFSDB cell name lookups to lower case if the DNS
1777 response is not all lower case.
1779 * 32-bit installers are now built with Visual Studio 2005
1781 * All vos.exe commands now support the "-noresolve" option which
1782 forces the display of IP addresses instead of DNS hostnames.
1784 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
1785 permits operations that do not require up to date volume
1786 location data to continue while waiting for the RPC to complete.
1787 Multiple threads waiting for a volume location update on the
1788 same volume set no longer trigger multiple calls VL_xxx RPCs.
1790 * Volume server references could be marked offline and never
1791 reset to the not_busy state.
1793 * Further optimizations to cm_Analyze() retry logic for
1794 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
1797 * Optimize the DNLC by applying Interlocked incrementing for
1798 statistics gathering and enabling greater use of read
1799 locks instead of write locks
1801 * Further optimize osi_Log operations when logging is disabled
1802 by removing the need for the function call overhead.
1804 * Add a new lock operation, Convert Read to Write, and apply
1805 it throughout the cache manager. This function provides a
1806 fast transition from a read lock to a write lock when the
1807 caller is the only reader.
1809 * Fix a timeout problem when opening files or request locks
1810 when the file in question is already write-locked by another
1813 * Fix a deadlock that can occur if "fs flushall" is executed
1814 while applications are writing to AFS.
1816 * The cache manager would re-read data from the file server
1817 after file truncation when obtaining buffers to write to the
1818 same file until the first StoreData RPC completes.
1820 * Fixes error handling in cm_NTCheckDelete()
1822 * Replaces cm_scache_t mutex with a read-write lock permitting
1823 additional parallel access.
1825 * Prevent an cm_scache_t reference undercount when evaluating
1826 symlinks containing @sys when none of the active sysnames
1827 match any of the available targets.
1829 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1832 * Set RxMaxMTU registry value to 0 which is the equivalent of
1833 it not being set at all.
1835 * Reduce contention between BkgDaemon threads
1837 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1840 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1841 objects to use the faster Interlocked operations which avoid
1842 the need to hold write locks.
1844 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1845 non-fatal error that forces a retry against another service
1848 * Fix a crash in smb_WriteData that would occur if an application
1849 opened the file as read-only and then sent a write request anyway.
1852 * Respond to VL_NOENT errors by removing cm_volume_t object and
1853 setting it to be the first object to be recycled. This is not
1854 only a more efficient use of resources but it also prevents
1855 repeated attempts to query the VLDB server for the volume from
1856 the CheckOfflineVolumes routine.
1858 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1861 * A first cut at a cache manager statistics monitor that can be used
1862 to determine the necessary cache parameters to support the
1863 application working set.
1865 Off by default, the performance package can be activated by setting
1866 "daemonPerformanceTuningInterval" in the service Parameters key.
1867 As with the other daemon interval values the unit is in seconds.
1869 At service start and each succeeding interval the cache manager
1870 will write statistics to %TEMP%\afsd_performance.txt showing the
1871 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1872 The FID statistics keep track of all FIDs seen by the cache manager
1873 during the service session whether or not they are backed by any
1874 live object in the cache.
1876 These statistics are not stored in the cache file.
1878 * Fix windows event logging broken in 1.5.33.
1880 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1881 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1882 as well as when computing the server preference list.
1884 * When merging status after a StoreData operation prevent data
1885 buffers from being discarded that we would prefer to keep.
1886 (Bug introduced in 1.5.33.)
1888 * Modify installers the set RxMaxMTU to 0.
1890 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1891 dialog, the Background Daemons and Service Threads fields were
1894 * Increase default number of Background Daemons to 4.
1896 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1897 cm_volume_t objects to be recycled.
1899 * Windows Error Reporting has shown a number of previously unseen
1900 stack corruption errors in the Explorer Shell extension. The
1901 minidump appeared to indicate the problem was in ParseAcl().
1902 Reviewing the function it appears that numerous things could go
1903 wrong if invalid input was provided. Added error checking and
1904 data validation. Hopefully this will address the problem.
1906 * Windows Error Reporting has shown crashes in aklog as a result
1907 of krb5_cc_get_default() failures not being handled. They are
1911 * The Rx library used a 32-bit type for sockets which was
1912 truncating the socket value on 64-bit Windows. This
1913 was not an issue in general but became an issue when
1914 Rx "hot thread" support was enabled.
1916 * Enable "Rx hot thread" support which ensures that while
1917 a received message is being processed that a new thread
1918 is allocated to receive the next message to arrive.
1920 * Increased parallel access to smb_rctLock protected
1921 resources. Reduce the number of times that the
1922 lock is dropped and re-obtained.
1924 * Increased parallel access to cm_scache_t and cm_buf_t
1925 resources by replacing the previous global write lock
1926 protected reference counting mechanism with one built
1927 upon the use of InterlockedIncrement and InterlockedDecrement.
1929 * Implement a hash for cm_fid_t to reduce comparison costs
1930 and inline the cm_FidCmp() function.
1932 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1933 hash which has a better distribution
1935 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1936 the most recent read request instead of the next chunkSize.
1937 cm_chunkSize can be dozens or hundreds of buffers. As a result
1938 too much time is spent performing the evaluation.
1940 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1941 lock is to prevent the creation of new buffers while a truncation
1942 is being performed. All references to bufCreateLock have been
1943 removed except in two places:
1944 i. a write-lock surrounding the function that calls buf_Truncate()
1945 ii. a read-lock within buf_GetNewLocked() that actually allocates
1948 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1949 CM_CONFIG_DATA_VERSION number which value be used to force the
1950 replacement of the cache file contents when incompatible changes
1951 are made between releases.
1953 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1954 cm_SyncOpCheckContinue
1956 * Avoid calls to multi_Rx if nconns == 0
1958 * Modify smb_WriteData to perform background writes based upon the
1959 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1960 boundaries. This will slow down writes from the SMB interface but
1961 will avoid the risk of the CIFS client disconnecting from the AFS
1962 client SMB server. The default value for 'smb_AsyncStoreSize'
1965 * SMB AsyncStore functionality is on by default but can be
1966 disabled using the registry value:
1968 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1970 smb_AsyncStoreSize can be adjusted with
1972 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1974 The value must be a multiple of the buffer block size
1975 and cannot be larger than the chunk size.
1977 * Do not leak a read lock on the bufCreateLock
1979 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1982 * In MergeStatus, only remove clean buffers from the hash table.
1984 * Avoid the need to update the data version number on each buffer
1985 associated with a scache when MergeStatus is called after a
1986 StoreData by maintaining a range of valid data versions as part
1987 of the cm_scache_t object.
1989 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1990 from cm_HaveAccess... We can safely avoid the call so do so.
1992 * Add "ForceAnonVLDB" service parameters registry value. Default
1993 is 0. Set to 1 to force the use of anonymous connections to
1994 the VLDB servers. This is useful when using krb5 tokens and the
1995 VLDB servers have not been updated.
1998 * Do not attempt to search for a cell name that matches one
1999 of the special share names "svrsvc", "wkssvc", "ipc$"
2001 * Issue permission denied warnings from afscreds if attempts
2002 to modify the startup shortcut or registry values fail.
2003 This happens under Vista UAC.
2005 * Do not return BAD_NETWORK_ERROR in response to attempts
2006 to create directories or files in the Freelance root.afs
2007 volume. Instead return ACCESS_DENIED.
2009 BAD_NETWORK_ERROR was being sent because the Freelance cell
2010 has no vldb servers. Short circuit the error by testing for
2011 freelance in CreateFile and MakeDir
2013 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
2014 to be opened and closed without actually existing. For now
2015 we treat them as equivalent to ioctls but this should
2018 * Fix RAP processing to not reject names that exist in the
2019 root.afs volume for GetInfo queries.
2021 * In afscreds, use GetUserNameEx to obtain the username instead
2022 of searching the registry. Try UserPrincipalName first and
2023 fallback to SamCompatible.
2025 * Stop linking to rpcns4.dll which is no longer supported on
2026 Vista and 2008. TaAfsAdmSvr is going to have to be
2029 * When processing dfs referral requests, if the sharename is
2030 a partial match and therefore will not be added automatically
2031 to the share list, make sure that we return an error.
2033 * Avoid a crash when logging VNOVOL errors returned from
2034 cm_CheckOfflineVolume.
2036 * When trace logging is disabled, do not permit osi_LogSaveString
2037 to perform the string copying. By disabling the string
2038 copies, throughput of the service more than doubles.
2040 * when checking offline volumes, update the cell vlserver info
2041 if required and don't perform a check if there are no servers
2044 * Add volumes in the vl_unknown state to the list of objects to be
2045 checked as offline or retried when an allbusy error is received
2048 * Add support for configuring rx_ExtraPackets
2050 * All calls to buf_Get() must be protected by a read lock on the
2051 cm_scache_t bufCreateLock
2053 * When a MergeStatus results in the invalidation of buffers, do
2054 not remove buffers from the hash tables that have active references.
2056 * Upon receiving a VNOVNODE error from a file server, be sure to
2057 invalidate any file locks. This prevents on-going attempts to
2058 extend the locks and will cause subsequent attempts to access
2059 the file to receive a invalid handle error.
2061 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
2062 to directory entry names.
2064 * No longer discard tokens when RXKADBADTICKET error is received.
2065 Instead return STATUS_NO_KERB_KEY error.
2068 * Add "cmdebug <host> -cellservdb" option and client service
2069 implementation permits querying the vldb server for each
2070 cell known to the cache manager. The output is in a form
2071 that permits it to be used as a CellServDB file.
2073 * Add Symlinks->Show option to the AFS Explorer Shell Context
2076 * Prevent buffer overrun in the cmd parser used by command
2077 line tools such as fs, vos, pts, etc.
2079 * The path ioctl operations have several issues:
2081 (1) the specified path for "fs examine, whereis, whichcell,
2082 flush" and so always has the follow symlinks and mount
2083 points semantics. This makes it impossible to determine
2084 what the FID of a symlink or mount point is.
2086 (2) "fs examine" out is not the result of a single pioctl
2087 operation but is actually the combined output of half
2088 a dozen operations. Path evaluation is an expensive
2089 operation. It would be faster if the caller could
2090 evaluate the FID first and then perform all of the
2091 rest of the operations by specifying the FID instead
2094 (3) fs output reports all objects as files. By adding a
2095 GetFileType pioctl more informative output can be provided
2096 that indicates what type of object the path evaluates to.
2098 (4) the Windows fs command includes a number of commands that
2099 do nothing but exist only because the Unix cache manager
2102 A new extendible data structure cm_ioctl_query_opts_t has been
2103 added which can be optionally specified with pioctls that do not
2104 already require input data. The first two fields of this structure
2105 are 'literal' and 'fid'. The literal field is used to indicate
2106 whether the last component of the path should be evaluated following
2107 symlinks and mount points.
2109 The fid field permits a fid to be specified.
2111 * A new GetFileType pioctl has been added. The type of objects are
2114 * A new "-literal" option is available for "fs examine, flush, whereis,
2117 * Unimplemented fs commands have been removed.
2119 * There are circumstances where a volume object is being accessed and
2120 the volume is marked indicating that the volume location information
2121 is out of date but where it is also pointless and perhaps dangerous
2122 to block waiting for the rpc to complete. One example is when
2123 processing the cmdebug requests. If we know that we are not going
2124 to use the volume object to contact a server, then we can now set
2125 the CM_GETVOL_FLAG_NO_RESET flag.
2127 * when moving up in the directory tree we search the recorded fid list
2128 to find a matching fid that we have already crossed. we must also
2129 reset the fid count based upon what we discover.
2131 * Cell names are published as share names. As a result they are searched
2132 for as part of the Dfs Referral evaluation. If share "foo" can not be
2133 reached, the CIFS client will fallback to searching for "fo" as well.
2135 Since the freelance client automatically adds a symlink for prefixes
2136 of the cell name, this results in both "foo" and "fo" being added to
2137 the freelance root.afs volume. This patch prevents that negative side
2140 * when tracking Kerberos credential cache names be sure to include
2141 the cache type prefix.
2143 * Microsoft has assigned OpenAFS a network type value
2145 * The algorithms that were used to produce a Netbios name were broken.
2146 The name that was produced when the hostname was too long was in fact
2147 longer than the maximum permitted netbios name.
2149 Also, the "NetbiosName" value was not used as a suffix when the loopback
2150 adapter was not installed. The hardcoded string "AFS" was used instead.
2152 * avoid another deadlock during server probes initiated by the ipaddr
2153 change daemon thread.
2155 * prevent the afs client service from crashing when fs newcell is
2158 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
2159 to \Program Files\OpenAFS\Client\Program in order to get them out
2160 of the %windir% tree
2162 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
2163 end STATUS_OUT_OF_QUOTA can be returned to the application.
2165 * Modify the search order for determining the location of CellServDB
2166 and other client configuration files.
2168 1. AFSCONF environment variable
2170 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
2171 4. use Program Files\OpenAFS\Client
2173 * More Vista Power Management tweaks
2175 * Close a socket handle leak in the kauth routines
2177 * Don't hold a cm_volume_t mutex across RPCs
2179 * Convert downTimes to use clock time instead of relative times.
2180 This permits correct comparisons with expiration times.
2182 * Update the Wix installer to properly install the prerequisites
2183 for translate_et.exe
2185 * Do not translate pioctl paths as they are not converted by the
2188 * Implement multi_rx probes for checking whether servers are up
2189 or down. Turns probes into a constant time activity instead
2190 of one dependent upon the number of known servers.
2192 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
2193 update of the volume location data and retry. VNOVOL means
2194 that the volume could not be attached. It doesn't mean that
2195 the volume is not on this server.
2197 * Modify the Explorer Shell AFS Context menu Symlink->Add to
2198 not attempt to validate the target path. There is no reliable
2201 * Improve performance of \\afs\<cell><mount-type><volume>\
2202 access by eliminating extraneous attempts to search the
2203 CellServDB file and DNS AFSDB records. Share names
2204 containing '%' or '#' are not valid cell names.
2206 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
2207 to unlock a range that is not currently locked.
2209 * Enable the 32-bit tools to read the 64-bit registry keys
2210 when executing in the WOW64 environment.
2214 * Fix buffer refcount leak introduced in 1.5.29
2216 * Prevent the NIM AFS Provider from crashing when configured to use
2217 Kerberos v4 for token acquisition in situations where no Kerberos v4
2218 support is available. For example, 64-bit KFW or sites that delete
2219 the Kerberos v4 DLL.
2221 * In afsd_init.log, support logging cache sizes greater than 2GB.
2223 * The volume status output of "fs examine" was being determined based
2224 upon the wrong error value. Instead of using the pioctl() return
2225 code, the errno value must be used.
2227 * Prevent invalid pts auto-registration attempts by aklog and the
2228 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
2230 * Add synchronization protection to all of the SMB Listener State
2231 variables. This prevents race conditions when the addition or
2232 loss of a network adapter takes place.
2234 * On Vista, there is a race condition between the restart of the
2235 afs client service and the network adapters. Ensure that the
2236 afs client service can handle switching between loopback and
2237 non-loopback modes of operation.
2239 * Fix FollowBackupPath to work with mount points that are not
2240 explicitly read/write.
2242 * Fix a deadlock that has been observed on Vista while resuming
2243 after at least four hours of sleep.
2245 * Add new HKLM registry configuration options that can be used
2246 to provide configuration hints to the NIM AFS Provider when
2247 creating new identities. See release notes for details on
2248 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
2251 * Speed up write buffer operations by holding a global read-lock
2252 instead of a write-lock when updating the buffer data version
2253 number. The data version is protect by the associated stat
2256 * Speed up read buffer operations by removing out of date buffers
2257 from the buffer hash tables.
2259 * Protect against a null pointer dereference in afslogon.dll
2260 if QueryAdHomePathFromSid() fails.
2262 * Remove a race condition while accessing private freelance data
2265 * Update the buffer data versions when Freelance directory data
2268 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
2271 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
2272 instead of returning zero length.
2274 * Return ERROR_RETRY when a server sends VBUSY.
2276 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
2277 a volstat plug-in is in use.
2279 * Enhance the volstat interface to support Dfs path mappings.
2281 * Fix "symlink list" to do the right thing for Dfs links
2283 * Fix callback expiration that was broken in 1.5.27.
2285 * Fix a crash at shutdown caused by the lack of synchronization
2286 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
2288 * Hold the cm_server_t mutex while walking or modifying the
2289 cm_server_vols_t list in order to avoid a race condition.
2291 * Ensure that the cm_server_t refCount is safely modified.
2293 * The pthread library implementations of pthread_cond_wait
2294 and pthread_cond_timedwait were failing to return with
2295 the mutex held when error conditions occurred. This could
2296 result in race conditions wherever they are used. One case
2297 in which problems were seen was the Rx event_handler().
2299 * Prevent cm_GetBuffer() from locking all buffers associated with
2300 a stat cache object before deciding if the current buffer is up
2303 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
2304 object for each offset that we will attempt to reserve.
2306 * Reduce the default number of background daemon threads used to
2307 prefetch or store to one in order to prevent two or more threads
2308 from stepping on each other.
2310 * Reduce the default chunk size to 256KB in order to prevent read
2311 timeouts on low bandwidth connections.
2313 * Fix a race condition that could result in threads becoming stranded
2314 when synchronizing multiple requests on the same stat cache object.
2317 * The GiveUpAllCallBacks functionality that was added in 1.5.21
2318 has been disabled. It has been disabled because it turns out
2319 that the RPC is improperly implemented in all file servers
2320 previous to 1.4.6. The RPC is executed without the proper
2321 locks being held resulting in data corruption and eventual file
2322 server failure. The functionality will be re-enabled in a
2325 * a couple of .readonly Volume Callback optimizations were
2327 - Apply most recent volume callback to all stat cache objects
2328 in the volume. This avoids premature callback expirations
2329 and unnecessary FetchStatus calls.
2330 - Add option to permit attempts to renew .readonly Volume
2331 Callbacks on a periodic basis to avoid invalidation of stat
2332 cache data. [Requires daemonCheckVolCBInterval be set to
2333 non-zero number of seconds. 1800 (30 minutes) is suggested.]
2335 * Adds support for 64-bit data version values. Previous releases
2336 only supported 32-bit values. Larger values sent by the file server
2337 would result in rollover.
2339 * aklog and asetkey no longer generate an exception if Kerberos for
2340 Windows is not installed.
2342 * If the LSA authentication functions report a temporary out of memory
2343 condition, fail the current authentication and do not cause the AFS
2344 client service to terminate unexpectedly.
2346 * When the AFS client service is shutting down, prevent the daemon check
2347 thread from executing additional checks which might attempt to access
2348 discarded locks or unmapped memory.
2350 * Correct a double free condition in the NIM AFS credentials provider.
2352 * Add "FollowBackupPath" service parameter registry value.
2353 This provides equivalent functionality to the UNIX afsd -backuptree
2354 option. When set to a non-zero value, normal mount points originating
2355 in a .backup volume will prefer a .backup volume.
2358 * Fix NSIS uninstall from the .MSI installer.
2360 * Add Stack Frame logging for AMD64
2362 * Add support for EWOULDBLOCK to cm_Analyze. If the file
2363 server returns EWOULDBLOCK, retry the request every two
2364 seconds for up to the RDRtimeout.
2366 * Reorganize the order of the includes to ensure that EWOULDBLOCK
2367 is not assigned the same value as EIO.
2369 * Update CellServDB to GCO Public 25 Oct 2007
2371 * Add additional validation and error handling code after
2372 each call to getSlot(). If an invalid slot is returned,
2373 return NONODE. If the invalid slot is returned when
2374 extracting a data node, invalidate the tree.
2376 * Modify compareKeys() to always perform a case-insensitive
2377 comparison and only perform a case sensistive comparison
2378 if the case-insensitive one matches. This ensures the
2379 ordering is consistently reported.
2381 * Add lock assertions to ensure that all calls are being
2382 performed with the correct locks being held. There have been
2383 some crash reports that provide stack data that does not appear
2384 to be possible unless there is a race. However, there are no
2385 obvious locations where the race is taking place and the test
2386 suite indicates that all of the correct locks are being held.
2387 We shall see what happens in the field.
2389 * For consistency replace all calls to findKey in which the range i
2390 s (1,numentries) with calls to getSlot().
2392 * Optimize the depth search loop by testing the slot value in the
2393 for statement instead of forcing the loop to be broken later.
2395 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
2396 There are a number of locations where locks are obtained, dropped,
2397 and reobtained. This reorganization attempts to accomplish several
2400 (1) be optimistic for the most common case so it will be fast
2402 (2) add consistency checks after each location where locks are
2403 dropped and re-obtained. If we lose a race in cm_BeginDirOp
2404 and the bplus tree is out of date, retry until we get to a
2405 consistent state that we can use.
2407 (3) Ensure that all operations take place with the correct locks.
2409 * One of the issues that has become a serious problem since the
2410 addition of the local directory updates is that although cm_SyncOp
2411 synchronizes operations, it does not preserve the order of requests.
2412 This has always been a problem in that it has been possible for a
2413 request to fail to complete due to its worker thread's bad luck.
2414 When a request takes longer than the Windows SMB Redirector's timeout,
2415 the SMB Redirector tears down the SMB virtual circuit.
2417 When using the local directory updates it is really important that
2418 the directory update operations complete in the order that they were
2419 sent to the file server. If they don't, then the local directory state
2420 and the file server state will not match and the local directory state
2421 must be discarded which in turn forces a new read of the entire
2422 directory contents over the network.
2424 This patch adds a new cm_scache_waiter_t object that is used to store
2425 the current thread, buffer, and syncop flags within a waiters queue
2426 on each cm_scache_t object. If a thread is forced to sleep in
2427 cm_SyncOp, upon waking it will check to see if there are any other
2428 threads waiting that are attempting to perform a similar task ahead
2429 of it in the queue. If yes, the thread goes back to sleep. If not,
2430 it goes ahead and enters the cm_SyncOp conflict resolution block.
2432 This patch has the additional side effect of reducing the number of
2433 competing threads that must obtain the cm_scache_t mutex and process
2434 the cm_SyncOp conflict resolution block. As a result, the overall CPU
2435 utilization of the service and the clock time associated with processing
2436 requests will be reduced.
2438 * assert that the cm_scache_t mutex is held when calling
2439 buf_ForceDataVersion().
2441 * rename findNode to leafNode in bplus_Lookup
2443 * replace all OutputDebugString calls with osi_LogX calls
2445 * modify bestMatch to special case the return values for leaf nodes.
2446 If an entry is above or below the values available in the leaf node
2447 return BTLOWER or BTUPPER instead of BTERROR.
2449 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
2450 if true convert to either slot 0 or Max and perform the insertion.
2451 This produces easier to read code when performing lookups.
2453 * modify lock_AssertXXX macros to call osi_assertx() and provide a
2454 descriptive message.
2456 * the difference between osi_assert() and osi_assertx() is whether or not
2457 a description is specified as part of the call. When no message is
2458 specified afsd_notifier() is called with a NULL msgp parameter. This
2459 results in a null pointer dereference during the EventLog() call.
2461 As a result, none of the information describing the assertion is written
2464 This commit sets a default message for use when no other message is
2465 provided and it replaces all calls to osi_assert with osi_assertx and
2468 * Migrate B+ tree search key into thread local storage
2471 * Fix an uninitialized variable in lana_IsLoopback() which
2472 could result in a random determination that a loopback
2473 adapter is not a loopback adapter. This would result in
2474 a computed Netbios service name of %MACHINE%-AFS instead
2475 of AFS. This is a serious problem if the value computed
2476 by the AFS Client Service and the client process pioctl
2477 calls produce different results as it will result in a
2478 communication failure.
2480 * Fix a crash in the B+ Tree code that results when a file
2481 rename only results in a change of case.
2483 * Fix the pioctl path parsing. The previous algorithm
2484 broke the current directory of the client into two parts,
2485 the TIDpath which represented the device root and the
2486 relative path. The problem is that this is not sufficient
2487 from proper processing of symlinks and mount points.
2489 What is required is that the last component of the path
2490 must not have the symlink or mount point be evaluated
2491 whereas symlink and mount point evaluation must be performed
2492 for the rest of the relative path. Therefore, the path is
2493 now split into three parts: TIDpath, intermediate path,
2496 "fs examine" now reports the data of the symlink or mount
2497 point object instead of the target and "symlink list" and
2498 "symlink remove" now works. The "Explorer Shell Extension"
2499 menu now is displayed for symlink objects.
2501 * More informative symlink.exe error messages.
2503 * When the cmd.exe "del" command is used to remove a symlink,
2504 not only was the link object removed, but the stat cache
2505 object for the target was marked deleted. This resulted
2506 in cache corruption which would require the destruction
2507 of the cache to correct.
2509 * Add undocumented registry value, "OfflineReadOnlyIsValid"
2510 which results in cached contents from read only volumes
2511 to be treated as valid when there is no callback registered
2512 and all servers are down or all instances are offline.
2514 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
2516 * Add documented registry value, "AcceptDottedPrincipalNames"
2517 which is used to permit the acquisition of AFS tokens
2518 using Kerberos v5 principal names which include a dot in
2519 the first component.
2521 {HKLM\HKCU}\Software\OpenAFS\Client
2523 * "fs uuid" does not require "AFS Client Administrator" group
2524 membership to display the current UUID value.
2527 * Update NSIS installer to version 2.30
2529 * Update Wix installer to version 2.0.5325
2531 * 64-bit KFW support:
2532 - Network Identity Manager plug-in
2534 - 64-bit krb5 enable afscreds.exe, afs server manager,
2537 * Enable B+ tree Directory Search by default.
2539 * NIM plug-in is now compatible with FILE: ccaches
2541 * Install instloop.exe into Client/Programs/ even though
2542 it isn't required. Permit it to be used by organizations
2543 that clone Windows systems.
2545 * On a suspend, only drop callbacks for servers that
2546 are online. If the server is offline, maintain the callback
2547 until the server is once again available.
2549 * Prevent against acquisition of a directory write lock when
2550 there are no updates to be performed. This was resulting
2551 in a write-lock leak which later on would lead to a deadlock.
2553 * Add configurable file pre-fetch mode. For files that match
2554 the specified file extensions, pre-fetch the entire file into
2557 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
2558 (MULTI_SZ) PrefetchExecutableExtensions
2560 * Fix the client_osi large integer support so that return values
2561 can be properly assigned to large integer variables.
2563 * Do not hold cm_scache_t mutex across cm_GetCell() calls
2564 since cm_GetCell() can block in a RPC.
2566 * Add short file name support to B+ tree searches.
2568 * Resolve a conflict between the desire to notify file servers
2569 that the client workstation is suspending and the desire to
2570 permit objects with valid callbacks when the client suspends
2571 to continue to use those objects from the cache if the client
2572 resumes without network connectivity.
2574 * Fix stack variables to ensure that all buffers storing paths
2575 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
2576 would cause a stack overflow condition.
2578 * On Vista, force afs_config.exe to run with the highest privileges
2579 available to the end user. For a user in the Administrators
2580 group running under UAC, this will cause afs_config.exe to be
2581 be run with full Administrators privileges after the user
2582 approves the privilege escalation.
2584 * More VNOVNODE handling issues. If a dirty buffer cannot be
2585 written due to VNOVNODE, clear the buffer and ignore the error.
2586 If a cm_scache_t object has been marked deleted due to a VNOVNODE
2587 error, the next time the smb_fid_t object that references it is
2588 used return either NOSUCHFILE or NOSUCHPATH and close the
2589 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
2590 result in an invalid handle error.
2592 * For Vista, set the RequestExecutionLevel of the NSIS installer
2595 * For the build environment, support the generation of manifests
2596 when using the CL1310 (VS .NET 2003) compiler/linker.
2598 * There was a bug in OutputDebugHexString() which could overflow
2599 a stack buffer resulting in unexpected process termination
2600 of afsd_service.exe if the input Security Token contained
2601 many bytes equivalent to the '%' char.
2603 * Add directory lookup statistics to "fs memdump" output.
2607 * Windows uses case-insensitive file name pattern matching
2608 but AFS is a case sensitive file system. The AFS3 directory
2609 format is block based, uses network byte order and
2610 includes a hash table for fast case sensitive lookups.
2611 This causes several problems for the Windows AFS client.
2612 (1) Traversing the directory blocks is cpu expensive
2613 (2) A hash table miss does not indicate that the desired
2614 entry does not exist.
2615 (3) Determining whether a non-ambiguous inexact match or
2616 the entry does not exist requires a linear traversal
2617 of the entire directory.
2618 These issues often result in 100% CPU utilization.
2620 These issues are addressed by building a modified B+ tree for
2621 each directory and then using the B+ tree for searches.
2623 This feature is disabled by default. To enable it set
2624 the "BPlusTrees" registry value to 0x01 in the AFS Client
2625 Service Parameters key.
2627 * Correct a null pointer dereference in the symlink recursion
2628 detection code. (Introduced in 1.5.22)
2630 * Add the Network Identity Manager AFS Provider to AMD64
2635 * A new registry value, "BlockSize", can now be used to adjust
2636 the size of the internal buffers used to store file data.
2637 The default is 4KB. When set, the values 1..1024 specify
2638 multiples of 4KB. Values > 1024 specify actual block sizes
2639 and must be multiples of 4KB. The actual block size in KB
2640 can not be larger than the chunk size which defaults to 1MB.
2642 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
2645 * pts.exe and ptserver.exe built with -DSUPERGROUPS
2647 * Service Manager shutdown hints have been added. The hint
2648 specifies that it may take the AFS client service up to two
2649 minutes to shutdown.
2651 * Fix the recursion detection feature introduced in 1.5.22.
2652 The previous implementation broke the evaluation of paths
2653 using ../path or ./path. It also broke the evaluation of
2654 symlinks to absolute paths. For example, /afs/cell/path.
2656 * Fix an uninitialized variable in cm_MakeDir that could result
2657 in crash if the requested directory could not be created.
2659 * Fix a the VolumeStatusNotification log messages to save
2660 the input string in the circular log buffer so garbarge
2661 won't be output to afsd.log when the log is dumped.
2664 * Increase the SMB Maximum Path length to 260 characters
2665 from 256 as that is what CIFS supports.
2667 * On 64-bit Windows, use a unique description string for
2668 the 32-bit AFS Shell Extension.
2672 * In the AFS Admin Library, replaced all of the
2673 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
2674 allows parameter checking and proper type conversions to be
2675 performed. This is crucial when passing parameters that are
2676 64-bit wide on systems in which 'long' is a 32-bit value.
2677 Otherwise, the parameter list on the stack is interpretted
2678 incorrectly and the function not only doesn't succeed but it
2679 might corrupt memory and if lucky crash the application.
2681 The only program that we care about that uses this library
2682 is the AFS Server Manager. Now it doesn't crash.
2684 * Fixed smb directory searches for non-wildcard names. The
2685 optimized function should not have set the bulk in progress
2686 flag. This would adversely affect later directory searches
2687 that required a bulk search.
2689 * Detect infinite recursions (up to 512 stat cache objects
2690 in a chain.) If an infinite recursion is detected, return
2691 an error to stop the CIFS client from querying us forever.
2693 * Executables executed out of AFS name space can have their
2694 pages swapped out by Windows. Those pages will be reloaded
2695 when they are required. However, if the cache manager no
2696 longer has a callback and all of the file servers with
2697 the volume data are down, then the application will crash
2698 due to the inability to read the required page.
2700 We now track the down time of each server. If the callback
2701 expires after the server that issued it went down and all
2702 servers are down, we delay the expiration to permit the data
2703 in the cache to be continue to be used by Windows or an
2706 * The Windows AFS client has always had very poor performance
2707 with regards to directory management. In the case of
2708 directory updates caused by create, rename,and delete operations,
2709 the AFS cache manager will now perform local directory updates
2710 when possible in order to avoid re-reading the directory from
2711 the file server. This will be done in whenever both of the
2712 following conditions are true:
2714 - the data version on the directory increased by one as a
2715 result of the operation
2717 - all of the directory buffers necessary to make the change
2718 are present in the cache
2720 * The error message in the AFS Control Panel describing the
2721 need to be a Windows Administrator has been replaced with
2722 a message describing the need to be in the AFS Admins Group.
2724 * If the vldb server list is empty when attempting to evaluate
2725 a cell mountpoint it was possible for a crash to occur. This
2726 can only happen if there is an existing mount point and the
2727 is a failure attempting to load the server list from the
2728 CellServDB file or via AFSDB queries.
2730 * dirty buffers only have their dirty bytes stored to the file
2731 server instead of the entire chunksize
2733 * the default chunksize has been increased from 128KB to 1MB
2738 * changed the enum values for cm_serverRef_t state info to use a
2739 private name space to avoid collisions (srv_)
2741 * added a srv_deleted state for cm_serverRef_t objects. This
2742 state is set when cm_FreeServerList() is called with the
2743 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
2744 not always delete the cm_serverRef_t from the list if it is
2745 still in use by another thread. the srv_deleted state means
2746 the object's contents are no longer valid and it must be
2747 skipped. It will be deleted the next time the object is
2748 freed and the refcount hits zero.
2750 * the srv_deleted state is also used when a file server reports
2751 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
2752 as offline. This is done to prevent additional usage of the
2753 stale vldb data while waiting for the update volume request
2756 * added a state field to the cm_volume_t object (enum volstate
2757 vl_ name space) that maintains the state of the volume based
2758 upon the states of all of the cm_serverRef_t and cm_server_t
2761 * modified cm_UpdateVolume() to set the state of the cm_volume_t
2762 RW, RO, and BK to either vl_alldown or vl_online. There can't
2763 be any other states because cm_UpdateVolume() destroys any
2764 previous knowledge we might have had regarding busy or offline
2767 * modified cm_UpdateVolume() to update the volume name in the
2768 cm_volume_t to the volume base name if the previous value was
2771 * modified cm_FollowMountPoint() to check to see if the volume
2772 name is a volume ID and if so call cm_GetVolumeByID instead
2773 of cm_GetVolumeByName. This ensures that volume IDs are always
2774 looked up as numeric values. There is no longer a need to
2775 maintain a separate cm_volume_t containing the string representation
2778 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
2779 The first flag is a "CREATE" flag which is set by all existing
2780 calls. The flag is not set by calls to cm_GetVolumeByID() from
2781 the server probe code when volume status is being updated. We
2782 do not want the server probe operation to result in additional
2783 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
2784 which is set when the server probe code updates the volume status.
2785 This flag will be used to prevent the server probe operation from
2786 changing the order of the least recently used queue.
2788 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
2789 allocated for a given set of normal, readonly, and backup volumes
2790 regardless of whether or not the volume is accessed via name or
2791 ID number. The cm_volume_t namep field is always the base name
2794 * Added a new volume state, vl_unknown. This state is used as
2795 the initial state for all cm_volume_t when the cache manager starts,
2796 for each cm_volume_t at creation, and for each cm_volume_t when
2797 recycling. The cache manager does not know the state of all
2798 volumes in the world, only those that are in the cache and for
2799 which it has queried the VLDB and hosting file servers.
2801 * modified cm_GetVolumeByName() to initialize the state of a
2802 volume to vl_unknown. The actual state will be set when a
2803 cm_VolumeUpdate() call completes successfully.
2805 * changed name of scache hash table variables to avoid ambiguity
2806 when adding hash tables for volumes
2808 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
2810 * modified cm_UpdateVolume() to handle the case in which there is
2811 no RW volume but there is are RO volumes for a given base name.
2812 This is done by querying for the ".readonly" volume name if the
2813 base name does not exist in the VLDB. We never query for the
2814 .backup name because under the current usage model a .backup
2815 volume may only exist on the server that the read-write volume
2816 is located. If there is no RW volume, there can be no .backup.
2818 * Added four hash tables for cm_volume_t objects to improve the
2819 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2820 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2821 BKID. Three ID hash tables are necessary as long as it is
2822 desireable to maintain a single cm_volume_t containing all
2823 of the related RW, RO, and BK volume data. Having the RW and
2824 RO volume data in the same object is necessary for the
2825 implementation of cm_GetROVolumeID() which returns either the
2826 RO or RW ID depending upon the existence of RO volume instances.
2828 * Added a volume LRU queue so that volume reuse becomes fairer.
2829 This does not replace the all Volumes list which is used when
2830 it is desireable to walk a list of all the volumes whose order
2831 is not going to change out from underneath you which makes it
2832 safe to drop the cm_volumeLock.
2834 * handles volume hash table updates where volume name to
2835 volume ID number changes. The volume name remains
2836 constant in the cm_volume_t. if a vos rename is performed,
2837 the name of the volume will change and the volume IDs will be
2838 updated. Subsequent access to the old volume ID will create a
2839 new cm_volume_t with the new name.
2841 * Added a daemon thread operation to query the state of volumes
2842 listed as busy or offline. cm_CheckBusyVolumes() calls
2843 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2844 or vl_offline. If the volume is now online, the status on the
2845 volume is updated. The default period is 600 seconds. This can
2846 be configured with the BusyVolumeCheckInterval registry value.
2848 * Added prototype for smb_IoctlPrepareRead() which was missing a
2849 return type in the function definition.
2851 * Added volume id lists to the cm_server_t. These lists are
2852 allocated in blocks of ~32 IDs. When a cm_PingServer()
2853 detects a change in server state, the state of the cm_volume_t
2856 * Added volID to the cm_serverRef_t object. volID is used
2857 to identify the volume for which the object is a referral.
2858 cm_FreeServerList() uses the volID to remove the volume
2859 from the cm_server_t.
2861 * In cm_Analyze, when VNOVOL or VMOVED are received,
2862 call cm_ForceVolumeUpdate() to force a refresh of the volume
2865 * Added cm_VolumeStatusNotification() which is used at the moment
2866 to log volume status changes to the trace log. It will also
2867 be used as the access point to the File System Filter driver
2868 notification engine.
2870 * Added an all cm_scache_t list to cm_data. This replaces the use
2871 of the stat cache LRU queue when we need to enumerate all
2872 entries. The LRU list order is not static and when using it to
2873 enumerate all entries it can result in items being missed or
2874 items being processed more than once.
2876 * Modified cm_Analyze(). Instead of reseting the busy or offline
2877 state of a volume and forcing a retry of the operation
2878 cm_Analyze will defer to the background daemon thread that will
2879 update the state once every 600 seconds.
2881 * Added the automatic generation of a Freelance ".root" read-write
2882 mountpoint that refers to the root.afs volume of the workstation
2883 cellname at the time the mountpoint is created.
2885 * re-write cm_Analyze to make better use of the known volume
2886 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2889 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2890 busy volumes will be reset to srv_non_busy by the function
2891 but there is no mechanism for querying the busy state other
2892 than by attempting to access the resource.
2894 * cm_Analyze will query the state of an offline volume before
2895 deciding whether or not to retry when all volume instances
2898 * the algorithm used after an IP address list change detection
2899 has been modified. we now check the state of the VLDB servers,
2900 then reset the rx connections, and finally check the state of
2901 the file servers. Previously we flushed the rx connections
2902 and then checked the servers in whatever random order they
2903 were first discovered.
2905 * fix the volume status reported by "fs examine"
2907 * various improvements to the AFS Network Identity Manager Provider.
2908 - Don't access freed objects after the window has been destroyed.
2909 - Write back the AFS cell list in the identity configuration
2910 even when its empty.
2911 - Disable all use of the provider when AFS is disabled for a
2912 particular identity.
2914 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2915 alias instead of the full cellname, double check the full cell
2916 name and see if we have a matching cell before allocating a new
2919 * Dump more data with "fs memdump". All volumes and scaches and
2922 * Reverse the order of the all cells list. This ensures that
2923 the workstation cell is always first in the list.
2925 * If we recognize the server from which a callback revoke was
2926 received, then we can ensure that we only invalidate the status
2927 information for cm_scache_t objects belonging to that cell.
2929 * When building for amd64, use APPVER 5.02 as we do not support
2930 64-bit Windows 2000 and explicitly state the machine type in
2933 * no longer treat VMWare network adapters as equivalent to a
2934 loopback adapter. For some unknown reason, publishing the "AFS"
2935 netbios name on them frequently results in a name collision
2938 * Use read locks whenever possible instead of write locks when
2939 walking the allCellsp list. This permits better use of multiple
2942 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2943 is not present on the directory optimization. It would report
2944 mount points and directory objects for which there was no
2945 status information as a zero length file. Instead we can use
2946 the vnode number of the object to determine if it should be
2947 considered to be a directory or a file. If it is odd it is a
2948 directory; even it is a file.
2950 * Give Back Callbacks to the file servers in response to suspend
2951 and shutdown power management events. Only give up callbacks
2952 to UP servers and do so with a very short RX dead timeout.
2955 * Fix a race condition in the SMB FID cleanup code that could
2956 result in a stat cache mutex never being released.
2958 * Fix the graphics files that are installed as part of the
2959 IBM Administrators Guide.
2961 * Change the behavior of pioctl path parsing to not follow
2962 symlinks or mount points during evaluation. This results in
2963 the FID being the symlink or mount point and not the object
2964 they point to. This is consistent with the behavior of the
2967 * Enable AFSDB record lookup for command line tools such as
2970 * CellServDB update 2007-07-05
2972 * Added a Volume Status Plug-in Interface that permits third parties
2973 to receive notifications when the volume state changes occur.
2975 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2977 * Prevent the afsprocmgmt.dll from being unloaded while the
2978 background thread it starts is running.
2980 * Avoid a race condition when closing smb_fid_t objects.
2983 * fix install path of IBM Quick Start Guide when installed by
2986 * fix the client so that it can write into directories with
2987 only 'li' permissions.
2989 * fix installation of 32-bit Shell Extension for 64-bit systems.
2990 Install it under its own HKCR keys.
2992 * modify the wix installer to remove and upgrade install packages
2993 with the same version number.
2995 * when flushing volume location data every two hours or in
2996 response to "fs checkvolumes", also flush the mount point
2997 links in order to force their re-evaluation.
2999 * aklog was copying an uninitialized buffer as part of the
3000 foreign cell automatic pts registration
3002 * when obtaining directory listings, check for read permission
3003 on the directory. If read permission does not exist, fake the
3004 status information for the directory entries. This improves
3005 performance, prevents the client from pounding the file server,
3006 and prevents leakage of cached status information.
3008 * "fs getcacheparms" output was incorrect.
3010 * NetIdMgr AFS Provider should only automatically add token
3011 acquisition for cell foo.com if the realm of the cell matches
3012 the realm of the identity.
3014 * Fix a null pointer dereference in cm_MergeStatus()
3018 * network identity manager provider:
3020 - Correct for the fact that pr_CreateUser tampers with the
3021 contents of the username input buffer.
3023 - Take into account that the krb4 credential data structure
3024 does not have separate realm fields for the user and service.
3025 When krb524 is used with a cross-realm ticket, the realm
3026 describes the service not the user.
3028 * Many integrated logon changes:
3030 - Fix eventlog reporting. Do not attempt to log an event
3031 if the event source registration fails. Use DebugEvent0
3032 instead of DebugEvent when there are no parameters.
3034 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
3037 - Fix the reading of Domain specific configuration for
3038 LogonScript and TheseCells. Previously the dwSize value was
3039 being overwritten so that subsequent RegQueryValueEx call
3042 - Fix a memory leak in the TheseCells reading code.
3044 - Add support for Domain specific "Realm" specification. The
3045 realm is the realm to be appended to the username. When
3046 logging in as a domain or to the local machine, the specified
3047 "Domain" name is not going to be a valid realm name.
3049 - Construct a proper principal name based upon the domain
3050 specified realm for use in obtaining tokens with KFW.
3052 - If the domain specified "TheseCells" list includes the
3053 default cell, do not obtain tokens twice.
3055 * An unprotected call to the Kerberos v5 function,
3056 krb5_free_context(), can result in an invalid memory access
3057 exception if the contents of the krb5 profile, krb5.ini, cannot
3058 be parsed correctly. The error is located within the afskfw
3059 library which is used by afscreds.exe, afssvrmgr.exe, and
3060 afslogon.dll. When the error is triggered within afslogon.dll
3061 a blue screen condition can occur when the winlogon.exe system
3062 process terminates unexpectedly.
3066 * Do not stop the afs client service with an assert
3067 if the cell name cannot be determined when responding
3068 to a "fs whereis" command.
3070 * Integrated Logon (afslogon.dll). No longer attempt
3071 to preserve Kerberos 5 tickets for the logon session
3072 as that job is now supported by KFW 3.1 and above.
3074 * The Network Identity Manager plug-in will attempt to
3075 remove expired service tickets from the cache.
3079 * CellServDB update - GCO Public CellServDB 09 Mar 2007
3081 * AFS Server Manager and AFS Account Manager data cache
3082 moved from HKCU\Software\TransarcCorporation to
3083 HKCU\Software\OpenAFS to prevent crashes due to data
3086 * Reduce the default UP server probe time from 10 minutes
3087 to 5 minutes to match UNIX. This will not solve the NAT
3088 port mapping timeout issue because most NATs default to
3089 a value less that 5 minutes and this value ensures that
3090 the afs client server will start probes 5 minutes after
3091 the last probe set ended. The probe period for any one
3092 server is therefore longer than 5 minutes. The more
3093 servers in the list the longer the time between probes.
3095 * Before marking a server as down in cm_Analyze, retry the
3096 request using a new rx connection to the server. If the
3097 cache manager is behind a NAT, it is possible that the
3098 response from the server is being sent to the original
3099 address:port from which the rx connection was established.
3100 This will appear to the cache manager as the server being
3101 down when in fact it is not. Forcing the use of a new rx
3102 connection will result in the server responding to the
3103 most recent address:port assigned to the client by the NAT.
3105 If the server is in fact down, it will simply take a little
3106 longer to notice it.
3110 * Removing a directory that contained entries was failing with
3111 the wrong error code being returned to the CIFS client.
3115 * rxdebug exponential back-off. don't take down a server that
3116 is already struggling to respond.
3118 * Network Identity Manager plug-in registers help on the Help
3119 menu if it is supported by NIM. (KFW 3.2 and above)
3121 * Plug-n-play networking did not always work after coming back
3122 from a Windows Suspend. The AFS client service failed to
3123 re-compute the LAN adapter to bind to and instead attempted
3124 to bind to all LAN adapters. This would polute the public
3125 network with the name "AFS" which could lead to conflicts
3128 * Build installers using Microsoft Hotfixes provided at
3129 http://tinyurl.com/243svk for Visual Studio C RunTime Library
3130 Daylight Savings Time issues.
3132 * "fs checkserver" does not require admin privileges unless
3133 the probe interval is being changed.
3137 * A reference leak was discovered in the mountpoint traversal
3138 functionality which resulted in cm_volume_t objects not becoming
3139 eligible for re-use. This will trigger a panic condition when
3140 more than the maximum number of volumes is accessed.
3142 * Fixed the success test for the "fs memdump" operation. Success
3143 is now reported as such. Added cm_volume_t object information
3146 * A data corruption error has been fixed. Write requests
3147 queued for background processing were not retried upon failure.
3148 This leads to inconsistencies between the contents of the file
3149 in the local cache and the contents of the file on the file
3152 The implemented fix is to re-queue any request that has failed
3153 due to timeout, volume busy, volume offline, server busy, or
3154 server offline. Requests are not processed while all servers
3155 for a volume are inaccessible. When servers become accessible
3156 the background request will be retried.
3158 In addition, if a background write fails, do not mark the buffers
3159 as clean. Failure to preserve the dirty flag prevents the buffers
3160 from being written to the file server in the future.
3162 This fix does not address the possibility that the credentials
3163 necessary to perform a write request might expire between the
3164 time the request is queued and when the request can be processed
3165 due to server accessibility.
3167 * The Freelance volume does not support per user ACLs. The
3168 cm_HaveAccessRights function did not handle this special case.
3169 Requests for privileges greater than read|lookup would result
3170 in afsd_service.exe entering an infinite loop. ISOBuster is
3171 known to cause this behavior.
3173 * When deleting files, the SMB Server would attempt to obtain
3174 the access permissions for the parent directory instead of
3175 for the file itself. This would result in an incorrect
3176 assessment of whether or not the file can be deleted.
3178 * The Network Identity Manager AFS plug-in has had its string
3179 tables updated for consistency with the language used by NIM.
3180 All credentials are referred to as credentials instead of tokens.
3181 Drop-down combo boxes now in fact drop down.
3183 * The AFS Salvager was not properly being built on Windows.
3184 This is relevant only if the AFS servers are in use.
3186 * The default signal handlers for the AFS servers incorrectly
3187 processed SIGQUIT. Instead of stopping the server, an
3188 exception was raised.
3190 * Fix another volume refcount leak. This one was triggered by
3191 an inability to contact the vlservers in order to update the
3194 * Fix checklist control used by the drive mappings tabs in
3195 afscreds.exe and afs_control.exe.
3197 * Improve output of "fs memdump" related to buffer management.
3199 * Fix refcount leak of cm_buf_t objects when prefetching buffers
3200 already in the process of being fetched.
3202 * Improve cache performance by computing the buffer hash table
3203 size based upon the number of buffers so that the average
3204 number of entries in a bucket is seven.
3206 * Improve cache performance of read/write data by locally
3207 updating the dataVersion of all buffers associated with the
3208 file on each StoreData or StoreStatus operation provided that
3209 the dataVersion was incremented by one and the dataVersion of
3210 the buffer matched the original dataVersion of the cached
3211 status information for the file.
3213 * When recycling cm_scache_t objects, make sure that we
3214 clear all of the previous fields. It is no longer the
3215 case that we can assume that the stat cache object is being
3216 obtained only after a successful FetchStatus call.
3217 InlineBulkStatus will allocate stat cache objects for which
3218 the no status information is available.
3220 * When a file is being opened with the OPEN_ALWAYS disposition,
3221 do not require write permissions when the file already exists
3222 unless of course the file is being opened requesting write
3225 * When a file has been opened for read-only access, if the
3226 application requests an exclusive lock, obtain a read-lock
3227 instead of a write-lock. The purpose of the exclusive lock
3228 in this situation is to prevent changes to the file and the
3229 full file read lock will provide this guarrantee.
3234 * Enforce Unix Mode Write bit to prevent file deletion of read only
3237 * The CIFS READ and LIST permissions use the same value 0x0001.
3238 When a directory is being accessed, check for PRSFS_LOOKUP
3239 instead of PRSFS_READ.
3243 * Fix computations of short names of mount points. They were
3244 broken as part of the directory search optimization. (1.5.13a)
3246 * When all buffers are in use, periodically sleep the threads
3247 waiting for a buffer to become free while the global buffer
3248 lock is not being held in order to permit those threads that
3249 can free buffers to do so.
3251 * Power Management improvements. Maintain a global flag that
3252 specifies whether or not the service is in a suspend state.
3253 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
3254 lana is no longer valid. Instead, stop the listener threads
3255 and if all listener threads are stopped, reset the lana_list.
3256 Allow the cm_Daemon() thread to periodically check the state
3257 of the smb listeners. If they are all stopped and the service
3258 is not suspended, attempt to restart them. If there are no valid
3259 lanas, return to the stopped state.
3261 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
3262 the user is permitted to obtain read or write locks. This function
3263 would obtain the lock and then drop it returning whether or not
3264 the lock could be obtained. If the lock was in fact required,
3265 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
3266 of course being that this pattern results in three RPCs to the
3267 file server (lock, unlock, lock). This is reduced to one RPC by
3268 implementing cm_CheckNTOpenDone() which frees the allocated byte
3269 range lock from cm_CheckNTOpen() after the cm_Lock() call is
3272 * Optimize the SMB FindFirst processing for the case in which there
3273 are no wildcards. If a cm_Lookup succeeds, generate a response
3274 from the stat cache entry we have a callback for. Its cheaper
3275 than walking each buffer of directory data associated with the
3276 directory stat cache entry when we have thousands of directory
3279 * Increase the SMB Ioctl MaxData size to support larger token sizes
3281 * Disable autorun of the AFS Server Wizard when the AFS Server is
3284 * Add AFS Server service dependencies on tcpip and PNP_TDI
3286 * Remove the AFS Server Configuration Tool shortcut
3288 * Add Freelance mode awareness to AFS Server Configuration Tool
3292 * Unlike previous versions of the OS, Vista performs a shutdown on
3293 the Microsoft Loopback adapter just like it would on a real adapter.
3294 This causes the smb_Listener threads to trigger a panic during a
3295 suspend/hibernate power event.
3297 The fix is to unbind from the network adapters in response to a
3298 suspend/hibernate power event and then rebind to the adapters
3299 when a resume power event is received. Note that the resume events
3300 are not reliably delivered so it is possible the afs service will
3301 not be accessible. However, this is the best we can do.
3303 * don't allow environment variables we set in afslogon.dll to be
3304 inherited by child processes
3306 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
3309 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
3310 thread walking the all buffer list periodically searching for dirty
3311 buffers to write to the file server. This had several negative
3312 results. The algorithm ate up ever increasing amounts of CPU time
3313 even when AFS is idle as the size of the cache increases. Also,
3314 buffers were written to the file server in an order based upon the
3315 original buffer allocation which has nothing to do with the order
3316 in which the buffers became dirty.
3318 The new algorithm maintains a dirty buffer list. Items are added
3319 when the buffer is originally marked dirty. A buffer is only
3320 removed from the list by the buf_IncrSyncer when the buffer is no
3321 longer dirty. If the list is empty the thread goes back to sleep
3322 immediately without additional processing requirements.
3326 * Enable renaming of directories from the Explorer Shell when
3327 Offline Folders is enabled
3329 * An error is reported in response to the SMB ReceiveCoreFlush
3330 request when an error has not occurred.
3334 * optimize dirty buffer flushes in response to suspend operations
3335 do not flush if only loopback adapter is installed
3336 do no retry dirty buffer writes
3338 * do not reset netbios when performing loopback checks. doing
3339 so will break the existing lana bind
3341 * remove a deadlock in smb3.c QPathInfo
3345 * Add NetIDMgr AFS Plug-in
3347 * "Show Tray Icon" checkbox removed from afs_config.exe
3349 * Disable DNS Registrations for the Loopback Adapter and make sure
3350 that Netbios is turned on.
3352 * A reference count leak on stat cache objects was fixed
3354 * For debugging, tread ids of writers are now tracked on locks
3356 * Corrected an incorrect lock that was being held but not
3357 freed in a rare error case
3359 * Changed trace log end of line to CR-LF so it can be read
3362 * modify the buffer synchronization thread to write all dirty
3363 buffers and sleep only when there is no work to be done
3365 * improve behavior of cache manager when large numbers of
3366 temporary files are being created by multiple clients within
3369 * do not permit dirty buffers that belong to deleted files
3370 to be written to the file server. its a waste of bandwidth
3372 * do not leak stat cache references when the file server
3375 * afscreds: do not permit an attempt to obtain tokens if the
3376 afs service could not be started. this was causing stack
3377 overflows on Vista after resuming from a suspend
3381 * better handling of dot directory
3383 * fs sq and fs mkm now behave as on UNIX
3385 * add license to installers
3387 * Return Invalid Parameter error for unknown InfoLevels in
3388 SMB Tran2 Search Directory.
3390 * Correctly use WSAStartup and WSACleanup within the RX library
3392 * Check that afsd_service.exe is running before performing
3395 * Force AFS Shell Extension to load resource DLL using
3396 afs_shl_ext.dll as a base name.
3398 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
3399 OpenAFS\Client\Program
3401 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
3403 * Do not panic if the maximum number of volume entries are in use
3404 and one of them can be recycled.
3408 * Fix the code that reads the backconnnectionhostnames value
3409 so that there is not an uninitialized variable
3413 * integrated logon could cause winlogin.exe to blue screen windows
3414 if it was called with a domain that does not exist on the machine.
3415 this is possible during a sysprep.
3417 * corrections to cm_buf_t handling to ensure that buffers are
3418 removed from queues before they are freed and that holds are
3419 obtained on the next buffer in a chain before releasing the
3422 * do not generate MessageBox dialogs on network errors. they
3423 cause more trouble then they are worth.
3425 * move GlobalAutoMapper into a separate thread so that it does
3426 not prevent service startup if the WNetAddConnection2 call
3429 * restrict hard dead and connection timeouts to the CIFS Session
3430 Timeout value. This is being done to try to prevent virtual
3431 circuits from being destroyed prematurely.
3433 * Fix the code that writes the backconnectionhostnames value
3434 to ensure that the data buffer is written with the correct
3437 (there were no windows specific changes between 1.5.4 and 1.5.5)
3441 * Corrections to the Integrated Logon functionality that permits
3442 Kerberos 5 TGTs to be communicated into the user session. The
3443 transfer is now more secure and succeeds when the user does not
3444 have read permission to the %SystemRoot%\TEMP directory.
3446 * NSIS installer updated to version 2.18
3448 * WiX installer updated to version 2.0.4221.0
3450 * Updates to the IFS functionality for 64-bit correctness and
3451 permit compilation. The IFS functionality works for pioctl()
3452 operations but not much else at this point.
3454 * A major re-write of the OpenAFS implementation of the CIFS Tran2
3455 query and set functions.
3457 * Ignore the Windows CIFS client session timeout entirely. Timeout
3458 only after the AFS Hard Dead Timeout period expires.
3462 * move headers, libraries, and samples from OpenAFS\Client to
3465 * remove the auto-registration of AFS IDs for foreign cells
3466 from integrated logon. this permits afslogon.dll to avoid
3467 initializing the RX library. afslogon.dll is a network
3468 provider library and as such it is loaded into all applications
3469 which perform a CreateProcessAsUser call. Initializing the
3470 RX library spawns background threads to listen for inbound
3471 packets. When the network provider is unloaded from the
3472 process, the RX library does not cleanup the threads. If they
3473 continue to execute after the library is unloaded, the process
3476 * implement inline bulk status calls. this will significantly
3477 improve the performance of directory listings.
3479 * prevent a crash when evaluating mountpoints to volumes that do not
3482 * new command: fs uuid [-generate]
3483 displays the current uuid and allows a new one to be generated
3484 on the fly without restarting the afs client service.
3486 * fix the 64-bit installers to properly document the install
3489 * move help files to the proper location so they can be located by
3490 the system tray tool and the control panel.
3492 * "over quota" and "disk full" errors were not being reported to the
3495 * largefile support has been added. Files greater than 31-bits can no
3496 longer be written to file servers which do not support large files.
3497 When file servers support large files, files up to 64-bits can be
3500 * cm_SyncOp could return with the input buffer unlocked upon error
3501 which would result in a panic condition. This bug was introduced
3504 * vlserver failover when freelance mode did not work. This prevented
3505 the root.afs volume from being loaded if the first vlserver could
3510 * A bug in the CIFS Standard Info response packet caused directories
3511 to appear as standard files. This prevented Microsoft Vista from
3512 being able to walk the AFS directory tree.
3514 * Mount points must be reported as directories in response to
3515 Query Path Info operations.
3517 * A deadlock was detected when performing "fs flushall" if the
3518 file server reports VNOVNODE. The scp->createBufferLock is
3519 already held by the current thread and there is no mechanism
3520 to propagate the knowledge. Therefore, an alternate mechanism
3521 for clearing the cache must be developed. A new function
3522 cm_RecycleSCache(scp,flags) has been extracted from
3523 cm_GetNewSCache(). This function performs the task of recycling
3524 an cm_scache_t object. When called from cm_GetNewSCache() with
3525 no flags the expectation is that there are no associated buffers
3526 that are queued to be read or written. When called from
3527 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
3528 flag, any queued buffers will be de-queued and marked as if
3529 the operations were performed so the data can be discarded.
3531 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
3533 * In response to VNOVNODE, the parent is only discarded if the current
3534 cm_scache_t is not a directory.
3536 * In the Ioctl FlushFile and FlushVolume functions, there were no
3537 checks to protect against flushing the Freelance SCache entries.
3539 * In FlushFile, the wrong cm_scache_t object was being released.
3541 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
3542 recycled. Choose a new entry if cm_RecycleSCache fails.
3544 * Speed up the performance of the cache manager by not
3545 holding the smb_fid_t mutex across calls to cm_SyncOp
3548 * Ensure that all smb_fid_t flag references are protected.
3550 * Remove a deadlock between smb_fid_t mutex and
3551 smb_rctLock introduced in 1.4.1.
3553 * Fix aklog AFSID lookup to use not send the realm
3554 name when the realm is the local realm for the
3555 cell. The PT Server doesn't strip the realm
3556 name when it is the local realm.
3558 * Treat "\\afs\*." as an alias for "\\afs\all"
3560 * Dynamically adjust the priority of server threads
3561 based upon the age of the CIFS request that is
3562 being processed. The priority is bumped one level
3563 for every 15 seconds of processing time.
3565 * The original openafs contribution mis-used the
3566 cm_GetCallback function. This function should only
3567 be called as a side effect of a call to
3568 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
3569 was being called results in a FetchStatus RPC being
3570 performed even when the client already has a valid
3571 callback and does so in a manner that destroys the
3572 synchronization of the threads that are calling
3573 cm_SyncOp. If nothing else this results in a
3574 significant performance penalty.
3576 * More changes to cleanup of smb_vc_t objects in
3577 order to prevent race conditions.
3579 * Reduce the amount of Power Management event logging
3582 * If the client detects that the IP address has
3583 changed, force the use of new RX connections. This
3584 enables the file server to recognize that the client
3587 * Remove a race condition and deadlock associated
3588 with cleanup of smb_vc_t objects.
3590 * Fix the ANSI filename support. One function that
3591 required it was missing the translations.
3593 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
3596 * Add a DllMain() function to afspthread.dll. Perform
3597 cleanup of thread local storage, rx queues, critical
3598 sections, event handles, etc. This was implemented
3599 to avoid crashes in afspthread.dll when applications
3600 which load/unload the dll attempt to terminate.
3602 * Lock request behavior has been altered. As of this
3603 release a locally allocated lock will be used whenever
3604 the volume is RO or the requesting user's permissions
3605 are 'rl'. This will allow end users to execute programs
3606 from RW volumes when they do not have the 'k' privilege.
3608 Note that a recent discovery indicates that the 'k' bit
3609 is only supposed to control the ability to obtain read
3610 locks. Write locks are to be controlled by the 'w' and
3613 * Code has been added to enable the cache manager to be
3614 able to obtain the AFSID of the authenticated user.
3615 This is required to be able to determine when "Insert"
3616 privileges should apply. AFS considers a file to be
3617 a new file whenever the owner's AFSID matches the
3618 AFSID of the user. The current implementation relies
3619 on the name specified during the SetToken call to be
3620 correct. This will be replaced in the future with a
3621 new RPC that allows the AFSID of the authenticated
3622 user to be obtained.
3624 * A 32-bit-tools installer for use on 64-bit systems
3625 is now available. These tools are for use with 32-bit
3626 versions of Kerberos for Windows or JPSoftware's 4NT.
3628 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
3630 * Remove the use of the "active_vcp" variable. This variable
3631 no longer serves any useful purpose and was the cause of
3632 race conditions that produced under counts of the smb_vc_t
3635 * Change default minidump output to include data segments
3637 * restore access to cm_IoctlGetVolumeStatus and
3638 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
3639 Still need to figure out why they are being called
3640 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
3642 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
3644 * Update build system to include the "sys_name" in the
3647 * optimize the number of calls to obtain the smb_user_t for a
3648 smb virtual circuit.
3650 * add logging of server addresses, status and capabilities when
3653 * prevent multiple threads from pinging the same server simultaneously.
3654 instead, threads block and then use the result of the ping that was
3655 already taking place.
3657 * Add registry values to allow the defaults for daemon checks to be altered
3658 at startup. Log values to afsd_init.log.
3660 daemonCheckDownInterval
3661 daemonCheckUpInterval
3662 daemonCheckVolInterval
3663 daemonCheckCBInterval
3664 daemonCheckLockInterval
3665 daemonCheckTokenInterval
3667 * Add EventLog messages for transitions in fs crypt state.
3669 * Byte range locking (when locks will be obtained from the
3670 is still under discussion)
3672 * 64-bit Windows operating systems now supported
3673 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
3674 This is a native 64-bit implementation of OpenAFS for Windows.
3675 There is currently no 64-bit version of KFW and therefore
3676 there is no (easy) method for obtaining AFS credentials.
3677 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
3678 to obtain AFS tokens but there is not for the feint of
3681 * A faster implementation of the fcrypt algorithms improves
3684 * The afsd_service.exe now supplies Event Log Message Text.
3685 The event log no longer generates warning text due to lack
3686 of message configuration.
3688 * When afs servers are unreachable and network stack returns
3689 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
3690 mark the attempted server as being down. This allows for
3691 very fast response to the calling application. This change
3692 has a pronounced improvement when the Windows Explorer is
3693 used on disconnected mobile computers.
3695 * Added DebugOutputString debugging of the RX library
3697 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
3699 * Fix the rx-lwp implementation to randomly generate rx call
3700 identifiers which are supposed to be unique at a given point
3701 in time. This affects some of the command line tools used
3702 for debugging but not the AFS Client Service itself which
3703 uses the rx-pthreads implementation.
3705 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
3706 Windows although they are not distributed
3708 * don't generate an exception when logging unexpected SMB packets
3710 * the data structures used by cm_IoctlGetVolumeStatus and
3711 cm_IoctlSetVolumeStatus are not compatible with the CIFS
3712 Get/Set File Information calls so don't bother executing them.
3716 * Fix data written to registry as part of BackConnectionHostnames
3719 * Fix initialization bug when allocating cm_server_t objects that
3720 would cause servers to appear to be down after each allocation
3721 or renewal. Server would appear up after the next check servers
3726 * Windows will return WSAECONNRESET to the next receive packet
3727 operation on a socket if an ICMP packet is received in response
3728 to a packet that was sent and could not be delivered because
3729 the port was unreachable. We have to treat this as
3730 WSAEWOULDBLOCK so that we can retry the select() check and
3731 not mark the connection as bad.
3733 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
3735 * The client manager maintains a least recently used list of
3736 all the ACLs. If an ACL is already the most recently used
3737 do not bother to remove it from the list and place it on the
3740 * As part of the process of adding a new server, set the
3741 server is down flag and then perform a ping. If the server
3742 is up, the flag will be reset. If the server is down, the
3743 client will not timeout unnecessarily while attempting to
3746 * If a Mountpoint String is the empty string, return PATH NOT
3749 * Modify behavior of the Network Provider to avoid querying the
3750 profile location if integrated logon is known to be disabled.
3752 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
3753 will always be preserved regardless of whether or not integrated
3754 logon was used to obtain tokens. When unset, tokens will be
3755 destroyed regardless of whether integrated logon was used to
3756 obtain the tokens unless the profile is stored in AFS.
3758 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
3759 message is being delivered even when the user is not logging
3760 off of the computer.
3762 * Fix a deadlock condition introduced in rc6.
3764 * Cell names are case-insensitive. Since the founding of OpenAFS
3765 there has been a bug that treated "cell", "Cell", and "CELL" as
3766 different cells. This would result in what appeared to be
3767 random disappearance of user tokens or a failure to authenticate
3768 even when tokens were listed by the "tokens" command.
3770 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
3772 * Additional auditing of the smb code revealed that smb ioctl
3773 operations were failing to free smb_fid_t objects
3775 * The File Identifiers assigned to smb_fid_t objects could be
3776 given the value 65535 which is 0xFFFF which is also the
3777 INVALID_HANDLE value. Both 0 and 65535 are now treated as
3780 * Audited all assignments to fields and ensured that they are
3781 protected by obtaining the relevant mutex object.
3783 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
3784 if both the current cm_scache_t object and its parent are
3785 both returning VNOVNODE.
3787 * Removed all the logic associated with the 'dead_vcp' variable
3788 which must have been added for debugging because its used
3789 was never thread safe and it truly served no purpose.
3791 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
3793 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
3794 registry settings. Default is ON and 120 seconds. If set to
3795 OFF, tokens will not be destroyed.
3797 * Fixed a number of race conditions and reference count problems
3798 associated with global queues.
3800 * Added a method to detect abandoned smb virtual circuits so their
3801 associated file handles, locks, user tokens, and smb session
3802 resources can be released.
3804 * Added a mechanism to allow user tokens to be preserved
3805 during integrated login and be freed after logoff.
3807 * Check access permissions to the file before the file is
3810 * More improvements to the handling of SMB Virtual Circuits
3811 when SMB sessions are closed.
3813 * Correct a race condition introduced in rc4 that could result
3814 in a crash of afsd_service.exe.
3816 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3818 * Properly handle pre-mature SMB session closures.
3819 File handles, tree connections, and user sessions are now
3820 freed when the virtual circuit is closed.
3822 * smb_username_t and smb_user_t objects are now properly reference
3823 counted and freed as appropriate.
3825 * disable to use of AFS file locks. byte range locking is still
3826 implemented but all locks are managed locally. This prevents
3827 multiple processes from editing the same file on the same
3828 machine but it does prevent multiple processes on separate
3829 machines from modifying the file.
3831 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3833 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3834 the same server parameter would result in a panic condition
3835 if the server had not already been contacted by the AFS
3838 * Better warnings are written to the afsd.log file when
3839 the Windows RPC Protocols are not properly configured as
3840 is often the case on Windows 2000.
3842 * fix the procmgmt library to restore C RTL signal handlers when
3843 it is being detached from the process prior to process termination.
3845 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3847 * close a lock leak in the rx library that could be triggered while
3850 * prevent removal of drive mappings in response to afscreds.exe -M
3852 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3853 strings and flush the stat cache entry.
3855 * prevent server objects from being freed if user preferences are set.
3857 * Update WIX support to 2.0.2419.0
3859 * Release all locks on the server at service shutdown.
3861 * A down server should be marked up as long as it responds to the client.
3862 It doesn't matter if the server returned an error as long as the error
3863 is something other than VRESTARTING. If there is such an error, all
3864 connections on that server should be marked for replacement.
3866 * Change the check "up" servers period from once per hour to once every
3867 ten minutes. This matches the behavior of the Unix cache manager.
3869 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3870 * Byte Range Locking support has been added. Architectural documentation
3871 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3872 made to enforce Windows mandatory locking model using locally managed
3873 byte range locks with full file locks obtained from the file server.
3874 Read-only volumes provide the user with an implied locking privilege.
3875 Read-write volumes require that a user be granted the locking privilege
3876 in order for a lock to be obtained. As most Windows applications use
3877 Shared file access modes, they require the ability to obtain locks in
3878 order to access files.
3880 * Implement error translation functions to convert Universal Error Codes
3881 output by modern servers to local errno and winsock error values.
3882 This fixes the error message problems using pioctl() calls from fs.exe
3883 and the AFS Shell Extension.
3885 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3886 * Allow the AFS Client Service to set the firewall configuration by
3887 retrying until after Windows determines the boot period is over.
3889 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3892 * Detect IP address changes from within the AFS Client Service for the
3893 purpose of checking down servers sooner
3895 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3896 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3897 user that there is a Kerberos key problem. Previously the user would
3898 receive no feedback when there was something wrong.
3900 * Fix the large integer library support to properly handle values greater
3903 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3904 * Updated CellServDB
3906 * Fixed file name pattern matching. Question marks can match nothing
3907 if found at the end of a component.
3909 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3910 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3912 * The MSI installer now has a new Product ID for each build. This
3913 is to allow for automated upgrades. The Upgrade ID is remaining
3916 * Volumes that were cached and then subsequently were salvaged,
3917 moved or otherwise taken temporarily offline (as opposed to being
3918 busy) would become unreachable. A distinction is now made between
3919 a volume being offline and a server being down.
3921 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3923 * Fixed several AFS Shell Extension issues:
3924 - Checking Server Status would only work if a specific cell was
3926 - If servers were down, the Server Down dialog would write to
3927 invalid memory and cause explorer.exe to crash
3928 - Default Push Buttons were not specified for many dialogs
3929 - The extension would be unloaded by the Explorer Shell and
3930 never be reloaded under some circumstances.
3932 * Fixed the AFS System Tray menu to disappear if another Window
3933 was selected before a menu item was selected
3935 * Added volume owner and group information to "fs examine" output
3937 * Added fs and registry support for enabling/disabling/clearing
3938 rx statistics gathering.
3940 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3942 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3943 GUIDs for each platform.
3945 * Documentation directory structure and default installation rules
3948 * New OpenAFS for Windows HTML Release Notes installed
3949 as part of the documentation set.
3951 * Add thread id to rx debugging messages
3953 * When all servers are offline, return a bad network path error
3954 immediately. This is necessary to allow Explorer to do the
3955 right thing and not hang. The background daemon thread will
3956 check the down servers every 30 seconds to see if they are
3959 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3961 * The AFSCache file has been modified to store both the
3962 serial number of the volume on which the file resides
3963 and the SID of the local machine. If either change,
3964 a new UUID is generated.
3966 * add "fs flushall" command that forces all data buffers
3969 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3971 * add support for '/' instead of '\' in pioctl() calls
3973 * Apply AFS Client Admins group protect to AFS Shell Extension
3975 * Add support for \\afs\<name> to most AFS Shell Extension
3976 commands except the symlink methods.
3978 * when installing the Microsoft Loopback Adapter, enable
3979 MS Client for Networks and prevent an install failure
3980 by not calling CoInitialize twice in the same thread.
3982 * reload cell vldb values from the CellServDB every two
3983 hours in case it changes
3985 * When updating cell information from DNS, be sure to set
3988 * Add support to allow use of \\AFS\<foo> where <foo> is
3989 either a mount point or symlink. As <foo> is normally
3990 treated as a share name, we transform it into \\AFS\all\<foo>
3993 * Init 'code' to prevent false errors when integrated logon
3994 is disabled and the service is not running
3996 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3998 * Removed trace log messages that were hampering performance.
4000 * Fixed a deadlock that was being triggered by editing Word
4001 documents stored within AFS with WinWord 2003.
4003 * Bit 3 of the TraceOption registry value is now used to set the
4004 default for "fs trace" in the non-Debug builds. The new default
4005 is off for release builds and on for debug builds.
4007 * Bit 2 of the TraceOption registry value can be used to turn on
4008 real-time output of debug log entries to the Windows Debug
4009 Output monitor interface. This data can be viewed with tools
4010 such as Sysinternal's DbgView.
4012 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
4013 * When the cache manager reports ALLBUSY or ALLOFFLINE for
4014 a resource we should not be returning status codes such
4015 as Network Busy or Remote Host Down because these errors
4016 will imply that the AFS Client Service is Busy or Down
4017 and that is simply not the case. Instead we will return
4018 Bad Network Path as the path is temporarily not available.
4020 Instead of returning Sharing Paused when there is a
4021 timeout error, return Timeout. Once again, the AFS Client
4022 Service is not paused.
4024 * afscreds.exe would display an Obtain Creds dialog when
4025 the expired credentials reminder was triggered even
4026 if there was no network path to the KDC. This is prevented
4027 by adding KDC probe logic to the reminder thread.
4029 * afscreds.exe would display expired tokens no differently
4030 than unexpired ones. This would make it difficult for the
4031 user to distiguish when the tokens were expired. For the
4032 English build added a new resource string "(expired) that
4033 is displayed instead of the expiration time.
4036 * "fs wscell" when executed in freelance mode will return the
4037 name of the cell configured in the registry. The root.afs
4038 volume is not loaded from this cell, but it is used for the
4039 default for aklog and integrated login.
4041 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
4042 no longer work on \\AFS\all when freelance is being used
4043 unless the user is a member of the "AFS Client Admins" Windows
4046 * some more checks performed during persistent cache validation.
4049 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
4050 down servers for the current cell not all cells.
4052 * If the scache LRU list becomes corrupted, the AFS Client Service
4053 might enter an infinite loop at startup. Detect the corruption
4054 and discard the cache contents in this case.
4056 * Fixed the Explorer Shell's remove mountpoint functionality.
4058 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
4059 obtain a mutex lock on a cm_scache_t object whose lock is held by
4060 a thread attempting to obtain the cm_aclLock.
4063 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
4064 misinterpretted as "csail.mit.edu" when the resolver
4065 is configured to append ".mit.edu" to failed DNS queries.
4067 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
4068 that can be used to force the preservation of user tokens upon logout.
4070 * Update the NSIS install scripts to use NSIS 2.07. This release adds
4071 recommended installation categories: AFS Client, AFS Administrator,
4072 AFS Server, AFS Development Kit. Each category includes a different
4073 default subset of the OpenAFS components.
4075 The OpenAFS logo is now associated with the NSIS uninstall entry
4076 in the Add/Remove Programs control panel.
4078 * The user name associated with AFS tokens when obtained with
4079 integrated login, the afs systray tool, or aklog will always
4080 include the full Kerberos 5 user principal name regardless of
4081 whether or not the cell is local to the realm.
4083 * Modify integrated login so that it does not enter an infinite
4084 loop if the service is not set to auto start.
4086 * Added asetkey.exe used to set a Kerberos 5 key for use by the
4089 * Added uninstall.exe to Wix installation
4091 * More modifications to algorithms used to wake sleeping threads.
4095 * Changes to the algorithms used to wake threads when they are
4096 put to sleep because of conflicts over buffer operations.
4099 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
4100 the process of producing AFS tokens. The tokens are stored within
4101 the AFS Client Service but the Kerberos 5 tickets have been discarded.
4102 New functionality has been added to temporarily copy the tickets to
4103 a file ccache during the short window between the NPLogonNotify and
4104 WinLogon Logon Event. The Logon Event handler starts a process as
4105 user, afscpcc.exe, that copies the tickets into the default ccache.
4106 The file ccache is then destroyed by the event handler.
4108 * During synchronization operations on stat cache entries, if threads
4109 are forced to sleep it was possible that the threads would not be
4110 woken under heavy load.
4112 * Enforce the delete on close flag specified by the NTCreateX()
4115 * A race condition in the rx library was fixed that resulted in
4116 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
4117 was cleared when threads were still waiting to use the
4118 connection. Also, fixed a potential case in which the
4119 connection could be destroyed while threads were waiting to use it.
4121 * The DNLC freelist has been seen to become corrupted with still
4122 active entries being stored on the freelist. Changes were made
4123 to perform a better job of cleaning entries before placing them
4124 onto the freelist; marking them with the correct magic value when
4125 purging the list; and allowing the list to be purged on startup
4126 if the cache validation fails.
4128 * Windows Crash Reporting does not get triggered for afsd_service.exe
4129 because it provides its own top level exception handler. This patch
4130 places an explicit call to ReportCrash() on platforms which support
4133 If you configure Crash Reporting via AD Policy to capture crash
4134 reports within the domain, then you will receive this output.
4136 * Generate MiniDumps independent of the Windows Crash Reporting.
4138 (a) If an exception occurs,a minidump will be generated at the path
4139 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
4140 can be set to any other type via use of the
4143 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
4145 (b) MiniDumps can be generated at any time using the "fs minidump"
4146 command. This allows you to generate a minidump without
4147 stopping the process or even requiring the use of a debugger
4150 * an RX connection which reports bad ticket now treats it the same way
4151 the expired ticket is treated. The ticket is removed from the
4152 connection and a new anonymous connection is established.
4154 * fs memdump now requires that the caller be in the AFS Client Admins
4157 * additional debugging information is provided in cm_Analyze when the
4158 error code is ignored.
4160 * Fix a regression introduced into the pioctl() client code which
4161 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
4162 MIT KFW is not installed. (1.3.82a)
4164 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
4167 * Fix a race condition in cm_EndCallbackGrantingCall() that
4168 could leave 'nrevp' pointing to freed memory. In the process,
4169 optimize the processing to ensure that the applications monitoring
4170 the status of the stat cache entry are only notified at most
4173 * Conditionalize access to the Explorer Shell AFS->Mount Point
4174 ->Remove menu item. Disable the item if the selected item(s)
4175 are not mount points.
4177 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
4178 Local Storage. This eliminates thread contention on the
4179 global RX free packet queue. This should improve performance
4180 on multi-CPU systems.
4182 * Fix the IP address filtering code to properly load/unload
4183 the IPHelper DLL so that Windows XP/2003 and beyond systems
4184 do not have to manually probe the registry.
4186 * CleanupACLEnt() was not being called consistently with
4187 the cm_scache_t object referred to by the back pointer
4188 mutex locked. This could in very rare conditions lead to
4189 an invalid memory access.
4191 * Added a script command to the msi installer "afs_replace"
4192 which can be used to replace one version of OpenAFS with
4193 another without requiring an uninstall and reinstall if
4194 the installed msi is no longer in its original location.
4196 * Apparently the problem with multi-domain forests with cross-
4197 realm trusts to non-Windows realms was not entirely solved.
4198 The authentication to the AFS SMB service failed because
4199 the wrong name was being used. Using ASU as an example,
4200 the authentication was being performed with the name
4201 "QAAD\user" (an account in the forest root) and not
4202 "user@ASU.EDU (the MIT Kerberos principal used to login with)
4204 The solution was to add an additional dependency on KFW
4205 in order or to be able to easily obtain the client principal
4206 name stored in the MSLSA ccache TGT. This information is
4207 used in two locations:
4209 - the pioctl() function
4211 - a new WinLogon Event Handler for the "logon" event.
4213 The pioctl function will now be able to use the correct
4214 name when calling WNetAddConnection2() and the "logon"
4215 event handler will now be able to call WNetAddConnection2().
4216 The hope is that the "logon" event handler will be called
4217 before the profile is loaded but I have not guarrantee
4221 * Fixed a locking error in cm_TryBulkStat() which had the
4222 potential of deadlocking the system for the length of time
4223 it takes to perform a bulk status fetch operation.
4225 * Replaced time conversion code (UnixTime <-> FILETIME) to
4226 be completely arithmetic instead of relying on a bizarre
4227 algorithm involving a variety of C RTL time functions.
4228 This has the side effect that UnixTime and FILETIME which
4229 are both stored in UTC are interpretted as UTC throughout
4230 the year. Windows will apply the same localization to AFS
4231 as it does to NTFS. Applications which rely on the ability
4232 to sync files between the two file systems will no longer
4233 see the timestamps of files in AFS change an hour relative
4234 to the files stored in NTFS or Windows based backup devices.
4236 * Fixed a invalid memory access under a bizarre circumstance.
4237 Windows will allow a physical mass media device to be
4238 installed via Plug N Play to the system and will assign it
4239 the lowest available drive letter. This is true even if the
4240 drive letter is currently assigned in the user session to
4241 a network device via NET USE (or its equivalent.) When this
4242 happens, queries sent to the CIFS server will contain invalid
4243 data. This invalid data was not being caught by the AFS
4244 Client Service and was resulting in a crash when Freelance
4247 * Fixed a reference count error when registering callbacks
4248 on a stat cache object if the callback was already registered
4250 * Add a case to cm_Analyze() in order to handle the case where
4251 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
4252 the server list for the volume is empty. In this case, force
4253 an update of the volume info and retry.
4255 * Insert a missing cm_EndCallbackGrantingCall() which could
4256 result in threads waiting for a callback to complete to never
4259 * In the persistent cache, there is a maximum number of volume
4260 entries. Allow volume entries to be reused if the maximum
4261 number have been allocated and their reference count is zero.
4263 * If we already have a dead virtual connection object, don't
4264 ignore the fact that additional dead vc objects must be taken
4267 * Removed a deadlock condition introduced in the summer of 2004.
4268 Do not hold mx locks around calls to RX functions including
4269 rx_NewCall(). That is what reference counts on the rx_connection
4272 * Fixed an initialization error in afslogon.dll which could
4273 result in random behavior including a failure to terminate
4274 the NPLogon function. (1.3.80b)
4276 * Fixed an error preventing the use of SMB authentication
4277 on Windows 2000. (1.3.80a)
4280 * Updated CellServDB to Public CellServDB 16 Mar 2005
4282 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
4283 misinterpretted as "csail.mit.edu" when the resolver
4284 is configured to append ".mit.edu" to failed DNS queries.
4286 * Fixed another case in which the client would replace connections
4287 to the host when it was not necessary. In this case the
4288 situation would occur if cryptall was on and the connections
4289 were unauthenticated due to lack of an appropriate token.
4291 * OpenAFS for Windows has failed to work at sites which are
4292 utilizing a cross-realm trust between an MIT/Heimdal realm
4293 and a multi-domain Windows forest when the workstation being
4294 accessed is not located in the root domain. This is caused
4295 by a bug in the workstation which was triggered after the
4296 introduction of Windows 2003 Server. When the bug is triggered,
4297 the workstation attempts to authenticate users to afsd_service.exe
4298 by contacting the Domain Controller instead of using the
4299 LSA loopback authentication mechanism.
4301 One of the reasons this bug occurs is because the workstation
4302 does not have a reliable way of knowing that the service whose
4303 netbios name is "AFS" is located on the workstation. This will
4304 be fixed starting in Longhorn Beta 1 by Microsoft. The
4305 "BackConnectionHostNames" registry value will be used to
4306 indicate that the authentications to that service name should
4307 be performed using the loopback authentication mechanism.
4309 In the meantime, when Logon Caching is enabled, we can force
4310 afsd_service.exe to authenticate using the logon cache before
4311 contacting the Domain Controller. This will work with both
4312 password and smart card based logons.
4314 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
4315 flag should not be cleared if a volume's server reference is
4316 marked as "offline". In the case where all of the volume's
4317 servers are either "down" or the volumes are "offline", we want
4318 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
4319 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
4320 should never occur in practice.
4322 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
4323 5 seconds, clear the server down and volume busy flags, and
4324 then force an update from the VLDB. This allows the client to
4325 update the location of a volume if the reason for it being
4326 marked offline is because it is being moved. Calls to
4327 cm_ConnByMServers() will be retried either until success or
4328 the RDRtimeout period is reached.
4330 * Correct the Power Management code behavior in response to
4331 Standby, Suspend, and Shutdown events. Instead of flushing the
4332 buffers associated with the mounted SMB submounts, simply write
4333 all buffers which are marked dirty.
4335 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
4336 provides a list of cells other than the default cell for which
4337 tokens should be obtained using the default Kerberos principal.
4338 This functionality is logon domain specific and is only available
4339 when using KFW for authentication.
4341 * Fixed FindNext Invalid Handle error caused by over agressive
4342 attempts at garbage collecting dirSearch entries when the
4343 dirSearch ID wraps from the maximum value to 1.
4345 * Add support for registry defined server preferences for VLDB and
4346 File servers. See registry.txt for details.
4348 * Increased default cache size to 96MB and default number of cache
4351 * Fixed refCount leaks related to directory lookups and pioctl calls.
4353 * Callbacks revoked during a race condition with an attempt to
4354 obtain the same callback no longer result in an inaccessible
4357 * New command line tool:
4359 afsdacl : Set or reset the DACL to allow starting or stopping
4360 the afsd service by any ordinary user.
4362 Usage : afsdacl [-set | -reset] [-show]
4363 -set : Sets the DACL
4364 -reset : Reset the DACL
4365 -show : Show current DACL (SDSF)
4367 * IP addresses are no longer queried once at startup. Instead IP
4368 addresses are obtained as needed. Loopback adapter addresses are
4369 no longer published to the server.
4371 * Pay attention to the MIT KFW registry configuration for automatic
4372 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
4374 * Fix (once again) case-sensitive comparisons which was apparently
4377 * Activate support for vos listvol -format
4379 * Population of the "AFS Client Admins" group failed on non-English
4380 versions of Windows because the name of the "Administrators" group
4381 is localized. Now we lookup the name of the group by using the
4382 Administrators Group Alias Relative ID.
4384 * Multiple cell token acquisition within afscreds.exe was broken in the
4385 case of Kerberos 5 cross realm authentication. Instead of contacting
4386 the KDC associated with the cell's realm, afscreds would obtain a
4387 token from the KDC of the user's realm. This would result in
4388 "invalid kvno errors" while authenticating to the AFS servers.
4389 Unauthenticated access would work. The symptoms would vary based upon
4390 whether or not the VLDB servers had been contacted using unauthenticated
4391 connections prior to the user obtaining tokens.
4393 * The list of ACL entries was becoming corrupted. It appears as if
4394 an ACL when it expires was not being appropriately cleaned up. In
4395 fact, it was left in the list of ACLs associated with the scache
4398 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
4399 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
4400 and "amd64_win64" for amd 64-bit processors.
4402 * When executing executables, Windows provides a mechanisms by which
4403 the normal search for DLLs can be bypassed. If a file foo.exe is
4404 being executed and there exists a file or directory "foo.exe.local"
4405 (the contents are ignored if a file), then Windows will search for
4406 DLLs first in the "foo.exe.local" directory and second in the
4407 directory in which the "foo.exe" file is located.
4409 Previous releases of OAFW would improperly return
4410 STATUS_NOT_A_DIRECTORY instead of the expected
4411 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
4412 to terminate the search for the DLL causing the execution of the
4413 application to fail.
4415 * Rework the reference counting of smb_vc_t objects. The references
4416 stored in the waiting locks were not counted. This could result in
4417 an assertion if the reference count drops to zero.
4419 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
4420 when running in freelance mode.
4422 * Added support for persistent caching of file contents, stat entries,
4423 volume data, cell data, and name lookup data. When the registry
4424 value, "NonPersistentCaching", is not defined or set to 0 the file
4425 specified by "CachePath" becomes a persistent cache file. The
4426 size of the cache file is computed at run time. It has a maximum
4427 size of approximately 1.2GB. The contents of the file will be
4428 validated according to the rules specified by the "ValidateCache"
4431 * Added support for UUIDs. UUIDs are kept for as long as the
4432 cache file is intact.
4434 * Added cmdebug.exe and support cache manager debugging callback
4435 interfaces in afsd_service.exe.
4437 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
4438 [-addrs] [-cache] [-help]
4439 Where: -long print all info
4440 -addrs print only host interfaces
4441 -cache print only cache configuration
4443 * Symlinks to \\AFS\[all\]... will now be treated the same as
4444 symlinks to /afs/.... However, please use /afs/... as otherwise
4445 the symlinks won't work on Unix.
4447 * Correct a problem with local CellID allocation for cells whose
4448 server list is obtained via DNS instead of a CellServDB file. If
4449 the DNS information expires the CellID assigned to the entry will
4450 be changed. This causes all of the cm_scache_t objects which refer
4451 to the old cellID number to become useless. Attempts to access
4452 files or directories with cm_scache_t objects using the old cellID
4453 will fail since the server list cannot be obtained.
4455 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
4456 be triggered if the volume referenced by the cm_scache_t object is
4459 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
4460 in addition to SERVICE_ACCEPT_STOP.
4462 * Move RPC shutdown until after the SMB and RX shutdown procedures
4463 complete. Block until RPC shutdown is complete.
4465 * Modify afslogon.dll (integrated logon) to wait for service if its
4466 state is START_PENDING. If the timeout period occurs, reset to
4467 the retry interval and not the sleep interval.
4469 * When renewing the server list for a cell obtained via DNS AFSDB
4470 records, the cm_cell_t entry must be removed from the list of all
4471 cells. Otherwise, the list of cells will be corrupted.
4473 * In the dcache and scache modules, use the cm_scache_t dataVersion
4474 instead of the cm_fakeDirVersion.
4476 * Synchronize fs.c with the unix version.
4478 * The variable used to determine whether a file or virtual memory
4479 mapped cache is used was not properly initialized to a default
4480 value. If the registry setting "NonPersistentCaching" was not
4481 set, the choice would be random. Properly initialized to be
4484 * The memory mapped view was never unmapped before closing the file
4485 at service shutdown. This is now properly cleaned up.
4487 * Default location of Cache file is now %TEMP%\AFSCache
4491 * A bug affecting new installations of 1.3.75/76 would result in
4492 the creating of incorrect mountpoints in the freelance root.afs
4493 volume for the default cell.
4495 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
4496 of "root.cell", you have been affected by the bug. To correct the
4497 problem, execute the following commands:
4499 fs rmmount \\afs\all\<cellname>
4500 fs rmmount \\afs\all\.<cellname>
4502 fs mkmount \\afs\all\<cellname> root.cell <cellname>
4503 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
4507 * A bug has been fixed in the auto-generation of Freelance root.afs
4508 symlinks which produced random entries in the \\AFS\all directory.
4510 * Support has been added for multi-homed servers
4513 * Added a new registry value, "StoreAnsiFilenames", which can be used
4514 to force the use of ANSI character sets instead of OEM Code Pages.
4515 This feature is useful when users require the ability to create
4516 filenames with 8-bit characters and need to access the files from
4517 both Latin-1 based Unix systems as well as from Windows.
4519 Activation of this feature will prevent access to files stored with
4520 8-bit OEM characters.
4522 * Shutdown all SMB threads in a synchronized manner when stopping the
4525 * There is currently a maximum cache size of 1.3GB. The limit is imposed
4526 by the largest contiguous block of unused memory within the 2GB process
4527 space which can be assigned to the memory mapped file. Unfortunately,
4528 when the executable digital signature verification code is activated
4529 Windows sees fit to further segment the process memory which in turn
4530 reduces the size of the maximum cache file to less then 800MB. If
4531 larger cache sizes are desired, a new registry value should be set:
4533 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
4535 Setting this value will disable the runtime verification of digital
4536 signatures on afsd_service.exe and the afs dlls which it loads. It
4537 will not disable the the version number check on those same files.
4538 The signature verification is not a security messure and is only meant
4539 to enhance the ability to afsd_service.exe to detect potential
4540 destablizing mixtures of DLLs from incompatible distributions.
4542 Added code to auto-disable the signature verification check if
4543 the desired cache size is greater then 700MB.
4545 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
4546 used to verify the validity of the afsd_service.exe binary
4547 as well as each of the AFS DLLs loaded by the service. Not only
4548 must the digital signature be valid but the signatures of the
4549 DLL must be signed by the same entity as the service.
4551 * Implement new functions: cm_freelanceMountPointExists and
4552 cm_freelanceSymlinkExists. Use them along with other validity
4553 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
4554 ensure that name collisions do not occur and that empty strings
4555 are not valid file names.
4557 A symlink may not have a name which would resolve to a valid
4558 cell name. Doing so would prevent access to the cell.
4560 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
4561 re-initialization code. The reference counts of the fake root.afs
4562 volume scache object(s) would become invalid when the mount point
4563 or symlink lists were altered.
4565 * Add registry entries to provide mappings from the afsdsbmt.ini
4566 to the new locations for applications which count on the use
4567 of the old Profile file APIs. These apps are likely to fail
4568 if the user does not have administrator privileges and the
4569 registry is locked down.
4571 * The afs_config.exe submounts dialog had two errors.
4572 First, attempts to remove entries failed because the registry
4573 key was being opened without KEY_WRITE privileges.
4574 Second, when editing a submount entry, changing the name
4575 would add a new key and leave the original one in place.
4576 Now the original submount will be removed if its name is
4579 * In recent months there have been several incidents in which
4580 users have experienced problems starting or accessing
4581 afsd_service.exe and after significant effort has been spent
4582 it has turned out that they have two versions of AFS on the
4583 machine or an inconsistent set of DLLs.
4585 Code has now been added to afsd_service.exe which will walk
4586 the list of modules loaded by afsd_service.exe and validate
4587 that the version of the AFS DLLs matches the version of the
4588 afsd_service.exe executable. If they do not match the service
4591 * When Freelance mode is enabled and there is no registry
4592 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
4593 will attempt to import the afs_freelance.ini file contents.
4594 If the file does not exist, it was creating a dummy file
4595 with a r/o and r/w entry for the default cell and then
4596 importing those values.
4598 This process has been changed. The temporary file is no
4599 longer created. Also, both the OpenAFS Client install
4600 directory as well as %WINDIR% are checked for previous
4601 afs_freelance.ini files.
4603 * Added support for VL_GetEntryByNameN(). Still need to add
4604 support for VL_GetEntryByNameU() for multi-homed support.
4606 * Fix a deadlock situation in afscreds.exe when canceling an
4607 auto-generated Obtain Tokens dialog
4610 * Fix the locking of objects during Directory Searches in the
4611 SMB/CIFS server. The failure to properly lock the reference
4612 counts was resulting in the premature freeing of smb_dirSearch_t
4613 objects while they were still in use by the SMB/CIFS client.
4614 This does not solve the "Invalid Handle" problem.
4616 * Fix Find Cell By Name pioctl call to return a valid cell
4617 name for the Freelance fake root.afs volume.
4618 "Freelance.Local.Root".
4620 * Fix the Explorer Shell Extension Symlinks->Add operation.
4621 The dialog template was missing and the link destination
4622 string was too short.
4624 * Add support for symlinks to Freelance root.afs volume
4625 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
4626 <number> = "<linkname>:<relative-path>."
4627 Use symlink.exe to create, list, or remove
4629 * Remove the fallback to the use of KFW's KRB4 library when obtaining
4630 tokens. We never obtain KRB4 tickets.
4632 * Fix AFS Client Configuration Control Panel to support new SysName
4635 * Fix a bug in afsd_service.exe which could result in the SysName
4636 not being read from the registry.
4639 * Add code to block the issuance of AFS tokens by aklog.exe or
4640 afscreds.exe when the Kerberos 5 principal name contains a dot.
4642 * Modify the IsAdmin() function to always treat the local SYSTEM
4643 account as an AFS client administrator. Affects fs.exe and
4646 * Modify the internal handling of Quota Exceeded errors
4648 * Upgrade all reference count fields in the Windows cache manager
4649 and the osi library to use unsigned long instead of signed short.
4650 A similar fix has been applied to the afs rpc (rx) library.
4652 * fix the Windows cache manager to prevent it from replacing the
4653 rx_connection object associated with the cm_conn_t object on each
4654 and every operation if "fs crypt" was set. This explains the
4655 dramatic performance difference when crypt is used vs clear.
4656 The problem: 'cryptall', a boolean flag indicating whether or not
4657 "fs crypt" is set, was being compared to the rx_connection
4658 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
4659 1 != 2 and therefore the rx_connection was always destroyed
4660 and replaced on each and every operation.
4662 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
4663 It is not safe for the cm_conn_t object to not be locked because
4664 rx_DestroyConnection might be called from another thread if:
4665 - the user's tokens have changed or expired
4666 - the crypt mode has changed
4668 This fix appears to have also taken care of the problems associated
4669 with Overlapped Writes resulting in Delayed Write errors.
4671 * fix NSIS installer's AdminGroup.exe to properly create and
4672 remove groups when given -create or -remove. The string comparison
4675 * fs sysname now accepts a list of sysname values
4677 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
4678 DWORD which when set to a non-zero value will cause error message
4679 text to be output to stderr from the pioctl() routine. Useful in
4680 debugging failures of fs.exe, tokens.exe, etc.
4682 * added a test to the power management code to only perform a
4683 flush operation if there is at least one network adapter which
4684 is not a loopback adapter.
4686 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
4687 "EnableKFW". This value will not be read if the key
4688 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
4689 value under that key does not.
4691 * provide mechanisms to force the use of krb524d for Kerberos 5
4692 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
4693 there is a new registry value "Use524" and for aklog.exe a new
4694 command line parameter "-m".
4696 * Fix the pattern matching algorithm to properly match patterns
4699 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
4700 which is used by both the original function and the new
4701 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
4702 creation of HardLinks in addition to Renaming. smb_Link()
4703 is a new function which creates HardLinks via cm_Link().
4704 cm_Link() is a new vnodeops function which creates links
4707 smb_ReceiveNTRename() does not support the File Copy and
4708 Move Cluster Information operations described in its interface.
4709 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
4711 * When opening files via symlinks, we should follow the symlinks
4712 until we reach the actual file stat cache entry. The stat cache
4713 entry of the file should then be stored in the FID instead of
4714 stat scache entry of the symlink.
4716 * return bad operation errors for all unimplemented functions
4717 even if we do not know the functions exist.
4719 * Log bad packets and unknown operation packets to the trace log
4721 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
4724 * Update list of known CIFS operations to include all those listed
4725 in CIFS-TR-1p00_FINAL.pdf.
4727 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
4728 to support the REG_EXPAND_SZ type.
4731 * A new Windows authorization group "AFS Client Admins" is now
4732 created and populated with the members of the "Administrators"
4733 group. The group is used to determine which accounts on the
4734 machine may be used to modify the AFS Client Configuration via
4735 the UI and command line tools. afs_config.exe, fs.exe,
4737 * Modify the WinLogon Logoff Event Handler to query NT4 domain
4738 controllers for the remote profile path if Active Directory
4739 services are not available.
4741 * Fix aklog.exe to not add the AFS ID to the username
4743 * PTS registration of new users to foreign cells has been added to
4746 * The cm_Daemon thread is used to perform checks for
4747 down servers, up servers, volumes, callback expirations,
4748 lock maintenance and token expiration. Due to a gaff in
4749 larger integer division the thread never performed any
4750 work. Instead the current time computation would always
4751 be less then the trigger times. This had an adverse affect
4752 on the client's ability to maintain communication with servers,
4753 keep volumes up to date, and flush user tokens and acls
4754 when they have expired. This was broken when the 1.3 branch
4755 was modified to support VC7 which no longer included
4758 * An initialization problem with the Freelance code was
4759 detected while fixing the callbackRequest. The cm_rootSCachep
4760 object is obtained during afsd_InitDaemons() but the callback
4761 information is incomplete. The callback information will not
4762 be obtained until cm_MergeStatus is called from within
4763 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
4764 test for the conditions under which the callback information
4767 * Reports have been filed indicating that callbacks were
4768 being lost. An examination of the code indicated that the
4769 cm_server_t objects were not being properly reference
4770 counted by the cm_scache_t and cm_callbackRequest_t objects.
4771 In particular, the cm_server_t objects may have been freed
4772 from beneath the cm_conn_t objects.
4774 All of the reference counting is now done via the functions:
4779 this improves the ability to track the referrals.
4781 Each cm_BeginCallbackGranting Call now allocates a reference
4782 to the cm_server_t. The cm_EndCallbackGrantingCall either
4783 frees the reference or transfers it to the cm_scache_t
4784 cbServerp field. These are then appropriately tracked
4785 through the cm_Analyze call.
4787 * Ensure that the dnlc hash table is the same size as the
4788 dir name hash table (as per original author's note).
4789 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
4790 for compatibility with 64-bit systems.
4792 * fix smb_ApplyV3DirListPatches to properly apply the hidden
4793 attribute to dotfiles when the infoLevel < 0x101 and
4794 cm_SyncOp has failed.
4796 * Fix the Freelance registry initialization code. There
4797 was a possibility that some systems could end up with
4798 garbage in the registry during a clean install.
4801 * file and directory names beginning with "." will now be given the
4802 hidden attribute when the volume access is anonymous. this matches
4803 the behavior when the volume access is via an authenticated user.
4805 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
4806 key. When a change occurs mark the root.afs data as invalid and
4807 for it to be reloaded on the next access. This allows administrators
4808 to modify the mount point list without restarting the service.
4810 The freelance client used to provide a fake modification time for
4811 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
4812 Added code to extract the last modification time of the Freelance
4813 registry key and use that instead. The time now represents the
4814 most recent mount point change.
4816 * PTS registration of new users to foreign cells has been added to
4819 * Additional Cache Control and Credential Manager options have been
4820 added to the WiX installer. See deployment guide for details.
4822 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4824 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4827 * The loopback installation code may have had a problem updating the
4828 %ETC%\HOSTS file which could have resulted in a premature failure.
4829 Work around code has been added for the case where the file cannot
4832 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4833 because Windows sends 64K blocks when using overlapped writes.
4835 * The default number of server threads was increased from 4 to 25 to
4836 better handle overlapped writes.
4838 * The "AfscredsShortcutParams" registry value was not being properly
4839 loaded by afscreds.exe. Therefore, the default value was always being
4840 used instead of the value set by the installer.
4842 * Windows XP provides downgrade attack detection to prevent an attacker
4843 from being able to force the use of NTLM simply by disrupting
4844 communication with the KDC. This attack cannot exist between the
4845 Windows CIFS client and the AFS Client Service. Therefore, when a
4846 downgrade has been detected the afs pioctl library will force the
4847 establishment of a new CIFS connection using NTLM.
4849 * A locking error was discovered surrounding all references to volume
4850 server lists within the cm_cell.c source file.
4852 * The logged into Windows username was incorrect on Terminal Server
4855 * A new registry value "NonPersistentCaching" was added to the service
4856 parameters key. When set to a non-zero value, the afs cache is stored
4857 in the Windows paging file. There are two limitations to choosing
4859 1. when persistent caching is implemented it won't work with
4860 this flag set since there will be nothing to persist.
4861 2. with this flag set the initial paging allocation cannot be
4862 changed while the service is running
4864 * An initialization bug was discovered in aklog.exe which affected users
4865 who have a domain name for their afs servers which could not be mapped
4869 * afs_config.exe now validates cell names against DNS in addition
4870 to the CellServDB file.
4872 * In order to allow the freelance client to connect to a volume with ID
4873 equal to 1 on the default cell we changed the fake root.afs volume ID
4874 once again. This time we choose 0xFFFFFFFF. In addition, we change
4875 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4876 It will now be impossible for a volume ID to match that of another
4877 cell unless the client is connected to 0xFFFFFFFD cells. That should
4878 be enough room for growth.
4880 * Fix "fs mkmount" command to work with UNC paths and when
4881 started from non-AFS drives. It is now possible to create a mount
4882 point in the freelance fake root.afs volume with the command
4884 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4888 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4889 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4891 * The algorithm used to re-attempt access to the servers associated with
4892 a volume has been altered to properly address the case in which all
4893 servers have been marked down. The previous algorithm did not reset
4894 the server's down flags so the servers were never actually retried.
4895 This caused a problem with active volumes if the network connectivity
4896 was lost as could be the case with a network cable removal, wireless
4897 drop, or laptop hibernation. With the fix volume access is restored
4898 almost instantenously when network connectivity becomes available.
4900 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4901 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4902 listing of AFS submounts and freelance mount points. Support for
4903 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4904 significantly enhances the behavior of AFS volumes within the Explorer
4905 Shell. For instance, "AFS" now shows up as server in the Explorer
4906 with each submount or freelance mount point visible as a share.
4907 The right click menu in each folder now works with full functionality
4908 on a consistent basis.
4910 * The network provider can be configured to have different behavior
4911 depending on the domain that the user logs into. These settings are
4912 only relevant when using integrated login. A domain refers to an
4913 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4914 local machine (i.e. local account logins). The domain name that is
4915 used for selecting the domain would be the domain that is passed into
4916 the NPLogonNotify function of the network provider. (see registry.txt
4919 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4920 "Authentication Cell" which may be used to specify a default
4921 authentication cell for afscreds.exe which is different from
4922 the default cell for the AFS Client Service daemon.
4924 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4925 afslogon.dll moved to %WINDIR%\System32\.
4926 New registry entries added to register the dll for Winlogon events.
4928 The logoff event will now force a call to ktc_ForgetAllTokens()
4929 using the context of the user being logged off as long as the
4930 user's profile is not loaded from within AFS. If the profile
4931 was loaded from AFS we can't release the tokens since the Logoff
4932 event is triggered prior to the profile being written back to
4933 the its source location. This is now performed in an XP SP2
4936 * Windows XP SP2 Internet Connection Firewall interoperability
4939 * The %WINDIR%\afsdsbmt.ini contains four sections:
4940 Submounts, Drive Mappings, Active Maps and CSC Policies.
4941 The Submounts and CSC policies are now stored in the registry under
4942 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4943 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4944 The Drive Mappings and Active Maps are stored in the registry under
4945 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4946 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4948 There is no automatic migration of this data as it would be impossible
4949 to consistently migrate data to user profiles which may not be active
4950 when the machine is updated.
4952 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4953 fake root.afs volume. For the same reasons as for the cellservdb file,
4954 this information should not be in %WINDIR%. This information is now
4955 kept under the registry key
4956 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4958 The data from the afs_freelance.ini file will be automatically
4959 migrated to the registry on first execution of afsd_service.exe
4961 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4962 troublesome for several reasons. One, it is confusing for those who
4963 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4964 Two, this file is not a Windows Profile formatted file. Three,
4965 applications should not be reading or writing to %WINDIR%. It causes
4966 problems for Windows Terminal Server.
4968 The new location of CellServDB will be the OpenAFS Client install
4969 directory which is by default C:\Program Files\OpenAFS\Client and can
4970 be determined by querying the registry for
4971 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4973 The existing afsdcell.ini will be migrated by the NSIS installer.
4974 The Wix installer must still be updated to do the same.
4976 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4977 High Security mode; and to add Terminal Services compatibility registry
4978 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4979 configuration files in %WINDIR%.
4981 * Add support for authenticated SMB connections. This will remove
4982 the need for high security mode in most situations. Both NTLM
4983 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4984 only NTLM can be used even though Kerberos is now supported. The
4985 reason is that it is not possible to construct a service principal
4986 which is unique to each individual machine.
4988 SMB Extended Auth does not work on XP SP2 unless one of two registry
4989 modifications are made:
4991 (1) To disable the check for matching host names on loopback connections
4992 set this key. This does not require a reboot:
4994 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4995 "DisableLoopbackCheck"=dword:00000001
4997 (2) To add the AFS SMB/CIFS service name to an approved list. This
4998 does require a reboot:
5000 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
5001 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
5003 afsd_service.exe will automatically add the current Netbios Name
5004 to the BackConnectionHostNames list and then temporarily disable
5005 the loopback check for one cycle of startup/shutdown of the service.
5006 We assume most folks do not start/stop without a reboot so this
5007 will be adequate in most cases.
5009 * Fix security hole in afslogon.dll which allowed passwords to be
5010 sent in clear text to the KDC in a misformed principal name.
5012 * Fix cm_GetCell() to properly handle expired dns entries
5015 * If Freelance mode is active and the afs_freelance.ini
5016 file does not exist, do not create an empty file.
5017 Instead create a file containing ro and rw mountpoints
5018 to the default cell using the standard conventions.
5020 * Modify the Freelance support to handle the ability
5021 to create rw mount points in the fake root.afs volume.
5023 * Changed the RPC mechanism used for token setting from
5024 named pipes to local. Use of named pipes can be restored
5025 by setting the environment variable AFS_RPC_PROTSEQ to
5028 Named pipes were required when a Windows 9x system was
5029 using a NT system in gateway mode which is incompatible
5030 with our use of local loopback adapters.
5032 * In afscreds.exe, if a username of the form user@REALM is
5033 specified and no password is specified, do not perform a
5034 kinit operation. Only perform the aklog functionality.
5036 * Add a new registry value which allows the number of processors
5037 on which afsd_service.exe executes to be restricted. Valid
5038 values are 1..numOfProcessors
5040 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
5044 * A second MSI based installer option is now available.
5046 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
5047 request tickets for user/instance@REALM instead of just user@REALM
5049 * Modify the Power Management Notify routine to wait for the Hard Dead
5050 timeout period instead of a fixed 19 seconds. With the longer timeout
5051 periods Hibernation and Standby could never succeed when network
5052 connectivity is not available.
5054 * The following fs.exe commands are now restricted to Administrator:
5055 - checkservers with a non-zero timer value
5058 - sysname with a new sysname list
5067 setting the default sysname for a machine should be done via the
5068 registry and not via "fs sysname".
5070 * NSIS installer adds options to install Debugging Symbols
5071 and the Microsoft Loopback Adapter; the user is now also
5072 given the ability to select the afscreds.exe startup options.
5074 * Build system modified to generate symbols for FREE (aka RELEASE)
5075 builds as well as CHECKED (aka DEBUG) builds
5077 * Sites which have a volume ID of 0x20000001 assigned to their
5078 root.cell volumes have been experiencing problems with accessing
5079 the root.cell volume of their cell when Freelance mode has been
5080 active. This was because 0x20000001 was assigned to the fake
5081 root.afs volume created by freelance. The fake volume id is
5082 now set to 0x00000001 to prevent conflicts.
5084 * The timeout logic in the AFS Client Service has been wrong
5085 for sometime. It is based on two different assumptions.
5086 First, the SMB client timeout is a fix value as was the case
5087 with OS/2 Lan Manager. This assumption is incorrect. The
5088 SMB timeout in Windows is a dynamic value computed based upon
5089 a fixed minimum timeout to which is added time based upon the
5090 size of the request and the performance characteristics of
5091 the connection. Second, it is the responsibility of the
5092 SMB Server to enforce the timeout requirements of the client.
5093 This is untrue. The SMB Server cannot be expected to know
5094 the requirements of the client. More importantly, if the
5095 SMB server uses the SMB client timeout as a value to restrict
5096 its behavior as an RX client, the performance characteristics
5097 of the local SMB session would be used to prematurely terminate
5098 WAN connections with significantly different performance
5101 The timeout logic has therefore been modified in the following
5103 . the Lan Manager Workstation (SMB) Session Timeout is used only
5104 as a basis for configuring the Connection Dead Timeout
5105 and Hard Dead Timeout values. The Connection Dead Timeout
5106 must be at least 15 seconds longer than the SMB Timeout
5107 and the Hard Dead Timeout must be at least double the
5108 Connection Dead Timeout.
5109 . New registry entries have been added to allow the Connection
5110 Dead Timeout and Hard Dead Timeout values independent of the
5111 Lan Manager Workstation Session Timeout
5112 . The test to enforce the SMB Client Timeout has been removed.
5114 One of the side-effects of removing the enforcement of the SMB
5115 Client Timeout is that regardless of whether or not the SMB client
5116 is available to receive the response (and how would the SMB server
5117 know) the RX protocol response can be used to update the AFS
5118 Client Service state for ready access by future SMB client
5121 This should be the end of the "Server paused or restarting messages"
5123 * Add support for arbitrary UNC paths to the pioctl() support.
5124 This enables the fs commands as well as the AFS Shell Extension
5125 to work correctly when UNC paths are being used.
5127 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
5128 DNS if the cell configuration cannot be determined via CellServDB
5130 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
5131 are ever set more than once at a time
5133 * Fix the management of lists of cm_cell_t structures when using
5134 DNS to lookup cell information. The previous code would fail to
5135 reuse the same cellID for a cell if DNS was used more than once
5136 for a given cell name. When the ttl expired, a single cm_cell_t
5137 could be inserted into the cm_allCellsp list more than once
5138 producing a loop. In addition, the vlServerp list belonging to
5139 the cell was not freed resulting in improper refCounting of the
5142 * Add DNS support to cm_IoctlNewCell() which previous only examined
5145 * Add cm_FreeServer() function and call it from cm_FreeServerList()
5146 to properly garbage collect cm_server_t objects
5148 * Add numVCs variable to smb.c to track the number of smb_vc_t
5149 objects created and use it to initialize the vcID field which
5150 previously was set to 0 in all objects resulting in FindByID
5153 * Fixed DNS lookups to work consistently throughout the OpenAFS
5154 product instead of just from within the afsd_service.exe
5156 * Added a runtime check to ensure that AFS Client Service SMB
5157 Server is accepting connections before attempting to mount
5160 * Read IP addresses for volume servers out of the CellServDB
5161 file if gethostbyname() on the hostname fails.
5163 * Fix getcellconfig() to populate both the Hostnames as well
5164 as IP addresses when loading cell data via DNS
5166 * Increase the Connection Dead Time to 50 from 20 seconds
5167 Increase the Hard Dead Time to 120 from 40 seconds
5168 (matches the Unix values)
5170 * Fixed an assertion validating the number of allocated NCBs
5172 * Fixed the build environment to consistently build for
5173 Windows 2000 and above. (APPVER = 5.0)
5175 * Fixed rx_debug to properly validate the receipt in incoming
5176 data with select() and recvfrom(). Do not copy data out of
5177 the socket buffer unless success is indicated.
5180 * afsd_service.exe will now display a message box to the
5181 desktop when it terminates due to an IP Address Change.
5183 * installer no longer deletes AFS Server configuration data
5186 * installer generates a warning dialog if the RPC service
5187 is not properly configured
5189 * installer compressed with lzma instead of bzip2
5191 * afsd_service.exe shutdown crash solved once and for all
5193 * reference counting of smb_vc_t data structures improved
5195 * name space collision of smb_fid_t event objects corrected
5197 * the output of "fs memdump" is now written to
5198 %WINDIR%\TEMP\afsd_alloc.log
5200 * the file TaAfsApp_1033.dll is now properly installed allowing
5201 the User Manager to start
5203 * a new algorithm is used for computing filename pattern matches
5205 * afscreds.exe now accepts user names containing instance
5208 * Fix the Directory Name Lookup Cache to be case-sensitive.
5209 This is crucial in environments in which a Windows client
5210 is accessing a directory with more than one filename that
5211 differs only by case. If the directory contains "FOO"
5212 and "Foo". You want "DEL Foo" to delete the correct one.
5213 We still have a problem in that "DEL foo" will delete a
5214 random filename. This will be addressed in a future release.
5216 * Fix afscreds.exe -M option (renewMaps) to work when High
5217 Security mode is off. Also, remember to disable the ActiveMap
5218 flag in afsdsbmt.ini when a drive mapping is removed.
5220 * Updates to NSIS installer script. AFS Server configuration
5221 data will not be destroyed on un-install or re-install.
5222 Use a better compression algorithm.
5224 * afslogon.dll now uses KFW to obtain tokens when available
5226 * afslogon.dll when given an all uppercase username will
5227 attempt to authenticate with both the uppercase name
5228 and an all lowercase variation
5230 * DST modification removed. The fix appears to make things
5231 worse after a reboot of the machine.
5233 * fs.exe: added "cscpolicy" which is used to
5234 change client side caching policy for AFS shares
5236 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
5237 [-documents] [-disable] [-help]
5239 * Several uninitialized variables have been initialized
5241 * It is now possible to obtain tokens using cross realm
5242 Kerberos within afscreds.exe:
5244 user: jaltman@ATHENA.MIT.EDU
5246 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
5247 will will in turn be used to obtain afs@DEMENTIA.ORG.
5248 The resulting token will be stored with the display name
5249 jaltman@ATHENA.MIT.EDU@dementia.org
5251 * aklog.exe has been added to the client
5253 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
5254 [[-p | -path] pathname]
5258 -d gives debugging information.
5259 krb_realm is the kerberos realm of a cell.
5260 pathname is the name of a directory to which you wish to authenticate.
5261 -noprdb means don't try to determine AFS ID.
5262 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
5263 (default is Kerberos V)
5264 No commandline arguments means authenticate to the local cell.
5267 * All of the resource files have been restructured to adhere to
5268 a set of rules IBM implemented for loading string resources.
5269 These rules had either been forgotten or were not discovered
5270 by folks working on the OpenAFS sources. The end result was
5271 memory corruption. This is primary item which was preventing
5272 the AFS Server from working.
5274 * Increased the size of the maximum ticket size stored in a token
5275 from 344 bytes to 12,000. Increased the buffers used to convey
5276 messages between the pioctl() caller and the SMB Server from
5277 1000 bytes to 12,512. The code appeared to have been writing
5278 above the top of the stack by quite a few number of bytes.
5279 (The increased ticket size is necessary for the next item.)
5281 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
5282 Instead the raw Kerberos 5 ticket is used in its entirety. This
5283 is extremely important as it allows us to use pure Kerberos 5 KDCs
5284 as the source of the AFS authentication. The use of up to 12,000 byte
5285 tickets will allow tickets produced by all versions of Microsoft
5286 Active Directory to be used.
5287 - create a user account.
5288 - designate it DES only
5290 - specify its UPN to be "afs@realm"
5291 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
5293 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
5294 is special when using long file names. (you must use "*.*" and not "*")
5295 Instead only enforce it when performing 8dot3 searches.
5297 * Fixed the DST problem with creation times being set one hour ahead
5299 * Fixed the problem when using \\afs\cell-alias. For example,
5300 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
5301 for the alias name; instead simply expand the name. One of the
5302 symptoms of this problem was a loss of acquired tokens.
5304 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
5305 of strings. (Only English strings provided.)
5307 * Fixed the installer to properly replace in use files.
5309 * Fixed the build system to cleanup generated component version files
5311 * The release build compiled with MSVC 6.0 compiler to avoid the
5312 afsd_service.exe shutdown crash. This does not solve the problem
5313 but simply avoids it for the time being.
5317 * fix afslogon.dll to not corrupt memory when High Security mode
5320 * fix afsd_service.exe to not attempt to restore the stack when
5321 an exception occurs. (not safe in multi-threaded programs)
5323 * fix uninstaller to properly remove the CRT and MFC DLLs
5325 * remove a Message Box from afscreds.exe when getcellconfig()
5326 fails on a kerberos realm which is not a cell
5328 The following is a list of changes to the OpenAFS for Window client
5331 * "fs setserverprefs" will leave afsd service deadlocked
5333 * "vos listaddrs" will core dump
5335 * installer sets the appropriate keys to support Integrated Logon
5337 * installer disables the "Find Lana by Name" functionality as it
5338 was causing headaches for many users
5340 * fix the intermittent crash of the power management thread when
5341 shutting down the AFS Client Service
5343 * optimizes the obtain drive mount list functionality which is
5344 executed every time the mount tab in afscreds.exe and afs_config.exe
5345 are refreshed. (this happens a lot)
5347 * fix the service shutdown logic. add the STOP_PENDING state
5348 and do not accept additional service events after we declare
5351 The following is a list of changes to the OpenAFS for Window client
5354 * flexelint was run against the source tree and hundreds (perhaps
5355 thousands) of corrections were applied to ensure prototypes
5356 were in use; types were used consistently; variables were
5357 initialized; unused variables were removed; etc.
5359 * A wide variety of instrumentation was added including the
5360 ability to produce a stack trace from within afsd_service.exe
5363 * Dynamic configuration of the RDRtimeout value based upon the
5364 LanMan Workstation Session Timeout
5366 * The mount root no longer needs to be called "/afs". This
5367 is now set by a registry value "MountRoot" within the key
5368 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
5370 * The cell list is now only read out of afsdcell.ini when the
5371 file changes instead of each time a cell is resolved.
5373 * Thread synchronization was added to cm_server.c and ktc_nt.c
5375 * All calls to GlobalAlloc()/GlobalFree() were replaced with
5376 calloc()/free(). The Global functions were needed on Windows 3.x
5377 but have caused a variety of problems on the Win32 platforms.
5378 Avoiding them is highly recommended by several Microsoft
5379 Knowledgebase articles
5381 * Support for Symbolic Links added to the AFS Shell Extension
5383 * Added a registry value "OverlayEnabled" to determine if
5384 Shell Extension Overlays should be enabled.
5385 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
5387 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
5388 separate trees for checked and free builds. Build system supports
5389 a custom directory src\WINNT\extra which can be used as a grafting
5390 location of organization specific additions to the build tree.
5392 * New installer built using NSIS 2.0.
5394 * Named all kernel objects in order to allow them to be monitored
5395 with tools such as SysInternals' ProcExp.exe.
5397 * Introduced new EventLog framework for AFSD
5399 * Introduced Power Management interface to AFSD for Standby and
5400 Hibernate modes to allow cache to be flushed prior to network
5403 * Utilize Win32 DNSQuery API instead of internal routines. This
5404 allows DNS SRV queries to be sent to all current domain name
5405 servers. Not just one specified in an INI file. DNS is now
5408 * "NetbiosName" registry value may be used to specify a fixed
5409 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
5410 when the loopback adapter is in use. If you need to use the
5411 old notation with a loopback adapter installed specify a registry
5414 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
5416 * Refactor all modules which depend on LAN Adapter and NetbiosName
5417 determination in a new library: lanahelper.lib. This allows for
5418 consistent behavior throughout the product.
5420 * Move the afsd.log and afsd_init.log files to the directory specified
5421 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
5422 for services. Added the Date to the log entries.
5424 * New registry value "RxMaxMTU" used to limit the size of the RX
5425 packets sent by the AFS Client Service to the Server. In order
5426 to enable OpenAFS to work across the Cisco IPSec VPN the packet
5427 size must be restricted to 1264 or smaller. The latest NSIS
5428 installer sets a value of 1260 by default.
5430 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
5431 packets. This is not needed in order to work across the Cisco
5432 VPN but might be needed for other network environments. This
5433 value is not set by the NSIS installer.
5435 * New registry value "HideDotFiles" is used to apply the Hidden
5436 attribute to files whose names begin with a '.'. This value
5437 is set by the NSIS installer.
5439 * New registry value "MaxMpxRequests" allows the maximum number
5440 of multiplexed sessions to be configured at run time. This
5441 value is not set by the NSIS installer. The default value is
5444 * New registry value "MaxVCPerServer" allows the maxmimum number
5445 of VCs per server to be configured at run time. This value is
5446 not set by the NSIS installer. The default value is 100.
5448 * New registry value "AllSubmount" allows the "all" submount to
5449 be disabled by setting its value to 0x00.
5451 * Allow cells names to be valid mount points
5452 \\<netbiosName>\<cellname>
5454 * Store the active state of drive mappings in order for afscreds.exe
5455 to restore them upon startup
5457 * Add exception handling to generate a Stack Trace to the afsd_init.log
5458 file if one happens to occur.
5460 * Add lots of logging to help detect the cause of invalid SMB packets
5462 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
5463 conversion of Kerberos 5 "afs" service tickets. Supports auto-
5464 renewal of expiring tokens as long as afscreds.exe is running.
5466 * New afscreds.exe command line options:
5468 -M = renew drive maps
5469 -N = ip address change detection
5472 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
5473 determines whether or not MIT Kerberos for Windows should be used
5474 to obtain tokens via Kerberos 5 tickets.
5476 * New registry value "AfscredsShortcutParams" in
5477 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
5478 determines the command line parameters to be specified when "fixing"
5479 the AFS Shortcut in the user's startup folder.
5481 * The "ShowTrayIcon" registry value has been moved from
5482 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
5483 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
5485 * The <cell name> registry values used to store the token expiration
5486 reminders have been moved from
5487 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
5488 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
5490 * Obtain the Logon User Name from the Explorer key when available
5492 * new text document doc\txt\winnotes\registry.txt lists all registry
5493 values used by OpenAFS (excluding the AFS Server)
5495 * BUG: rx_securityClass objects were not properly reference
5496 counted and were never freed.
5498 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
5499 would be generated. The existence of a server does not imply
5500 that it is not down. If all of the servers for a cell are down
5501 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
5504 * BUG: the directory name lookup cache failed to free the entries
5505 in the cache when the name cache entries cycled. The entries
5506 in the cache would become dereferenced without being freed.
5508 * BUG: fs setserverprefs could be executed without Administrator
5511 * BUG: the number of allocated NCB objects (100) exceeded the number
5512 which could actually be waited upon by the kernel (64). Any objects
5513 which were utilized above the limit could never have event completions
5516 * BUG: smb_username_t objects were not being reference counted and
5517 were not properly freed.
5519 * BUG: smb_tid_t objects could under unusual circumstances be freed
5520 before they were no longer referenced.
5522 * BUG: smb_fid_t object pointer were frequently used even when
5523 their value could be NULL. They were not properly released and
5524 therefore they were never freed.
5526 * BUG: smb_packet_t data structures were not completely initialized
5529 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
5530 Denied" because that causes the Explorer Shell to try again until
5531 access is obtained. Instead return "Remote Resources" which allows
5532 the shell to move on and treat the error as transient.
5534 * BUG: when initializing the NCBreturns structure, separate Event objects
5535 were created for each NCB although a single Event object was supposed
5536 to be shared by all.
5538 * BUG: smb_dirSearch_t objects were not being properly referenced counted
5541 * BUG: smb_tran2Packet_t objects were not being properly referenced
5544 * BUG: directory path creation did not handle the case of multiple
5545 directories requiring creation in one attempt
5547 * BUG: SMB requests which required an Extended Response were ignored.
5548 This prevented some files from being written to AFS volumes.
5550 * BUG: character strings were being freed even after they were
5551 inserted into in use data structures
5553 * BUG: inconsistent usernames were used when High Security mode was
5554 enabled. (there is still much to do in this area)
5556 * BUG: pioctl() calls which require out of band RPC operations were
5557 susceptible to race conditions when performed by multiple processes
5559 * BUG: memory allocation and deallocation crossed instances of the
5560 C Runtime Library producing memory leakage and corruption in
5561 afscreds and the client configurator.