2 [there was no 1.6pre1 for Windows]
4 * vos commands now manipulate servers by UUID and can
5 recognize multi-homed servers.
7 * afs_config will not longer set the Tray Icon State
8 in the registry if the checkbox is not present in
9 the dialog. [RT 128591]
11 * Heimdal's roken utility library has been added
12 as \Program Files\Common\afsroken.dll
14 * When probing servers to determine if they are up or down
15 no longer issue an RXAFS_GetTime RPC to servers that indicated
16 that they do not understand the RXAFS_GetCapabilities RPC.
17 Since they responded it is known that they are up.
19 * AFS Explorer Shell Extension now works from folder
20 backgrounds. Overlays for mount points and symlinks
21 are present in the dll, but are not registered at present
24 * Do not use RankServerInterval registry value as the value for
25 PerformanceTuningInterval.
27 * CellServDB updated to 13 Dec 2010 release from grand.central.org
29 * Add "fs chmod" command and display current mode as part
30 of "fs examine" output.
32 * When the data version of a mountpoint or symlink changes,
33 the target string in the cm_scache_t object must be cleared.
34 Otherwise, the new target will not be queried.
36 * "fs checkservers" now includes vldb servers in the output
37 and only lists multi-homed servers once. A multi-homed
38 server that has at least one up interface is no longer
39 considered to be down.
41 * When asynchronously storing dirty data buffers to the
42 file server ensure that (a) the cm_scache_t object and
43 the cm_buf_t object are for the same File ID so that
44 locking and signalling work properly; and (b) if the
45 FID no longer exists on the file server, do not panic,
46 just discard the buffer.
48 * When processing VNOVOL, VMOVED and VOFFLINE errors perform
49 server comparisons by UUID or address and not simply by
50 cm_server_t pointer. Otherwise, server failover may not
53 * Do not preserve status information for cm_scache_t objects
54 when the issuing server is multi-homed.
56 * Giving up all callbacks when shutting down or suspending
57 the machine is now significantly faster due to the use
58 of an rx_multi implementation. (This functionality is
59 still off by default and must be activated by a registry
62 * Race conditions were possible when updating the state
63 of the cm_volume_t flags and when moving the volumes
64 within the least recently used list.
66 * Ensure that the lanahelper library does not perform a
67 NCBRESET of each lan adapter when enumerating the
68 current network bindings. Correcting this permits OpenAFS
69 to work on Windows 7 when the network adapter settings
72 * Fix creation of mount points and symlinks as \\AFS\xxxx
76 * Reimplement FetchData and StoreData operations to make
77 use of rx_Writev and rx_Readv which is faster than than
78 multiple calls to rx_Write and rx_Read. Throughput
79 improvements of up to 20% are possible.
81 * Many improvements to the rx rpc stack:
85 - Fast recovery is not triggered by packet resend timeouts
87 - Socket errors are now detected and reported
89 - avoid repetitive memory allocations
91 - avoid calling gettimeofday() for every packet
93 - reduced dependency on the call->lock
95 - When a call receive is completed, ack all of the packets
96 immediately. Do not wait for the first response packet
99 - Do not hold the call->lock during rx_Write* and rx_Read*
100 operations when the call fields in use can only be accessed
101 by the application thread.
103 - Reduce the default window size from 128 back to 32 because
104 the current overhead from walking packet queues in the rx
105 listener thread significantly hampers the ability to
106 process incoming packets from the network.
108 - Only backoff the peer timeout once for a series of naks.
110 - Release builds are lean and mean. All debugging and
111 packet tracking code is disabled.
113 * Version number is logged to the Windows Application Event
114 Log as part of the start pending message.
116 * MSI installers now backup and restore configuration
117 data during upgrades.
119 * Rx pmtu discovery has been disabled by default. Use
120 HKLM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
121 RxPMTUDiscovery "DWORD"
122 value to re-activate.
124 * cm_BkgDaemon event evaluation has been optimized to avoid
125 unnecessary overhead.
127 * If a volume is known to be readonly, then locally fail any
128 change request without sending it to the file server. This
129 avoids triggering the file server's abort threshold.
131 * Do not leak cm_volume_t objects from the volume LRU queue.
133 * Ensure that cm_NameI errors are acted upon promptly.
134 There are many cases in the SMB server where an error from cm_NameI()
135 was either ignored or not acted upon until several other operations
136 are performed that could result in the same error being repeated.
137 This is a mistake which did not have negative side effects until
138 additional checks for callback status were added recently.
140 At present, if a CM_ERROR_ACCESS error is returned and ignored,
141 subsequent attempts to operate on the same cm_scache_t will result
142 in additional queries to the file server that will also end in an
143 abort response. This can trigger the file server to delay responses
146 * Fix fs mount point and symlink create operations which were
147 failing due to improper directory separator termination.
149 * RXAFS_GetVolumeStatus can return VNOVOL, VMOVED, etc. In
150 order to process them and update volume state a fid must
151 be passed to cm_Analyze(). Use the volume root fid.
153 * Negative caching for volume lookups that fail with either
154 VL_NOENT or VL_BADNAME. The expiration time is five minutes.
155 This prevents volume lookup storms.
157 * Provide each volume lookup its own two hour lifetime instead
158 of reseting all volume location information every two hours.
159 This prevents the first smb path evaluation after the volume
160 location data has been discarded from potentially causing an
164 * When the SYSTEM account is in use, the SMB connection is
165 not always established as S-1-5-18. Sometimes it uses an
166 anonymous, S-1-5-7, connection. In both cases the SMB
167 authentication name is the nul string. To better identify
168 when an SMB connection is local system, the smb_username_t
169 is now stored as a SID instead of a user name. When a SID
170 is used, a new flag bit is set SMB_USERNAMEFLAG_SID.
172 In smb_SetToken if AFS_PIOCTL_AFSLOGON is set and the user
173 connection is not SYSTEM, the RPC SID is checked. If RPC
174 SID is SYSTEM, that supercedes the SMB authentication.
176 * Replace STATUS_IO_TIMEOUT with RPC_NT_SERVER_TOO_BUSY for
179 * Replace STATUS_PATH_NOT_FOUND with RPC_NT_SERVER_UNAVAILABLE
180 for CM_ERROR_ALLOFFLINE and CM_ERROR_ALLBUSY.
182 * Optimize behavior of buf_CleanAsyncLocked(). Avoid searching
183 for the cm_scache_t object by FID if it is already known.
185 * If the readonly file attribute is set (stored as a unix mode)
186 then a CreateFile operation should fail if the file is opened
187 for DELETE in combination with any other privilege.
189 * If the directory buffer contents are garbage we can crash
190 the service. Add some simple validation checks to ensure
191 that cm_dirEntry_t objects have the correct flag value and
192 that the name strings are not too long.
194 * If the bulkStat errorCode indicates that a particular object
195 is inaccessible due to a VIO error, we must update the server
196 status appropriately in order to permit failover.
198 * An RX_MSGSIZE error is returned by the new PMTU detection
199 code. It is critical that such an error result in a retry of
200 the operation that failed. Otherwise, the PMTU detection can't
201 work and the server will be marked down.
202 Secondly, it is important that such errors not leak to the
203 application layer. Map them to CM_ERROR_RETRY in all cases.
205 * Translate RX_RESTARTING to the same cache manager error code
208 * When logging server volume instance errors to the windows
209 application event log, be sure to log the cell as well.
210 Translating from server ip address is non-trivial. Make it
211 easier for administrators triaging issues to plug the volume
212 and cell info into vos commands.
214 * Since unix mode bits are represented in octal in most cases
215 make sure we log them that way.
217 * The cm_enforceTrailingDot() function failed if there was a
218 trailing dot followed by white space.
220 * Rx: Do not send RX_MSGSIZE if the PMTU size did not change.
223 * A crash could occur if a single letter share name was
224 dynamically evaluated as matching a cell name. This was
225 due to a failure to treat the comparison of two nul
226 strings as identical.
228 * Docbook validation is performed by xmllint during builds.
230 * Volume package bugs in the file server can result in VOFFLINE
231 being returned to the client instead of VNOVOL or VMOVED. As
232 a result the Unix CM treats VOFFLINE the same as VMOVED and VNOVOL.
233 The Windows client has not. As a result, bugs in the file server
234 can cause the Windows client to lose if the volume has in fact
235 been moved to another server.
237 As part of this change, the volume location list is updated prior
238 to the volume status being applied to the server from which the
241 * If the pages to be flushed are from a readonly or backup volume
242 they can't be dirty. In this case there is no need to stabilize
243 the pages before they are locked and recycled.
244 Stablilization is performed on the cm_scache_t object so do not
245 stabilize and unstablize for each cm_buf_t object.
247 * Over the years the processing of the Freelance callbacks have
248 added functionality that behaves much more like FetchStatus checks
249 to a file server. If the data version of the object has changed,
250 get the new data. Given that is the case, we can remove much of
251 the original refresh logic that is rather race prone. Say goodbye
252 to cm_fakeGettingCallback and cm_fakeDirCallback.
254 * When processing a pioctl path with either smb_ParseIoctlPath or
255 smb_ParseIoctlParent, cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK|GETSTATUS)
256 must be called on the cm_scache_t object to ensure that it is up
257 to date before we permit cm_Lookup or other operations to be performed
258 on it. Add the cm_SyncOp() call to smb_ParseIoctlPath and
259 smb_ParseIoctlParent to ensure it is done for all pioctl operations.
261 * The 32-bit tools installer was displaying client configuration wizard
262 pages. Since, the 32-bit tools share configuration with the AFS
263 service, the 32-bit tools installer should not be attempting to change
264 client configuration. Remove said wizard pages from the 32-bit tools
267 * cm_data.rootSCachep is a global pointer to the cm_scache_t that
268 represents the root.afs volume root directory. Throughout the
269 code this cm_scache_t was being used without ensuring that a
270 callback to the volume is in fact valid.
272 Under most circumstances this would not be a problem. However,
273 it is possible for a request to fail due to the lack of a callback
274 at a critical moment. Add a new function cm_rootScachep() that
275 attempts to ensure that a callback is present (if possible) prior
276 to use the cm_scache_t object.
278 * When the contents of the Freelance root directory changes the fake
279 directory buffers are updated and a fakeDirVersion is incremented.
280 The dataVersion of the cm_scache_t object is supposed to be updated
281 on the next access by performing a fake get callback request.
282 Unfortunately, this did not always occur because of a race. If another
283 Freelance object is updated first, the root directory object would
284 never successfully get a fake callback.
286 This patchset ensures that the generation of the fake directory
287 buffer content and the callback are obtained under the same set
288 of locks thereby removing the race.
290 * Enforce in afsd_service that tokens can be set for alternate
291 userids only if the SetToken pioctl is called from the
292 Local System account.
295 * Revise SMB QuerySecurityInfo Response for MS10-020
297 MS10-020 (http://support.microsoft.com/kb/980232) has caused
298 many problems for implementors of SMB 1.0 servers and applications
299 that call GetFileSecurity() without checking the return code to
300 determine if the call succeeded. The gist of the vulnerability
301 was that the SMB redirector would pass any buffer it received
302 to the application regardless of whether or not it was valid.
303 MS10-020 protects the applications by strictly validating the
304 SMB response data structure and the data in the security descriptor
307 The problem for SMB 1.0 server implementors is that there have
308 been at least three different protocol descriptions for
309 NT_TRANSACT_QUERY_SECURITY_DESC published over the last decade
310 and all of them are incomplete. Therefore, just about no one but
311 Microsoft has an SMB 1.0 server implementation that produces the
312 exact out that they are expecting to validate.
314 The end result is that in an attempt to protect applications from
315 crashing due to invalid input being passed in directly caused
316 dozens of applications to crash by not returning any security
317 descriptor data at all. Even when the applications didn't crash
318 they might not have been able to save their data. Cisco WAAS
319 and NetApp DataOnTap systems were most adversely affected and
320 they have had CIFS protocol licenses for many many years.
322 To fix OpenAFS here is what needed to be done:
324 1. Instead of returning a security descriptor that gives ownership
325 to the NUL SID, give it to the Everyone SID and set the flag
326 that states that everyone has full access.
328 2. Validate the input parameters. In particular, check to ensure
329 that the SMB file descriptor is valid and the file has not
332 3. Enforce the maximum output data and parameter counts.
334 4. Handle buffer overflow and buffertoosmall conditions
335 in the manner that Microsoft expects them to be handled.
336 In particular, note that the parameter data which is returned
337 in the SMB Data Region is not counted in the Data Count.
338 Even if MaxData is 0, we can still return parameters values
339 as long as MaxParm is large enough.
341 * Prevent use of AFSCache file contents if mapped to
344 * The Windows version of "fs newcell" did not accept any parameters
345 and behaved quite differently from the Unix version. Instead of
346 permitting new cell information to be added, the Windows version
347 simply forced the existing cell information to be reacquired.
349 This update adds a new pioctl, VIOCNEWCELL2, to support the
350 implementation of a Unix-style "fs newcell". The functionality
351 added here differs from the Unix version in the following ways:
353 1. "fs newcell" with no arguments is still accepted
354 in order to maintain compatibility with prior Windows
357 2. "fs newcell -cell <cell> -dns" instructs the cache manager
358 to add the new cell but obtain the vldb server info from
361 3. "fs newcell -cell <cell> ... -registry" instructs the cache
362 manager to add the new cell and also save the cell configuration
363 data in the registry for use the next time the service restarts.
365 4. The -vlport and -fsport options are accepted although the
366 -fsport value is currently unsupported by the cache manager.
368 * New registry value "FreelanceImportCellServDB" instructs Freelance
369 to create a mount point for every cell name listed within the
372 * Path MTU discovery for Rx is activated.
374 * Rx socket input buffer is converted to a circular buffer.
376 * Fix usage of cm_FreeServerList(). Do not set the server list
377 pointer to NULL after calling cm_FreeServerList(). Doing so
378 can result in a memory leak.
380 * Only enable Rx NAT pings on a single anonymous connection at a
383 * Fix cm_IoctlSkipQueryOptions() buffer management. Prevents a
384 potential read beyond end of memory buffer.
386 * Reduce requested privileges when reading registry CellServDB
387 to the minimum required.
389 * Add support for RPC Pipe Service NetWkstaGetInfo levels
390 101 and 102 which are called on Windows 7 and 2008-R2.
392 * Prevent integer overflow during quota percent used calculation
393 in Explorer Shell Extension (RT 126846)
395 * Generate a meaningful error if "fs listacls" or "fs setacls"
396 are executed on the Freelance root.afs volume.
398 * RXAFS_InlineBulkStat errors must be processed via cm_Analyze.
399 RXAFS_InlineBulkStatus does not return errors such as EACCES,
400 VNOVOL, VNOVNODE, VOFFLINE, VBUSY, VIO, VMOVED, etc. as an RPC return
401 code. Instead they are returned in the status info errorCode field
404 Traditionally, the error associated with the first FID in the query
405 list has been returned to the caller of cm_TryBulkStatRPC().
406 However, the error has never been processed through cm_Analyze()
407 which means that the per-vnode processing for VNOVNODE and the volume
408 global processing for VMOVED, VNOVOL, etc. has never been performed.
409 As a result, failover to other .readonly volume instances cannot occur,
410 volume moves will not be handled, and files that have been deleted
413 This patchset makes the following changes:
415 1. If an inline bulk operation has been performed and the inline
416 errorCode is a volume global error, then that error replaces
417 the RPC return code within the cm_Analyze() processing for
418 the RPC. This will affect whether or not a retry operation
421 2. The variable 'inlinebulk' is reset to 0 at the top of the
422 cm_Analyze() loop in case failover from an inlinebulk capable
423 file to an inlinebulk incapable file server takes place.
425 3. The FID that is passed into cm_Analyze() is not a real fid.
426 Instead it consists of the cell and volume but vnode = 0.
427 This ensures that the error (if any) is not applied to the
430 4. If an inline bulk operation was performed, prior to performing
431 the cm_MergeStatus() operation a vnode a check is made to
432 determine if an error was returned for that vnode. If so,
433 cm_Analyze() is called with no connection, a fake cm_req_t,
434 the fid, and the error. This permits cm_Analyze() processing
435 to be performed on the file.
437 * Show configuration pages for all types of MSI installations
439 The OpenAFS MSI installer wizard used to not show any configuration
440 pages for "Typical" and "Complete" installations. Setting the
441 workstation cell and logon options during installation required
442 selecting the "Custom" option. Many users choose the "Typical" option
443 during installation, and thus would never see the configuration pages.
444 Therefore, for these users, the workstation cell was being set to the
447 This patch makes the workstation cell and logon option configuration
448 pages visible to all types of installations (except silent
449 installations which show no UI).
451 * cm_LookupInternal creates Freelance mount points and symlinks
452 when queries cannot be found in the Freelance root.afs directory.
453 If the search name is a full cell name for which vldb information
454 can be obtained, then a mount point is added. If the search name
455 is a left-most substring or the full cell name with a dot appended
456 to it, then a symlink was created. This approach created a very
457 poluted Freelance name space.
459 This patchset makes the following changes:
461 1. Do not create symlinks with a dot appended to the cellname
463 2. Do not create symlinks where the left-most substring is not
464 a full dot separated component of the cellname.
466 3. Permit lookups to succeed when we would have created a
467 symlink in the past without creating the symlink.
469 * BPlus tree lookups are much faster than searching through
470 the native directory format on Windows because the case sensitive
471 hash tables cannot be used successfully. Permit BPlus trees
472 to be used except when called with cm_BPlusDirFoo as the action
473 function because cm_BPlusDirFoo is used to build the BPlus trees
474 from the native directory format.
476 * Symlinks are ending up in the Freelance root.afs directory that
477 end with a dot. Make sure it cannot happen.
479 * cm_FreelanceAddMount and cm_FreelanceAddSymlink is supposed to
480 return the allocated FID of the entry that was added. However,
481 cm_NameI is called to perform the lookup without forcing an update
482 of the Freelance fake directory. As a result the entry may not be
485 Force an update prior to calling cm_NameI() by using
486 cm_clearLocalMountPointChange() and cm_reInitLocalMountPoints()
489 * The Freelance fake root directory buffers were not zero-filled.
490 This results in random behavior that can cause the service to
491 terminate unexpectedly.
493 * The validation check for the response from the GetVolumeStatus
494 pioctl is incorrect. The response is not simply a VolumeStatus
495 structure but also several C strings appended to it.
497 * When flushing a file, we need to commit the file length changes
498 as well as the dirty buffers. Call cm_FSync instead of buf_CleanVnode
499 which is called by cm_FSync.
501 * Prevent rx_rpc_stats global lock from being a bottleneck in the
506 * Avoid a race when updating cell vldb server lists
507 that can result in a crash.
509 * Avoid a deadlock when managing CM_SCACHESYNC_STOREDATA
510 state operations for directory objects.
512 * Add new Windows Application Event log messages for
513 VBUSY, VRESTARTING, ALL_BUSY, ALL_OFFLINE, and ALL_DOWN.
514 Include message throttling to prevent the same message
515 from being logged repeatedly within a five second window.
517 * Reduce lock contention by waiting for cm_buf_t I/O
518 operations to complete before permitting cm_SetupStoreBIOD
519 to analyze a buffer for inclusion in a BIOD.
521 * Split the cm_buf_t flags field to separate the flags
522 that are protected by the cm_buf_t mutex from those
523 protected by the buf_globalLock. This eliminates the need
524 to hold both locks everytime the flags field is accessed.
525 Both locks were not held in the past resulting in race
526 conditions that could result in deadlocks.
528 * Add "vos setaddrs" command.
530 * Rx library lock contention avoidance between rx_NewCall and
533 * Rx library races due to inconsistent use of rx_connection
534 conn_data_lock to protect the flags field.
536 * Rx library inconsistent use of RX_CALL_TQ_WAIT which could
539 * Rx library must signal transmit queue waiters when flushing.
540 Otherwise, deadlocks can occur.
542 * In cm_UpdateVolumeLocation, avoid searching for a ".readonly"
543 volume on a numeric volume name.
545 * File buffer allocations whose offsets are beyond server EOF
546 should be locally allocated and zero filled. The file server
547 should not be issued a FetchData rpc which is guaranteed to
550 * Enable integrated logon to work with Windows 7/2008 when
551 user logons are performed with a non-Domain Kerberos principal.
553 * Add Protection Error messages to aklog output.
557 * Prevent the Explorer Shell extension from crashing if
558 symlink creation failed. (126406)
560 * A Rx level NAT ping has been implemented.
561 Add NatPingInterval registry value to
562 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
563 to permit Nat Ping to be enabled. The default value is 0 seconds.
565 * When a re-initialization is taking place, be sure to reset
566 cm_noLocalMountPoints to 0 in case someone deletes the "Freelance"
567 registry key out from underneath the service.
569 * Add krb5 error message translation to aklog, afscreds,
570 afslogon.dll, the network identity manager afs provider
573 * Mode bits aren't directly exposed by the Win32 API. We were leaving
574 them to default to 0777 when creating new files and directories.
575 This version introduces two configuration parameters;
576 'UnixModeFileDefault' and 'UnixModeDirDefault' which are DWORD
577 registry entries that are used to set the initial mode bits.
578 If the values are set to 0, then the behavior is identical to what we
581 * Minidump files are now produced with a timestamp appended
584 * An SMB request debugging monitor has been added. When activated
585 the monitor will automatically turn on trace logging if any SMB
586 request has required longer than 60 seconds to complete and will
587 then create a minidump every 60 seconds thereafter until the
592 * Restore use of DNS AFSDB and SRV records by kaserver clients.
596 * Avoid a potential Freelance deadlock during initial execution
597 of afsd_service.exe if the old ini file data has to be
600 * Three rx library corrections. (1) Idle data connection
601 processing could timeout if the send window filled and
602 took longer than the idle data timeout period for the
603 transmit window to re-open. (2) The transmit queue
604 could be emptied prematurely. A required check for the
605 queue being in use was forgotten. (3) The function that
606 is supposed to implement a wait for the transmit queue
607 to cease being busy failed to wait.
610 * Restore use of DNS AFSDB and SRV records which were
611 unintentionally disabled in 1.5.69
614 * Add a context menu to the NetIdMgr AFS Provider
617 * Prevent an empty directory Btree from being created and
618 marked as valid if cm_BPlusDirBuildTree fails.
620 * cm_BPlusEnumAlloc should not fail if the enumeration
621 contains zero entries.
623 * In cm_BPlusXXX functions, return ENOMEM if malloc() fails.
625 * Do not leak the cm_scache_t dirlock if cm_BPlusDirBuildTree
628 * fs examine should report owner and group ids as signed values.
630 * Protect buffers in smb_WriteData from simultaneous writes.
632 * Query network interface MTU values from the registry because
633 the IP Helper GetAdapterAddresses() fails to report a manually
636 * Improve performance by not dropping and reacquiring the rx_call
637 lock in rx_WriteProc* and rx_ReadProc*.
639 * When configuring the rx library, the network interface MTU
640 must be applied to both the maximum send size as well as the
641 maximum receive size as is done when rx_SetMaxMTU is called
644 There was no 1.5.67 release
647 * Convert afsauthent.dll to use the xdr_alloc and xdr_free
648 routines from afsrpc.dll in place of malloc and free.
649 This permits the two libraries to be built with different
650 C runtime libraries without crashing.
652 * Set the DOS Readonly attribute on a file/directory
653 whenever the unix mode combined with the mask 0200
654 is true. Previously there was a discrepency between
655 the mask used for testing for readonly behavior and
656 that used for setting the attribute.
658 * Disable AFSVolSync based .readonly "whole-volume callback"
659 support because the all file servers prior to 1.5.67
660 (and perhaps 1.4.12) do not properly assign a value to
661 the AFSVolSync structure in bulk status RPC responses.
663 * Improve the error output from aklog to output the value
664 from krb5 error_message() if the afs_com_err output
665 indicates an unknown value.
667 * Hold a lock on the cm_scache_t object when the smb
668 server is assigning a new clientModTime value.
670 * Permit custom build version numbers to be set
671 without modifying the src/config/NTMakefile.<platform>
674 * Add support for binary code signing with cross-signed
675 certificates and arbitrary versions of signtool.exe.
677 * Convert VBUSY and VRESTARTING to CM_ERROR_ALLBUSY and
678 do not permit them to be exposed to the smb redirector.
680 * Convert STATUS_TIMEOUT responses to STATUS_IO_TIMEOUT
681 to avoid confusion within the smb redirector.
683 * Fix the byte order assigned to port numbers associated
684 with AFSDB record lookups. They must be network byte
685 order not host byte order.
687 * Add dynamic server ranking based on RPC round trip
691 * Official support for Windows 7 and Server 2008 R2
693 * Improved Rx RTT computation following the lessons
696 * Prevent a file server bug (FetchData returning an
697 invalid length instead of zero) from causing an
698 "unexpected network error" when writing to files.
700 * Promote DNS SRV records as superior to DNS AFSDB
701 records. Support arbitrary port numbers for vldb
704 * Add "fs listacl -cmd" support.
706 * Add AFSVolSync based .readonly "whole-volume callback"
707 support. With this functionality, multiple objects from
708 a .readonly volume can have their status validated by
709 issuing a single RXAFS_FetchStatus RPC.
711 * Remove drive mapping functionality and service start/stop
714 * Remove drive mapping functionality from afs_config.exe.
716 * Adjust SMB error return codes to avoid returning
717 STATUS_TIMEOUT which results in the SMB redirector
720 * Network Identity Manager OpenAFS Provider now provides its
721 own "AFS lock" notification icon to report the status of
722 "have tokens, have no tokens, service not started, service
723 started but inaccessible". Hovering over the icon lists the
724 cells for which tokens exist (if any) and the OpenAFS version
725 number. Double-clicking executes the Network Identity Manager
728 * When merging the current status for an AFS object, ensure
729 that the volume status for that object is set to 'vl_online'.
731 * Permit cm_Analyze to request a retry even when the cm_req_t is
732 flagged CM_REQ_NORETRY if the RPC failed due to RX_CALL_DEAD.
733 The retry will force the use of a new connection.
735 * Add digital signatures to resource only dlls as required by
738 * Prevent pioctl calls from retrying indefinitely when a sharing
739 violation error occurs.
741 * Add {HKLM,HKCU}\SOFTWARE\OpenAFS\Client DWORD "ShowMountTab"
742 to restore access to drive mapping functionality in afscreds.exe
745 * Permit Rx minimum peer timeout, maximum receive window size and
746 maximum send window size to be configured via the registry.
748 * The MSI installer now sets the ARPINSTALLLOCATION as required by
751 * DNS lookups set the port numbers in network byte order within
754 * Replace use of the 16-bit compatibility API WinExec with
755 ShellExecuteEx in afs_config.exe and afscreds.exe. WinExec
756 is incompatible with Windows 7 / Server 2008 R2 UAC
760 * Short circuit background volume checks if a shutdown
761 or suspend request is received.
763 * Avoid contacting the file server in order to allocate
764 an empty buffer beyond the current length of the file.
766 * When a request for streams on a directory or mount
767 point object is received, do not offer a default stream.
768 This was affecting the saving of roaming profiles.
771 * Fix Windows 2000 compatibility. Do not call rand_s()
772 even if it is compatible with the C run time library
775 * Fix a data consistency error between the output of
776 NetWkstaGetInfo and NetServerGetInfo RPCs, specify the Lan
777 workstation group name "AFS", and report server name as
778 "AFS" instead of "\\AFS" when the caller asks for "\\AFS".
780 * Enable executables to be run from \\AFS on Windows 7.
781 Return "Name not found" instead of "File not found" when
782 a directory or file name cannot be found.
784 * Prevent cache manager from marking file server "down" when
785 the data returned in response to either RXAFS_FetchData64
786 or RXAFS_StoreData64 is invalid.
788 * Add pioctl data validation to the AFS Explorer Shell extension.
791 * After 1.5.62 afslogon.dll would no longer crash but it
792 also would not load "domain" specific configurations.
795 * An uninitialized variable was present in the symlink
796 recursion detection loop.
798 * Reverse addition of unique per file GUID during
799 Create/Open responses. Inclusion of the GUID prevents
800 Cygwin applications (and others) from accessing //afs.
802 * Treat "filename::$DATA" as a synonym for "filename".
803 This addresses the recent report of roaming profiles
804 failing to save back to the file server.
806 * When given a choice between an error from rx_Write()
807 or RXAFS_EndStoreData(), use the rx_Write() error.
808 This ensures that over quota and access denied errors
811 * If an error occurs during RXAFS_StoreData operations,
812 apply that error to all of the buffers in the BIOD.
813 This ensures that failures will not be retried if they
814 will only fail again.
816 * Do not perform a DNS lookup for a cellname if the
817 cellname prefix is _._AFS_IOCTL_._. This will avoid
818 DNS lookups when the query is for
819 _._AFS_IOCTL_._.F7E5F580200909010061TTTTNT7TT.{10E39A49-4531-4496-A08E-842D4C440D20}.
821 * Fix Freelance root.afs volume object callback processing.
822 Object synchronization and status merging bugs.
824 * Fix Explorer Shell "invalid parameter" bug introduced
827 * Log RXKAD errors to the trace log.
829 * Prevent simultaneous pioctl calls from multiple processes
830 from stepping on each other's data.
832 * Prevent simultaneous pioctl calls from crashing the
835 * Add data validation to pioctl output processing in
836 ktc_GetToken and ktc_ListTokens.
838 * Add data validation to pioctl output processing in
839 fs.exe and symlink.exe.
842 * IMPORTANT: New support for DCE RPC Services: SRVSVC and WKSSVC.
843 No longer will browsing \\AFS produce truncated share names.
845 * Registry specified server preferences did not result in
846 the CM_SERVERFLAG_PREF_SET flag being applied to the
849 * Avoid unnecessary DNS lookups of share names as cells
850 when it is known that the name cannot be a cell name.
851 Any name that does not contain a dot is skipped.
853 * When processing Dfs Referral requests, do not return an
854 error if the server is down or busy; if the volume is
855 offline; the cell vldb cannot be reached; etc.
856 These paths are still \\AFS paths and so the client
857 should not be sent to search elsewhere for resolution.
859 * CRITICAL: Prevent the cm_Daemon thread from terminating when
860 the machine enters the suspend state. IF the cm_Daemon thread
861 dies, there is nothing to execute down server checks.
863 * Prevent use of smb_StartedLock before initialization which
864 can result in a panic in the lock package.
866 * Improve error checking in afslogon.dll in order to prevent
867 NULL pointer dereferences within GetLogonDomainOptions()
868 if the Lsa operations fail.
870 * Add support to the build system to automatically update
871 a Microsoft Symbol Store.
873 * Do not access the cm_conn_t after executing the cm_Analyze
874 loops. Doing so can result in an access to memory that has
877 * CRITICAL: If a StoreData request offset is not aligned on the
878 buffer module blockSize, (offset % blockSize) bytes of dirty
879 data will not be written to the file server.
881 * CRITICAL: If a StoreData64 request is sent to a file server
882 that does not support large files (>2GB) and the client is
883 unaware that the server has no large file support, the first
884 chunksize worth of data will not be written to the file server.
885 This bug was introduced in 1.5.3. It affects all IBM AFS file
886 servers and all OpenAFS file servers older than 1.4.0.
889 * If a file server becomes inaccessible while the cache manager has
890 dirty buffers to write, the afsd_service buf_IncrSync thread can
891 attempt to use 100% of the cpu.
893 * Fix the locking that protects the cm_cell_t lists
894 (all, name, id, free).
896 * Remove cm_cell_t objects from the name and id hash tables before
897 placing them on the free list.
899 * Fix "fs newcell" which was broken in 1.5.60.
901 * Do not attempt to synchronize dirty buffers if the associated
902 volume is known to be unavailable.
904 * Log to the Windows Event Log dirty buffers that cannot be written
905 during the shutdown process.
907 * Modify behavior of a Freelance mountpoint target that does not
908 specify a cell. Instead of assuming the target volume is in the
909 Freelance.Local cell, use the workstation "Cell" specified in the
910 registry. A mountpoint target of "#root.cell." will now mean the
911 root.cell volume in the workstation cell for the current session.
912 If the workstation cell changes from "athena.mit.edu" to
913 "andrew.cmu.edu", the referenced volume will also change without
914 requiring that the mount point targets be altered.
916 * Avoid false infinite loop errors when validating the cm_cell_t
917 free list during startup.
919 * Fix two more locations where xdr_free() should be used instead of
920 free() to permit a checked built of afsd_service.exe to work with
921 release builds of afsrpc.dll.
923 * Add cm_FindServerByUuid(). Re-implement RXAFS_InitCallBackState3()
924 to permit the server Uuid to be used to lookup the server object
925 and from that determine the cell. This permits callbacks that are
926 received from alternate addresses to be processed with a known server
927 object. Previously a request from an unknown server would clear all
928 callbacks from all cells.
930 * Fix a bug that prevented optimal performance when using a non-zero
931 value for 'daemonCheckVolCBInterval'. As a reminder, when
932 "daemonCheckVolCBInterval" is set to a non-zero value, all .readonly
933 volume callbacks are automatically renewed 90 minutes before their
936 * Fix automatic ranking of vldb servers whose values are obtained from
939 * Add logging for RX CALL DEAD errors which are generated if the remote
942 * Add failover for RX CALL TIMEOUT errors when the volume is readonly
943 or the call is to a vldb server.
945 * Add registry based cell search functionality to NetIdMgr,
946 afs_config.exe, and klog.exe.
948 * afsconf_GetCellInfo() has been modified to perform gethostbyname()
949 lookups on the host names in the CellServDB instead of using the
950 specified IP addresses. This provides aklog, pts, vos, etc. the same
951 CellServDB behavior that the Windows Cache Manager uses.
953 * Completely a behind the scenes change. The snprintf() routines have
954 been completely rewritten.
956 * When updating the stat cache entry callback of a .readonly object
957 from the volume group object, update the file server reference to
958 ensure it matches the most update to date callback.
960 * Add proper support for processing callbacks from multi-homed file
961 servers. Instead of comparing servers by cm_server_t pointer,
962 compare them by UUID when the UUID is known.
964 * During a shutdown short circuit the offline volume check daemon
967 * Properly track link counts during cm_Link and cm_Unlink operations.
969 * Return the error code of RXAFS_FetchData / RXAFS_StoreData in
970 preference to an error code reported by rx_EndCall.
972 * Add "PerFileAccessCheck" registry value to permit testing against
973 experimental file servers that include per-file acl support. This
974 value is intentionally undocumented. It is not to be used by
975 production environment deployments.
977 * Log a warning to trace log if a volume group is marked as a DFS
978 File Set. The Windows OpenAFS client does not include DFS support.
980 * Fix a bug introduced in 1.5.60 that prevents the afs netidmgr
981 provider from obtaining tokens when referrals are in play.
983 * Add "fs chown" and "fs chgrp" commands to permit the owner and
984 group of objects stored in AFS to be set from Windows.
986 * Avoid performing background daemon operations when the machine is
987 going into suspend mode.
989 * Perform offline volume checks in most recently used order.
991 * Prevent crash when a data version for a cache object goes backwards.
993 * fs setquota, fs setcachesize, vos setfields, and vos create now
994 accept human readable orders of magnitude. (K, M, G)
996 * fs listquota fixed to permit large quota sizes to be displayed.
998 * rx packet resend and data packets sent counts were incorrect.
1000 * Multi-thread safe library versions are now being generated and used.
1001 mtafsubik.lib, mtafsutil.lib, mtafsvldb.lib, mtafsvol.lib.
1003 * Microsoft SMB Redirector (mrxsmb.sys) support for ExtendedSessTimeout
1004 values are now available on XP through Windows 7. Add functionality
1005 to autodetect if such support is present on the machine. If so,
1006 configure it if necessary and dynamically adjust the AFS Rx timeout
1009 * Correct another race condition in the Rx library that could result
1010 in an unexpected panic while freeing the Rx call iovq.
1013 * A fix to the pioctl library to support drive substitution
1014 to UNC paths. (SUBST <d:> <\\afs\cell\path>).
1016 * On April 9th Microsoft released a Hot Fix for Windows Server 2003 SP2
1017 that corrects a deadlock in the smb redirector and also adds new
1018 functionality that permits the AFS SMB server to be given a longer
1019 timeout than is normally the case. New functionality has been added
1020 to configure these additional LanmanWorkstation\Parameter values.
1021 (This functionality has been backported to XP SP3 and is scheduled
1022 to be released on June 5th.)
1024 * The BackConnectionHostNames registry value configuration was broken
1025 when dynamic re-establishment of Netbios Name registrations was added.
1028 * Hidden vos commands are revealed.
1030 * If the "DisableLoopbackCheck" registry value is set, do not unset
1031 it during the same service session.
1033 * Reorganize code that prevents multiple Store operations to the same
1036 * Modify IsPathInAfs test in Explorer Shell Extension and fs.exe to
1037 permit broken symlinks to be treated as being in AFS.
1039 * Fix vos commands that output 64-bit integer values.
1041 * Cygwin Import Libraries are provided in the SDK for all OpenAFS DLLs.
1042 Permits building cygwin applications against OpenAFS libraries.
1044 * OpenAFS Release Notes, Administrator Guide and User Guide now installed
1045 as Windows HTML Help (.CHM) files.
1047 * NSIS installer does a much better job of cleaning up files left
1048 over from previous installs.
1050 * Fix RT#124787, a race condition between "fs flush <dir>", "fs flushvolume",
1051 or "fs flushall" and on-going directory operations that can result in
1052 afsd_service.exe crashing.
1054 * Add support for DNS SRV records in place of AFSDB records.Â
1055 _afs3-vlserver._udp.<cellname>. Priority field is used.Â
1056 Weight is currently ignored.
1058 * Add a method of specifying Client CellServDB information within the
1059 registry that can be used to either override the CellServDB file or
1060 force the use of DNS lookups for a given cell.
1062 The registry schema is as follows:
1064 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\
1065 "LinkedCell" REG_SZ "[cellname]"
1066 "Description" REG_SZ "[comment]"
1067 "ForceDNS" DWORD {0,1}
1069 HKLM\SOFTWARE\OpenAFS\{Client,Server}\CellServDB\[cellname]\[servername]\
1070 "HostName" REG_SZ "[hostname]" Default: [servername]
1071 "IPv4Address" REG_SZ "[address]" Used only if gethostbyname() fails.
1072 "IPv6Address" REG_SZ "[address]" <future>
1073 "Comment" REG_SZ "[comment]"
1074 "Rank" DWORD "0..65535" Default: 0
1075 "Clone" DWORD "{0,1}" <future: server only>
1076 "vlserver" DWORD "7003" <future>
1077 "ptserver" DWORD ... <future>
1079 ForceDNS is implied non-zero if there are no [servername]
1080 keys under the [cellname] key. Otherwise, ForceDNS is zero.
1081 If [servername] keys are specified and none of them evaluate
1082 to a valid server configuration, the return code is success.
1083 This prevents failover to the CellServDB file or DNS.
1085 Only one of "HostName" or "IPv4Address" is required. "HostName"
1086 is optional if [servername] is the host name.
1088 * Extend registry based CellServDB functionality to
1089 afsconf_GetCellInfo() interface used by aklog and many
1090 other command line utilities.
1092 * libafsconf.dll moved from OpenAFS\Client\Program to OpenAFS\Common
1093 as it is now used by both client and server components.
1096 * PriorityClass of afsd_service.exe process raised to
1097 "High" to match the priority of the system services
1098 that are waiting for afsd_service.exe to respond.
1099 Configurable using [TransarcAFSDaemon\Parameters]
1100 DWORD "PriorityClass" value. (use values from Win32
1103 * Change error reporting of CM_ERROR_BADSHARENAME to
1104 use the correct error (NTSTATUS_BAD_NETWORK_PATH).
1105 The error we were sending would cause the smb
1106 redirector to drop the connection.
1108 * Change the error reporting of CM_ERROR_ALLOFFLINE
1109 and CM_ERROR_ALLDOWN to return NTSTATUS_PATH_NOT_FOUND.
1110 Previous we would return Bad Network Path which
1111 would cause the smb redirector to cancel all file
1112 handles on the same "share".
1114 * A debug version of afsd_service.exe can now be used
1115 with a release version of afsrpc.dll (or vice versa).
1116 afsrpc.dll now implements and exports xdr_free().
1118 * Removed all unprotected OutputDebugString() calls as
1119 those calls can block for up to 10 seconds if there
1120 is contention for the globally shared memory or if
1121 the debugger fails to read data in a timely manner.
1123 * Visual Studio 2008, WDK 6.1 and DDK 6.1 are now
1124 supported for builds. The production releases still
1128 * Discovered that XP/2003 supports a variant on recvmsg
1129 called WSARecvMsg. The input parameters are mostly
1130 compatible with the Posix recvmsg. This will save
1131 multiple memcpy() calls per received packet.
1133 * Discovered that Vista/2008 supports a variant on
1134 sendmsg called WSASendMsg. The input parameters
1135 are mostly compatible with the Posix sendmsg. This
1136 will save multiple memcpy() calls per sent packet.
1138 * Exporting rx_ReadProc32, rx_WriteProc32 from afsrpc.dll
1139 and utilizing rx_ReadProc32 from within the cache manager
1140 as part of the FetchData operations. rx_ReadProc32 and
1141 rx_WriteProc32 are optimized versions of rx_ReadProc and
1142 rx_WriteProc for use when reading 32-bit integer values.
1144 * The Network Identity Manager AFS provider now reports
1145 progress events. These provide a better user experience
1146 with Network Identity Manager 2.0.
1148 * [RT 124293] A race condition exists which permits the
1149 scp field of the an smb_fid_t object to become invalid
1150 after a request on the smb_fid_t has begun. Remove the
1151 race by protecting all access to the scp field with the
1152 smb_fid_t mx mutex and obtaining a local reference on
1153 the cm_scache_t object for the length of the request.
1155 * [RT 124276] If the vldb is out of sync with the contents
1156 of the file servers, afsd_service will retry too many
1157 times when a file server reports a volume as not being
1158 present. Now if the list reported by the vldb is the
1159 same as the previously seen list, then the retry is
1162 * [RT 124276] Read-only volume failover was broken in
1163 1.5.53 whenever accessing a volume results in VNOVOL or
1166 * [RT 124276] Prior to 1.3.70 the volume server reference
1167 list was not reference counted and would be prematurely
1168 freed while in use. When reference counting was added
1169 in 1.3.70 a bug was introduced that could result in
1170 service reference list corruption.
1172 * Add Windows Application Event Log warning messages for
1173 "Client SMB MPX value too large" and
1174 "Client SMB Buffer Size too small".
1176 * Renaming of files across directory boundaries would result
1177 in an invalid handle error when attempting to access the
1178 files after the move.
1180 * Fix the handling of Tran2 Set Path Info RPCs. Do not fail
1181 when a smb file descriptor cannot be found. The whole point
1182 of using a Path Info function is because an smb file
1183 descriptor wasn't allocated.
1185 * More edge cases in which dynamic addition of Freelance
1186 root.afs entries would get the wrong FID or where the
1187 root.afs directory would not be refreshed.
1189 * Buffer overflow could occur if the workstation cell name was
1190 longer than 64 characters. Crashes could occur in
1191 afscreds.exe, afslogon.dll, and afsd_service.exe.
1193 * VNOSERVICE and VOFFLINE errors were leaking and were exposed
1196 * Improve consistency of the bad data version value used for
1199 * Log file server uuid values as part of the cm_server object
1200 when available. Dump the cm_server object list in response
1203 * Optimize the performance of resetting access control lists
1204 when tokens are set or removed.
1206 * Fix vos.exe so that it can work with very large volume
1209 * Remove symlink recursion tests and increase max symlink count
1213 * Undo the "UAC manifest fix" applied to afs_config.exe.
1215 * Fixed some additional race conditions in the Rx library
1217 * Ensure that Freelance allocation of vnodes follow
1218 the AFS convention of odd vnodes are directories
1219 and everything else is an even vnode.
1221 * Add Freelance logic to mount point and symlink
1222 evaluation functions.
1224 * Enhance smb_ParseASCIIBlock() so that it can handle
1225 all of the STRING formats defined by the CIFS Technical
1228 * Validate the output of smb_ParseASCIIBlock() in all
1229 callers. Return CM_ERROR_BADSMB if the STRING field
1230 cannot be parsed. CM_ERROR_BADSMB will cause the
1231 contents of the packet to be logged.
1233 * If multiple SMB Raw Write operations were taking
1234 place at the same time, there could be data corruption
1235 because unique event objects were not generated for
1236 each Netbios receive operation.
1240 * Fix a race in rxi_WriteProc which could result
1241 in a freed or in use packet being passed to
1242 rxi_PrepareSendPacket if the call is currently
1243 in a transmit wait state. Passing a freed packet
1244 will result in a panic. Passing an in use packet
1245 might result in a call failure or perhaps data
1248 * Kerberos referrals broke the NIM AFS provider
1249 logic for automated configuration of the workstation
1250 cell with a new identity. Now if a referrals
1251 response is returned, assume the cell belongs to the
1254 * The interface to buf_CleanAsync() and
1255 buf_CleanAsyncLocked() has been changed to permit
1256 errors to be propagated to the caller.
1258 * During an enumeration of a directory, if the btree
1259 has not yet been generated, return an error.
1261 * When mapping CM_ERROR values to NTSTATUS values
1262 do not treat 0 as an error.
1264 * Avoid a cm_scache_t object reference count underflow
1265 when renaming via the SMB interface.
1267 * smb_IsMask() can be called with a non-8.3 input buffer.
1268 Separate smb_IsMask() into smb_Is8Dot3Mask() for 8.3
1269 input buffers and smb_IsMask() for non-8.3 input buffers.
1271 * Add cm_buf_t error state to the "fs memdump" output.
1273 * pioctl drive mapping checks need to take into account
1274 Global Drive Mappings and SUBST assignments.
1276 * Rename operations should not be permitted if the
1277 destination name collides with an existing name that
1278 differs only by case UNLESS the collision is the object
1279 that is being renamed.
1281 * Increase the maximum number of background daemon threads
1282 to 64. Leave the default at 4.
1284 * Set the FILE_UNICODE_ON_DISK flag when reporting the
1285 file system characteristics.
1287 * Report the maximum component name length as 255 even
1288 for Unicode as part of the file system characteristics.
1290 * Add RX debugging information to "fs memdump" output
1293 * Refcount the cm_server_t objects returned by
1296 * Fix the UAC manifest applied to afs_config.exe
1298 * rxdebug now indicates if any RX_CONN flag is set
1299 not just some of them
1301 * Change the mapping of CM_ERROR_WOULDBLOCK from
1302 NTSTATUS_LOCK_NOT_GRANTED to NTSTATUS_CANT_WAIT
1306 * Includes #GCO Public CellServDB 07 Nov 2008
1308 * Prevent separate cell entries from being created for
1309 "foobar.org" and "foobar.org." which can happen if
1310 an AFSDB record is found instead of a CellServDB
1313 * Fix a panic when checking the SMB virtual circuit list.
1314 If a VC is destroyed while the list is being checked
1315 it is possible for the next object to become invalid.
1317 * Increased logging to afsd_init.log of netbios name
1320 * Fix a DNLC panic when a UTF-16 string that cannot be
1321 converted to a utf-8 string is provided as a filename.
1323 * In the NSIS installer, delete old afslogon.dll/afscpcc.exe
1324 from %windir%\system32 and fix the WinLogon\Notify\AfsLogon
1325 location to refer directly to the new location in
1326 %programfiles%\openafs\client\program
1328 * Be more aggressive about responding to Netbios
1329 NRC_BRIDGE/NRC_NOWILD errors. Rebind the Netbios name to
1330 the LANA within the listener thread without performing a
1333 * Use the WNet Enumeration of all connected drives as a filter
1334 in the pioctl GetIoctlHandle operation. This should remove
1335 the last situations in which there was a delay when opening
1336 an Explorer Shell properties menu with the AFS Shell
1339 * Fix the parsing of CellServDB in Windows to handle linked
1340 cells as is done on Unix.
1342 * Add support for linked cells to the cache manager,
1343 netidmgr afs provider, and aklog. When two cells are
1344 linked, a volume lookup that fails in one cell is looked
1345 up in the other. This can be used for construction of
1346 test cells that override specific volumes; cell renaming;
1349 * Increase the length of cell and realm names that can
1350 be entered in the NIM AFS provider.
1352 * Actively prevent a 32-bit afsd_service.exe from
1353 executing in the WOW64 environment.
1355 * Protect against invalid IP address values in CellServDB
1360 * Fix an Rx panic caused by an attempt to double free
1361 an rx_packet when an error occurs while writing the
1364 * Fix an exception cause by a race condition when
1365 processing multiple file change notifications registered
1366 by the Explorer Shell or other applications.
1368 * Modify the Rx debug interface to use var_args and
1369 permit Rx debugging to be used on 64-bit Windows
1370 which previously had it disabled.
1374 * Automated lock hierarchy enforcement added to the
1375 osi base lock package. This permits detection of lock
1376 violations that might but often does not result in
1377 deadlocks. Enabled by default in the checked builds.
1378 May be activated in release builds via the registry
1379 value "LockOrderValidation" in the TransarcAFSDaemon
1382 When a violation is detection an assertion is thrown
1383 generating a dump file and a message indicating which
1384 locks were obtained out of order is logged to the
1385 Windows Application Event Log.
1387 * Attempt to ensure that a dump file will be generated
1388 whenever the service panics.
1390 * Removed unused lock cm_bufGetMutex
1392 * Many out of order lock acquisitions were detected
1393 using the lock order verifier and were corrected.
1395 * A race condition within the function path cm_GetSCache() ->
1396 cm_GetNewSCache() -> cm_RecycleSCache() permitted a
1397 cm_scache_t object to be issued simultaneously to two
1398 threads. This would eventually result in a panic due
1399 to the resulting undercount.
1401 * Fix interpretation of the empty string as the ioctl path
1402 to mean the current directory. "fs lsm", "symlink list",
1403 etc. now return a "not a ..." error instead of "does not
1406 * Add SMB_STRF_SRCNULTERM flag to smb_ParseStringXXX functions.
1407 to indicate that the input string is a nul terminated string.
1408 Assign ti when input strings are nul terminated instead of
1409 length counted. (from 1.5.50)
1411 * The CIFS protocol specification for handling NT_TRANSACT_CREATE
1412 does not match the observed behavor. The 'nameLength' is specified
1413 in bytes, not in characters. Fix the implementation to match
1414 the observed behavior. (1.5.50)
1416 * The cm_HaveAccessRights() attempt at deadlock avoidance by
1417 calling lock_TryRead() on the parent directory cm_scache_t rw-lock
1418 does not avoid the deadlock. Avoid the deadlock by enforcing the
1419 lock order of the lowest vnode first. Then remove the infinite
1420 loop avoidance in cm_SyncOp() that was returning an unwarranted
1421 access denied error.
1423 * File Server lock synchronization was not propertly enforced in
1424 smb_ReceiveNTCreateX() and smb_ReceiveNTTranCreate().
1426 * Another change to the short name generator. Plus and equal are
1427 not valid characters in a short name according to the
1428 RtlIsNameLegalDOS8Dot3() in kernel mode. They have been replaced
1429 with left and right brace in order to minimize the impact.
1431 * Fix a race condition that can result in a heap overwrite error if
1432 the number of known file servers or vldb servers increases during
1433 the preparation phase of a cm_CheckServers() operation.
1435 * If the SMB Authentication would fail, a LSA allocated memory
1436 buffer would be leaked.
1438 * Rewrite the cm_cell_t allocation code to permit cell vldb server
1439 lookups to fail without leaking cm_cell_t objects.
1441 * Add "fs setcrypt auth" to permit the use of rxkad authentication
1442 only mode. Not documented.
1444 * Initialize Rx mutex and lock objects once and only once as part of
1445 the DLL load process.
1447 * When reporting the Volume Name "AFS" do not include the terminating
1448 NUL in the reported length.
1450 * Refresh Volume Location information upon receipt of a VNOVOL error
1451 from the file server and avoid a race condition that prevented the
1452 updated volume location information from being destroyed immediately
1453 after it was acquired.
1455 * In response to VNOVOL and VMOVED, remove the volume id from the server's
1456 volume list so that the volume's up and down status is not affected
1457 when the machine is shutdown.
1459 * Many fixes to the Rx stack related to packet allocation and deallocation;
1460 Rx debugging; Rx packet tracking across queues; and Rx packet loss caused
1461 by overwriting the rx_call currentPacket pointer when it was already
1462 referring to a packet.
1464 * Rx Jumbograms are now off by default. To re-enable, add "RxJumbo" = "1"
1465 in the TransarcAFSDeamon/Parameters key.
1467 * "fs flush", "fs flushvolume", and "fs flushall" now destroy locally
1468 constructed directory B+ trees for flushed directory objects.
1470 * When cached pages for directory objects have been locally modified, do
1471 not permit the locally modified pages to be mixed with pages subsequently
1472 obtained from the file server. Doing so can result in directory
1475 * If Ubix errors are received in response to VL_xxx operations record
1476 the error on the server instance and permit the operation request to
1477 be retried in case another server is able to answer the query. We have
1478 seen at least one instance of a cell in which a VL server returned
1479 UNOQUORUM (possible database truncation) when the other servers were
1480 responding successfully.
1483 * Optimize the assignment of write errors to dirty
1484 buffers by applying the first error to all buffers
1485 belonging to the same object instead of trying each
1488 * Add resource version info to xstat_fs_test and xstat_cm_test
1490 * Fix a delete vs delete[] usage problem in the WiX custom
1493 * When creating a mount point, validate the volume name to
1494 make sure it exists.
1496 * When processing a mount point, pay attention to VL_BADNAME
1499 * Remove a race condition from cm_volume_t recycling that could
1500 have resulted in two threads making use of the same object at
1503 * Fix a reference count leak on smb_vc_t objects introduced in
1504 1.5.50. Permit the objects to be freed once again.
1506 * Fix the output of "fs memdump" for smb_vc_t objects that was
1509 * Add routines for debugging smb refcount errors.
1511 * When shutting down the cache manager, do not permit AFS
1512 callback requests to be processed. Doing so can result in
1513 invalid memory accesses.
1515 * Do not recycle a deleted cm_scache_t object unless the
1516 refCount is 0. Doing so can result in a deadlock and
1517 invalid memory access.
1519 * Minimize the number of RPCs sent to the file server for
1520 directory bulk status requests. Only request status for
1521 objects that we are actually reporting on and only if we
1522 do not already have status for the object and then make
1523 sure that we return as much as we can at once via the SMB
1524 interface but not so much that it exceeds the RDR timeout
1526 * Enable bos restricted operations
1528 * Fix the creation of submounts used by the afscreds and afs_config
1531 * Avoid repeated retries when attempting to obtain access rights.
1532 If the access rights were obtained successfully but do not
1533 satisfy the request, fail the request.
1535 * Fix error code reported when attempting delete a file on a
1536 readonly volume or one that is marked with the readonly DOS
1539 * Fix a short name truncation bug.
1541 * Fix a heap corruption error when reading the CellServDB file
1544 * Fix directory listings when the directory has "lookup" permission
1545 only and the directory has never been previously accessed.
1547 * Fix several tracking issues for cm_scache_t rw locks.
1549 * Ensure that all buffers used to store cell names are the
1552 * Restore Windows 2000 support (broken in 1.5.50).
1553 LOCALE_INVARIANT was introduced in XP.
1555 * Add RxUdpBufSize registry value. Default is 256K.
1557 * Avoid multiple deadlocks between cm_freelanceLock and cm_scacheLock
1559 * Do not include trailing NULs when returning file names in
1560 a directory search response.
1562 * Prevent RX from crashing the afsd_service by attempting
1563 to allocate to many buffers at once when memory is fragmented.
1565 * Avoid a deadlock when freeing cm_server_t objects
1567 * Optimize search cell file. Do not search to end of file
1568 if the cell has already been found.
1570 * Remove explorer shell extension delay on local disks.
1572 * Automatically pre-allocate 64 Rx packets per thread
1574 * Add smb lock requests and stat cache lock allocations to
1575 the output from fs memdump.
1578 * Fix Large File Support (broken in 1.5.50)
1580 * Fix "fs setacl" (broken in 1.5.50)
1582 * Fix pioctl path evaluation of drive letter
1583 mapped paths. (broken in 1.5.50)
1585 * Fix installation of VC8 Runtime library in the EXE
1586 Retail installer. (broken in 1.5.50)
1588 * Avoid a potential deadlock during Volume Location
1589 updates if all the file servers are unreachable at
1590 the time the VL update succeeds.
1592 * Ensure that rx connection object is references
1595 * Fix propagation of over quota errors during cm_FSync()
1598 Since 1.5.39 [1.5.50 released 16 July 2008]
1599 * There were no releases between 1.5.39 and 1.5.50.
1601 * FIRST RELEASE WITH UNICODE SUPPORT
1603 * Pioctl interfaces to the cache manager refactored
1604 to provide layering between the SMB specific code
1605 and the general purpose ioctl operation.
1607 * Garbage collect dead SMB virtual circuits as soon
1608 as they are no longer being referenced. This avoids
1609 problems with outstanding locks not being dropped
1610 when the virtual circuit becomes invalid.
1612 * Remove the IBM Administration Reference documentation
1613 and replace it with the OpenAFS Command Reference
1616 * Avoid calling rx_SetDeadTime and rx_SetHardDeadTime
1617 functions each time a connection is about to be used.
1618 Do not hold a lock on the rx connection object while
1619 it is being selected. This avoids a race between
1620 threads attempting to set the timeout values and
1621 removes a bottleneck that was hampering performance.
1623 * Ensure that the smb directory attribute is set
1624 for all directory objects.
1626 * Add vs2008 support to the NSIS installer scripts
1628 * Replace the VC Runtime EXE installer with the MSI
1629 installer in the NSIS installer scripts
1631 * Properly delete the VC7.1 C Runtime libraries.
1633 Since 1.5.36 [1.5.39 released 23 June 2008]
1634 * There were no 1.5.37 or 1.5.38 releases for Windows
1636 * Use Visual Studio version of vsprintf() instead of
1637 home grown version that resulted in linking conflicts
1639 * Prevent buserver, ptserver and fileserver from crashing
1640 due to non-portable strftime() format strings.
1642 * Prevent server crashes caused by rx_InitMorePackets()
1643 being called prior to allocation of all of the required
1644 synchronization objects.
1646 * Fixed a memory leak in the internal B+ tree directory.
1647 For objects with non-8.3 names, the 'longname' memory
1648 allocation would be leaked during object deletion or
1651 * Fixed a memory leak in the background daemon cm_CheckServer()
1652 routine. The memory allocated to store the server list
1655 * Obtain a missing lock around a call to cm_RemoveSCacheFromHashTable().
1657 * Correct an abstraction layer violation. cm_scache_t objects
1658 should be marked deleted in cm_Unlink() and cm_RemoveDir() and not
1659 in smb_CloseFID(). Cleanup of deleted cm_scache_t objects should be
1660 performed in cm_ReleaseSCache() when the reference count hits zero.
1662 * Prototype cm_AdjustScacheLRU() and re-implement it using osi_QAddH().
1664 * Do not reference 'smb_logp' in cm_vnodeops.c. Use 'afsd_logp'
1667 * Ensure that scache object 'nextp' pointers are set to NULL
1668 when the objects are removed from the hash table linked list.
1670 * If a dirty buffer cannot be written to the file server because
1671 of an access denied error, mark the error in the cm_buf_t
1672 structure so that it can be dealt with instead of retrying
1676 * Update CellServDB file to GCO Public 23 Apr 2008
1678 * Fix a cm_buf_t reference count leak when attempts to write
1679 dirty buffers to the file server from within cm_IncrSyncer()
1682 * Prevent udebug from crashing.
1684 * Activate RX Idle Timeouts. If the file server is busy for more
1685 then 30 seconds, attempt to failover to another server without
1686 marking the busy server down.
1688 * Another VNOVNODE issue fixed. When writing a dirty buffer
1689 to the file server, if VNOVNODE is received, mark all buffers
1690 as invalid without further attempts to contact the file server.
1692 * Improved performance on high latency links.
1695 * Fix a bug preventing the re-initialization of the Freelance
1696 root.cell contents when a change is made.
1698 * Properly return CM_ERROR_NOSUCHFILE when removing a symlink
1699 or mountpoint that does not exist.
1701 * aklog, afscreds, integrated logon, and the NIM AFS credential
1702 provider will all attempt to first acquire an afs service ticket
1703 using the service principal: afs/<cell>@<CLIENT_REALM>. This
1704 permits the proper detection of the cell realm when multiple
1705 realms are local realms for the cell.
1707 * Fix fs commands that re-use file identifiers instead of forcing
1708 a path evaluation each time. This was broken in 1.5.33 when
1709 hashes were added to the cm_fid_t structure.
1711 * Fix symlink and mountpoint removal and then recreation when
1712 applied to the fake Freelance root.afs volume.
1714 * Fix SMB change notifications for the Freelance root.afs volume
1715 when creating or removing symlinks and mount points.
1717 * Force DNS AFSDB cell name lookups to lower case if the DNS
1718 response is not all lower case.
1720 * 32-bit installers are now built with Visual Studio 2005
1722 * All vos.exe commands now support the "-noresolve" option which
1723 forces the display of IP addresses instead of DNS hostnames.
1725 * No longer hold the cm_volume_t mutex across VL_xxx RPCs. This
1726 permits operations that do not require up to date volume
1727 location data to continue while waiting for the RPC to complete.
1728 Multiple threads waiting for a volume location update on the
1729 same volume set no longer trigger multiple calls VL_xxx RPCs.
1731 * Volume server references could be marked offline and never
1732 reset to the not_busy state.
1734 * Further optimizations to cm_Analyze() retry logic for
1735 CM_ERROR_OFFLINE and CM_ERROR_ALLBUSY error handling.
1738 * Optimize the DNLC by applying Interlocked incrementing for
1739 statistics gathering and enabling greater use of read
1740 locks instead of write locks
1742 * Further optimize osi_Log operations when logging is disabled
1743 by removing the need for the function call overhead.
1745 * Add a new lock operation, Convert Read to Write, and apply
1746 it throughout the cache manager. This function provides a
1747 fast transition from a read lock to a write lock when the
1748 caller is the only reader.
1750 * Fix a timeout problem when opening files or request locks
1751 when the file in question is already write-locked by another
1754 * Fix a deadlock that can occur if "fs flushall" is executed
1755 while applications are writing to AFS.
1757 * The cache manager would re-read data from the file server
1758 after file truncation when obtaining buffers to write to the
1759 same file until the first StoreData RPC completes.
1761 * Fixes error handling in cm_NTCheckDelete()
1763 * Replaces cm_scache_t mutex with a read-write lock permitting
1764 additional parallel access.
1766 * Prevent an cm_scache_t reference undercount when evaluating
1767 symlinks containing @sys when none of the active sysnames
1768 match any of the available targets.
1770 * Prevent the leak of a cm_scache_t rwlock if associated symlink
1773 * Set RxMaxMTU registry value to 0 which is the equivalent of
1774 it not being set at all.
1776 * Reduce contention between BkgDaemon threads
1778 * Prevent CM_SCACHEFLAG_ASYNCSTORE from being reset on a write
1781 * Convert reference counts on cm_volume_t, cm_conn_t, cm_server_t
1782 objects to use the faster Interlocked operations which avoid
1783 the need to hold write locks.
1785 * Permit rxkad errors other than RXKADEXPIRED to be treated as a
1786 non-fatal error that forces a retry against another service
1789 * Fix a crash in smb_WriteData that would occur if an application
1790 opened the file as read-only and then sent a write request anyway.
1793 * Respond to VL_NOENT errors by removing cm_volume_t object and
1794 setting it to be the first object to be recycled. This is not
1795 only a more efficient use of resources but it also prevents
1796 repeated attempts to query the VLDB server for the volume from
1797 the CheckOfflineVolumes routine.
1799 * Change the defaults for RxEnablePeerStats and RxEnableProcessStats
1802 * A first cut at a cache manager statistics monitor that can be used
1803 to determine the necessary cache parameters to support the
1804 application working set.
1806 Off by default, the performance package can be activated by setting
1807 "daemonPerformanceTuningInterval" in the service Parameters key.
1808 As with the other daemon interval values the unit is in seconds.
1810 At service start and each succeeding interval the cache manager
1811 will write statistics to %TEMP%\afsd_performance.txt showing the
1812 relative usage of cm_scache_t, cm_volume_t and cm_buf_t objects.
1813 The FID statistics keep track of all FIDs seen by the cache manager
1814 during the service session whether or not they are backed by any
1815 live object in the cache.
1817 These statistics are not stored in the cache file.
1819 * Fix windows event logging broken in 1.5.33.
1821 * Eliminate multiple calls to syscfg_GetIFInfo() which is used to
1822 populate the IP address list for WhoAreYou/TellMeAboutYourself RPCs
1823 as well as when computing the server preference list.
1825 * When merging status after a StoreData operation prevent data
1826 buffers from being discarded that we would prefer to keep.
1827 (Bug introduced in 1.5.33.)
1829 * Modify installers the set RxMaxMTU to 0.
1831 * In the AFS Configuration Panel on the Advanced->Miscellaneous
1832 dialog, the Background Daemons and Service Threads fields were
1835 * Increase default number of Background Daemons to 4.
1837 * Remove cm_volume_t reference from the cm_scache_t object. Permit
1838 cm_volume_t objects to be recycled.
1840 * Windows Error Reporting has shown a number of previously unseen
1841 stack corruption errors in the Explorer Shell extension. The
1842 minidump appeared to indicate the problem was in ParseAcl().
1843 Reviewing the function it appears that numerous things could go
1844 wrong if invalid input was provided. Added error checking and
1845 data validation. Hopefully this will address the problem.
1847 * Windows Error Reporting has shown crashes in aklog as a result
1848 of krb5_cc_get_default() failures not being handled. They are
1852 * The Rx library used a 32-bit type for sockets which was
1853 truncating the socket value on 64-bit Windows. This
1854 was not an issue in general but became an issue when
1855 Rx "hot thread" support was enabled.
1857 * Enable "Rx hot thread" support which ensures that while
1858 a received message is being processed that a new thread
1859 is allocated to receive the next message to arrive.
1861 * Increased parallel access to smb_rctLock protected
1862 resources. Reduce the number of times that the
1863 lock is dropped and re-obtained.
1865 * Increased parallel access to cm_scache_t and cm_buf_t
1866 resources by replacing the previous global write lock
1867 protected reference counting mechanism with one built
1868 upon the use of InterlockedIncrement and InterlockedDecrement.
1870 * Implement a hash for cm_fid_t to reduce comparison costs
1871 and inline the cm_FidCmp() function.
1873 * Redefine the BUF_HASH and BUF_FILEHASH in terms of the cm_fid_t
1874 hash which has a better distribution
1876 * Modify cm_ConsiderPrefetch to evaluate the amount of data in
1877 the most recent read request instead of the next chunkSize.
1878 cm_chunkSize can be dozens or hundreds of buffers. As a result
1879 too much time is spent performing the evaluation.
1881 * Fix the usage of cm_scache_t bufCreateLock. The purpose of this
1882 lock is to prevent the creation of new buffers while a truncation
1883 is being performed. All references to bufCreateLock have been
1884 removed except in two places:
1885 i. a write-lock surrounding the function that calls buf_Truncate()
1886 ii. a read-lock within buf_GetNewLocked() that actually allocates
1889 * Modify the CM_CONFIG_DATA_MAGIC value to include a
1890 CM_CONFIG_DATA_VERSION number which value be used to force the
1891 replacement of the cache file contents when incompatible changes
1892 are made between releases.
1894 * CM_SCACHESYNC_ASYNCSTORE should not be ordered by
1895 cm_SyncOpCheckContinue
1897 * Avoid calls to multi_Rx if nconns == 0
1899 * Modify smb_WriteData to perform background writes based upon the
1900 crossing of 'smb_AsyncStoreSize' boundaries instead of cm_chunkSize
1901 boundaries. This will slow down writes from the SMB interface but
1902 will avoid the risk of the CIFS client disconnecting from the AFS
1903 client SMB server. The default value for 'smb_AsyncStoreSize'
1906 * SMB AsyncStore functionality is on by default but can be
1907 disabled using the registry value:
1909 HKLM\SOFTWARE\OpenAFS\Client DWORD "EnableSMBAsyncStore"
1911 smb_AsyncStoreSize can be adjusted with
1913 HKLM\SOFTWARE\OpenAFS\Client DWORD "SMBAsyncStoreSize"
1915 The value must be a multiple of the buffer block size
1916 and cannot be larger than the chunk size.
1918 * Do not leak a read lock on the bufCreateLock
1920 * Do not forget to call cm_SyncOpDone if cm_SyncOp succeeded
1923 * In MergeStatus, only remove clean buffers from the hash table.
1925 * Avoid the need to update the data version number on each buffer
1926 associated with a scache when MergeStatus is called after a
1927 StoreData by maintaining a range of valid data versions as part
1928 of the cm_scache_t object.
1930 * Profiling shows large numbers of blocked calls in cm_HoldSCache
1931 from cm_HaveAccess... We can safely avoid the call so do so.
1933 * Add "ForceAnonVLDB" service parameters registry value. Default
1934 is 0. Set to 1 to force the use of anonymous connections to
1935 the VLDB servers. This is useful when using krb5 tokens and the
1936 VLDB servers have not been updated.
1939 * Do not attempt to search for a cell name that matches one
1940 of the special share names "svrsvc", "wkssvc", "ipc$"
1942 * Issue permission denied warnings from afscreds if attempts
1943 to modify the startup shortcut or registry values fail.
1944 This happens under Vista UAC.
1946 * Do not return BAD_NETWORK_ERROR in response to attempts
1947 to create directories or files in the Freelance root.afs
1948 volume. Instead return ACCESS_DENIED.
1950 BAD_NETWORK_ERROR was being sent because the Freelance cell
1951 has no vldb servers. Short circuit the error by testing for
1952 freelance in CreateFile and MakeDir
1954 * Treat srvsvc, wkssvc and ipc$ as pseudo files. Permit them
1955 to be opened and closed without actually existing. For now
1956 we treat them as equivalent to ioctls but this should
1959 * Fix RAP processing to not reject names that exist in the
1960 root.afs volume for GetInfo queries.
1962 * In afscreds, use GetUserNameEx to obtain the username instead
1963 of searching the registry. Try UserPrincipalName first and
1964 fallback to SamCompatible.
1966 * Stop linking to rpcns4.dll which is no longer supported on
1967 Vista and 2008. TaAfsAdmSvr is going to have to be
1970 * When processing dfs referral requests, if the sharename is
1971 a partial match and therefore will not be added automatically
1972 to the share list, make sure that we return an error.
1974 * Avoid a crash when logging VNOVOL errors returned from
1975 cm_CheckOfflineVolume.
1977 * When trace logging is disabled, do not permit osi_LogSaveString
1978 to perform the string copying. By disabling the string
1979 copies, throughput of the service more than doubles.
1981 * when checking offline volumes, update the cell vlserver info
1982 if required and don't perform a check if there are no servers
1985 * Add volumes in the vl_unknown state to the list of objects to be
1986 checked as offline or retried when an allbusy error is received
1989 * Add support for configuring rx_ExtraPackets
1991 * All calls to buf_Get() must be protected by a read lock on the
1992 cm_scache_t bufCreateLock
1994 * When a MergeStatus results in the invalidation of buffers, do
1995 not remove buffers from the hash tables that have active references.
1997 * Upon receiving a VNOVNODE error from a file server, be sure to
1998 invalidate any file locks. This prevents on-going attempts to
1999 extend the locks and will cause subsequent attempts to access
2000 the file to receive a invalid handle error.
2002 * Prevent smb_ReceiveTran2SearchDir from adding terminating NULs
2003 to directory entry names.
2005 * No longer discard tokens when RXKADBADTICKET error is received.
2006 Instead return STATUS_NO_KERB_KEY error.
2009 * Add "cmdebug <host> -cellservdb" option and client service
2010 implementation permits querying the vldb server for each
2011 cell known to the cache manager. The output is in a form
2012 that permits it to be used as a CellServDB file.
2014 * Add Symlinks->Show option to the AFS Explorer Shell Context
2017 * Prevent buffer overrun in the cmd parser used by command
2018 line tools such as fs, vos, pts, etc.
2020 * The path ioctl operations have several issues:
2022 (1) the specified path for "fs examine, whereis, whichcell,
2023 flush" and so always has the follow symlinks and mount
2024 points semantics. This makes it impossible to determine
2025 what the FID of a symlink or mount point is.
2027 (2) "fs examine" out is not the result of a single pioctl
2028 operation but is actually the combined output of half
2029 a dozen operations. Path evaluation is an expensive
2030 operation. It would be faster if the caller could
2031 evaluate the FID first and then perform all of the
2032 rest of the operations by specifying the FID instead
2035 (3) fs output reports all objects as files. By adding a
2036 GetFileType pioctl more informative output can be provided
2037 that indicates what type of object the path evaluates to.
2039 (4) the Windows fs command includes a number of commands that
2040 do nothing but exist only because the Unix cache manager
2043 A new extendible data structure cm_ioctl_query_opts_t has been
2044 added which can be optionally specified with pioctls that do not
2045 already require input data. The first two fields of this structure
2046 are 'literal' and 'fid'. The literal field is used to indicate
2047 whether the last component of the path should be evaluated following
2048 symlinks and mount points.
2050 The fid field permits a fid to be specified.
2052 * A new GetFileType pioctl has been added. The type of objects are
2055 * A new "-literal" option is available for "fs examine, flush, whereis,
2058 * Unimplemented fs commands have been removed.
2060 * There are circumstances where a volume object is being accessed and
2061 the volume is marked indicating that the volume location information
2062 is out of date but where it is also pointless and perhaps dangerous
2063 to block waiting for the rpc to complete. One example is when
2064 processing the cmdebug requests. If we know that we are not going
2065 to use the volume object to contact a server, then we can now set
2066 the CM_GETVOL_FLAG_NO_RESET flag.
2068 * when moving up in the directory tree we search the recorded fid list
2069 to find a matching fid that we have already crossed. we must also
2070 reset the fid count based upon what we discover.
2072 * Cell names are published as share names. As a result they are searched
2073 for as part of the Dfs Referral evaluation. If share "foo" can not be
2074 reached, the CIFS client will fallback to searching for "fo" as well.
2076 Since the freelance client automatically adds a symlink for prefixes
2077 of the cell name, this results in both "foo" and "fo" being added to
2078 the freelance root.afs volume. This patch prevents that negative side
2081 * when tracking Kerberos credential cache names be sure to include
2082 the cache type prefix.
2084 * Microsoft has assigned OpenAFS a network type value
2086 * The algorithms that were used to produce a Netbios name were broken.
2087 The name that was produced when the hostname was too long was in fact
2088 longer than the maximum permitted netbios name.
2090 Also, the "NetbiosName" value was not used as a suffix when the loopback
2091 adapter was not installed. The hardcoded string "AFS" was used instead.
2093 * avoid another deadlock during server probes initiated by the ipaddr
2094 change daemon thread.
2096 * prevent the afs client service from crashing when fs newcell is
2099 * In the NSIS installer, move afslogon.dll, afscpcc.exe and afs_cpa.cpl
2100 to \Program Files\OpenAFS\Client\Program in order to get them out
2101 of the %windir% tree
2103 * EDQUOT == WSAEDQUOT. Define it in the right places so that in the
2104 end STATUS_OUT_OF_QUOTA can be returned to the application.
2106 * Modify the search order for determining the location of CellServDB
2107 and other client configuration files.
2109 1. AFSCONF environment variable
2111 3. NEW - use All Users\AppData\OpenAFS\Client only if CellServDB exists
2112 4. use Program Files\OpenAFS\Client
2114 * More Vista Power Management tweaks
2116 * Close a socket handle leak in the kauth routines
2118 * Don't hold a cm_volume_t mutex across RPCs
2120 * Convert downTimes to use clock time instead of relative times.
2121 This permits correct comparisons with expiration times.
2123 * Update the Wix installer to properly install the prerequisites
2124 for translate_et.exe
2126 * Do not translate pioctl paths as they are not converted by the
2129 * Implement multi_rx probes for checking whether servers are up
2130 or down. Turns probes into a constant time activity instead
2131 of one dependent upon the number of known servers.
2133 * Treat VNOVOL errors the same as VOFFLINE. Do not force an
2134 update of the volume location data and retry. VNOVOL means
2135 that the volume could not be attached. It doesn't mean that
2136 the volume is not on this server.
2138 * Modify the Explorer Shell AFS Context menu Symlink->Add to
2139 not attempt to validate the target path. There is no reliable
2142 * Improve performance of \\afs\<cell><mount-type><volume>\
2143 access by eliminating extraneous attempts to search the
2144 CellServDB file and DNS AFSDB records. Share names
2145 containing '%' or '#' are not valid cell names.
2147 * Return STATUS_RANGE_NOT_LOCKED if the application attempts
2148 to unlock a range that is not currently locked.
2150 * Enable the 32-bit tools to read the 64-bit registry keys
2151 when executing in the WOW64 environment.
2155 * Fix buffer refcount leak introduced in 1.5.29
2157 * Prevent the NIM AFS Provider from crashing when configured to use
2158 Kerberos v4 for token acquisition in situations where no Kerberos v4
2159 support is available. For example, 64-bit KFW or sites that delete
2160 the Kerberos v4 DLL.
2162 * In afsd_init.log, support logging cache sizes greater than 2GB.
2164 * The volume status output of "fs examine" was being determined based
2165 upon the wrong error value. Instead of using the pioctl() return
2166 code, the errno value must be used.
2168 * Prevent invalid pts auto-registration attempts by aklog and the
2169 NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
2171 * Add synchronization protection to all of the SMB Listener State
2172 variables. This prevents race conditions when the addition or
2173 loss of a network adapter takes place.
2175 * On Vista, there is a race condition between the restart of the
2176 afs client service and the network adapters. Ensure that the
2177 afs client service can handle switching between loopback and
2178 non-loopback modes of operation.
2180 * Fix FollowBackupPath to work with mount points that are not
2181 explicitly read/write.
2183 * Fix a deadlock that has been observed on Vista while resuming
2184 after at least four hours of sleep.
2186 * Add new HKLM registry configuration options that can be used
2187 to provide configuration hints to the NIM AFS Provider when
2188 creating new identities. See release notes for details on
2189 HKLM\SOFTWARE\OpenAFS\Client\Realms\.
2192 * Speed up write buffer operations by holding a global read-lock
2193 instead of a write-lock when updating the buffer data version
2194 number. The data version is protect by the associated stat
2197 * Speed up read buffer operations by removing out of date buffers
2198 from the buffer hash tables.
2200 * Protect against a null pointer dereference in afslogon.dll
2201 if QueryAdHomePathFromSid() fails.
2203 * Remove a race condition while accessing private freelance data
2206 * Update the buffer data versions when Freelance directory data
2209 * Complete the implementation of smb_ReceiveTran2GetDFSReferral().
2212 * Return EAS_NOT_SUPPORTED in response to EA read/store/query
2213 instead of returning zero length.
2215 * Return ERROR_RETRY when a server sends VBUSY.
2217 * Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
2218 a volstat plug-in is in use.
2220 * Enhance the volstat interface to support Dfs path mappings.
2222 * Fix "symlink list" to do the right thing for Dfs links
2224 * Fix callback expiration that was broken in 1.5.27.
2226 * Fix a crash at shutdown caused by the lack of synchronization
2227 within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
2229 * Hold the cm_server_t mutex while walking or modifying the
2230 cm_server_vols_t list in order to avoid a race condition.
2232 * Ensure that the cm_server_t refCount is safely modified.
2234 * The pthread library implementations of pthread_cond_wait
2235 and pthread_cond_timedwait were failing to return with
2236 the mutex held when error conditions occurred. This could
2237 result in race conditions wherever they are used. One case
2238 in which problems were seen was the Rx event_handler().
2240 * Prevent cm_GetBuffer() from locking all buffers associated with
2241 a stat cache object before deciding if the current buffer is up
2244 * Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
2245 object for each offset that we will attempt to reserve.
2247 * Reduce the default number of background daemon threads used to
2248 prefetch or store to one in order to prevent two or more threads
2249 from stepping on each other.
2251 * Reduce the default chunk size to 256KB in order to prevent read
2252 timeouts on low bandwidth connections.
2254 * Fix a race condition that could result in threads becoming stranded
2255 when synchronizing multiple requests on the same stat cache object.
2258 * The GiveUpAllCallBacks functionality that was added in 1.5.21
2259 has been disabled. It has been disabled because it turns out
2260 that the RPC is improperly implemented in all file servers
2261 previous to 1.4.6. The RPC is executed without the proper
2262 locks being held resulting in data corruption and eventual file
2263 server failure. The functionality will be re-enabled in a
2266 * a couple of .readonly Volume Callback optimizations were
2268 - Apply most recent volume callback to all stat cache objects
2269 in the volume. This avoids premature callback expirations
2270 and unnecessary FetchStatus calls.
2271 - Add option to permit attempts to renew .readonly Volume
2272 Callbacks on a periodic basis to avoid invalidation of stat
2273 cache data. [Requires daemonCheckVolCBInterval be set to
2274 non-zero number of seconds. 1800 (30 minutes) is suggested.]
2276 * Adds support for 64-bit data version values. Previous releases
2277 only supported 32-bit values. Larger values sent by the file server
2278 would result in rollover.
2280 * aklog and asetkey no longer generate an exception if Kerberos for
2281 Windows is not installed.
2283 * If the LSA authentication functions report a temporary out of memory
2284 condition, fail the current authentication and do not cause the AFS
2285 client service to terminate unexpectedly.
2287 * When the AFS client service is shutting down, prevent the daemon check
2288 thread from executing additional checks which might attempt to access
2289 discarded locks or unmapped memory.
2291 * Correct a double free condition in the NIM AFS credentials provider.
2293 * Add "FollowBackupPath" service parameter registry value.
2294 This provides equivalent functionality to the UNIX afsd -backuptree
2295 option. When set to a non-zero value, normal mount points originating
2296 in a .backup volume will prefer a .backup volume.
2299 * Fix NSIS uninstall from the .MSI installer.
2301 * Add Stack Frame logging for AMD64
2303 * Add support for EWOULDBLOCK to cm_Analyze. If the file
2304 server returns EWOULDBLOCK, retry the request every two
2305 seconds for up to the RDRtimeout.
2307 * Reorganize the order of the includes to ensure that EWOULDBLOCK
2308 is not assigned the same value as EIO.
2310 * Update CellServDB to GCO Public 25 Oct 2007
2312 * Add additional validation and error handling code after
2313 each call to getSlot(). If an invalid slot is returned,
2314 return NONODE. If the invalid slot is returned when
2315 extracting a data node, invalidate the tree.
2317 * Modify compareKeys() to always perform a case-insensitive
2318 comparison and only perform a case sensistive comparison
2319 if the case-insensitive one matches. This ensures the
2320 ordering is consistently reported.
2322 * Add lock assertions to ensure that all calls are being
2323 performed with the correct locks being held. There have been
2324 some crash reports that provide stack data that does not appear
2325 to be possible unless there is a race. However, there are no
2326 obvious locations where the race is taking place and the test
2327 suite indicates that all of the correct locks are being held.
2328 We shall see what happens in the field.
2330 * For consistency replace all calls to findKey in which the range i
2331 s (1,numentries) with calls to getSlot().
2333 * Optimize the depth search loop by testing the slot value in the
2334 for statement instead of forcing the loop to be broken later.
2336 * Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
2337 There are a number of locations where locks are obtained, dropped,
2338 and reobtained. This reorganization attempts to accomplish several
2341 (1) be optimistic for the most common case so it will be fast
2343 (2) add consistency checks after each location where locks are
2344 dropped and re-obtained. If we lose a race in cm_BeginDirOp
2345 and the bplus tree is out of date, retry until we get to a
2346 consistent state that we can use.
2348 (3) Ensure that all operations take place with the correct locks.
2350 * One of the issues that has become a serious problem since the
2351 addition of the local directory updates is that although cm_SyncOp
2352 synchronizes operations, it does not preserve the order of requests.
2353 This has always been a problem in that it has been possible for a
2354 request to fail to complete due to its worker thread's bad luck.
2355 When a request takes longer than the Windows SMB Redirector's timeout,
2356 the SMB Redirector tears down the SMB virtual circuit.
2358 When using the local directory updates it is really important that
2359 the directory update operations complete in the order that they were
2360 sent to the file server. If they don't, then the local directory state
2361 and the file server state will not match and the local directory state
2362 must be discarded which in turn forces a new read of the entire
2363 directory contents over the network.
2365 This patch adds a new cm_scache_waiter_t object that is used to store
2366 the current thread, buffer, and syncop flags within a waiters queue
2367 on each cm_scache_t object. If a thread is forced to sleep in
2368 cm_SyncOp, upon waking it will check to see if there are any other
2369 threads waiting that are attempting to perform a similar task ahead
2370 of it in the queue. If yes, the thread goes back to sleep. If not,
2371 it goes ahead and enters the cm_SyncOp conflict resolution block.
2373 This patch has the additional side effect of reducing the number of
2374 competing threads that must obtain the cm_scache_t mutex and process
2375 the cm_SyncOp conflict resolution block. As a result, the overall CPU
2376 utilization of the service and the clock time associated with processing
2377 requests will be reduced.
2379 * assert that the cm_scache_t mutex is held when calling
2380 buf_ForceDataVersion().
2382 * rename findNode to leafNode in bplus_Lookup
2384 * replace all OutputDebugString calls with osi_LogX calls
2386 * modify bestMatch to special case the return values for leaf nodes.
2387 If an entry is above or below the values available in the leaf node
2388 return BTLOWER or BTUPPER instead of BTERROR.
2390 * In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
2391 if true convert to either slot 0 or Max and perform the insertion.
2392 This produces easier to read code when performing lookups.
2394 * modify lock_AssertXXX macros to call osi_assertx() and provide a
2395 descriptive message.
2397 * the difference between osi_assert() and osi_assertx() is whether or not
2398 a description is specified as part of the call. When no message is
2399 specified afsd_notifier() is called with a NULL msgp parameter. This
2400 results in a null pointer dereference during the EventLog() call.
2402 As a result, none of the information describing the assertion is written
2405 This commit sets a default message for use when no other message is
2406 provided and it replaces all calls to osi_assert with osi_assertx and
2409 * Migrate B+ tree search key into thread local storage
2412 * Fix an uninitialized variable in lana_IsLoopback() which
2413 could result in a random determination that a loopback
2414 adapter is not a loopback adapter. This would result in
2415 a computed Netbios service name of %MACHINE%-AFS instead
2416 of AFS. This is a serious problem if the value computed
2417 by the AFS Client Service and the client process pioctl
2418 calls produce different results as it will result in a
2419 communication failure.
2421 * Fix a crash in the B+ Tree code that results when a file
2422 rename only results in a change of case.
2424 * Fix the pioctl path parsing. The previous algorithm
2425 broke the current directory of the client into two parts,
2426 the TIDpath which represented the device root and the
2427 relative path. The problem is that this is not sufficient
2428 from proper processing of symlinks and mount points.
2430 What is required is that the last component of the path
2431 must not have the symlink or mount point be evaluated
2432 whereas symlink and mount point evaluation must be performed
2433 for the rest of the relative path. Therefore, the path is
2434 now split into three parts: TIDpath, intermediate path,
2437 "fs examine" now reports the data of the symlink or mount
2438 point object instead of the target and "symlink list" and
2439 "symlink remove" now works. The "Explorer Shell Extension"
2440 menu now is displayed for symlink objects.
2442 * More informative symlink.exe error messages.
2444 * When the cmd.exe "del" command is used to remove a symlink,
2445 not only was the link object removed, but the stat cache
2446 object for the target was marked deleted. This resulted
2447 in cache corruption which would require the destruction
2448 of the cache to correct.
2450 * Add undocumented registry value, "OfflineReadOnlyIsValid"
2451 which results in cached contents from read only volumes
2452 to be treated as valid when there is no callback registered
2453 and all servers are down or all instances are offline.
2455 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
2457 * Add documented registry value, "AcceptDottedPrincipalNames"
2458 which is used to permit the acquisition of AFS tokens
2459 using Kerberos v5 principal names which include a dot in
2460 the first component.
2462 {HKLM\HKCU}\Software\OpenAFS\Client
2464 * "fs uuid" does not require "AFS Client Administrator" group
2465 membership to display the current UUID value.
2468 * Update NSIS installer to version 2.30
2470 * Update Wix installer to version 2.0.5325
2472 * 64-bit KFW support:
2473 - Network Identity Manager plug-in
2475 - 64-bit krb5 enable afscreds.exe, afs server manager,
2478 * Enable B+ tree Directory Search by default.
2480 * NIM plug-in is now compatible with FILE: ccaches
2482 * Install instloop.exe into Client/Programs/ even though
2483 it isn't required. Permit it to be used by organizations
2484 that clone Windows systems.
2486 * On a suspend, only drop callbacks for servers that
2487 are online. If the server is offline, maintain the callback
2488 until the server is once again available.
2490 * Prevent against acquisition of a directory write lock when
2491 there are no updates to be performed. This was resulting
2492 in a write-lock leak which later on would lead to a deadlock.
2494 * Add configurable file pre-fetch mode. For files that match
2495 the specified file extensions, pre-fetch the entire file into
2498 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
2499 (MULTI_SZ) PrefetchExecutableExtensions
2501 * Fix the client_osi large integer support so that return values
2502 can be properly assigned to large integer variables.
2504 * Do not hold cm_scache_t mutex across cm_GetCell() calls
2505 since cm_GetCell() can block in a RPC.
2507 * Add short file name support to B+ tree searches.
2509 * Resolve a conflict between the desire to notify file servers
2510 that the client workstation is suspending and the desire to
2511 permit objects with valid callbacks when the client suspends
2512 to continue to use those objects from the cache if the client
2513 resumes without network connectivity.
2515 * Fix stack variables to ensure that all buffers storing paths
2516 can handle AFSPATHMAX (1024) characters. Paths of 256 characters
2517 would cause a stack overflow condition.
2519 * On Vista, force afs_config.exe to run with the highest privileges
2520 available to the end user. For a user in the Administrators
2521 group running under UAC, this will cause afs_config.exe to be
2522 be run with full Administrators privileges after the user
2523 approves the privilege escalation.
2525 * More VNOVNODE handling issues. If a dirty buffer cannot be
2526 written due to VNOVNODE, clear the buffer and ignore the error.
2527 If a cm_scache_t object has been marked deleted due to a VNOVNODE
2528 error, the next time the smb_fid_t object that references it is
2529 used return either NOSUCHFILE or NOSUCHPATH and close the
2530 smb_fid_t object. Subsequent attempts to use the smb_fid_t will
2531 result in an invalid handle error.
2533 * For Vista, set the RequestExecutionLevel of the NSIS installer
2536 * For the build environment, support the generation of manifests
2537 when using the CL1310 (VS .NET 2003) compiler/linker.
2539 * There was a bug in OutputDebugHexString() which could overflow
2540 a stack buffer resulting in unexpected process termination
2541 of afsd_service.exe if the input Security Token contained
2542 many bytes equivalent to the '%' char.
2544 * Add directory lookup statistics to "fs memdump" output.
2548 * Windows uses case-insensitive file name pattern matching
2549 but AFS is a case sensitive file system. The AFS3 directory
2550 format is block based, uses network byte order and
2551 includes a hash table for fast case sensitive lookups.
2552 This causes several problems for the Windows AFS client.
2553 (1) Traversing the directory blocks is cpu expensive
2554 (2) A hash table miss does not indicate that the desired
2555 entry does not exist.
2556 (3) Determining whether a non-ambiguous inexact match or
2557 the entry does not exist requires a linear traversal
2558 of the entire directory.
2559 These issues often result in 100% CPU utilization.
2561 These issues are addressed by building a modified B+ tree for
2562 each directory and then using the B+ tree for searches.
2564 This feature is disabled by default. To enable it set
2565 the "BPlusTrees" registry value to 0x01 in the AFS Client
2566 Service Parameters key.
2568 * Correct a null pointer dereference in the symlink recursion
2569 detection code. (Introduced in 1.5.22)
2571 * Add the Network Identity Manager AFS Provider to AMD64
2576 * A new registry value, "BlockSize", can now be used to adjust
2577 the size of the internal buffers used to store file data.
2578 The default is 4KB. When set, the values 1..1024 specify
2579 multiples of 4KB. Values > 1024 specify actual block sizes
2580 and must be multiples of 4KB. The actual block size in KB
2581 can not be larger than the chunk size which defaults to 1MB.
2583 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
2586 * pts.exe and ptserver.exe built with -DSUPERGROUPS
2588 * Service Manager shutdown hints have been added. The hint
2589 specifies that it may take the AFS client service up to two
2590 minutes to shutdown.
2592 * Fix the recursion detection feature introduced in 1.5.22.
2593 The previous implementation broke the evaluation of paths
2594 using ../path or ./path. It also broke the evaluation of
2595 symlinks to absolute paths. For example, /afs/cell/path.
2597 * Fix an uninitialized variable in cm_MakeDir that could result
2598 in crash if the requested directory could not be created.
2600 * Fix a the VolumeStatusNotification log messages to save
2601 the input string in the circular log buffer so garbarge
2602 won't be output to afsd.log when the log is dumped.
2605 * Increase the SMB Maximum Path length to 260 characters
2606 from 256 as that is what CIFS supports.
2608 * On 64-bit Windows, use a unique description string for
2609 the 32-bit AFS Shell Extension.
2613 * In the AFS Admin Library, replaced all of the
2614 Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
2615 allows parameter checking and proper type conversions to be
2616 performed. This is crucial when passing parameters that are
2617 64-bit wide on systems in which 'long' is a 32-bit value.
2618 Otherwise, the parameter list on the stack is interpretted
2619 incorrectly and the function not only doesn't succeed but it
2620 might corrupt memory and if lucky crash the application.
2622 The only program that we care about that uses this library
2623 is the AFS Server Manager. Now it doesn't crash.
2625 * Fixed smb directory searches for non-wildcard names. The
2626 optimized function should not have set the bulk in progress
2627 flag. This would adversely affect later directory searches
2628 that required a bulk search.
2630 * Detect infinite recursions (up to 512 stat cache objects
2631 in a chain.) If an infinite recursion is detected, return
2632 an error to stop the CIFS client from querying us forever.
2634 * Executables executed out of AFS name space can have their
2635 pages swapped out by Windows. Those pages will be reloaded
2636 when they are required. However, if the cache manager no
2637 longer has a callback and all of the file servers with
2638 the volume data are down, then the application will crash
2639 due to the inability to read the required page.
2641 We now track the down time of each server. If the callback
2642 expires after the server that issued it went down and all
2643 servers are down, we delay the expiration to permit the data
2644 in the cache to be continue to be used by Windows or an
2647 * The Windows AFS client has always had very poor performance
2648 with regards to directory management. In the case of
2649 directory updates caused by create, rename,and delete operations,
2650 the AFS cache manager will now perform local directory updates
2651 when possible in order to avoid re-reading the directory from
2652 the file server. This will be done in whenever both of the
2653 following conditions are true:
2655 - the data version on the directory increased by one as a
2656 result of the operation
2658 - all of the directory buffers necessary to make the change
2659 are present in the cache
2661 * The error message in the AFS Control Panel describing the
2662 need to be a Windows Administrator has been replaced with
2663 a message describing the need to be in the AFS Admins Group.
2665 * If the vldb server list is empty when attempting to evaluate
2666 a cell mountpoint it was possible for a crash to occur. This
2667 can only happen if there is an existing mount point and the
2668 is a failure attempting to load the server list from the
2669 CellServDB file or via AFSDB queries.
2671 * dirty buffers only have their dirty bytes stored to the file
2672 server instead of the entire chunksize
2674 * the default chunksize has been increased from 128KB to 1MB
2679 * changed the enum values for cm_serverRef_t state info to use a
2680 private name space to avoid collisions (srv_)
2682 * added a srv_deleted state for cm_serverRef_t objects. This
2683 state is set when cm_FreeServerList() is called with the
2684 CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
2685 not always delete the cm_serverRef_t from the list if it is
2686 still in use by another thread. the srv_deleted state means
2687 the object's contents are no longer valid and it must be
2688 skipped. It will be deleted the next time the object is
2689 freed and the refcount hits zero.
2691 * the srv_deleted state is also used when a file server reports
2692 either VNOVOL or VMOVED instead of marking the cm_serverRef_t
2693 as offline. This is done to prevent additional usage of the
2694 stale vldb data while waiting for the update volume request
2697 * added a state field to the cm_volume_t object (enum volstate
2698 vl_ name space) that maintains the state of the volume based
2699 upon the states of all of the cm_serverRef_t and cm_server_t
2702 * modified cm_UpdateVolume() to set the state of the cm_volume_t
2703 RW, RO, and BK to either vl_alldown or vl_online. There can't
2704 be any other states because cm_UpdateVolume() destroys any
2705 previous knowledge we might have had regarding busy or offline
2708 * modified cm_UpdateVolume() to update the volume name in the
2709 cm_volume_t to the volume base name if the previous value was
2712 * modified cm_FollowMountPoint() to check to see if the volume
2713 name is a volume ID and if so call cm_GetVolumeByID instead
2714 of cm_GetVolumeByName. This ensures that volume IDs are always
2715 looked up as numeric values. There is no longer a need to
2716 maintain a separate cm_volume_t containing the string representation
2719 * Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
2720 The first flag is a "CREATE" flag which is set by all existing
2721 calls. The flag is not set by calls to cm_GetVolumeByID() from
2722 the server probe code when volume status is being updated. We
2723 do not want the server probe operation to result in additional
2724 turnover in the cached volume data. The second flag is NO_LRU_UPDATE
2725 which is set when the server probe code updates the volume status.
2726 This flag will be used to prevent the server probe operation from
2727 changing the order of the least recently used queue.
2729 * Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
2730 allocated for a given set of normal, readonly, and backup volumes
2731 regardless of whether or not the volume is accessed via name or
2732 ID number. The cm_volume_t namep field is always the base name
2735 * Added a new volume state, vl_unknown. This state is used as
2736 the initial state for all cm_volume_t when the cache manager starts,
2737 for each cm_volume_t at creation, and for each cm_volume_t when
2738 recycling. The cache manager does not know the state of all
2739 volumes in the world, only those that are in the cache and for
2740 which it has queried the VLDB and hosting file servers.
2742 * modified cm_GetVolumeByName() to initialize the state of a
2743 volume to vl_unknown. The actual state will be set when a
2744 cm_VolumeUpdate() call completes successfully.
2746 * changed name of scache hash table variables to avoid ambiguity
2747 when adding hash tables for volumes
2749 * fix a buffer overrun in sys\pioctl_nt.c pioctl().
2751 * modified cm_UpdateVolume() to handle the case in which there is
2752 no RW volume but there is are RO volumes for a given base name.
2753 This is done by querying for the ".readonly" volume name if the
2754 base name does not exist in the VLDB. We never query for the
2755 .backup name because under the current usage model a .backup
2756 volume may only exist on the server that the read-write volume
2757 is located. If there is no RW volume, there can be no .backup.
2759 * Added four hash tables for cm_volume_t objects to improve the
2760 search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
2761 cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
2762 BKID. Three ID hash tables are necessary as long as it is
2763 desireable to maintain a single cm_volume_t containing all
2764 of the related RW, RO, and BK volume data. Having the RW and
2765 RO volume data in the same object is necessary for the
2766 implementation of cm_GetROVolumeID() which returns either the
2767 RO or RW ID depending upon the existence of RO volume instances.
2769 * Added a volume LRU queue so that volume reuse becomes fairer.
2770 This does not replace the all Volumes list which is used when
2771 it is desireable to walk a list of all the volumes whose order
2772 is not going to change out from underneath you which makes it
2773 safe to drop the cm_volumeLock.
2775 * handles volume hash table updates where volume name to
2776 volume ID number changes. The volume name remains
2777 constant in the cm_volume_t. if a vos rename is performed,
2778 the name of the volume will change and the volume IDs will be
2779 updated. Subsequent access to the old volume ID will create a
2780 new cm_volume_t with the new name.
2782 * Added a daemon thread operation to query the state of volumes
2783 listed as busy or offline. cm_CheckBusyVolumes() calls
2784 RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
2785 or vl_offline. If the volume is now online, the status on the
2786 volume is updated. The default period is 600 seconds. This can
2787 be configured with the BusyVolumeCheckInterval registry value.
2789 * Added prototype for smb_IoctlPrepareRead() which was missing a
2790 return type in the function definition.
2792 * Added volume id lists to the cm_server_t. These lists are
2793 allocated in blocks of ~32 IDs. When a cm_PingServer()
2794 detects a change in server state, the state of the cm_volume_t
2797 * Added volID to the cm_serverRef_t object. volID is used
2798 to identify the volume for which the object is a referral.
2799 cm_FreeServerList() uses the volID to remove the volume
2800 from the cm_server_t.
2802 * In cm_Analyze, when VNOVOL or VMOVED are received,
2803 call cm_ForceVolumeUpdate() to force a refresh of the volume
2806 * Added cm_VolumeStatusNotification() which is used at the moment
2807 to log volume status changes to the trace log. It will also
2808 be used as the access point to the File System Filter driver
2809 notification engine.
2811 * Added an all cm_scache_t list to cm_data. This replaces the use
2812 of the stat cache LRU queue when we need to enumerate all
2813 entries. The LRU list order is not static and when using it to
2814 enumerate all entries it can result in items being missed or
2815 items being processed more than once.
2817 * Modified cm_Analyze(). Instead of reseting the busy or offline
2818 state of a volume and forcing a retry of the operation
2819 cm_Analyze will defer to the background daemon thread that will
2820 update the state once every 600 seconds.
2822 * Added the automatic generation of a Freelance ".root" read-write
2823 mountpoint that refers to the root.afs volume of the workstation
2824 cellname at the time the mountpoint is created.
2826 * re-write cm_Analyze to make better use of the known volume
2827 status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
2830 * renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
2831 busy volumes will be reset to srv_non_busy by the function
2832 but there is no mechanism for querying the busy state other
2833 than by attempting to access the resource.
2835 * cm_Analyze will query the state of an offline volume before
2836 deciding whether or not to retry when all volume instances
2839 * the algorithm used after an IP address list change detection
2840 has been modified. we now check the state of the VLDB servers,
2841 then reset the rx connections, and finally check the state of
2842 the file servers. Previously we flushed the rx connections
2843 and then checked the servers in whatever random order they
2844 were first discovered.
2846 * fix the volume status reported by "fs examine"
2848 * various improvements to the AFS Network Identity Manager Provider.
2849 - Don't access freed objects after the window has been destroyed.
2850 - Write back the AFS cell list in the identity configuration
2851 even when its empty.
2852 - Disable all use of the provider when AFS is disabled for a
2853 particular identity.
2855 * Fix cm_cell_t leak. When evaluating a mountpoint that uses an
2856 alias instead of the full cellname, double check the full cell
2857 name and see if we have a matching cell before allocating a new
2860 * Dump more data with "fs memdump". All volumes and scaches and
2863 * Reverse the order of the all cells list. This ensures that
2864 the workstation cell is always first in the list.
2866 * If we recognize the server from which a callback revoke was
2867 received, then we can ensure that we only invalidate the status
2868 information for cm_scache_t objects belonging to that cell.
2870 * When building for amd64, use APPVER 5.02 as we do not support
2871 64-bit Windows 2000 and explicitly state the machine type in
2874 * no longer treat VMWare network adapters as equivalent to a
2875 loopback adapter. For some unknown reason, publishing the "AFS"
2876 netbios name on them frequently results in a name collision
2879 * Use read locks whenever possible instead of write locks when
2880 walking the allCellsp list. This permits better use of multiple
2883 * In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
2884 is not present on the directory optimization. It would report
2885 mount points and directory objects for which there was no
2886 status information as a zero length file. Instead we can use
2887 the vnode number of the object to determine if it should be
2888 considered to be a directory or a file. If it is odd it is a
2889 directory; even it is a file.
2891 * Give Back Callbacks to the file servers in response to suspend
2892 and shutdown power management events. Only give up callbacks
2893 to UP servers and do so with a very short RX dead timeout.
2896 * Fix a race condition in the SMB FID cleanup code that could
2897 result in a stat cache mutex never being released.
2899 * Fix the graphics files that are installed as part of the
2900 IBM Administrators Guide.
2902 * Change the behavior of pioctl path parsing to not follow
2903 symlinks or mount points during evaluation. This results in
2904 the FID being the symlink or mount point and not the object
2905 they point to. This is consistent with the behavior of the
2908 * Enable AFSDB record lookup for command line tools such as
2911 * CellServDB update 2007-07-05
2913 * Added a Volume Status Plug-in Interface that permits third parties
2914 to receive notifications when the volume state changes occur.
2916 * Ensure that all writes to the afsd_alloc.log file use CR-LF.
2918 * Prevent the afsprocmgmt.dll from being unloaded while the
2919 background thread it starts is running.
2921 * Avoid a race condition when closing smb_fid_t objects.
2924 * fix install path of IBM Quick Start Guide when installed by
2927 * fix the client so that it can write into directories with
2928 only 'li' permissions.
2930 * fix installation of 32-bit Shell Extension for 64-bit systems.
2931 Install it under its own HKCR keys.
2933 * modify the wix installer to remove and upgrade install packages
2934 with the same version number.
2936 * when flushing volume location data every two hours or in
2937 response to "fs checkvolumes", also flush the mount point
2938 links in order to force their re-evaluation.
2940 * aklog was copying an uninitialized buffer as part of the
2941 foreign cell automatic pts registration
2943 * when obtaining directory listings, check for read permission
2944 on the directory. If read permission does not exist, fake the
2945 status information for the directory entries. This improves
2946 performance, prevents the client from pounding the file server,
2947 and prevents leakage of cached status information.
2949 * "fs getcacheparms" output was incorrect.
2951 * NetIdMgr AFS Provider should only automatically add token
2952 acquisition for cell foo.com if the realm of the cell matches
2953 the realm of the identity.
2955 * Fix a null pointer dereference in cm_MergeStatus()
2959 * network identity manager provider:
2961 - Correct for the fact that pr_CreateUser tampers with the
2962 contents of the username input buffer.
2964 - Take into account that the krb4 credential data structure
2965 does not have separate realm fields for the user and service.
2966 When krb524 is used with a cross-realm ticket, the realm
2967 describes the service not the user.
2969 * Many integrated logon changes:
2971 - Fix eventlog reporting. Do not attempt to log an event
2972 if the event source registration fails. Use DebugEvent0
2973 instead of DebugEvent when there are no parameters.
2975 - Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
2978 - Fix the reading of Domain specific configuration for
2979 LogonScript and TheseCells. Previously the dwSize value was
2980 being overwritten so that subsequent RegQueryValueEx call
2983 - Fix a memory leak in the TheseCells reading code.
2985 - Add support for Domain specific "Realm" specification. The
2986 realm is the realm to be appended to the username. When
2987 logging in as a domain or to the local machine, the specified
2988 "Domain" name is not going to be a valid realm name.
2990 - Construct a proper principal name based upon the domain
2991 specified realm for use in obtaining tokens with KFW.
2993 - If the domain specified "TheseCells" list includes the
2994 default cell, do not obtain tokens twice.
2996 * An unprotected call to the Kerberos v5 function,
2997 krb5_free_context(), can result in an invalid memory access
2998 exception if the contents of the krb5 profile, krb5.ini, cannot
2999 be parsed correctly. The error is located within the afskfw
3000 library which is used by afscreds.exe, afssvrmgr.exe, and
3001 afslogon.dll. When the error is triggered within afslogon.dll
3002 a blue screen condition can occur when the winlogon.exe system
3003 process terminates unexpectedly.
3007 * Do not stop the afs client service with an assert
3008 if the cell name cannot be determined when responding
3009 to a "fs whereis" command.
3011 * Integrated Logon (afslogon.dll). No longer attempt
3012 to preserve Kerberos 5 tickets for the logon session
3013 as that job is now supported by KFW 3.1 and above.
3015 * The Network Identity Manager plug-in will attempt to
3016 remove expired service tickets from the cache.
3020 * CellServDB update - GCO Public CellServDB 09 Mar 2007
3022 * AFS Server Manager and AFS Account Manager data cache
3023 moved from HKCU\Software\TransarcCorporation to
3024 HKCU\Software\OpenAFS to prevent crashes due to data
3027 * Reduce the default UP server probe time from 10 minutes
3028 to 5 minutes to match UNIX. This will not solve the NAT
3029 port mapping timeout issue because most NATs default to
3030 a value less that 5 minutes and this value ensures that
3031 the afs client server will start probes 5 minutes after
3032 the last probe set ended. The probe period for any one
3033 server is therefore longer than 5 minutes. The more
3034 servers in the list the longer the time between probes.
3036 * Before marking a server as down in cm_Analyze, retry the
3037 request using a new rx connection to the server. If the
3038 cache manager is behind a NAT, it is possible that the
3039 response from the server is being sent to the original
3040 address:port from which the rx connection was established.
3041 This will appear to the cache manager as the server being
3042 down when in fact it is not. Forcing the use of a new rx
3043 connection will result in the server responding to the
3044 most recent address:port assigned to the client by the NAT.
3046 If the server is in fact down, it will simply take a little
3047 longer to notice it.
3051 * Removing a directory that contained entries was failing with
3052 the wrong error code being returned to the CIFS client.
3056 * rxdebug exponential back-off. don't take down a server that
3057 is already struggling to respond.
3059 * Network Identity Manager plug-in registers help on the Help
3060 menu if it is supported by NIM. (KFW 3.2 and above)
3062 * Plug-n-play networking did not always work after coming back
3063 from a Windows Suspend. The AFS client service failed to
3064 re-compute the LAN adapter to bind to and instead attempted
3065 to bind to all LAN adapters. This would polute the public
3066 network with the name "AFS" which could lead to conflicts
3069 * Build installers using Microsoft Hotfixes provided at
3070 http://tinyurl.com/243svk for Visual Studio C RunTime Library
3071 Daylight Savings Time issues.
3073 * "fs checkserver" does not require admin privileges unless
3074 the probe interval is being changed.
3078 * A reference leak was discovered in the mountpoint traversal
3079 functionality which resulted in cm_volume_t objects not becoming
3080 eligible for re-use. This will trigger a panic condition when
3081 more than the maximum number of volumes is accessed.
3083 * Fixed the success test for the "fs memdump" operation. Success
3084 is now reported as such. Added cm_volume_t object information
3087 * A data corruption error has been fixed. Write requests
3088 queued for background processing were not retried upon failure.
3089 This leads to inconsistencies between the contents of the file
3090 in the local cache and the contents of the file on the file
3093 The implemented fix is to re-queue any request that has failed
3094 due to timeout, volume busy, volume offline, server busy, or
3095 server offline. Requests are not processed while all servers
3096 for a volume are inaccessible. When servers become accessible
3097 the background request will be retried.
3099 In addition, if a background write fails, do not mark the buffers
3100 as clean. Failure to preserve the dirty flag prevents the buffers
3101 from being written to the file server in the future.
3103 This fix does not address the possibility that the credentials
3104 necessary to perform a write request might expire between the
3105 time the request is queued and when the request can be processed
3106 due to server accessibility.
3108 * The Freelance volume does not support per user ACLs. The
3109 cm_HaveAccessRights function did not handle this special case.
3110 Requests for privileges greater than read|lookup would result
3111 in afsd_service.exe entering an infinite loop. ISOBuster is
3112 known to cause this behavior.
3114 * When deleting files, the SMB Server would attempt to obtain
3115 the access permissions for the parent directory instead of
3116 for the file itself. This would result in an incorrect
3117 assessment of whether or not the file can be deleted.
3119 * The Network Identity Manager AFS plug-in has had its string
3120 tables updated for consistency with the language used by NIM.
3121 All credentials are referred to as credentials instead of tokens.
3122 Drop-down combo boxes now in fact drop down.
3124 * The AFS Salvager was not properly being built on Windows.
3125 This is relevant only if the AFS servers are in use.
3127 * The default signal handlers for the AFS servers incorrectly
3128 processed SIGQUIT. Instead of stopping the server, an
3129 exception was raised.
3131 * Fix another volume refcount leak. This one was triggered by
3132 an inability to contact the vlservers in order to update the
3135 * Fix checklist control used by the drive mappings tabs in
3136 afscreds.exe and afs_control.exe.
3138 * Improve output of "fs memdump" related to buffer management.
3140 * Fix refcount leak of cm_buf_t objects when prefetching buffers
3141 already in the process of being fetched.
3143 * Improve cache performance by computing the buffer hash table
3144 size based upon the number of buffers so that the average
3145 number of entries in a bucket is seven.
3147 * Improve cache performance of read/write data by locally
3148 updating the dataVersion of all buffers associated with the
3149 file on each StoreData or StoreStatus operation provided that
3150 the dataVersion was incremented by one and the dataVersion of
3151 the buffer matched the original dataVersion of the cached
3152 status information for the file.
3154 * When recycling cm_scache_t objects, make sure that we
3155 clear all of the previous fields. It is no longer the
3156 case that we can assume that the stat cache object is being
3157 obtained only after a successful FetchStatus call.
3158 InlineBulkStatus will allocate stat cache objects for which
3159 the no status information is available.
3161 * When a file is being opened with the OPEN_ALWAYS disposition,
3162 do not require write permissions when the file already exists
3163 unless of course the file is being opened requesting write
3166 * When a file has been opened for read-only access, if the
3167 application requests an exclusive lock, obtain a read-lock
3168 instead of a write-lock. The purpose of the exclusive lock
3169 in this situation is to prevent changes to the file and the
3170 full file read lock will provide this guarrantee.
3175 * Enforce Unix Mode Write bit to prevent file deletion of read only
3178 * The CIFS READ and LIST permissions use the same value 0x0001.
3179 When a directory is being accessed, check for PRSFS_LOOKUP
3180 instead of PRSFS_READ.
3184 * Fix computations of short names of mount points. They were
3185 broken as part of the directory search optimization. (1.5.13a)
3187 * When all buffers are in use, periodically sleep the threads
3188 waiting for a buffer to become free while the global buffer
3189 lock is not being held in order to permit those threads that
3190 can free buffers to do so.
3192 * Power Management improvements. Maintain a global flag that
3193 specifies whether or not the service is in a suspend state.
3194 Do not panic if Netbios() returns NRC_BRIDGE meaning that the
3195 lana is no longer valid. Instead, stop the listener threads
3196 and if all listener threads are stopped, reset the lana_list.
3197 Allow the cm_Daemon() thread to periodically check the state
3198 of the smb listeners. If they are all stopped and the service
3199 is not suspended, attempt to restart them. If there are no valid
3200 lanas, return to the stopped state.
3202 * CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
3203 the user is permitted to obtain read or write locks. This function
3204 would obtain the lock and then drop it returning whether or not
3205 the lock could be obtained. If the lock was in fact required,
3206 CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
3207 of course being that this pattern results in three RPCs to the
3208 file server (lock, unlock, lock). This is reduced to one RPC by
3209 implementing cm_CheckNTOpenDone() which frees the allocated byte
3210 range lock from cm_CheckNTOpen() after the cm_Lock() call is
3213 * Optimize the SMB FindFirst processing for the case in which there
3214 are no wildcards. If a cm_Lookup succeeds, generate a response
3215 from the stat cache entry we have a callback for. Its cheaper
3216 than walking each buffer of directory data associated with the
3217 directory stat cache entry when we have thousands of directory
3220 * Increase the SMB Ioctl MaxData size to support larger token sizes
3222 * Disable autorun of the AFS Server Wizard when the AFS Server is
3225 * Add AFS Server service dependencies on tcpip and PNP_TDI
3227 * Remove the AFS Server Configuration Tool shortcut
3229 * Add Freelance mode awareness to AFS Server Configuration Tool
3233 * Unlike previous versions of the OS, Vista performs a shutdown on
3234 the Microsoft Loopback adapter just like it would on a real adapter.
3235 This causes the smb_Listener threads to trigger a panic during a
3236 suspend/hibernate power event.
3238 The fix is to unbind from the network adapters in response to a
3239 suspend/hibernate power event and then rebind to the adapters
3240 when a resume power event is received. Note that the resume events
3241 are not reliably delivered so it is possible the afs service will
3242 not be accessible. However, this is the best we can do.
3244 * don't allow environment variables we set in afslogon.dll to be
3245 inherited by child processes
3247 * force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
3250 * The old dirty buffer synchronization algorithm had a buf_IncrSyncer
3251 thread walking the all buffer list periodically searching for dirty
3252 buffers to write to the file server. This had several negative
3253 results. The algorithm ate up ever increasing amounts of CPU time
3254 even when AFS is idle as the size of the cache increases. Also,
3255 buffers were written to the file server in an order based upon the
3256 original buffer allocation which has nothing to do with the order
3257 in which the buffers became dirty.
3259 The new algorithm maintains a dirty buffer list. Items are added
3260 when the buffer is originally marked dirty. A buffer is only
3261 removed from the list by the buf_IncrSyncer when the buffer is no
3262 longer dirty. If the list is empty the thread goes back to sleep
3263 immediately without additional processing requirements.
3267 * Enable renaming of directories from the Explorer Shell when
3268 Offline Folders is enabled
3270 * An error is reported in response to the SMB ReceiveCoreFlush
3271 request when an error has not occurred.
3275 * optimize dirty buffer flushes in response to suspend operations
3276 do not flush if only loopback adapter is installed
3277 do no retry dirty buffer writes
3279 * do not reset netbios when performing loopback checks. doing
3280 so will break the existing lana bind
3282 * remove a deadlock in smb3.c QPathInfo
3286 * Add NetIDMgr AFS Plug-in
3288 * "Show Tray Icon" checkbox removed from afs_config.exe
3290 * Disable DNS Registrations for the Loopback Adapter and make sure
3291 that Netbios is turned on.
3293 * A reference count leak on stat cache objects was fixed
3295 * For debugging, tread ids of writers are now tracked on locks
3297 * Corrected an incorrect lock that was being held but not
3298 freed in a rare error case
3300 * Changed trace log end of line to CR-LF so it can be read
3303 * modify the buffer synchronization thread to write all dirty
3304 buffers and sleep only when there is no work to be done
3306 * improve behavior of cache manager when large numbers of
3307 temporary files are being created by multiple clients within
3310 * do not permit dirty buffers that belong to deleted files
3311 to be written to the file server. its a waste of bandwidth
3313 * do not leak stat cache references when the file server
3316 * afscreds: do not permit an attempt to obtain tokens if the
3317 afs service could not be started. this was causing stack
3318 overflows on Vista after resuming from a suspend
3322 * better handling of dot directory
3324 * fs sq and fs mkm now behave as on UNIX
3326 * add license to installers
3328 * Return Invalid Parameter error for unknown InfoLevels in
3329 SMB Tran2 Search Directory.
3331 * Correctly use WSAStartup and WSACleanup within the RX library
3333 * Check that afsd_service.exe is running before performing
3336 * Force AFS Shell Extension to load resource DLL using
3337 afs_shl_ext.dll as a base name.
3339 * (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
3340 OpenAFS\Client\Program
3342 * Fix SMB_SET_FILE_END_OF_FILE_LENGTH
3344 * Do not panic if the maximum number of volume entries are in use
3345 and one of them can be recycled.
3349 * Fix the code that reads the backconnnectionhostnames value
3350 so that there is not an uninitialized variable
3354 * integrated logon could cause winlogin.exe to blue screen windows
3355 if it was called with a domain that does not exist on the machine.
3356 this is possible during a sysprep.
3358 * corrections to cm_buf_t handling to ensure that buffers are
3359 removed from queues before they are freed and that holds are
3360 obtained on the next buffer in a chain before releasing the
3363 * do not generate MessageBox dialogs on network errors. they
3364 cause more trouble then they are worth.
3366 * move GlobalAutoMapper into a separate thread so that it does
3367 not prevent service startup if the WNetAddConnection2 call
3370 * restrict hard dead and connection timeouts to the CIFS Session
3371 Timeout value. This is being done to try to prevent virtual
3372 circuits from being destroyed prematurely.
3374 * Fix the code that writes the backconnectionhostnames value
3375 to ensure that the data buffer is written with the correct
3378 (there were no windows specific changes between 1.5.4 and 1.5.5)
3382 * Corrections to the Integrated Logon functionality that permits
3383 Kerberos 5 TGTs to be communicated into the user session. The
3384 transfer is now more secure and succeeds when the user does not
3385 have read permission to the %SystemRoot%\TEMP directory.
3387 * NSIS installer updated to version 2.18
3389 * WiX installer updated to version 2.0.4221.0
3391 * Updates to the IFS functionality for 64-bit correctness and
3392 permit compilation. The IFS functionality works for pioctl()
3393 operations but not much else at this point.
3395 * A major re-write of the OpenAFS implementation of the CIFS Tran2
3396 query and set functions.
3398 * Ignore the Windows CIFS client session timeout entirely. Timeout
3399 only after the AFS Hard Dead Timeout period expires.
3403 * move headers, libraries, and samples from OpenAFS\Client to
3406 * remove the auto-registration of AFS IDs for foreign cells
3407 from integrated logon. this permits afslogon.dll to avoid
3408 initializing the RX library. afslogon.dll is a network
3409 provider library and as such it is loaded into all applications
3410 which perform a CreateProcessAsUser call. Initializing the
3411 RX library spawns background threads to listen for inbound
3412 packets. When the network provider is unloaded from the
3413 process, the RX library does not cleanup the threads. If they
3414 continue to execute after the library is unloaded, the process
3417 * implement inline bulk status calls. this will significantly
3418 improve the performance of directory listings.
3420 * prevent a crash when evaluating mountpoints to volumes that do not
3423 * new command: fs uuid [-generate]
3424 displays the current uuid and allows a new one to be generated
3425 on the fly without restarting the afs client service.
3427 * fix the 64-bit installers to properly document the install
3430 * move help files to the proper location so they can be located by
3431 the system tray tool and the control panel.
3433 * "over quota" and "disk full" errors were not being reported to the
3436 * largefile support has been added. Files greater than 31-bits can no
3437 longer be written to file servers which do not support large files.
3438 When file servers support large files, files up to 64-bits can be
3441 * cm_SyncOp could return with the input buffer unlocked upon error
3442 which would result in a panic condition. This bug was introduced
3445 * vlserver failover when freelance mode did not work. This prevented
3446 the root.afs volume from being loaded if the first vlserver could
3451 * A bug in the CIFS Standard Info response packet caused directories
3452 to appear as standard files. This prevented Microsoft Vista from
3453 being able to walk the AFS directory tree.
3455 * Mount points must be reported as directories in response to
3456 Query Path Info operations.
3458 * A deadlock was detected when performing "fs flushall" if the
3459 file server reports VNOVNODE. The scp->createBufferLock is
3460 already held by the current thread and there is no mechanism
3461 to propagate the knowledge. Therefore, an alternate mechanism
3462 for clearing the cache must be developed. A new function
3463 cm_RecycleSCache(scp,flags) has been extracted from
3464 cm_GetNewSCache(). This function performs the task of recycling
3465 an cm_scache_t object. When called from cm_GetNewSCache() with
3466 no flags the expectation is that there are no associated buffers
3467 that are queued to be read or written. When called from
3468 cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
3469 flag, any queued buffers will be de-queued and marked as if
3470 the operations were performed so the data can be discarded.
3472 * CM_ERROR codes were logged as "unknown" by cm_Analyze.
3474 * In response to VNOVNODE, the parent is only discarded if the current
3475 cm_scache_t is not a directory.
3477 * In the Ioctl FlushFile and FlushVolume functions, there were no
3478 checks to protect against flushing the Freelance SCache entries.
3480 * In FlushFile, the wrong cm_scache_t object was being released.
3482 * In cm_GetNewSCache, do not allow Freelance SCache entries to be
3483 recycled. Choose a new entry if cm_RecycleSCache fails.
3485 * Speed up the performance of the cache manager by not
3486 holding the smb_fid_t mutex across calls to cm_SyncOp
3489 * Ensure that all smb_fid_t flag references are protected.
3491 * Remove a deadlock between smb_fid_t mutex and
3492 smb_rctLock introduced in 1.4.1.
3494 * Fix aklog AFSID lookup to use not send the realm
3495 name when the realm is the local realm for the
3496 cell. The PT Server doesn't strip the realm
3497 name when it is the local realm.
3499 * Treat "\\afs\*." as an alias for "\\afs\all"
3501 * Dynamically adjust the priority of server threads
3502 based upon the age of the CIFS request that is
3503 being processed. The priority is bumped one level
3504 for every 15 seconds of processing time.
3506 * The original openafs contribution mis-used the
3507 cm_GetCallback function. This function should only
3508 be called as a side effect of a call to
3509 cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
3510 was being called results in a FetchStatus RPC being
3511 performed even when the client already has a valid
3512 callback and does so in a manner that destroys the
3513 synchronization of the threads that are calling
3514 cm_SyncOp. If nothing else this results in a
3515 significant performance penalty.
3517 * More changes to cleanup of smb_vc_t objects in
3518 order to prevent race conditions.
3520 * Reduce the amount of Power Management event logging
3523 * If the client detects that the IP address has
3524 changed, force the use of new RX connections. This
3525 enables the file server to recognize that the client
3528 * Remove a race condition and deadlock associated
3529 with cleanup of smb_vc_t objects.
3531 * Fix the ANSI filename support. One function that
3532 required it was missing the translations.
3534 Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
3537 * Add a DllMain() function to afspthread.dll. Perform
3538 cleanup of thread local storage, rx queues, critical
3539 sections, event handles, etc. This was implemented
3540 to avoid crashes in afspthread.dll when applications
3541 which load/unload the dll attempt to terminate.
3543 * Lock request behavior has been altered. As of this
3544 release a locally allocated lock will be used whenever
3545 the volume is RO or the requesting user's permissions
3546 are 'rl'. This will allow end users to execute programs
3547 from RW volumes when they do not have the 'k' privilege.
3549 Note that a recent discovery indicates that the 'k' bit
3550 is only supposed to control the ability to obtain read
3551 locks. Write locks are to be controlled by the 'w' and
3554 * Code has been added to enable the cache manager to be
3555 able to obtain the AFSID of the authenticated user.
3556 This is required to be able to determine when "Insert"
3557 privileges should apply. AFS considers a file to be
3558 a new file whenever the owner's AFSID matches the
3559 AFSID of the user. The current implementation relies
3560 on the name specified during the SetToken call to be
3561 correct. This will be replaced in the future with a
3562 new RPC that allows the AFSID of the authenticated
3563 user to be obtained.
3565 * A 32-bit-tools installer for use on 64-bit systems
3566 is now available. These tools are for use with 32-bit
3567 versions of Kerberos for Windows or JPSoftware's 4NT.
3569 Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
3571 * Remove the use of the "active_vcp" variable. This variable
3572 no longer serves any useful purpose and was the cause of
3573 race conditions that produced under counts of the smb_vc_t
3576 * Change default minidump output to include data segments
3578 * restore access to cm_IoctlGetVolumeStatus and
3579 cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
3580 Still need to figure out why they are being called
3581 for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
3583 Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
3585 * Update build system to include the "sys_name" in the
3588 * optimize the number of calls to obtain the smb_user_t for a
3589 smb virtual circuit.
3591 * add logging of server addresses, status and capabilities when
3594 * prevent multiple threads from pinging the same server simultaneously.
3595 instead, threads block and then use the result of the ping that was
3596 already taking place.
3598 * Add registry values to allow the defaults for daemon checks to be altered
3599 at startup. Log values to afsd_init.log.
3601 daemonCheckDownInterval
3602 daemonCheckUpInterval
3603 daemonCheckVolInterval
3604 daemonCheckCBInterval
3605 daemonCheckLockInterval
3606 daemonCheckTokenInterval
3608 * Add EventLog messages for transitions in fs crypt state.
3610 * Byte range locking (when locks will be obtained from the
3611 is still under discussion)
3613 * 64-bit Windows operating systems now supported
3614 (XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
3615 This is a native 64-bit implementation of OpenAFS for Windows.
3616 There is currently no 64-bit version of KFW and therefore
3617 there is no (easy) method for obtaining AFS credentials.
3618 32-bit KFW 3.0 can be used with 32-bit OAFW libraries
3619 to obtain AFS tokens but there is not for the feint of
3622 * A faster implementation of the fcrypt algorithms improves
3625 * The afsd_service.exe now supplies Event Log Message Text.
3626 The event log no longer generates warning text due to lack
3627 of message configuration.
3629 * When afs servers are unreachable and network stack returns
3630 a WSAEHOSTUNREACHABLE error, the AFS client will immediately
3631 mark the attempted server as being down. This allows for
3632 very fast response to the calling application. This change
3633 has a pronounced improvement when the Windows Explorer is
3634 used on disconnected mobile computers.
3636 * Added DebugOutputString debugging of the RX library
3638 Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
3640 * Fix the rx-lwp implementation to randomly generate rx call
3641 identifiers which are supposed to be unique at a given point
3642 in time. This affects some of the command line tools used
3643 for debugging but not the AFS Client Service itself which
3644 uses the rx-pthreads implementation.
3646 * xstats_fs_test.exe and xstats_cm_test.exe are now built for
3647 Windows although they are not distributed
3649 * don't generate an exception when logging unexpected SMB packets
3651 * the data structures used by cm_IoctlGetVolumeStatus and
3652 cm_IoctlSetVolumeStatus are not compatible with the CIFS
3653 Get/Set File Information calls so don't bother executing them.
3657 * Fix data written to registry as part of BackConnectionHostnames
3660 * Fix initialization bug when allocating cm_server_t objects that
3661 would cause servers to appear to be down after each allocation
3662 or renewal. Server would appear up after the next check servers
3667 * Windows will return WSAECONNRESET to the next receive packet
3668 operation on a socket if an ICMP packet is received in response
3669 to a packet that was sent and could not be delivered because
3670 the port was unreachable. We have to treat this as
3671 WSAEWOULDBLOCK so that we can retry the select() check and
3672 not mark the connection as bad.
3674 Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
3676 * The client manager maintains a least recently used list of
3677 all the ACLs. If an ACL is already the most recently used
3678 do not bother to remove it from the list and place it on the
3681 * As part of the process of adding a new server, set the
3682 server is down flag and then perform a ping. If the server
3683 is up, the flag will be reset. If the server is down, the
3684 client will not timeout unnecessarily while attempting to
3687 * If a Mountpoint String is the empty string, return PATH NOT
3690 * Modify behavior of the Network Provider to avoid querying the
3691 profile location if integrated logon is known to be disabled.
3693 * Correct behavior of 'LogoffPreserveTokens'. When set, tokens
3694 will always be preserved regardless of whether or not integrated
3695 logon was used to obtain tokens. When unset, tokens will be
3696 destroyed regardless of whether integrated logon was used to
3697 obtain the tokens unless the profile is stored in AFS.
3699 * Disable LogoffTransferToken functionality. The SMB V3 Logoff
3700 message is being delivered even when the user is not logging
3701 off of the computer.
3703 * Fix a deadlock condition introduced in rc6.
3705 * Cell names are case-insensitive. Since the founding of OpenAFS
3706 there has been a bug that treated "cell", "Cell", and "CELL" as
3707 different cells. This would result in what appeared to be
3708 random disappearance of user tokens or a failure to authenticate
3709 even when tokens were listed by the "tokens" command.
3711 Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
3713 * Additional auditing of the smb code revealed that smb ioctl
3714 operations were failing to free smb_fid_t objects
3716 * The File Identifiers assigned to smb_fid_t objects could be
3717 given the value 65535 which is 0xFFFF which is also the
3718 INVALID_HANDLE value. Both 0 and 65535 are now treated as
3721 * Audited all assignments to fields and ensured that they are
3722 protected by obtaining the relevant mutex object.
3724 * Re-wrote the VNOVNODE handling to avoid a potential deadlock
3725 if both the current cm_scache_t object and its parent are
3726 both returning VNOVNODE.
3728 * Removed all the logic associated with the 'dead_vcp' variable
3729 which must have been added for debugging because its used
3730 was never thread safe and it truly served no purpose.
3732 Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
3734 * Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
3735 registry settings. Default is ON and 120 seconds. If set to
3736 OFF, tokens will not be destroyed.
3738 * Fixed a number of race conditions and reference count problems
3739 associated with global queues.
3741 * Added a method to detect abandoned smb virtual circuits so their
3742 associated file handles, locks, user tokens, and smb session
3743 resources can be released.
3745 * Added a mechanism to allow user tokens to be preserved
3746 during integrated login and be freed after logoff.
3748 * Check access permissions to the file before the file is
3751 * More improvements to the handling of SMB Virtual Circuits
3752 when SMB sessions are closed.
3754 * Correct a race condition introduced in rc4 that could result
3755 in a crash of afsd_service.exe.
3757 Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
3759 * Properly handle pre-mature SMB session closures.
3760 File handles, tree connections, and user sessions are now
3761 freed when the virtual circuit is closed.
3763 * smb_username_t and smb_user_t objects are now properly reference
3764 counted and freed as appropriate.
3766 * disable to use of AFS file locks. byte range locking is still
3767 implemented but all locks are managed locally. This prevents
3768 multiple processes from editing the same file on the same
3769 machine but it does prevent multiple processes on separate
3770 machines from modifying the file.
3772 Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
3774 * A fix to "fs setserverprefs -vlserver". Multiple calls with
3775 the same server parameter would result in a panic condition
3776 if the server had not already been contacted by the AFS
3779 * Better warnings are written to the afsd.log file when
3780 the Windows RPC Protocols are not properly configured as
3781 is often the case on Windows 2000.
3783 * fix the procmgmt library to restore C RTL signal handlers when
3784 it is being detached from the process prior to process termination.
3786 Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
3788 * close a lock leak in the rx library that could be triggered while
3791 * prevent removal of drive mappings in response to afscreds.exe -M
3793 * add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
3794 strings and flush the stat cache entry.
3796 * prevent server objects from being freed if user preferences are set.
3798 * Update WIX support to 2.0.2419.0
3800 * Release all locks on the server at service shutdown.
3802 * A down server should be marked up as long as it responds to the client.
3803 It doesn't matter if the server returned an error as long as the error
3804 is something other than VRESTARTING. If there is such an error, all
3805 connections on that server should be marked for replacement.
3807 * Change the check "up" servers period from once per hour to once every
3808 ten minutes. This matches the behavior of the Unix cache manager.
3810 Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
3811 * Byte Range Locking support has been added. Architectural documentation
3812 is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
3813 made to enforce Windows mandatory locking model using locally managed
3814 byte range locks with full file locks obtained from the file server.
3815 Read-only volumes provide the user with an implied locking privilege.
3816 Read-write volumes require that a user be granted the locking privilege
3817 in order for a lock to be obtained. As most Windows applications use
3818 Shared file access modes, they require the ability to obtain locks in
3819 order to access files.
3821 * Implement error translation functions to convert Universal Error Codes
3822 output by modern servers to local errno and winsock error values.
3823 This fixes the error message problems using pioctl() calls from fs.exe
3824 and the AFS Shell Extension.
3826 Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
3827 * Allow the AFS Client Service to set the firewall configuration by
3828 retrying until after Windows determines the boot period is over.
3830 * Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
3833 * Detect IP address changes from within the AFS Client Service for the
3834 purpose of checking down servers sooner
3836 * If the file server or vldb server report RXKADUNKNOWNKEY, return
3837 SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
3838 user that there is a Kerberos key problem. Previously the user would
3839 receive no feedback when there was something wrong.
3841 * Fix the large integer library support to properly handle values greater
3844 Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
3845 * Updated CellServDB
3847 * Fixed file name pattern matching. Question marks can match nothing
3848 if found at the end of a component.
3850 Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
3851 * Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
3853 * The MSI installer now has a new Product ID for each build. This
3854 is to allow for automated upgrades. The Upgrade ID is remaining
3857 * Volumes that were cached and then subsequently were salvaged,
3858 moved or otherwise taken temporarily offline (as opposed to being
3859 busy) would become unreachable. A distinction is now made between
3860 a volume being offline and a server being down.
3862 Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
3864 * Fixed several AFS Shell Extension issues:
3865 - Checking Server Status would only work if a specific cell was
3867 - If servers were down, the Server Down dialog would write to
3868 invalid memory and cause explorer.exe to crash
3869 - Default Push Buttons were not specified for many dialogs
3870 - The extension would be unloaded by the Explorer Shell and
3871 never be reloaded under some circumstances.
3873 * Fixed the AFS System Tray menu to disappear if another Window
3874 was selected before a menu item was selected
3876 * Added volume owner and group information to "fs examine" output
3878 * Added fs and registry support for enabling/disabling/clearing
3879 rx statistics gathering.
3881 Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
3883 * New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
3884 GUIDs for each platform.
3886 * Documentation directory structure and default installation rules
3889 * New OpenAFS for Windows HTML Release Notes installed
3890 as part of the documentation set.
3892 * Add thread id to rx debugging messages
3894 * When all servers are offline, return a bad network path error
3895 immediately. This is necessary to allow Explorer to do the
3896 right thing and not hang. The background daemon thread will
3897 check the down servers every 30 seconds to see if they are
3900 Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
3902 * The AFSCache file has been modified to store both the
3903 serial number of the volume on which the file resides
3904 and the SID of the local machine. If either change,
3905 a new UUID is generated.
3907 * add "fs flushall" command that forces all data buffers
3910 Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
3912 * add support for '/' instead of '\' in pioctl() calls
3914 * Apply AFS Client Admins group protect to AFS Shell Extension
3916 * Add support for \\afs\<name> to most AFS Shell Extension
3917 commands except the symlink methods.
3919 * when installing the Microsoft Loopback Adapter, enable
3920 MS Client for Networks and prevent an install failure
3921 by not calling CoInitialize twice in the same thread.
3923 * reload cell vldb values from the CellServDB every two
3924 hours in case it changes
3926 * When updating cell information from DNS, be sure to set
3929 * Add support to allow use of \\AFS\<foo> where <foo> is
3930 either a mount point or symlink. As <foo> is normally
3931 treated as a share name, we transform it into \\AFS\all\<foo>
3934 * Init 'code' to prevent false errors when integrated logon
3935 is disabled and the service is not running
3937 Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
3939 * Removed trace log messages that were hampering performance.
3941 * Fixed a deadlock that was being triggered by editing Word
3942 documents stored within AFS with WinWord 2003.
3944 * Bit 3 of the TraceOption registry value is now used to set the
3945 default for "fs trace" in the non-Debug builds. The new default
3946 is off for release builds and on for debug builds.
3948 * Bit 2 of the TraceOption registry value can be used to turn on
3949 real-time output of debug log entries to the Windows Debug
3950 Output monitor interface. This data can be viewed with tools
3951 such as Sysinternal's DbgView.
3953 Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
3954 * When the cache manager reports ALLBUSY or ALLOFFLINE for
3955 a resource we should not be returning status codes such
3956 as Network Busy or Remote Host Down because these errors
3957 will imply that the AFS Client Service is Busy or Down
3958 and that is simply not the case. Instead we will return
3959 Bad Network Path as the path is temporarily not available.
3961 Instead of returning Sharing Paused when there is a
3962 timeout error, return Timeout. Once again, the AFS Client
3963 Service is not paused.
3965 * afscreds.exe would display an Obtain Creds dialog when
3966 the expired credentials reminder was triggered even
3967 if there was no network path to the KDC. This is prevented
3968 by adding KDC probe logic to the reminder thread.
3970 * afscreds.exe would display expired tokens no differently
3971 than unexpired ones. This would make it difficult for the
3972 user to distiguish when the tokens were expired. For the
3973 English build added a new resource string "(expired) that
3974 is displayed instead of the expiration time.
3977 * "fs wscell" when executed in freelance mode will return the
3978 name of the cell configured in the registry. The root.afs
3979 volume is not loaded from this cell, but it is used for the
3980 default for aklog and integrated login.
3982 * "fs mkmount, fs rmmount, symlink make, symlink remove" will
3983 no longer work on \\AFS\all when freelance is being used
3984 unless the user is a member of the "AFS Client Admins" Windows
3987 * some more checks performed during persistent cache validation.
3990 * Optimize calls to cm_CheckServer from cm_Analyze. Only check
3991 down servers for the current cell not all cells.
3993 * If the scache LRU list becomes corrupted, the AFS Client Service
3994 might enter an infinite loop at startup. Detect the corruption
3995 and discard the cache contents in this case.
3997 * Fixed the Explorer Shell's remove mountpoint functionality.
3999 * Fixed a deadlock caused by the holder of cm_aclLock attempting to
4000 obtain a mutex lock on a cm_scache_t object whose lock is held by
4001 a thread attempting to obtain the cm_aclLock.
4004 * Really fixed DNS AFSDB queries to ensure that "csail" cannot be
4005 misinterpretted as "csail.mit.edu" when the resolver
4006 is configured to append ".mit.edu" to failed DNS queries.
4008 * Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
4009 that can be used to force the preservation of user tokens upon logout.
4011 * Update the NSIS install scripts to use NSIS 2.07. This release adds
4012 recommended installation categories: AFS Client, AFS Administrator,
4013 AFS Server, AFS Development Kit. Each category includes a different
4014 default subset of the OpenAFS components.
4016 The OpenAFS logo is now associated with the NSIS uninstall entry
4017 in the Add/Remove Programs control panel.
4019 * The user name associated with AFS tokens when obtained with
4020 integrated login, the afs systray tool, or aklog will always
4021 include the full Kerberos 5 user principal name regardless of
4022 whether or not the cell is local to the realm.
4024 * Modify integrated login so that it does not enter an infinite
4025 loop if the service is not set to auto start.
4027 * Added asetkey.exe used to set a Kerberos 5 key for use by the
4030 * Added uninstall.exe to Wix installation
4032 * More modifications to algorithms used to wake sleeping threads.
4036 * Changes to the algorithms used to wake threads when they are
4037 put to sleep because of conflicts over buffer operations.
4040 * The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
4041 the process of producing AFS tokens. The tokens are stored within
4042 the AFS Client Service but the Kerberos 5 tickets have been discarded.
4043 New functionality has been added to temporarily copy the tickets to
4044 a file ccache during the short window between the NPLogonNotify and
4045 WinLogon Logon Event. The Logon Event handler starts a process as
4046 user, afscpcc.exe, that copies the tickets into the default ccache.
4047 The file ccache is then destroyed by the event handler.
4049 * During synchronization operations on stat cache entries, if threads
4050 are forced to sleep it was possible that the threads would not be
4051 woken under heavy load.
4053 * Enforce the delete on close flag specified by the NTCreateX()
4056 * A race condition in the rx library was fixed that resulted in
4057 a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
4058 was cleared when threads were still waiting to use the
4059 connection. Also, fixed a potential case in which the
4060 connection could be destroyed while threads were waiting to use it.
4062 * The DNLC freelist has been seen to become corrupted with still
4063 active entries being stored on the freelist. Changes were made
4064 to perform a better job of cleaning entries before placing them
4065 onto the freelist; marking them with the correct magic value when
4066 purging the list; and allowing the list to be purged on startup
4067 if the cache validation fails.
4069 * Windows Crash Reporting does not get triggered for afsd_service.exe
4070 because it provides its own top level exception handler. This patch
4071 places an explicit call to ReportCrash() on platforms which support
4074 If you configure Crash Reporting via AD Policy to capture crash
4075 reports within the domain, then you will receive this output.
4077 * Generate MiniDumps independent of the Windows Crash Reporting.
4079 (a) If an exception occurs,a minidump will be generated at the path
4080 %TEMP%\afsd.dmp. The type of minidump defaults to Normal but
4081 can be set to any other type via use of the
4084 [HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
4086 (b) MiniDumps can be generated at any time using the "fs minidump"
4087 command. This allows you to generate a minidump without
4088 stopping the process or even requiring the use of a debugger
4091 * an RX connection which reports bad ticket now treats it the same way
4092 the expired ticket is treated. The ticket is removed from the
4093 connection and a new anonymous connection is established.
4095 * fs memdump now requires that the caller be in the AFS Client Admins
4098 * additional debugging information is provided in cm_Analyze when the
4099 error code is ignored.
4101 * Fix a regression introduced into the pioctl() client code which
4102 will cause a crash in winlogon.exe, explorer.exe and fs.exe if
4103 MIT KFW is not installed. (1.3.82a)
4105 * Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
4108 * Fix a race condition in cm_EndCallbackGrantingCall() that
4109 could leave 'nrevp' pointing to freed memory. In the process,
4110 optimize the processing to ensure that the applications monitoring
4111 the status of the stat cache entry are only notified at most
4114 * Conditionalize access to the Explorer Shell AFS->Mount Point
4115 ->Remove menu item. Disable the item if the selected item(s)
4116 are not mount points.
4118 * Activate AFS RPC (RX) Free Packet Queues stored in Thread
4119 Local Storage. This eliminates thread contention on the
4120 global RX free packet queue. This should improve performance
4121 on multi-CPU systems.
4123 * Fix the IP address filtering code to properly load/unload
4124 the IPHelper DLL so that Windows XP/2003 and beyond systems
4125 do not have to manually probe the registry.
4127 * CleanupACLEnt() was not being called consistently with
4128 the cm_scache_t object referred to by the back pointer
4129 mutex locked. This could in very rare conditions lead to
4130 an invalid memory access.
4132 * Added a script command to the msi installer "afs_replace"
4133 which can be used to replace one version of OpenAFS with
4134 another without requiring an uninstall and reinstall if
4135 the installed msi is no longer in its original location.
4137 * Apparently the problem with multi-domain forests with cross-
4138 realm trusts to non-Windows realms was not entirely solved.
4139 The authentication to the AFS SMB service failed because
4140 the wrong name was being used. Using ASU as an example,
4141 the authentication was being performed with the name
4142 "QAAD\user" (an account in the forest root) and not
4143 "user@ASU.EDU (the MIT Kerberos principal used to login with)
4145 The solution was to add an additional dependency on KFW
4146 in order or to be able to easily obtain the client principal
4147 name stored in the MSLSA ccache TGT. This information is
4148 used in two locations:
4150 - the pioctl() function
4152 - a new WinLogon Event Handler for the "logon" event.
4154 The pioctl function will now be able to use the correct
4155 name when calling WNetAddConnection2() and the "logon"
4156 event handler will now be able to call WNetAddConnection2().
4157 The hope is that the "logon" event handler will be called
4158 before the profile is loaded but I have not guarrantee
4162 * Fixed a locking error in cm_TryBulkStat() which had the
4163 potential of deadlocking the system for the length of time
4164 it takes to perform a bulk status fetch operation.
4166 * Replaced time conversion code (UnixTime <-> FILETIME) to
4167 be completely arithmetic instead of relying on a bizarre
4168 algorithm involving a variety of C RTL time functions.
4169 This has the side effect that UnixTime and FILETIME which
4170 are both stored in UTC are interpretted as UTC throughout
4171 the year. Windows will apply the same localization to AFS
4172 as it does to NTFS. Applications which rely on the ability
4173 to sync files between the two file systems will no longer
4174 see the timestamps of files in AFS change an hour relative
4175 to the files stored in NTFS or Windows based backup devices.
4177 * Fixed a invalid memory access under a bizarre circumstance.
4178 Windows will allow a physical mass media device to be
4179 installed via Plug N Play to the system and will assign it
4180 the lowest available drive letter. This is true even if the
4181 drive letter is currently assigned in the user session to
4182 a network device via NET USE (or its equivalent.) When this
4183 happens, queries sent to the CIFS server will contain invalid
4184 data. This invalid data was not being caught by the AFS
4185 Client Service and was resulting in a crash when Freelance
4188 * Fixed a reference count error when registering callbacks
4189 on a stat cache object if the callback was already registered
4191 * Add a case to cm_Analyze() in order to handle the case where
4192 cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
4193 the server list for the volume is empty. In this case, force
4194 an update of the volume info and retry.
4196 * Insert a missing cm_EndCallbackGrantingCall() which could
4197 result in threads waiting for a callback to complete to never
4200 * In the persistent cache, there is a maximum number of volume
4201 entries. Allow volume entries to be reused if the maximum
4202 number have been allocated and their reference count is zero.
4204 * If we already have a dead virtual connection object, don't
4205 ignore the fact that additional dead vc objects must be taken
4208 * Removed a deadlock condition introduced in the summer of 2004.
4209 Do not hold mx locks around calls to RX functions including
4210 rx_NewCall(). That is what reference counts on the rx_connection
4213 * Fixed an initialization error in afslogon.dll which could
4214 result in random behavior including a failure to terminate
4215 the NPLogon function. (1.3.80b)
4217 * Fixed an error preventing the use of SMB authentication
4218 on Windows 2000. (1.3.80a)
4221 * Updated CellServDB to Public CellServDB 16 Mar 2005
4223 * Fixed DNS AFSDB queries to ensure that "csail" cannot be
4224 misinterpretted as "csail.mit.edu" when the resolver
4225 is configured to append ".mit.edu" to failed DNS queries.
4227 * Fixed another case in which the client would replace connections
4228 to the host when it was not necessary. In this case the
4229 situation would occur if cryptall was on and the connections
4230 were unauthenticated due to lack of an appropriate token.
4232 * OpenAFS for Windows has failed to work at sites which are
4233 utilizing a cross-realm trust between an MIT/Heimdal realm
4234 and a multi-domain Windows forest when the workstation being
4235 accessed is not located in the root domain. This is caused
4236 by a bug in the workstation which was triggered after the
4237 introduction of Windows 2003 Server. When the bug is triggered,
4238 the workstation attempts to authenticate users to afsd_service.exe
4239 by contacting the Domain Controller instead of using the
4240 LSA loopback authentication mechanism.
4242 One of the reasons this bug occurs is because the workstation
4243 does not have a reliable way of knowing that the service whose
4244 netbios name is "AFS" is located on the workstation. This will
4245 be fixed starting in Longhorn Beta 1 by Microsoft. The
4246 "BackConnectionHostNames" registry value will be used to
4247 indicate that the authentications to that service name should
4248 be performed using the loopback authentication mechanism.
4250 In the meantime, when Logon Caching is enabled, we can force
4251 afsd_service.exe to authenticate using the logon cache before
4252 contacting the Domain Controller. This will work with both
4253 password and smart card based logons.
4255 * The allDown logic in cm_ConnByMServers() was wrong. The allDown
4256 flag should not be cleared if a volume's server reference is
4257 marked as "offline". In the case where all of the volume's
4258 servers are either "down" or the volumes are "offline", we want
4259 cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
4260 instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
4261 should never occur in practice.
4263 In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
4264 5 seconds, clear the server down and volume busy flags, and
4265 then force an update from the VLDB. This allows the client to
4266 update the location of a volume if the reason for it being
4267 marked offline is because it is being moved. Calls to
4268 cm_ConnByMServers() will be retried either until success or
4269 the RDRtimeout period is reached.
4271 * Correct the Power Management code behavior in response to
4272 Standby, Suspend, and Shutdown events. Instead of flushing the
4273 buffers associated with the mounted SMB submounts, simply write
4274 all buffers which are marked dirty.
4276 * Added support for "TheseCells" to afslogon.dll. "TheseCells"
4277 provides a list of cells other than the default cell for which
4278 tokens should be obtained using the default Kerberos principal.
4279 This functionality is logon domain specific and is only available
4280 when using KFW for authentication.
4282 * Fixed FindNext Invalid Handle error caused by over agressive
4283 attempts at garbage collecting dirSearch entries when the
4284 dirSearch ID wraps from the maximum value to 1.
4286 * Add support for registry defined server preferences for VLDB and
4287 File servers. See registry.txt for details.
4289 * Increased default cache size to 96MB and default number of cache
4292 * Fixed refCount leaks related to directory lookups and pioctl calls.
4294 * Callbacks revoked during a race condition with an attempt to
4295 obtain the same callback no longer result in an inaccessible
4298 * New command line tool:
4300 afsdacl : Set or reset the DACL to allow starting or stopping
4301 the afsd service by any ordinary user.
4303 Usage : afsdacl [-set | -reset] [-show]
4304 -set : Sets the DACL
4305 -reset : Reset the DACL
4306 -show : Show current DACL (SDSF)
4308 * IP addresses are no longer queried once at startup. Instead IP
4309 addresses are obtained as needed. Loopback adapter addresses are
4310 no longer published to the server.
4312 * Pay attention to the MIT KFW registry configuration for automatic
4313 importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
4315 * Fix (once again) case-sensitive comparisons which was apparently
4318 * Activate support for vos listvol -format
4320 * Population of the "AFS Client Admins" group failed on non-English
4321 versions of Windows because the name of the "Administrators" group
4322 is localized. Now we lookup the name of the group by using the
4323 Administrators Group Alias Relative ID.
4325 * Multiple cell token acquisition within afscreds.exe was broken in the
4326 case of Kerberos 5 cross realm authentication. Instead of contacting
4327 the KDC associated with the cell's realm, afscreds would obtain a
4328 token from the KDC of the user's realm. This would result in
4329 "invalid kvno errors" while authenticating to the AFS servers.
4330 Unauthenticated access would work. The symptoms would vary based upon
4331 whether or not the VLDB servers had been contacted using unauthenticated
4332 connections prior to the user obtaining tokens.
4334 * The list of ACL entries was becoming corrupted. It appears as if
4335 an ACL when it expires was not being appropriately cleaned up. In
4336 fact, it was left in the list of ACLs associated with the scache
4339 * Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
4340 for 32-bit x86 systems. The default for itanium will be "ia64_win64"
4341 and "amd64_win64" for amd 64-bit processors.
4343 * When executing executables, Windows provides a mechanisms by which
4344 the normal search for DLLs can be bypassed. If a file foo.exe is
4345 being executed and there exists a file or directory "foo.exe.local"
4346 (the contents are ignored if a file), then Windows will search for
4347 DLLs first in the "foo.exe.local" directory and second in the
4348 directory in which the "foo.exe" file is located.
4350 Previous releases of OAFW would improperly return
4351 STATUS_NOT_A_DIRECTORY instead of the expected
4352 STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
4353 to terminate the search for the DLL causing the execution of the
4354 application to fail.
4356 * Rework the reference counting of smb_vc_t objects. The references
4357 stored in the waiting locks were not counted. This could result in
4358 an assertion if the reference count drops to zero.
4360 * "fs wscell" returns "Freelance.Local.Root" for the workstation cell
4361 when running in freelance mode.
4363 * Added support for persistent caching of file contents, stat entries,
4364 volume data, cell data, and name lookup data. When the registry
4365 value, "NonPersistentCaching", is not defined or set to 0 the file
4366 specified by "CachePath" becomes a persistent cache file. The
4367 size of the cache file is computed at run time. It has a maximum
4368 size of approximately 1.2GB. The contents of the file will be
4369 validated according to the rules specified by the "ValidateCache"
4372 * Added support for UUIDs. UUIDs are kept for as long as the
4373 cache file is intact.
4375 * Added cmdebug.exe and support cache manager debugging callback
4376 interfaces in afsd_service.exe.
4378 Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
4379 [-addrs] [-cache] [-help]
4380 Where: -long print all info
4381 -addrs print only host interfaces
4382 -cache print only cache configuration
4384 * Symlinks to \\AFS\[all\]... will now be treated the same as
4385 symlinks to /afs/.... However, please use /afs/... as otherwise
4386 the symlinks won't work on Unix.
4388 * Correct a problem with local CellID allocation for cells whose
4389 server list is obtained via DNS instead of a CellServDB file. If
4390 the DNS information expires the CellID assigned to the entry will
4391 be changed. This causes all of the cm_scache_t objects which refer
4392 to the old cellID number to become useless. Attempts to access
4393 files or directories with cm_scache_t objects using the old cellID
4394 will fail since the server list cannot be obtained.
4396 * Correct deadlock condition in cm_EndCallbackGrantingCall which can
4397 be triggered if the volume referenced by the cm_scache_t object is
4400 * The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
4401 in addition to SERVICE_ACCEPT_STOP.
4403 * Move RPC shutdown until after the SMB and RX shutdown procedures
4404 complete. Block until RPC shutdown is complete.
4406 * Modify afslogon.dll (integrated logon) to wait for service if its
4407 state is START_PENDING. If the timeout period occurs, reset to
4408 the retry interval and not the sleep interval.
4410 * When renewing the server list for a cell obtained via DNS AFSDB
4411 records, the cm_cell_t entry must be removed from the list of all
4412 cells. Otherwise, the list of cells will be corrupted.
4414 * In the dcache and scache modules, use the cm_scache_t dataVersion
4415 instead of the cm_fakeDirVersion.
4417 * Synchronize fs.c with the unix version.
4419 * The variable used to determine whether a file or virtual memory
4420 mapped cache is used was not properly initialized to a default
4421 value. If the registry setting "NonPersistentCaching" was not
4422 set, the choice would be random. Properly initialized to be
4425 * The memory mapped view was never unmapped before closing the file
4426 at service shutdown. This is now properly cleaned up.
4428 * Default location of Cache file is now %TEMP%\AFSCache
4432 * A bug affecting new installations of 1.3.75/76 would result in
4433 the creating of incorrect mountpoints in the freelance root.afs
4434 volume for the default cell.
4436 If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
4437 of "root.cell", you have been affected by the bug. To correct the
4438 problem, execute the following commands:
4440 fs rmmount \\afs\all\<cellname>
4441 fs rmmount \\afs\all\.<cellname>
4443 fs mkmount \\afs\all\<cellname> root.cell <cellname>
4444 fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
4448 * A bug has been fixed in the auto-generation of Freelance root.afs
4449 symlinks which produced random entries in the \\AFS\all directory.
4451 * Support has been added for multi-homed servers
4454 * Added a new registry value, "StoreAnsiFilenames", which can be used
4455 to force the use of ANSI character sets instead of OEM Code Pages.
4456 This feature is useful when users require the ability to create
4457 filenames with 8-bit characters and need to access the files from
4458 both Latin-1 based Unix systems as well as from Windows.
4460 Activation of this feature will prevent access to files stored with
4461 8-bit OEM characters.
4463 * Shutdown all SMB threads in a synchronized manner when stopping the
4466 * There is currently a maximum cache size of 1.3GB. The limit is imposed
4467 by the largest contiguous block of unused memory within the 2GB process
4468 space which can be assigned to the memory mapped file. Unfortunately,
4469 when the executable digital signature verification code is activated
4470 Windows sees fit to further segment the process memory which in turn
4471 reduces the size of the maximum cache file to less then 800MB. If
4472 larger cache sizes are desired, a new registry value should be set:
4474 HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
4476 Setting this value will disable the runtime verification of digital
4477 signatures on afsd_service.exe and the afs dlls which it loads. It
4478 will not disable the the version number check on those same files.
4479 The signature verification is not a security messure and is only meant
4480 to enhance the ability to afsd_service.exe to detect potential
4481 destablizing mixtures of DLLs from incompatible distributions.
4483 Added code to auto-disable the signature verification check if
4484 the desired cache size is greater then 700MB.
4486 * Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
4487 used to verify the validity of the afsd_service.exe binary
4488 as well as each of the AFS DLLs loaded by the service. Not only
4489 must the digital signature be valid but the signatures of the
4490 DLL must be signed by the same entity as the service.
4492 * Implement new functions: cm_freelanceMountPointExists and
4493 cm_freelanceSymlinkExists. Use them along with other validity
4494 checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
4495 ensure that name collisions do not occur and that empty strings
4496 are not valid file names.
4498 A symlink may not have a name which would resolve to a valid
4499 cell name. Doing so would prevent access to the cell.
4501 * Add missing cm_HoldSCacheNoLock call to Freelance mount point
4502 re-initialization code. The reference counts of the fake root.afs
4503 volume scache object(s) would become invalid when the mount point
4504 or symlink lists were altered.
4506 * Add registry entries to provide mappings from the afsdsbmt.ini
4507 to the new locations for applications which count on the use
4508 of the old Profile file APIs. These apps are likely to fail
4509 if the user does not have administrator privileges and the
4510 registry is locked down.
4512 * The afs_config.exe submounts dialog had two errors.
4513 First, attempts to remove entries failed because the registry
4514 key was being opened without KEY_WRITE privileges.
4515 Second, when editing a submount entry, changing the name
4516 would add a new key and leave the original one in place.
4517 Now the original submount will be removed if its name is
4520 * In recent months there have been several incidents in which
4521 users have experienced problems starting or accessing
4522 afsd_service.exe and after significant effort has been spent
4523 it has turned out that they have two versions of AFS on the
4524 machine or an inconsistent set of DLLs.
4526 Code has now been added to afsd_service.exe which will walk
4527 the list of modules loaded by afsd_service.exe and validate
4528 that the version of the AFS DLLs matches the version of the
4529 afsd_service.exe executable. If they do not match the service
4532 * When Freelance mode is enabled and there is no registry
4533 key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
4534 will attempt to import the afs_freelance.ini file contents.
4535 If the file does not exist, it was creating a dummy file
4536 with a r/o and r/w entry for the default cell and then
4537 importing those values.
4539 This process has been changed. The temporary file is no
4540 longer created. Also, both the OpenAFS Client install
4541 directory as well as %WINDIR% are checked for previous
4542 afs_freelance.ini files.
4544 * Added support for VL_GetEntryByNameN(). Still need to add
4545 support for VL_GetEntryByNameU() for multi-homed support.
4547 * Fix a deadlock situation in afscreds.exe when canceling an
4548 auto-generated Obtain Tokens dialog
4551 * Fix the locking of objects during Directory Searches in the
4552 SMB/CIFS server. The failure to properly lock the reference
4553 counts was resulting in the premature freeing of smb_dirSearch_t
4554 objects while they were still in use by the SMB/CIFS client.
4555 This does not solve the "Invalid Handle" problem.
4557 * Fix Find Cell By Name pioctl call to return a valid cell
4558 name for the Freelance fake root.afs volume.
4559 "Freelance.Local.Root".
4561 * Fix the Explorer Shell Extension Symlinks->Add operation.
4562 The dialog template was missing and the link destination
4563 string was too short.
4565 * Add support for symlinks to Freelance root.afs volume
4566 Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
4567 <number> = "<linkname>:<relative-path>."
4568 Use symlink.exe to create, list, or remove
4570 * Remove the fallback to the use of KFW's KRB4 library when obtaining
4571 tokens. We never obtain KRB4 tickets.
4573 * Fix AFS Client Configuration Control Panel to support new SysName
4576 * Fix a bug in afsd_service.exe which could result in the SysName
4577 not being read from the registry.
4580 * Add code to block the issuance of AFS tokens by aklog.exe or
4581 afscreds.exe when the Kerberos 5 principal name contains a dot.
4583 * Modify the IsAdmin() function to always treat the local SYSTEM
4584 account as an AFS client administrator. Affects fs.exe and
4587 * Modify the internal handling of Quota Exceeded errors
4589 * Upgrade all reference count fields in the Windows cache manager
4590 and the osi library to use unsigned long instead of signed short.
4591 A similar fix has been applied to the afs rpc (rx) library.
4593 * fix the Windows cache manager to prevent it from replacing the
4594 rx_connection object associated with the cm_conn_t object on each
4595 and every operation if "fs crypt" was set. This explains the
4596 dramatic performance difference when crypt is used vs clear.
4597 The problem: 'cryptall', a boolean flag indicating whether or not
4598 "fs crypt" is set, was being compared to the rx_connection
4599 cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
4600 1 != 2 and therefore the rx_connection was always destroyed
4601 and replaced on each and every operation.
4603 Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
4604 It is not safe for the cm_conn_t object to not be locked because
4605 rx_DestroyConnection might be called from another thread if:
4606 - the user's tokens have changed or expired
4607 - the crypt mode has changed
4609 This fix appears to have also taken care of the problems associated
4610 with Overlapped Writes resulting in Delayed Write errors.
4612 * fix NSIS installer's AdminGroup.exe to properly create and
4613 remove groups when given -create or -remove. The string comparison
4616 * fs sysname now accepts a list of sysname values
4618 * added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
4619 DWORD which when set to a non-zero value will cause error message
4620 text to be output to stderr from the pioctl() routine. Useful in
4621 debugging failures of fs.exe, tokens.exe, etc.
4623 * added a test to the power management code to only perform a
4624 flush operation if there is at least one network adapter which
4625 is not a loopback adapter.
4627 * Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
4628 "EnableKFW". This value will not be read if the key
4629 HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
4630 value under that key does not.
4632 * provide mechanisms to force the use of krb524d for Kerberos 5
4633 ticket to AFS token conversion. For afslogon.dll and afscreds.exe
4634 there is a new registry value "Use524" and for aklog.exe a new
4635 command line parameter "-m".
4637 * Fix the pattern matching algorithm to properly match patterns
4640 * smb_ReceiveCoreRename() was factored to produce smb_Rename()
4641 which is used by both the original function and the new
4642 smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
4643 creation of HardLinks in addition to Renaming. smb_Link()
4644 is a new function which creates HardLinks via cm_Link().
4645 cm_Link() is a new vnodeops function which creates links
4648 smb_ReceiveNTRename() does not support the File Copy and
4649 Move Cluster Information operations described in its interface.
4650 ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
4652 * When opening files via symlinks, we should follow the symlinks
4653 until we reach the actual file stat cache entry. The stat cache
4654 entry of the file should then be stored in the FID instead of
4655 stat scache entry of the symlink.
4657 * return bad operation errors for all unimplemented functions
4658 even if we do not know the functions exist.
4660 * Log bad packets and unknown operation packets to the trace log
4662 * Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
4665 * Update list of known CIFS operations to include all those listed
4666 in CIFS-TR-1p00_FINAL.pdf.
4668 * Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
4669 to support the REG_EXPAND_SZ type.
4672 * A new Windows authorization group "AFS Client Admins" is now
4673 created and populated with the members of the "Administrators"
4674 group. The group is used to determine which accounts on the
4675 machine may be used to modify the AFS Client Configuration via
4676 the UI and command line tools. afs_config.exe, fs.exe,
4678 * Modify the WinLogon Logoff Event Handler to query NT4 domain
4679 controllers for the remote profile path if Active Directory
4680 services are not available.
4682 * Fix aklog.exe to not add the AFS ID to the username
4684 * PTS registration of new users to foreign cells has been added to
4687 * The cm_Daemon thread is used to perform checks for
4688 down servers, up servers, volumes, callback expirations,
4689 lock maintenance and token expiration. Due to a gaff in
4690 larger integer division the thread never performed any
4691 work. Instead the current time computation would always
4692 be less then the trigger times. This had an adverse affect
4693 on the client's ability to maintain communication with servers,
4694 keep volumes up to date, and flush user tokens and acls
4695 when they have expired. This was broken when the 1.3 branch
4696 was modified to support VC7 which no longer included
4699 * An initialization problem with the Freelance code was
4700 detected while fixing the callbackRequest. The cm_rootSCachep
4701 object is obtained during afsd_InitDaemons() but the callback
4702 information is incomplete. The callback information will not
4703 be obtained until cm_MergeStatus is called from within
4704 cm_GetCallback. Unfortunately, cm_SyncOp did not properly
4705 test for the conditions under which the callback information
4708 * Reports have been filed indicating that callbacks were
4709 being lost. An examination of the code indicated that the
4710 cm_server_t objects were not being properly reference
4711 counted by the cm_scache_t and cm_callbackRequest_t objects.
4712 In particular, the cm_server_t objects may have been freed
4713 from beneath the cm_conn_t objects.
4715 All of the reference counting is now done via the functions:
4720 this improves the ability to track the referrals.
4722 Each cm_BeginCallbackGranting Call now allocates a reference
4723 to the cm_server_t. The cm_EndCallbackGrantingCall either
4724 frees the reference or transfers it to the cm_scache_t
4725 cbServerp field. These are then appropriately tracked
4726 through the cm_Analyze call.
4728 * Ensure that the dnlc hash table is the same size as the
4729 dir name hash table (as per original author's note).
4730 Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
4731 for compatibility with 64-bit systems.
4733 * fix smb_ApplyV3DirListPatches to properly apply the hidden
4734 attribute to dotfiles when the infoLevel < 0x101 and
4735 cm_SyncOp has failed.
4737 * Fix the Freelance registry initialization code. There
4738 was a possibility that some systems could end up with
4739 garbage in the registry during a clean install.
4742 * file and directory names beginning with "." will now be given the
4743 hidden attribute when the volume access is anonymous. this matches
4744 the behavior when the volume access is via an authenticated user.
4746 * Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
4747 key. When a change occurs mark the root.afs data as invalid and
4748 for it to be reloaded on the next access. This allows administrators
4749 to modify the mount point list without restarting the service.
4751 The freelance client used to provide a fake modification time for
4752 the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
4753 Added code to extract the last modification time of the Freelance
4754 registry key and use that instead. The time now represents the
4755 most recent mount point change.
4757 * PTS registration of new users to foreign cells has been added to
4760 * Additional Cache Control and Credential Manager options have been
4761 added to the WiX installer. See deployment guide for details.
4763 * The CachePath setting is now optionally a REG_EXPAND_SZ type
4765 * The WiX installer has been upgraded. Version 2.0.1927.1 is now
4768 * The loopback installation code may have had a problem updating the
4769 %ETC%\HOSTS file which could have resulted in a premature failure.
4770 Work around code has been added for the case where the file cannot
4773 * The default max chunksize was increased from 15 (32K) to 17 (128K)
4774 because Windows sends 64K blocks when using overlapped writes.
4776 * The default number of server threads was increased from 4 to 25 to
4777 better handle overlapped writes.
4779 * The "AfscredsShortcutParams" registry value was not being properly
4780 loaded by afscreds.exe. Therefore, the default value was always being
4781 used instead of the value set by the installer.
4783 * Windows XP provides downgrade attack detection to prevent an attacker
4784 from being able to force the use of NTLM simply by disrupting
4785 communication with the KDC. This attack cannot exist between the
4786 Windows CIFS client and the AFS Client Service. Therefore, when a
4787 downgrade has been detected the afs pioctl library will force the
4788 establishment of a new CIFS connection using NTLM.
4790 * A locking error was discovered surrounding all references to volume
4791 server lists within the cm_cell.c source file.
4793 * The logged into Windows username was incorrect on Terminal Server
4796 * A new registry value "NonPersistentCaching" was added to the service
4797 parameters key. When set to a non-zero value, the afs cache is stored
4798 in the Windows paging file. There are two limitations to choosing
4800 1. when persistent caching is implemented it won't work with
4801 this flag set since there will be nothing to persist.
4802 2. with this flag set the initial paging allocation cannot be
4803 changed while the service is running
4805 * An initialization bug was discovered in aklog.exe which affected users
4806 who have a domain name for their afs servers which could not be mapped
4810 * afs_config.exe now validates cell names against DNS in addition
4811 to the CellServDB file.
4813 * In order to allow the freelance client to connect to a volume with ID
4814 equal to 1 on the default cell we changed the fake root.afs volume ID
4815 once again. This time we choose 0xFFFFFFFF. In addition, we change
4816 the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
4817 It will now be impossible for a volume ID to match that of another
4818 cell unless the client is connected to 0xFFFFFFFD cells. That should
4819 be enough room for growth.
4821 * Fix "fs mkmount" command to work with UNC paths and when
4822 started from non-AFS drives. It is now possible to create a mount
4823 point in the freelance fake root.afs volume with the command
4825 fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
4829 fs mkmount \\AFS\all\openafs.org root.cell openafs.org
4830 fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
4832 * The algorithm used to re-attempt access to the servers associated with
4833 a volume has been altered to properly address the case in which all
4834 servers have been marked down. The previous algorithm did not reset
4835 the server's down flags so the servers were never actually retried.
4836 This caused a problem with active volumes if the network connectivity
4837 was lost as could be the case with a network cable removal, wireless
4838 drop, or laptop hibernation. With the fix volume access is restored
4839 almost instantenously when network connectivity becomes available.
4841 * Support for SMB/CIFS browsing has been added to the AFS Client Service
4842 SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
4843 listing of AFS submounts and freelance mount points. Support for
4844 NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
4845 significantly enhances the behavior of AFS volumes within the Explorer
4846 Shell. For instance, "AFS" now shows up as server in the Explorer
4847 with each submount or freelance mount point visible as a share.
4848 The right click menu in each folder now works with full functionality
4849 on a consistent basis.
4851 * The network provider can be configured to have different behavior
4852 depending on the domain that the user logs into. These settings are
4853 only relevant when using integrated login. A domain refers to an
4854 Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
4855 local machine (i.e. local account logins). The domain name that is
4856 used for selecting the domain would be the domain that is passed into
4857 the NPLogonNotify function of the network provider. (see registry.txt
4860 * Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
4861 "Authentication Cell" which may be used to specify a default
4862 authentication cell for afscreds.exe which is different from
4863 the default cell for the AFS Client Service daemon.
4865 * Added a Logoff WinLogon Event Notification function to afslogon.dll.
4866 afslogon.dll moved to %WINDIR%\System32\.
4867 New registry entries added to register the dll for Winlogon events.
4869 The logoff event will now force a call to ktc_ForgetAllTokens()
4870 using the context of the user being logged off as long as the
4871 user's profile is not loaded from within AFS. If the profile
4872 was loaded from AFS we can't release the tokens since the Logoff
4873 event is triggered prior to the profile being written back to
4874 the its source location. This is now performed in an XP SP2
4877 * Windows XP SP2 Internet Connection Firewall interoperability
4880 * The %WINDIR%\afsdsbmt.ini contains four sections:
4881 Submounts, Drive Mappings, Active Maps and CSC Policies.
4882 The Submounts and CSC policies are now stored in the registry under
4883 [HKLM\SOFTWARE\OpenAFS\Client\Submounts]
4884 [HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
4885 The Drive Mappings and Active Maps are stored in the registry under
4886 [HKCU\SOFTWARE\OpenAFS\Client\Mappings]
4887 [HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
4889 There is no automatic migration of this data as it would be impossible
4890 to consistently migrate data to user profiles which may not be active
4891 when the machine is updated.
4893 * The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
4894 fake root.afs volume. For the same reasons as for the cellservdb file,
4895 this information should not be in %WINDIR%. This information is now
4896 kept under the registry key
4897 [HKLM\SOFTWARE\OpenAFS\Client\Freelance]
4899 The data from the afs_freelance.ini file will be automatically
4900 migrated to the registry on first execution of afsd_service.exe
4902 * Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
4903 troublesome for several reasons. One, it is confusing for those who
4904 expect the file to be named "CellServDB" instead of "afsdcell.ini".
4905 Two, this file is not a Windows Profile formatted file. Three,
4906 applications should not be reading or writing to %WINDIR%. It causes
4907 problems for Windows Terminal Server.
4909 The new location of CellServDB will be the OpenAFS Client install
4910 directory which is by default C:\Program Files\OpenAFS\Client and can
4911 be determined by querying the registry for
4912 [HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
4914 The existing afsdcell.ini will be migrated by the NSIS installer.
4915 The Wix installer must still be updated to do the same.
4917 * Change NSIS installer to use DNS by default; to remove Integrated Logon
4918 High Security mode; and to add Terminal Services compatibility registry
4919 entries to allow the OpenAFS tools to find the afsdcell.ini and other
4920 configuration files in %WINDIR%.
4922 * Add support for authenticated SMB connections. This will remove
4923 the need for high security mode in most situations. Both NTLM
4924 and Extended Security (GSS SPNEGO) modes are supported. Effectively,
4925 only NTLM can be used even though Kerberos is now supported. The
4926 reason is that it is not possible to construct a service principal
4927 which is unique to each individual machine.
4929 SMB Extended Auth does not work on XP SP2 unless one of two registry
4930 modifications are made:
4932 (1) To disable the check for matching host names on loopback connections
4933 set this key. This does not require a reboot:
4935 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
4936 "DisableLoopbackCheck"=dword:00000001
4938 (2) To add the AFS SMB/CIFS service name to an approved list. This
4939 does require a reboot:
4941 [HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
4942 "BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
4944 afsd_service.exe will automatically add the current Netbios Name
4945 to the BackConnectionHostNames list and then temporarily disable
4946 the loopback check for one cycle of startup/shutdown of the service.
4947 We assume most folks do not start/stop without a reboot so this
4948 will be adequate in most cases.
4950 * Fix security hole in afslogon.dll which allowed passwords to be
4951 sent in clear text to the KDC in a misformed principal name.
4953 * Fix cm_GetCell() to properly handle expired dns entries
4956 * If Freelance mode is active and the afs_freelance.ini
4957 file does not exist, do not create an empty file.
4958 Instead create a file containing ro and rw mountpoints
4959 to the default cell using the standard conventions.
4961 * Modify the Freelance support to handle the ability
4962 to create rw mount points in the fake root.afs volume.
4964 * Changed the RPC mechanism used for token setting from
4965 named pipes to local. Use of named pipes can be restored
4966 by setting the environment variable AFS_RPC_PROTSEQ to
4969 Named pipes were required when a Windows 9x system was
4970 using a NT system in gateway mode which is incompatible
4971 with our use of local loopback adapters.
4973 * In afscreds.exe, if a username of the form user@REALM is
4974 specified and no password is specified, do not perform a
4975 kinit operation. Only perform the aklog functionality.
4977 * Add a new registry value which allows the number of processors
4978 on which afsd_service.exe executes to be restricted. Valid
4979 values are 1..numOfProcessors
4981 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
4985 * A second MSI based installer option is now available.
4987 * Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
4988 request tickets for user/instance@REALM instead of just user@REALM
4990 * Modify the Power Management Notify routine to wait for the Hard Dead
4991 timeout period instead of a fixed 19 seconds. With the longer timeout
4992 periods Hibernation and Standby could never succeed when network
4993 connectivity is not available.
4995 * The following fs.exe commands are now restricted to Administrator:
4996 - checkservers with a non-zero timer value
4999 - sysname with a new sysname list
5008 setting the default sysname for a machine should be done via the
5009 registry and not via "fs sysname".
5011 * NSIS installer adds options to install Debugging Symbols
5012 and the Microsoft Loopback Adapter; the user is now also
5013 given the ability to select the afscreds.exe startup options.
5015 * Build system modified to generate symbols for FREE (aka RELEASE)
5016 builds as well as CHECKED (aka DEBUG) builds
5018 * Sites which have a volume ID of 0x20000001 assigned to their
5019 root.cell volumes have been experiencing problems with accessing
5020 the root.cell volume of their cell when Freelance mode has been
5021 active. This was because 0x20000001 was assigned to the fake
5022 root.afs volume created by freelance. The fake volume id is
5023 now set to 0x00000001 to prevent conflicts.
5025 * The timeout logic in the AFS Client Service has been wrong
5026 for sometime. It is based on two different assumptions.
5027 First, the SMB client timeout is a fix value as was the case
5028 with OS/2 Lan Manager. This assumption is incorrect. The
5029 SMB timeout in Windows is a dynamic value computed based upon
5030 a fixed minimum timeout to which is added time based upon the
5031 size of the request and the performance characteristics of
5032 the connection. Second, it is the responsibility of the
5033 SMB Server to enforce the timeout requirements of the client.
5034 This is untrue. The SMB Server cannot be expected to know
5035 the requirements of the client. More importantly, if the
5036 SMB server uses the SMB client timeout as a value to restrict
5037 its behavior as an RX client, the performance characteristics
5038 of the local SMB session would be used to prematurely terminate
5039 WAN connections with significantly different performance
5042 The timeout logic has therefore been modified in the following
5044 . the Lan Manager Workstation (SMB) Session Timeout is used only
5045 as a basis for configuring the Connection Dead Timeout
5046 and Hard Dead Timeout values. The Connection Dead Timeout
5047 must be at least 15 seconds longer than the SMB Timeout
5048 and the Hard Dead Timeout must be at least double the
5049 Connection Dead Timeout.
5050 . New registry entries have been added to allow the Connection
5051 Dead Timeout and Hard Dead Timeout values independent of the
5052 Lan Manager Workstation Session Timeout
5053 . The test to enforce the SMB Client Timeout has been removed.
5055 One of the side-effects of removing the enforcement of the SMB
5056 Client Timeout is that regardless of whether or not the SMB client
5057 is available to receive the response (and how would the SMB server
5058 know) the RX protocol response can be used to update the AFS
5059 Client Service state for ready access by future SMB client
5062 This should be the end of the "Server paused or restarting messages"
5064 * Add support for arbitrary UNC paths to the pioctl() support.
5065 This enables the fs commands as well as the AFS Shell Extension
5066 to work correctly when UNC paths are being used.
5068 * Fix afscreds.exe (by updating afskfw.lib) to search for cells via
5069 DNS if the cell configuration cannot be determined via CellServDB
5071 * Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
5072 are ever set more than once at a time
5074 * Fix the management of lists of cm_cell_t structures when using
5075 DNS to lookup cell information. The previous code would fail to
5076 reuse the same cellID for a cell if DNS was used more than once
5077 for a given cell name. When the ttl expired, a single cm_cell_t
5078 could be inserted into the cm_allCellsp list more than once
5079 producing a loop. In addition, the vlServerp list belonging to
5080 the cell was not freed resulting in improper refCounting of the
5083 * Add DNS support to cm_IoctlNewCell() which previous only examined
5086 * Add cm_FreeServer() function and call it from cm_FreeServerList()
5087 to properly garbage collect cm_server_t objects
5089 * Add numVCs variable to smb.c to track the number of smb_vc_t
5090 objects created and use it to initialize the vcID field which
5091 previously was set to 0 in all objects resulting in FindByID
5094 * Fixed DNS lookups to work consistently throughout the OpenAFS
5095 product instead of just from within the afsd_service.exe
5097 * Added a runtime check to ensure that AFS Client Service SMB
5098 Server is accepting connections before attempting to mount
5101 * Read IP addresses for volume servers out of the CellServDB
5102 file if gethostbyname() on the hostname fails.
5104 * Fix getcellconfig() to populate both the Hostnames as well
5105 as IP addresses when loading cell data via DNS
5107 * Increase the Connection Dead Time to 50 from 20 seconds
5108 Increase the Hard Dead Time to 120 from 40 seconds
5109 (matches the Unix values)
5111 * Fixed an assertion validating the number of allocated NCBs
5113 * Fixed the build environment to consistently build for
5114 Windows 2000 and above. (APPVER = 5.0)
5116 * Fixed rx_debug to properly validate the receipt in incoming
5117 data with select() and recvfrom(). Do not copy data out of
5118 the socket buffer unless success is indicated.
5121 * afsd_service.exe will now display a message box to the
5122 desktop when it terminates due to an IP Address Change.
5124 * installer no longer deletes AFS Server configuration data
5127 * installer generates a warning dialog if the RPC service
5128 is not properly configured
5130 * installer compressed with lzma instead of bzip2
5132 * afsd_service.exe shutdown crash solved once and for all
5134 * reference counting of smb_vc_t data structures improved
5136 * name space collision of smb_fid_t event objects corrected
5138 * the output of "fs memdump" is now written to
5139 %WINDIR%\TEMP\afsd_alloc.log
5141 * the file TaAfsApp_1033.dll is now properly installed allowing
5142 the User Manager to start
5144 * a new algorithm is used for computing filename pattern matches
5146 * afscreds.exe now accepts user names containing instance
5149 * Fix the Directory Name Lookup Cache to be case-sensitive.
5150 This is crucial in environments in which a Windows client
5151 is accessing a directory with more than one filename that
5152 differs only by case. If the directory contains "FOO"
5153 and "Foo". You want "DEL Foo" to delete the correct one.
5154 We still have a problem in that "DEL foo" will delete a
5155 random filename. This will be addressed in a future release.
5157 * Fix afscreds.exe -M option (renewMaps) to work when High
5158 Security mode is off. Also, remember to disable the ActiveMap
5159 flag in afsdsbmt.ini when a drive mapping is removed.
5161 * Updates to NSIS installer script. AFS Server configuration
5162 data will not be destroyed on un-install or re-install.
5163 Use a better compression algorithm.
5165 * afslogon.dll now uses KFW to obtain tokens when available
5167 * afslogon.dll when given an all uppercase username will
5168 attempt to authenticate with both the uppercase name
5169 and an all lowercase variation
5171 * DST modification removed. The fix appears to make things
5172 worse after a reboot of the machine.
5174 * fs.exe: added "cscpolicy" which is used to
5175 change client side caching policy for AFS shares
5177 Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
5178 [-documents] [-disable] [-help]
5180 * Several uninitialized variables have been initialized
5182 * It is now possible to obtain tokens using cross realm
5183 Kerberos within afscreds.exe:
5185 user: jaltman@ATHENA.MIT.EDU
5187 Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
5188 will will in turn be used to obtain afs@DEMENTIA.ORG.
5189 The resulting token will be stored with the display name
5190 jaltman@ATHENA.MIT.EDU@dementia.org
5192 * aklog.exe has been added to the client
5194 Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
5195 [[-p | -path] pathname]
5199 -d gives debugging information.
5200 krb_realm is the kerberos realm of a cell.
5201 pathname is the name of a directory to which you wish to authenticate.
5202 -noprdb means don't try to determine AFS ID.
5203 -5 or -4 selects whether to use Kerberos V or Kerberos IV.
5204 (default is Kerberos V)
5205 No commandline arguments means authenticate to the local cell.
5208 * All of the resource files have been restructured to adhere to
5209 a set of rules IBM implemented for loading string resources.
5210 These rules had either been forgotten or were not discovered
5211 by folks working on the OpenAFS sources. The end result was
5212 memory corruption. This is primary item which was preventing
5213 the AFS Server from working.
5215 * Increased the size of the maximum ticket size stored in a token
5216 from 344 bytes to 12,000. Increased the buffers used to convey
5217 messages between the pioctl() caller and the SMB Server from
5218 1000 bytes to 12,512. The code appeared to have been writing
5219 above the top of the stack by quite a few number of bytes.
5220 (The increased ticket size is necessary for the next item.)
5222 * When obtaining AFS Tokens via KFW, krb524 is no longer required.
5223 Instead the raw Kerberos 5 ticket is used in its entirety. This
5224 is extremely important as it allows us to use pure Kerberos 5 KDCs
5225 as the source of the AFS authentication. The use of up to 12,000 byte
5226 tickets will allow tickets produced by all versions of Microsoft
5227 Active Directory to be used.
5228 - create a user account.
5229 - designate it DES only
5231 - specify its UPN to be "afs@realm"
5232 - assign a SPN of "afs/cellname" to the UPN with setspn.exe
5234 * Do not enforce the funky 8dot3 pattern matching rule that the first "."
5235 is special when using long file names. (you must use "*.*" and not "*")
5236 Instead only enforce it when performing 8dot3 searches.
5238 * Fixed the DST problem with creation times being set one hour ahead
5240 * Fixed the problem when using \\afs\cell-alias. For example,
5241 \\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
5242 for the alias name; instead simply expand the name. One of the
5243 symptoms of this problem was a loss of acquired tokens.
5245 * Fixed the AFS Shell Extension. The Symbolic Link menu was empty
5246 of strings. (Only English strings provided.)
5248 * Fixed the installer to properly replace in use files.
5250 * Fixed the build system to cleanup generated component version files
5252 * The release build compiled with MSVC 6.0 compiler to avoid the
5253 afsd_service.exe shutdown crash. This does not solve the problem
5254 but simply avoids it for the time being.
5258 * fix afslogon.dll to not corrupt memory when High Security mode
5261 * fix afsd_service.exe to not attempt to restore the stack when
5262 an exception occurs. (not safe in multi-threaded programs)
5264 * fix uninstaller to properly remove the CRT and MFC DLLs
5266 * remove a Message Box from afscreds.exe when getcellconfig()
5267 fails on a kerberos realm which is not a cell
5269 The following is a list of changes to the OpenAFS for Window client
5272 * "fs setserverprefs" will leave afsd service deadlocked
5274 * "vos listaddrs" will core dump
5276 * installer sets the appropriate keys to support Integrated Logon
5278 * installer disables the "Find Lana by Name" functionality as it
5279 was causing headaches for many users
5281 * fix the intermittent crash of the power management thread when
5282 shutting down the AFS Client Service
5284 * optimizes the obtain drive mount list functionality which is
5285 executed every time the mount tab in afscreds.exe and afs_config.exe
5286 are refreshed. (this happens a lot)
5288 * fix the service shutdown logic. add the STOP_PENDING state
5289 and do not accept additional service events after we declare
5292 The following is a list of changes to the OpenAFS for Window client
5295 * flexelint was run against the source tree and hundreds (perhaps
5296 thousands) of corrections were applied to ensure prototypes
5297 were in use; types were used consistently; variables were
5298 initialized; unused variables were removed; etc.
5300 * A wide variety of instrumentation was added including the
5301 ability to produce a stack trace from within afsd_service.exe
5304 * Dynamic configuration of the RDRtimeout value based upon the
5305 LanMan Workstation Session Timeout
5307 * The mount root no longer needs to be called "/afs". This
5308 is now set by a registry value "MountRoot" within the key
5309 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
5311 * The cell list is now only read out of afsdcell.ini when the
5312 file changes instead of each time a cell is resolved.
5314 * Thread synchronization was added to cm_server.c and ktc_nt.c
5316 * All calls to GlobalAlloc()/GlobalFree() were replaced with
5317 calloc()/free(). The Global functions were needed on Windows 3.x
5318 but have caused a variety of problems on the Win32 platforms.
5319 Avoiding them is highly recommended by several Microsoft
5320 Knowledgebase articles
5322 * Support for Symbolic Links added to the AFS Shell Extension
5324 * Added a registry value "OverlayEnabled" to determine if
5325 Shell Extension Overlays should be enabled.
5326 HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
5328 * New Build system to support VC6, VC.NET, VC.NET2003 compilers and
5329 separate trees for checked and free builds. Build system supports
5330 a custom directory src\WINNT\extra which can be used as a grafting
5331 location of organization specific additions to the build tree.
5333 * New installer built using NSIS 2.0.
5335 * Named all kernel objects in order to allow them to be monitored
5336 with tools such as SysInternals' ProcExp.exe.
5338 * Introduced new EventLog framework for AFSD
5340 * Introduced Power Management interface to AFSD for Standby and
5341 Hibernate modes to allow cache to be flushed prior to network
5344 * Utilize Win32 DNSQuery API instead of internal routines. This
5345 allows DNS SRV queries to be sent to all current domain name
5346 servers. Not just one specified in an INI file. DNS is now
5349 * "NetbiosName" registry value may be used to specify a fixed
5350 Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
5351 when the loopback adapter is in use. If you need to use the
5352 old notation with a loopback adapter installed specify a registry
5355 "NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
5357 * Refactor all modules which depend on LAN Adapter and NetbiosName
5358 determination in a new library: lanahelper.lib. This allows for
5359 consistent behavior throughout the product.
5361 * Move the afsd.log and afsd_init.log files to the directory specified
5362 by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
5363 for services. Added the Date to the log entries.
5365 * New registry value "RxMaxMTU" used to limit the size of the RX
5366 packets sent by the AFS Client Service to the Server. In order
5367 to enable OpenAFS to work across the Cisco IPSec VPN the packet
5368 size must be restricted to 1264 or smaller. The latest NSIS
5369 installer sets a value of 1260 by default.
5371 * New registry value "RxNoJumbo" to disable the use of Jumbo Rx
5372 packets. This is not needed in order to work across the Cisco
5373 VPN but might be needed for other network environments. This
5374 value is not set by the NSIS installer.
5376 * New registry value "HideDotFiles" is used to apply the Hidden
5377 attribute to files whose names begin with a '.'. This value
5378 is set by the NSIS installer.
5380 * New registry value "MaxMpxRequests" allows the maximum number
5381 of multiplexed sessions to be configured at run time. This
5382 value is not set by the NSIS installer. The default value is
5385 * New registry value "MaxVCPerServer" allows the maxmimum number
5386 of VCs per server to be configured at run time. This value is
5387 not set by the NSIS installer. The default value is 100.
5389 * New registry value "AllSubmount" allows the "all" submount to
5390 be disabled by setting its value to 0x00.
5392 * Allow cells names to be valid mount points
5393 \\<netbiosName>\<cellname>
5395 * Store the active state of drive mappings in order for afscreds.exe
5396 to restore them upon startup
5398 * Add exception handling to generate a Stack Trace to the afsd_init.log
5399 file if one happens to occur.
5401 * Add lots of logging to help detect the cause of invalid SMB packets
5403 * Enable Kerberos for Windows to be used to obtain AFS Tokens via
5404 conversion of Kerberos 5 "afs" service tickets. Supports auto-
5405 renewal of expiring tokens as long as afscreds.exe is running.
5407 * New afscreds.exe command line options:
5409 -M = renew drive maps
5410 -N = ip address change detection
5413 * New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
5414 determines whether or not MIT Kerberos for Windows should be used
5415 to obtain tokens via Kerberos 5 tickets.
5417 * New registry value "AfscredsShortcutParams" in
5418 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
5419 determines the command line parameters to be specified when "fixing"
5420 the AFS Shortcut in the user's startup folder.
5422 * The "ShowTrayIcon" registry value has been moved from
5423 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
5424 {HKCU,HKLM}SOFTWARE\OpenAFS\Client
5426 * The <cell name> registry values used to store the token expiration
5427 reminders have been moved from
5428 HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
5429 {HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
5431 * Obtain the Logon User Name from the Explorer key when available
5433 * new text document doc\txt\winnotes\registry.txt lists all registry
5434 values used by OpenAFS (excluding the AFS Server)
5436 * BUG: rx_securityClass objects were not properly reference
5437 counted and were never freed.
5439 * BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
5440 would be generated. The existence of a server does not imply
5441 that it is not down. If all of the servers for a cell are down
5442 return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
5445 * BUG: the directory name lookup cache failed to free the entries
5446 in the cache when the name cache entries cycled. The entries
5447 in the cache would become dereferenced without being freed.
5449 * BUG: fs setserverprefs could be executed without Administrator
5452 * BUG: the number of allocated NCB objects (100) exceeded the number
5453 which could actually be waited upon by the kernel (64). Any objects
5454 which were utilized above the limit could never have event completions
5457 * BUG: smb_username_t objects were not being reference counted and
5458 were not properly freed.
5460 * BUG: smb_tid_t objects could under unusual circumstances be freed
5461 before they were no longer referenced.
5463 * BUG: smb_fid_t object pointer were frequently used even when
5464 their value could be NULL. They were not properly released and
5465 therefore they were never freed.
5467 * BUG: smb_packet_t data structures were not completely initialized
5470 * BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
5471 Denied" because that causes the Explorer Shell to try again until
5472 access is obtained. Instead return "Remote Resources" which allows
5473 the shell to move on and treat the error as transient.
5475 * BUG: when initializing the NCBreturns structure, separate Event objects
5476 were created for each NCB although a single Event object was supposed
5477 to be shared by all.
5479 * BUG: smb_dirSearch_t objects were not being properly referenced counted
5482 * BUG: smb_tran2Packet_t objects were not being properly referenced
5485 * BUG: directory path creation did not handle the case of multiple
5486 directories requiring creation in one attempt
5488 * BUG: SMB requests which required an Extended Response were ignored.
5489 This prevented some files from being written to AFS volumes.
5491 * BUG: character strings were being freed even after they were
5492 inserted into in use data structures
5494 * BUG: inconsistent usernames were used when High Security mode was
5495 enabled. (there is still much to do in this area)
5497 * BUG: pioctl() calls which require out of band RPC operations were
5498 susceptible to race conditions when performed by multiple processes
5500 * BUG: memory allocation and deallocation crossed instances of the
5501 C Runtime Library producing memory leakage and corruption in
5502 afscreds and the client configurator.