jimzah/openafs
1
0
Fork 0
mirror of https://git.openafs.org/openafs.git synced 2025-01-31 21:47:45 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
openafs/doc/txt/winnotes/afs-changes-since-1.2.txt
Jeffrey Altman 2e4a7c47a1 DEVEL15-windows-notes-20080108 
LICENSE MIT

1.5.30
2008-01-08 16:32:31 +00:00

3353 lines
139 KiB
Plaintext
Raw Blame History

Since 1.5.29
* Fix buffer refcount leak introduced in 1.5.29
* Prevent the NIM AFS Provider from crashing when configured to use
Kerberos v4 for token acquisition in situations where no Kerberos v4
support is available. For example, 64-bit KFW or sites that delete
the Kerberos v4 DLL.
* In afsd_init.log, support logging cache sizes greater than 2GB.
* The volume status output of "fs examine" was being determined based
upon the wrong error value. Instead of using the pioctl() return
code, the errno value must be used.
* Prevent invalid pts auto-registration attempts by aklog and the
NIM afs provider when Kerberos referrals are in use. (krb5 1.6+)
* Add synchronization protection to all of the SMB Listener State
variables. This prevents race conditions when the addition or
loss of a network adapter takes place.
* On Vista, there is a race condition between the restart of the
afs client service and the network adapters. Ensure that the
afs client service can handle switching between loopback and
non-loopback modes of operation.
* Fix FollowBackupPath to work with mount points that are not
explicitly read/write.
* Fix a deadlock that has been observed on Vista while resuming
after at least four hours of sleep.
* Add new HKLM registry configuration options that can be used
to provide configuration hints to the NIM AFS Provider when
creating new identities. See release notes for details on
HKLM\SOFTWARE\OpenAFS\Client\Realms\.
Since 1.5.28
* Speed up write buffer operations by holding a global read-lock
instead of a write-lock when updating the buffer data version
number. The data version is protect by the associated stat
cache mutex.
* Speed up read buffer operations by removing out of date buffers
from the buffer hash tables.
* Protect against a null pointer dereference in afslogon.dll
if QueryAdHomePathFromSid() fails.
* Remove a race condition while accessing private freelance data
structures.
* Update the buffer data versions when Freelance directory data
changes.
* Complete the implementation of smb_ReceiveTran2GetDFSReferral().
Critical for Vista.
* Return EAS_NOT_SUPPORTED in response to EA read/store/query
instead of returning zero length.
* Return ERROR_RETRY when a server sends VBUSY.
* Return PATH_NOT_COVERED instead of BADSHARENAME for DfsLinks when
a volstat plug-in is in use.
* Enhance the volstat interface to support Dfs path mappings.
* Fix "symlink list" to do the right thing for Dfs links
* Fix callback expiration that was broken in 1.5.27.
* Fix a crash at shutdown caused by the lack of synchronization
within the daemon threads (cm_Daemon, cm_BkgDaemon, cm_IPAddrDaemon)
* Hold the cm_server_t mutex while walking or modifying the
cm_server_vols_t list in order to avoid a race condition.
* Ensure that the cm_server_t refCount is safely modified.
* The pthread library implementations of pthread_cond_wait
and pthread_cond_timedwait were failing to return with
the mutex held when error conditions occurred. This could
result in race conditions wherever they are used. One case
in which problems were seen was the Rx event_handler().
* Prevent cm_GetBuffer() from locking all buffers associated with
a stat cache object before deciding if the current buffer is up
to date.
* Fix cm_BkgPrefetch to actually ensure that there is a cm_buf_t
object for each offset that we will attempt to reserve.
* Reduce the default number of background daemon threads used to
prefetch or store to one in order to prevent two or more threads
from stepping on each other.
* Reduce the default chunk size to 256KB in order to prevent read
timeouts on low bandwidth connections.
* Fix a race condition that could result in threads becoming stranded
when synchronizing multiple requests on the same stat cache object.
Since 1.5.27
* The GiveUpAllCallBacks functionality that was added in 1.5.21
has been disabled. It has been disabled because it turns out
that the RPC is improperly implemented in all file servers
previous to 1.4.6. The RPC is executed without the proper
locks being held resulting in data corruption and eventual file
server failure. The functionality will be re-enabled in a
a few months.
* a couple of .readonly Volume Callback optimizations were
implemented:
- Apply most recent volume callback to all stat cache objects
in the volume. This avoids premature callback expirations
and unnecessary FetchStatus calls.
- Add option to permit attempts to renew .readonly Volume
Callbacks on a periodic basis to avoid invalidation of stat
cache data. [Requires daemonCheckVolCBInterval be set to
non-zero number of seconds. 1800 (30 minutes) is suggested.]
* Adds support for 64-bit data version values. Previous releases
only supported 32-bit values. Larger values sent by the file server
would result in rollover.
* aklog and asetkey no longer generate an exception if Kerberos for
Windows is not installed.
* If the LSA authentication functions report a temporary out of memory
condition, fail the current authentication and do not cause the AFS
client service to terminate unexpectedly.
* When the AFS client service is shutting down, prevent the daemon check
thread from executing additional checks which might attempt to access
discarded locks or unmapped memory.
* Correct a double free condition in the NIM AFS credentials provider.
* Add "FollowBackupPath" service parameter registry value.
This provides equivalent functionality to the UNIX afsd -backuptree
option. When set to a non-zero value, normal mount points originating
in a .backup volume will prefer a .backup volume.
Since 1.5.26
* Fix NSIS uninstall from the .MSI installer.
* Add Stack Frame logging for AMD64
* Add support for EWOULDBLOCK to cm_Analyze. If the file
server returns EWOULDBLOCK, retry the request every two
seconds for up to the RDRtimeout.
* Reorganize the order of the includes to ensure that EWOULDBLOCK
is not assigned the same value as EIO.
* Update CellServDB to GCO Public 25 Oct 2007
* Add additional validation and error handling code after
each call to getSlot(). If an invalid slot is returned,
return NONODE. If the invalid slot is returned when
extracting a data node, invalidate the tree.
* Modify compareKeys() to always perform a case-insensitive
comparison and only perform a case sensistive comparison
if the case-insensitive one matches. This ensures the
ordering is consistently reported.
* Add lock assertions to ensure that all calls are being
performed with the correct locks being held. There have been
some crash reports that provide stack data that does not appear
to be possible unless there is a race. However, there are no
obvious locations where the race is taking place and the test
suite indicates that all of the correct locks are being held.
We shall see what happens in the field.
* For consistency replace all calls to findKey in which the range i
s (1,numentries) with calls to getSlot().
* Optimize the depth search loop by testing the slot value in the
for statement instead of forcing the loop to be broken later.
* Reorganize the locking for cm_BeginDirOp and cm_EndDirOp.
There are a number of locations where locks are obtained, dropped,
and reobtained. This reorganization attempts to accomplish several
things:
(1) be optimistic for the most common case so it will be fast
(2) add consistency checks after each location where locks are
dropped and re-obtained. If we lose a race in cm_BeginDirOp
and the bplus tree is out of date, retry until we get to a
consistent state that we can use.
(3) Ensure that all operations take place with the correct locks.
* One of the issues that has become a serious problem since the
addition of the local directory updates is that although cm_SyncOp
synchronizes operations, it does not preserve the order of requests.
This has always been a problem in that it has been possible for a
request to fail to complete due to its worker thread's bad luck.
When a request takes longer than the Windows SMB Redirector's timeout,
the SMB Redirector tears down the SMB virtual circuit.
When using the local directory updates it is really important that
the directory update operations complete in the order that they were
sent to the file server. If they don't, then the local directory state
and the file server state will not match and the local directory state
must be discarded which in turn forces a new read of the entire
directory contents over the network.
This patch adds a new cm_scache_waiter_t object that is used to store
the current thread, buffer, and syncop flags within a waiters queue
on each cm_scache_t object. If a thread is forced to sleep in
cm_SyncOp, upon waking it will check to see if there are any other
threads waiting that are attempting to perform a similar task ahead
of it in the queue. If yes, the thread goes back to sleep. If not,
it goes ahead and enters the cm_SyncOp conflict resolution block.
This patch has the additional side effect of reducing the number of
competing threads that must obtain the cm_scache_t mutex and process
the cm_SyncOp conflict resolution block. As a result, the overall CPU
utilization of the service and the clock time associated with processing
requests will be reduced.
* assert that the cm_scache_t mutex is held when calling
buf_ForceDataVersion().
* rename findNode to leafNode in bplus_Lookup
* replace all OutputDebugString calls with osi_LogX calls
* modify bestMatch to special case the return values for leaf nodes.
If an entry is above or below the values available in the leaf node
return BTLOWER or BTUPPER instead of BTERROR.
* In insert and delete operations check for BTLOWER/BTUPPER and isleaf,
if true convert to either slot 0 or Max and perform the insertion.
This produces easier to read code when performing lookups.
* modify lock_AssertXXX macros to call osi_assertx() and provide a
descriptive message.
* the difference between osi_assert() and osi_assertx() is whether or not
a description is specified as part of the call. When no message is
specified afsd_notifier() is called with a NULL msgp parameter. This
results in a null pointer dereference during the EventLog() call.
As a result, none of the information describing the assertion is written
to the log file.
This commit sets a default message for use when no other message is
provided and it replaces all calls to osi_assert with osi_assertx and
adds descriptions.
* Migrate B+ tree search key into thread local storage
Since 1.5.25
* Fix an uninitialized variable in lana_IsLoopback() which
could result in a random determination that a loopback
adapter is not a loopback adapter. This would result in
a computed Netbios service name of %MACHINE%-AFS instead
of AFS. This is a serious problem if the value computed
by the AFS Client Service and the client process pioctl
calls produce different results as it will result in a
communication failure.
* Fix a crash in the B+ Tree code that results when a file
rename only results in a change of case.
* Fix the pioctl path parsing. The previous algorithm
broke the current directory of the client into two parts,
the TIDpath which represented the device root and the
relative path. The problem is that this is not sufficient
from proper processing of symlinks and mount points.
What is required is that the last component of the path
must not have the symlink or mount point be evaluated
whereas symlink and mount point evaluation must be performed
for the rest of the relative path. Therefore, the path is
now split into three parts: TIDpath, intermediate path,
and last component.
"fs examine" now reports the data of the symlink or mount
point object instead of the target and "symlink list" and
"symlink remove" now works. The "Explorer Shell Extension"
menu now is displayed for symlink objects.
* More informative symlink.exe error messages.
* When the cmd.exe "del" command is used to remove a symlink,
not only was the link object removed, but the stat cache
object for the target was marked deleted. This resulted
in cache corruption which would require the destruction
of the cache to correct.
* Add undocumented registry value, "OfflineReadOnlyIsValid"
which results in cached contents from read only volumes
to be treated as valid when there is no callback registered
and all servers are down or all instances are offline.
HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
* Add documented registry value, "AcceptDottedPrincipalNames"
which is used to permit the acquisition of AFS tokens
using Kerberos v5 principal names which include a dot in
the first component.
{HKLM\HKCU}\Software\OpenAFS\Client
* "fs uuid" does not require "AFS Client Administrator" group
membership to display the current UUID value.
Since 1.5.24
* Update NSIS installer to version 2.30
* Update Wix installer to version 2.0.5325
* 64-bit KFW support:
- Network Identity Manager plug-in
- 64-bit aklog
- 64-bit krb5 enable afscreds.exe, afs server manager,
etc.
* Enable B+ tree Directory Search by default.
* NIM plug-in is now compatible with FILE: ccaches
* Install instloop.exe into Client/Programs/ even though
it isn't required. Permit it to be used by organizations
that clone Windows systems.
* On a suspend, only drop callbacks for servers that
are online. If the server is offline, maintain the callback
until the server is once again available.
* Prevent against acquisition of a directory write lock when
there are no updates to be performed. This was resulting
in a write-lock leak which later on would lead to a deadlock.
* Add configurable file pre-fetch mode. For files that match
the specified file extensions, pre-fetch the entire file into
the cache.
HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
(MULTI_SZ) PrefetchExecutableExtensions
* Fix the client_osi large integer support so that return values
can be properly assigned to large integer variables.
* Do not hold cm_scache_t mutex across cm_GetCell() calls
since cm_GetCell() can block in a RPC.
* Add short file name support to B+ tree searches.
* Resolve a conflict between the desire to notify file servers
that the client workstation is suspending and the desire to
permit objects with valid callbacks when the client suspends
to continue to use those objects from the cache if the client
resumes without network connectivity.
* Fix stack variables to ensure that all buffers storing paths
can handle AFSPATHMAX (1024) characters. Paths of 256 characters
would cause a stack overflow condition.
* On Vista, force afs_config.exe to run with the highest privileges
available to the end user. For a user in the Administrators
group running under UAC, this will cause afs_config.exe to be
be run with full Administrators privileges after the user
approves the privilege escalation.
* More VNOVNODE handling issues. If a dirty buffer cannot be
written due to VNOVNODE, clear the buffer and ignore the error.
If a cm_scache_t object has been marked deleted due to a VNOVNODE
error, the next time the smb_fid_t object that references it is
used return either NOSUCHFILE or NOSUCHPATH and close the
smb_fid_t object. Subsequent attempts to use the smb_fid_t will
result in an invalid handle error.
* For Vista, set the RequestExecutionLevel of the NSIS installer
to "admin".
* For the build environment, support the generation of manifests
when using the CL1310 (VS .NET 2003) compiler/linker.
* There was a bug in OutputDebugHexString() which could overflow
a stack buffer resulting in unexpected process termination
of afsd_service.exe if the input Security Token contained
many bytes equivalent to the '%' char.
* Add directory lookup statistics to "fs memdump" output.
Since 1.5.23
* Windows uses case-insensitive file name pattern matching
but AFS is a case sensitive file system. The AFS3 directory
format is block based, uses network byte order and
includes a hash table for fast case sensitive lookups.
This causes several problems for the Windows AFS client.
(1) Traversing the directory blocks is cpu expensive
(2) A hash table miss does not indicate that the desired
entry does not exist.
(3) Determining whether a non-ambiguous inexact match or
the entry does not exist requires a linear traversal
of the entire directory.
These issues often result in 100% CPU utilization.
These issues are addressed by building a modified B+ tree for
each directory and then using the B+ tree for searches.
This feature is disabled by default. To enable it set
the "BPlusTrees" registry value to 0x01 in the AFS Client
Service Parameters key.
* Correct a null pointer dereference in the symlink recursion
detection code. (Introduced in 1.5.22)
* Add the Network Identity Manager AFS Provider to AMD64
builds.
Since 1.5.22
* A new registry value, "BlockSize", can now be used to adjust
the size of the internal buffers used to store file data.
The default is 4KB. When set, the values 1..1024 specify
multiples of 4KB. Values > 1024 specify actual block sizes
and must be multiples of 4KB. The actual block size in KB
can not be larger than the chunk size which defaults to 1MB.
HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
(DWORD) BlockSize
* pts.exe and ptserver.exe built with -DSUPERGROUPS
* Service Manager shutdown hints have been added. The hint
specifies that it may take the AFS client service up to two
minutes to shutdown.
* Fix the recursion detection feature introduced in 1.5.22.
The previous implementation broke the evaluation of paths
using ../path or ./path. It also broke the evaluation of
symlinks to absolute paths. For example, /afs/cell/path.
* Fix an uninitialized variable in cm_MakeDir that could result
in crash if the requested directory could not be created.
* Fix a the VolumeStatusNotification log messages to save
the input string in the circular log buffer so garbarge
won't be output to afsd.log when the log is dumped.
* Increase the SMB Maximum Path length to 260 characters
from 256 as that is what CIFS supports.
* On 64-bit Windows, use a unique description string for
the 32-bit AFS Shell Extension.
Since 1.5.21
* In the AFS Admin Library, replaced all of the
Ubik_Call(XXX_RPC,...) calls with Ubik_XXX_RPC(...). This
allows parameter checking and proper type conversions to be
performed. This is crucial when passing parameters that are
64-bit wide on systems in which 'long' is a 32-bit value.
Otherwise, the parameter list on the stack is interpretted
incorrectly and the function not only doesn't succeed but it
might corrupt memory and if lucky crash the application.
The only program that we care about that uses this library
is the AFS Server Manager. Now it doesn't crash.
* Fixed smb directory searches for non-wildcard names. The
optimized function should not have set the bulk in progress
flag. This would adversely affect later directory searches
that required a bulk search.
* Detect infinite recursions (up to 512 stat cache objects
in a chain.) If an infinite recursion is detected, return
an error to stop the CIFS client from querying us forever.
* Executables executed out of AFS name space can have their
pages swapped out by Windows. Those pages will be reloaded
when they are required. However, if the cache manager no
longer has a callback and all of the file servers with
the volume data are down, then the application will crash
due to the inability to read the required page.
We now track the down time of each server. If the callback
expires after the server that issued it went down and all
servers are down, we delay the expiration to permit the data
in the cache to be continue to be used by Windows or an
application.
* The Windows AFS client has always had very poor performance
with regards to directory management. In the case of
directory updates caused by create, rename,and delete operations,
the AFS cache manager will now perform local directory updates
when possible in order to avoid re-reading the directory from
the file server. This will be done in whenever both of the
following conditions are true:
- the data version on the directory increased by one as a
result of the operation
- all of the directory buffers necessary to make the change
are present in the cache
* The error message in the AFS Control Panel describing the
need to be a Windows Administrator has been replaced with
a message describing the need to be in the AFS Admins Group.
* If the vldb server list is empty when attempting to evaluate
a cell mountpoint it was possible for a crash to occur. This
can only happen if there is an existing mount point and the
is a failure attempting to load the server list from the
CellServDB file or via AFSDB queries.
* dirty buffers only have their dirty bytes stored to the file
server instead of the entire chunksize
* the default chunksize has been increased from 128KB to 1MB
Since 1.5.20
* changed the enum values for cm_serverRef_t state info to use a
private name space to avoid collisions (srv_)
* added a srv_deleted state for cm_serverRef_t objects. This
state is set when cm_FreeServerList() is called with the
CM_FREESERVERLIST_DELETE flag set. cm_FreeServerList() may
not always delete the cm_serverRef_t from the list if it is
still in use by another thread. the srv_deleted state means
the object's contents are no longer valid and it must be
skipped. It will be deleted the next time the object is
freed and the refcount hits zero.
* the srv_deleted state is also used when a file server reports
either VNOVOL or VMOVED instead of marking the cm_serverRef_t
as offline. This is done to prevent additional usage of the
stale vldb data while waiting for the update volume request
to complete.
* added a state field to the cm_volume_t object (enum volstate
vl_ name space) that maintains the state of the volume based
upon the states of all of the cm_serverRef_t and cm_server_t
objects.
* modified cm_UpdateVolume() to set the state of the cm_volume_t
RW, RO, and BK to either vl_alldown or vl_online. There can't
be any other states because cm_UpdateVolume() destroys any
previous knowledge we might have had regarding busy or offline
volume status
* modified cm_UpdateVolume() to update the volume name in the
cm_volume_t to the volume base name if the previous value was
a volume ID.
* modified cm_FollowMountPoint() to check to see if the volume
name is a volume ID and if so call cm_GetVolumeByID instead
of cm_GetVolumeByName. This ensures that volume IDs are always
looked up as numeric values. There is no longer a need to
maintain a separate cm_volume_t containing the string representation
of the ID value.
* Added a flags parameter to cm_GetVolumeByName() and cm_GetVolumeByID().
The first flag is a "CREATE" flag which is set by all existing
calls. The flag is not set by calls to cm_GetVolumeByID() from
the server probe code when volume status is being updated. We
do not want the server probe operation to result in additional
turnover in the cached volume data. The second flag is NO_LRU_UPDATE
which is set when the server probe code updates the volume status.
This flag will be used to prevent the server probe operation from
changing the order of the least recently used queue.
* Modified cm_GetVolumeByName to ensure that only one cm_volume_t is
allocated for a given set of normal, readonly, and backup volumes
regardless of whether or not the volume is accessed via name or
ID number. The cm_volume_t namep field is always the base name
of the volume.
* Added a new volume state, vl_unknown. This state is used as
the initial state for all cm_volume_t when the cache manager starts,
for each cm_volume_t at creation, and for each cm_volume_t when
recycling. The cache manager does not know the state of all
volumes in the world, only those that are in the cache and for
which it has queried the VLDB and hosting file servers.
* modified cm_GetVolumeByName() to initialize the state of a
volume to vl_unknown. The actual state will be set when a
cm_VolumeUpdate() call completes successfully.
* changed name of scache hash table variables to avoid ambiguity
when adding hash tables for volumes
* fix a buffer overrun in sys\pioctl_nt.c pioctl().
* modified cm_UpdateVolume() to handle the case in which there is
no RW volume but there is are RO volumes for a given base name.
This is done by querying for the ".readonly" volume name if the
base name does not exist in the VLDB. We never query for the
.backup name because under the current usage model a .backup
volume may only exist on the server that the read-write volume
is located. If there is no RW volume, there can be no .backup.
* Added four hash tables for cm_volume_t objects to improve the
search time of cm_GetVolumeByID(), cm_GetVolumeByName() and
cm_ForceUpdateVolume(). One each for Name, RWID, ROID, and
BKID. Three ID hash tables are necessary as long as it is
desireable to maintain a single cm_volume_t containing all
of the related RW, RO, and BK volume data. Having the RW and
RO volume data in the same object is necessary for the
implementation of cm_GetROVolumeID() which returns either the
RO or RW ID depending upon the existence of RO volume instances.
* Added a volume LRU queue so that volume reuse becomes fairer.
This does not replace the all Volumes list which is used when
it is desireable to walk a list of all the volumes whose order
is not going to change out from underneath you which makes it
safe to drop the cm_volumeLock.
* handles volume hash table updates where volume name to
volume ID number changes. The volume name remains
constant in the cm_volume_t. if a vos rename is performed,
the name of the volume will change and the volume IDs will be
updated. Subsequent access to the old volume ID will create a
new cm_volume_t with the new name.
* Added a daemon thread operation to query the state of volumes
listed as busy or offline. cm_CheckBusyVolumes() calls
RXAFS_GetVolumeStatus() for each volume ID that is marked vl_busy
or vl_offline. If the volume is now online, the status on the
volume is updated. The default period is 600 seconds. This can
be configured with the BusyVolumeCheckInterval registry value.
* Added prototype for smb_IoctlPrepareRead() which was missing a
return type in the function definition.
* Added volume id lists to the cm_server_t. These lists are
allocated in blocks of ~32 IDs. When a cm_PingServer()
detects a change in server state, the state of the cm_volume_t
is updated.
* Added volID to the cm_serverRef_t object. volID is used
to identify the volume for which the object is a referral.
cm_FreeServerList() uses the volID to remove the volume
from the cm_server_t.
* In cm_Analyze, when VNOVOL or VMOVED are received,
call cm_ForceVolumeUpdate() to force a refresh of the volume
location data.
* Added cm_VolumeStatusNotification() which is used at the moment
to log volume status changes to the trace log. It will also
be used as the access point to the File System Filter driver
notification engine.
* Added an all cm_scache_t list to cm_data. This replaces the use
of the stat cache LRU queue when we need to enumerate all
entries. The LRU list order is not static and when using it to
enumerate all entries it can result in items being missed or
items being processed more than once.
* Modified cm_Analyze(). Instead of reseting the busy or offline
state of a volume and forcing a retry of the operation
cm_Analyze will defer to the background daemon thread that will
update the state once every 600 seconds.
* Added the automatic generation of a Freelance ".root" read-write
mountpoint that refers to the root.afs volume of the workstation
cellname at the time the mountpoint is created.
* re-write cm_Analyze to make better use of the known volume
status. VL_Server queries cannot result in CM_ERROR_ALLOFFLINE
messages.
* renamed cm_CheckBusyVolumes to cm_CheckOfflineVolumes.
busy volumes will be reset to srv_non_busy by the function
but there is no mechanism for querying the busy state other
than by attempting to access the resource.
* cm_Analyze will query the state of an offline volume before
deciding whether or not to retry when all volume instances
are offline.
* the algorithm used after an IP address list change detection
has been modified. we now check the state of the VLDB servers,
then reset the rx connections, and finally check the state of
the file servers. Previously we flushed the rx connections
and then checked the servers in whatever random order they
were first discovered.
* fix the volume status reported by "fs examine"
* various improvements to the AFS Network Identity Manager Provider.
- Don't access freed objects after the window has been destroyed.
- Write back the AFS cell list in the identity configuration
even when its empty.
- Disable all use of the provider when AFS is disabled for a
particular identity.
* Fix cm_cell_t leak. When evaluating a mountpoint that uses an
alias instead of the full cellname, double check the full cell
name and see if we have a matching cell before allocating a new
one.
* Dump more data with "fs memdump". All volumes and scaches and
now cells.
* Reverse the order of the all cells list. This ensures that
the workstation cell is always first in the list.
* If we recognize the server from which a callback revoke was
received, then we can ensure that we only invalidate the status
information for cm_scache_t objects belonging to that cell.
* When building for amd64, use APPVER 5.02 as we do not support
64-bit Windows 2000 and explicitly state the machine type in
the link flags.
* no longer treat VMWare network adapters as equivalent to a
loopback adapter. For some unknown reason, publishing the "AFS"
netbios name on them frequently results in a name collision
error.
* Use read locks whenever possible instead of write locks when
walking the allCellsp list. This permits better use of multiple
CPUs.
* In 1.5.20 a bug was introduced by the fake status when PRSFS_READ
is not present on the directory optimization. It would report
mount points and directory objects for which there was no
status information as a zero length file. Instead we can use
the vnode number of the object to determine if it should be
considered to be a directory or a file. If it is odd it is a
directory; even it is a file.
* Give Back Callbacks to the file servers in response to suspend
and shutdown power management events. Only give up callbacks
to UP servers and do so with a very short RX dead timeout.
(10 seconds)
* Fix a race condition in the SMB FID cleanup code that could
result in a stat cache mutex never being released.
* Fix the graphics files that are installed as part of the
IBM Administrators Guide.
* Change the behavior of pioctl path parsing to not follow
symlinks or mount points during evaluation. This results in
the FID being the symlink or mount point and not the object
they point to. This is consistent with the behavior of the
UNIX client.
* Enable AFSDB record lookup for command line tools such as
"fs" and "vos".
* CellServDB update 2007-07-05
* Added a Volume Status Plug-in Interface that permits third parties
to receive notifications when the volume state changes occur.
* Ensure that all writes to the afsd_alloc.log file use CR-LF.
* Prevent the afsprocmgmt.dll from being unloaded while the
background thread it starts is running.
* Avoid a race condition when closing smb_fid_t objects.
Since 1.5.19
* fix install path of IBM Quick Start Guide when installed by
the Wix installer
* fix the client so that it can write into directories with
only 'li' permissions.
* fix installation of 32-bit Shell Extension for 64-bit systems.
Install it under its own HKCR keys.
* modify the wix installer to remove and upgrade install packages
with the same version number.
* when flushing volume location data every two hours or in
response to "fs checkvolumes", also flush the mount point
links in order to force their re-evaluation.
* aklog was copying an uninitialized buffer as part of the
foreign cell automatic pts registration
* when obtaining directory listings, check for read permission
on the directory. If read permission does not exist, fake the
status information for the directory entries. This improves
performance, prevents the client from pounding the file server,
and prevents leakage of cached status information.
* "fs getcacheparms" output was incorrect.
* NetIdMgr AFS Provider should only automatically add token
acquisition for cell foo.com if the realm of the cell matches
the realm of the identity.
* Fix a null pointer dereference in cm_MergeStatus()
Since 1.5.18
* network identity manager provider:
- Correct for the fact that pr_CreateUser tampers with the
contents of the username input buffer.
- Take into account that the krb4 credential data structure
does not have separate realm fields for the user and service.
When krb524 is used with a cross-realm ticket, the realm
describes the service not the user.
* Many integrated logon changes:
- Fix eventlog reporting. Do not attempt to log an event
if the event source registration fails. Use DebugEvent0
instead of DebugEvent when there are no parameters.
- Modify the LOOKUPKEYCHAIN macro to recognize ERROR_MORE_DATA
errors.
- Fix the reading of Domain specific configuration for
LogonScript and TheseCells. Previously the dwSize value was
being overwritten so that subsequent RegQueryValueEx call
would fail.
- Fix a memory leak in the TheseCells reading code.
- Add support for Domain specific "Realm" specification. The
realm is the realm to be appended to the username. When
logging in as a domain or to the local machine, the specified
"Domain" name is not going to be a valid realm name.
- Construct a proper principal name based upon the domain
specified realm for use in obtaining tokens with KFW.
- If the domain specified "TheseCells" list includes the
default cell, do not obtain tokens twice.
* An unprotected call to the Kerberos v5 function,
krb5_free_context(), can result in an invalid memory access
exception if the contents of the krb5 profile, krb5.ini, cannot
be parsed correctly. The error is located within the afskfw
library which is used by afscreds.exe, afssvrmgr.exe, and
afslogon.dll. When the error is triggered within afslogon.dll
a blue screen condition can occur when the winlogon.exe system
process terminates unexpectedly.
Since 1.5.17
* Do not stop the afs client service with an assert
if the cell name cannot be determined when responding
to a "fs whereis" command.
* Integrated Logon (afslogon.dll). No longer attempt
to preserve Kerberos 5 tickets for the logon session
as that job is now supported by KFW 3.1 and above.
* The Network Identity Manager plug-in will attempt to
remove expired service tickets from the cache.
Since 1.5.16
* CellServDB update - GCO Public CellServDB 09 Mar 2007
* AFS Server Manager and AFS Account Manager data cache
moved from HKCU\Software\TransarcCorporation to
HKCU\Software\OpenAFS to prevent crashes due to data
format changes.
* Reduce the default UP server probe time from 10 minutes
to 5 minutes to match UNIX. This will not solve the NAT
port mapping timeout issue because most NATs default to
a value less that 5 minutes and this value ensures that
the afs client server will start probes 5 minutes after
the last probe set ended. The probe period for any one
server is therefore longer than 5 minutes. The more
servers in the list the longer the time between probes.
* Before marking a server as down in cm_Analyze, retry the
request using a new rx connection to the server. If the
cache manager is behind a NAT, it is possible that the
response from the server is being sent to the original
address:port from which the rx connection was established.
This will appear to the cache manager as the server being
down when in fact it is not. Forcing the use of a new rx
connection will result in the server responding to the
most recent address:port assigned to the client by the NAT.
If the server is in fact down, it will simply take a little
longer to notice it.
Since 1.5.15
* Removing a directory that contained entries was failing with
the wrong error code being returned to the CIFS client.
* rxdebug -long
* rxdebug exponential back-off. don't take down a server that
is already struggling to respond.
* Network Identity Manager plug-in registers help on the Help
menu if it is supported by NIM. (KFW 3.2 and above)
* Plug-n-play networking did not always work after coming back
from a Windows Suspend. The AFS client service failed to
re-compute the LAN adapter to bind to and instead attempted
to bind to all LAN adapters. This would polute the public
network with the name "AFS" which could lead to conflicts
between machines.
* Build installers using Microsoft Hotfixes provided at
http://tinyurl.com/243svk for Visual Studio C RunTime Library
Daylight Savings Time issues.
* "fs checkserver" does not require admin privileges unless
the probe interval is being changed.
Since 1.5.14
* A reference leak was discovered in the mountpoint traversal
functionality which resulted in cm_volume_t objects not becoming
eligible for re-use. This will trigger a panic condition when
more than the maximum number of volumes is accessed.
* Fixed the success test for the "fs memdump" operation. Success
is now reported as such. Added cm_volume_t object information
to the dump file.
* A data corruption error has been fixed. Write requests
queued for background processing were not retried upon failure.
This leads to inconsistencies between the contents of the file
in the local cache and the contents of the file on the file
server.
The implemented fix is to re-queue any request that has failed
due to timeout, volume busy, volume offline, server busy, or
server offline. Requests are not processed while all servers
for a volume are inaccessible. When servers become accessible
the background request will be retried.
In addition, if a background write fails, do not mark the buffers
as clean. Failure to preserve the dirty flag prevents the buffers
from being written to the file server in the future.
This fix does not address the possibility that the credentials
necessary to perform a write request might expire between the
time the request is queued and when the request can be processed
due to server accessibility.
* The Freelance volume does not support per user ACLs. The
cm_HaveAccessRights function did not handle this special case.
Requests for privileges greater than read|lookup would result
in afsd_service.exe entering an infinite loop. ISOBuster is
known to cause this behavior.
* When deleting files, the SMB Server would attempt to obtain
the access permissions for the parent directory instead of
for the file itself. This would result in an incorrect
assessment of whether or not the file can be deleted.
* The Network Identity Manager AFS plug-in has had its string
tables updated for consistency with the language used by NIM.
All credentials are referred to as credentials instead of tokens.
Drop-down combo boxes now in fact drop down.
* The AFS Salvager was not properly being built on Windows.
This is relevant only if the AFS servers are in use.
* The default signal handlers for the AFS servers incorrectly
processed SIGQUIT. Instead of stopping the server, an
exception was raised.
* Fix another volume refcount leak. This one was triggered by
an inability to contact the vlservers in order to update the
file server list.
* Fix checklist control used by the drive mappings tabs in
afscreds.exe and afs_control.exe.
* Improve output of "fs memdump" related to buffer management.
* Fix refcount leak of cm_buf_t objects when prefetching buffers
already in the process of being fetched.
* Improve cache performance by computing the buffer hash table
size based upon the number of buffers so that the average
number of entries in a bucket is seven.
* Improve cache performance of read/write data by locally
updating the dataVersion of all buffers associated with the
file on each StoreData or StoreStatus operation provided that
the dataVersion was incremented by one and the dataVersion of
the buffer matched the original dataVersion of the cached
status information for the file.
* When recycling cm_scache_t objects, make sure that we
clear all of the previous fields. It is no longer the
case that we can assume that the stat cache object is being
obtained only after a successful FetchStatus call.
InlineBulkStatus will allocate stat cache objects for which
the no status information is available.
* When a file is being opened with the OPEN_ALWAYS disposition,
do not require write permissions when the file already exists
unless of course the file is being opened requesting write
access.
* When a file has been opened for read-only access, if the
application requests an exclusive lock, obtain a read-lock
instead of a write-lock. The purpose of the exclusive lock
in this situation is to prevent changes to the file and the
full file read lock will provide this guarrantee.
Since 1.5.13
* Enforce Unix Mode Write bit to prevent file deletion of read only
files.
* The CIFS READ and LIST permissions use the same value 0x0001.
When a directory is being accessed, check for PRSFS_LOOKUP
instead of PRSFS_READ.
Since 1.5.12
* Fix computations of short names of mount points. They were
broken as part of the directory search optimization. (1.5.13a)
* When all buffers are in use, periodically sleep the threads
waiting for a buffer to become free while the global buffer
lock is not being held in order to permit those threads that
can free buffers to do so.
* Power Management improvements. Maintain a global flag that
specifies whether or not the service is in a suspend state.
Do not panic if Netbios() returns NRC_BRIDGE meaning that the
lana is no longer valid. Instead, stop the listener threads
and if all listener threads are stopped, reset the lana_list.
Allow the cm_Daemon() thread to periodically check the state
of the smb listeners. If they are all stopped and the service
is not suspended, attempt to restart them. If there are no valid
lanas, return to the stopped state.
* CreateX and NTCreateX use cm_CheckNTOpen() to test whether or not
the user is permitted to obtain read or write locks. This function
would obtain the lock and then drop it returning whether or not
the lock could be obtained. If the lock was in fact required,
CreateX/NTCreateX would then obtain it with cm_Lock(). The problem
of course being that this pattern results in three RPCs to the
file server (lock, unlock, lock). This is reduced to one RPC by
implementing cm_CheckNTOpenDone() which frees the allocated byte
range lock from cm_CheckNTOpen() after the cm_Lock() call is
performed.
* Optimize the SMB FindFirst processing for the case in which there
are no wildcards. If a cm_Lookup succeeds, generate a response
from the stat cache entry we have a callback for. Its cheaper
than walking each buffer of directory data associated with the
directory stat cache entry when we have thousands of directory
entries to search.
* Increase the SMB Ioctl MaxData size to support larger token sizes
* Disable autorun of the AFS Server Wizard when the AFS Server is
installed.
* Add AFS Server service dependencies on tcpip and PNP_TDI
* Remove the AFS Server Configuration Tool shortcut
* Add Freelance mode awareness to AFS Server Configuration Tool
Since 1.5.11
* Unlike previous versions of the OS, Vista performs a shutdown on
the Microsoft Loopback adapter just like it would on a real adapter.
This causes the smb_Listener threads to trigger a panic during a
suspend/hibernate power event.
The fix is to unbind from the network adapters in response to a
suspend/hibernate power event and then rebind to the adapters
when a resume power event is received. Note that the resume events
are not reliably delivered so it is possible the afs service will
not be accessible. However, this is the best we can do.
* don't allow environment variables we set in afslogon.dll to be
inherited by child processes
* force an rx connection reset if VICECONNBAD or VICETOKENDEAD are
received
* The old dirty buffer synchronization algorithm had a buf_IncrSyncer
thread walking the all buffer list periodically searching for dirty
buffers to write to the file server. This had several negative
results. The algorithm ate up ever increasing amounts of CPU time
even when AFS is idle as the size of the cache increases. Also,
buffers were written to the file server in an order based upon the
original buffer allocation which has nothing to do with the order
in which the buffers became dirty.
The new algorithm maintains a dirty buffer list. Items are added
when the buffer is originally marked dirty. A buffer is only
removed from the list by the buf_IncrSyncer when the buffer is no
longer dirty. If the list is empty the thread goes back to sleep
immediately without additional processing requirements.
Since 1.5.10:
* Enable renaming of directories from the Explorer Shell when
Offline Folders is enabled
* An error is reported in response to the SMB ReceiveCoreFlush
request when an error has not occurred.
Since 1.5.9:
* optimize dirty buffer flushes in response to suspend operations
do not flush if only loopback adapter is installed
do no retry dirty buffer writes
* do not reset netbios when performing loopback checks. doing
so will break the existing lana bind
* remove a deadlock in smb3.c QPathInfo
Since 1.5.8:
* Add NetIDMgr AFS Plug-in
* "Show Tray Icon" checkbox removed from afs_config.exe
* Disable DNS Registrations for the Loopback Adapter and make sure
that Netbios is turned on.
* A reference count leak on stat cache objects was fixed
* For debugging, tread ids of writers are now tracked on locks
* Corrected an incorrect lock that was being held but not
freed in a rare error case
* Changed trace log end of line to CR-LF so it can be read
with notepad.exe
* modify the buffer synchronization thread to write all dirty
buffers and sleep only when there is no work to be done
* improve behavior of cache manager when large numbers of
temporary files are being created by multiple clients within
the same directory.
* do not permit dirty buffers that belong to deleted files
to be written to the file server. its a waste of bandwidth
* do not leak stat cache references when the file server
returns VNOVNODE
* afscreds: do not permit an attempt to obtain tokens if the
afs service could not be started. this was causing stack
overflows on Vista after resuming from a suspend
Since 1.5.7:
* better handling of dot directory
* fs sq and fs mkm now behave as on UNIX
* add license to installers
* Return Invalid Parameter error for unknown InfoLevels in
SMB Tran2 Search Directory.
* Correctly use WSAStartup and WSACleanup within the RX library
* Check that afsd_service.exe is running before performing
a pioctl call.
* Force AFS Shell Extension to load resource DLL using
afs_shl_ext.dll as a base name.
* (Wix) install afs_shl_ext_XXXX.dll and afs_creds_XXXX.dll in
OpenAFS\Client\Program
* Fix SMB_SET_FILE_END_OF_FILE_LENGTH
* Do not panic if the maximum number of volume entries are in use
and one of them can be recycled.
Since 1.5.6:
* Fix the code that reads the backconnnectionhostnames value
so that there is not an uninitialized variable
Since 1.5.4/1.5.5:
* integrated logon could cause winlogin.exe to blue screen windows
if it was called with a domain that does not exist on the machine.
this is possible during a sysprep.
* corrections to cm_buf_t handling to ensure that buffers are
removed from queues before they are freed and that holds are
obtained on the next buffer in a chain before releasing the
current one.
* do not generate MessageBox dialogs on network errors. they
cause more trouble then they are worth.
* move GlobalAutoMapper into a separate thread so that it does
not prevent service startup if the WNetAddConnection2 call
does not return.
* restrict hard dead and connection timeouts to the CIFS Session
Timeout value. This is being done to try to prevent virtual
circuits from being destroyed prematurely.
* Fix the code that writes the backconnectionhostnames value
to ensure that the data buffer is written with the correct
length.
(there were no windows specific changes between 1.5.4 and 1.5.5)
Since 1.5.3:
* Corrections to the Integrated Logon functionality that permits
Kerberos 5 TGTs to be communicated into the user session. The
transfer is now more secure and succeeds when the user does not
have read permission to the %SystemRoot%\TEMP directory.
* NSIS installer updated to version 2.18
* WiX installer updated to version 2.0.4221.0
* Updates to the IFS functionality for 64-bit correctness and
permit compilation. The IFS functionality works for pioctl()
operations but not much else at this point.
* A major re-write of the OpenAFS implementation of the CIFS Tran2
query and set functions.
* Ignore the Windows CIFS client session timeout entirely. Timeout
only after the AFS Hard Dead Timeout period expires.
Since 1.5.2:
* move headers, libraries, and samples from OpenAFS\Client to
OpenAFS\SDK
* remove the auto-registration of AFS IDs for foreign cells
from integrated logon. this permits afslogon.dll to avoid
initializing the RX library. afslogon.dll is a network
provider library and as such it is loaded into all applications
which perform a CreateProcessAsUser call. Initializing the
RX library spawns background threads to listen for inbound
packets. When the network provider is unloaded from the
process, the RX library does not cleanup the threads. If they
continue to execute after the library is unloaded, the process
will crash.
* implement inline bulk status calls. this will significantly
improve the performance of directory listings.
* prevent a crash when evaluating mountpoints to volumes that do not
exist
* new command: fs uuid [-generate]
displays the current uuid and allows a new one to be generated
on the fly without restarting the afs client service.
* fix the 64-bit installers to properly document the install
location
* move help files to the proper location so they can be located by
the system tray tool and the control panel.
* "over quota" and "disk full" errors were not being reported to the
caller.
* largefile support has been added. Files greater than 31-bits can no
longer be written to file servers which do not support large files.
When file servers support large files, files up to 64-bits can be
read or written.
* cm_SyncOp could return with the input buffer unlocked upon error
which would result in a panic condition. This bug was introduced
in 1.4.1.
* vlserver failover when freelance mode did not work. This prevented
the root.afs volume from being loaded if the first vlserver could
not be reached.
Since 1.5.1:
* A bug in the CIFS Standard Info response packet caused directories
to appear as standard files. This prevented Microsoft Vista from
being able to walk the AFS directory tree.
* Mount points must be reported as directories in response to
Query Path Info operations.
* A deadlock was detected when performing "fs flushall" if the
file server reports VNOVNODE. The scp->createBufferLock is
already held by the current thread and there is no mechanism
to propagate the knowledge. Therefore, an alternate mechanism
for clearing the cache must be developed. A new function
cm_RecycleSCache(scp,flags) has been extracted from
cm_GetNewSCache(). This function performs the task of recycling
an cm_scache_t object. When called from cm_GetNewSCache() with
no flags the expectation is that there are no associated buffers
that are queued to be read or written. When called from
cm_Analyze() with the CM_SCACHE_RECYCLEFLAG_DESTROY_BUFFERS
flag, any queued buffers will be de-queued and marked as if
the operations were performed so the data can be discarded.
* CM_ERROR codes were logged as "unknown" by cm_Analyze.
* In response to VNOVNODE, the parent is only discarded if the current
cm_scache_t is not a directory.
* In the Ioctl FlushFile and FlushVolume functions, there were no
checks to protect against flushing the Freelance SCache entries.
* In FlushFile, the wrong cm_scache_t object was being released.
* In cm_GetNewSCache, do not allow Freelance SCache entries to be
recycled. Choose a new entry if cm_RecycleSCache fails.
* Speed up the performance of the cache manager by not
holding the smb_fid_t mutex across calls to cm_SyncOp
and AFS RPCs.
* Ensure that all smb_fid_t flag references are protected.
* Remove a deadlock between smb_fid_t mutex and
smb_rctLock introduced in 1.4.1.
* Fix aklog AFSID lookup to use not send the realm
name when the realm is the local realm for the
cell. The PT Server doesn't strip the realm
name when it is the local realm.
* Treat "\\afs\*." as an alias for "\\afs\all"
* Dynamically adjust the priority of server threads
based upon the age of the CIFS request that is
being processed. The priority is bumped one level
for every 15 seconds of processing time.
* The original openafs contribution mis-used the
cm_GetCallback function. This function should only
be called as a side effect of a call to
cm_SyncOp(CM_SCACHESYNC_NEEDCALLBACK). The way it
was being called results in a FetchStatus RPC being
performed even when the client already has a valid
callback and does so in a manner that destroys the
synchronization of the threads that are calling
cm_SyncOp. If nothing else this results in a
significant performance penalty.
* More changes to cleanup of smb_vc_t objects in
order to prevent race conditions.
* Reduce the amount of Power Management event logging
to afsd_init.log
* If the client detects that the IP address has
changed, force the use of new RX connections. This
enables the file server to recognize that the client
has moved.
* Remove a race condition and deadlock associated
with cleanup of smb_vc_t objects.
* Fix the ANSI filename support. One function that
required it was missing the translations.
Changes not present in 1.4.1 and since 1.5.0: 1.5.1 released
(24 April 2006)
* Add a DllMain() function to afspthread.dll. Perform
cleanup of thread local storage, rx queues, critical
sections, event handles, etc. This was implemented
to avoid crashes in afspthread.dll when applications
which load/unload the dll attempt to terminate.
* Lock request behavior has been altered. As of this
release a locally allocated lock will be used whenever
the volume is RO or the requesting user's permissions
are 'rl'. This will allow end users to execute programs
from RW volumes when they do not have the 'k' privilege.
Note that a recent discovery indicates that the 'k' bit
is only supposed to control the ability to obtain read
locks. Write locks are to be controlled by the 'w' and
'i' privilege bits.
* Code has been added to enable the cache manager to be
able to obtain the AFSID of the authenticated user.
This is required to be able to determine when "Insert"
privileges should apply. AFS considers a file to be
a new file whenever the owner's AFSID matches the
AFSID of the user. The current implementation relies
on the name specified during the SetToken call to be
correct. This will be replaced in the future with a
new RPC that allows the AFSID of the authenticated
user to be obtained.
* A 32-bit-tools installer for use on 64-bit systems
is now available. These tools are for use with 32-bit
versions of Kerberos for Windows or JPSoftware's 4NT.
Since 1.4.1 rc10: 1.4.1 final released (9 April 2006)
* Remove the use of the "active_vcp" variable. This variable
no longer serves any useful purpose and was the cause of
race conditions that produced under counts of the smb_vc_t
references.
* Change default minidump output to include data segments
* restore access to cm_IoctlGetVolumeStatus and
cm_IoctlSetVolumeStatus calls for AFS pioctl operations.
Still need to figure out why they are being called
for SMB/CIFS IOCTL operations instead of Trans2GetFileInfo.
Changes not present in 1.4.1: 1.5.0 released (16 Feb 2006)
* Update build system to include the "sys_name" in the
obj and dest paths.
* optimize the number of calls to obtain the smb_user_t for a
smb virtual circuit.
* add logging of server addresses, status and capabilities when
pinging the server.
* prevent multiple threads from pinging the same server simultaneously.
instead, threads block and then use the result of the ping that was
already taking place.
* Add registry values to allow the defaults for daemon checks to be altered
at startup. Log values to afsd_init.log.
daemonCheckDownInterval
daemonCheckUpInterval
daemonCheckVolInterval
daemonCheckCBInterval
daemonCheckLockInterval
daemonCheckTokenInterval
* Add EventLog messages for transitions in fs crypt state.
* Byte range locking (when locks will be obtained from the
is still under discussion)
* 64-bit Windows operating systems now supported
(XP SP2 64-bit, 2003 SP1 64-bit, 2003 R2 64-bit)
This is a native 64-bit implementation of OpenAFS for Windows.
There is currently no 64-bit version of KFW and therefore
there is no (easy) method for obtaining AFS credentials.
32-bit KFW 3.0 can be used with 32-bit OAFW libraries
to obtain AFS tokens but there is not for the feint of
heart.
* A faster implementation of the fcrypt algorithms improves
performance.
* The afsd_service.exe now supplies Event Log Message Text.
The event log no longer generates warning text due to lack
of message configuration.
* When afs servers are unreachable and network stack returns
a WSAEHOSTUNREACHABLE error, the AFS client will immediately
mark the attempted server as being down. This allows for
very fast response to the calling application. This change
has a pronounced improvement when the Windows Explorer is
used on disconnected mobile computers.
* Added DebugOutputString debugging of the RX library
Since 1.4.1 rc9: 1.4.1 rc10 released (10 March 2006)
* Fix the rx-lwp implementation to randomly generate rx call
identifiers which are supposed to be unique at a given point
in time. This affects some of the command line tools used
for debugging but not the AFS Client Service itself which
uses the rx-pthreads implementation.
* xstats_fs_test.exe and xstats_cm_test.exe are now built for
Windows although they are not distributed
* don't generate an exception when logging unexpected SMB packets
* the data structures used by cm_IoctlGetVolumeStatus and
cm_IoctlSetVolumeStatus are not compatible with the CIFS
Get/Set File Information calls so don't bother executing them.
Since 1.4.1 rc8:
* Fix data written to registry as part of BackConnectionHostnames
data value
* Fix initialization bug when allocating cm_server_t objects that
would cause servers to appear to be down after each allocation
or renewal. Server would appear up after the next check servers
cycle.
Since 1.4.1 rc7:
* Windows will return WSAECONNRESET to the next receive packet
operation on a socket if an ICMP packet is received in response
to a packet that was sent and could not be delivered because
the port was unreachable. We have to treat this as
WSAEWOULDBLOCK so that we can retry the select() check and
not mark the connection as bad.
Since 1.4.1 rc6: 1.4.1 rc7 released (14 Feb 2006)
* The client manager maintains a least recently used list of
all the ACLs. If an ACL is already the most recently used
do not bother to remove it from the list and place it on the
front.
* As part of the process of adding a new server, set the
server is down flag and then perform a ping. If the server
is up, the flag will be reset. If the server is down, the
client will not timeout unnecessarily while attempting to
perform real work.
* If a Mountpoint String is the empty string, return PATH NOT
FOUND.
* Modify behavior of the Network Provider to avoid querying the
profile location if integrated logon is known to be disabled.
* Correct behavior of 'LogoffPreserveTokens'. When set, tokens
will always be preserved regardless of whether or not integrated
logon was used to obtain tokens. When unset, tokens will be
destroyed regardless of whether integrated logon was used to
obtain the tokens unless the profile is stored in AFS.
* Disable LogoffTransferToken functionality. The SMB V3 Logoff
message is being delivered even when the user is not logging
off of the computer.
* Fix a deadlock condition introduced in rc6.
* Cell names are case-insensitive. Since the founding of OpenAFS
there has been a bug that treated "cell", "Cell", and "CELL" as
different cells. This would result in what appeared to be
random disappearance of user tokens or a failure to authenticate
even when tokens were listed by the "tokens" command.
Since 1.4.1 rc5: 1.4.1 rc6 released (2 Feb 2006)
* Additional auditing of the smb code revealed that smb ioctl
operations were failing to free smb_fid_t objects
* The File Identifiers assigned to smb_fid_t objects could be
given the value 65535 which is 0xFFFF which is also the
INVALID_HANDLE value. Both 0 and 65535 are now treated as
reserved values.
* Audited all assignments to fields and ensured that they are
protected by obtaining the relevant mutex object.
* Re-wrote the VNOVNODE handling to avoid a potential deadlock
if both the current cm_scache_t object and its parent are
both returning VNOVNODE.
* Removed all the logic associated with the 'dead_vcp' variable
which must have been added for debugging because its used
was never thread safe and it truly served no purpose.
Since 1.4.1 rc4: 1.4.1 rc5 released (25 Jan 2006)
* Re-enabled LogoffTokenTransfer and LogoffTokenTransferTimeout
registry settings. Default is ON and 120 seconds. If set to
OFF, tokens will not be destroyed.
* Fixed a number of race conditions and reference count problems
associated with global queues.
* Added a method to detect abandoned smb virtual circuits so their
associated file handles, locks, user tokens, and smb session
resources can be released.
* Added a mechanism to allow user tokens to be preserved
during integrated login and be freed after logoff.
* Check access permissions to the file before the file is
truncated.
* More improvements to the handling of SMB Virtual Circuits
when SMB sessions are closed.
* Correct a race condition introduced in rc4 that could result
in a crash of afsd_service.exe.
Since 1.4.1 rc3: 1.4.1 rc4 released (11 Jan 2006)
* Properly handle pre-mature SMB session closures.
File handles, tree connections, and user sessions are now
freed when the virtual circuit is closed.
* smb_username_t and smb_user_t objects are now properly reference
counted and freed as appropriate.
* disable to use of AFS file locks. byte range locking is still
implemented but all locks are managed locally. This prevents
multiple processes from editing the same file on the same
machine but it does prevent multiple processes on separate
machines from modifying the file.
Since 1.4.1 rc2: 1.4.1 RC3 released (4 Jan 2006)
* A fix to "fs setserverprefs -vlserver". Multiple calls with
the same server parameter would result in a panic condition
if the server had not already been contacted by the AFS
client service.
* Better warnings are written to the afsd.log file when
the Windows RPC Protocols are not properly configured as
is often the case on Windows 2000.
* fix the procmgmt library to restore C RTL signal handlers when
it is being detached from the process prior to process termination.
Since 1.4.1 rc1: 1.4.1 RC2 released (30 Nov 2005)
* close a lock leak in the rx library that could be triggered while
processing aborts
* prevent removal of drive mappings in response to afscreds.exe -M
* add logic to process VNOVNODE in cm_Analyze. Force re-evaluation of symlink
strings and flush the stat cache entry.
* prevent server objects from being freed if user preferences are set.
* Update WIX support to 2.0.2419.0
* Release all locks on the server at service shutdown.
* A down server should be marked up as long as it responds to the client.
It doesn't matter if the server returned an error as long as the error
is something other than VRESTARTING. If there is such an error, all
connections on that server should be marked for replacement.
* Change the check "up" servers period from once per hour to once every
ten minutes. This matches the behavior of the Unix cache manager.
Post 1.4.0 revisions: 1.4.1 RC1 released (21 Oct 2005)
* Byte Range Locking support has been added. Architectural documentation
is located within src/WINNT/afsd/cm_vnodeops.c. Best efforts are
made to enforce Windows mandatory locking model using locally managed
byte range locks with full file locks obtained from the file server.
Read-only volumes provide the user with an implied locking privilege.
Read-write volumes require that a user be granted the locking privilege
in order for a lock to be obtained. As most Windows applications use
Shared file access modes, they require the ability to obtain locks in
order to access files.
* Implement error translation functions to convert Universal Error Codes
output by modern servers to local errno and winsock error values.
This fixes the error message problems using pioctl() calls from fs.exe
and the AFS Shell Extension.
Since 1.4.0 rc8: 1.4.0 final tagged (20 Oct 2005); released (31 Oct 2005)
* Allow the AFS Client Service to set the firewall configuration by
retrying until after Windows determines the boot period is over.
* Do not attempt to configure the firewall on Windows 2000, XP pre-SP2,
or 2003 pre-SP1
* Detect IP address changes from within the AFS Client Service for the
purpose of checking down servers sooner
* If the file server or vldb server report RXKADUNKNOWNKEY, return
SEC_E_NO_KERB_KEY to Windows. This provides an indication to the
user that there is a Kerberos key problem. Previously the user would
receive no feedback when there was something wrong.
* Fix the large integer library support to properly handle values greater
than 32-bits
Since 1.4.0 rc7: 1.4.0 rc8 released (15 Oct 2005)
* Updated CellServDB
* Fixed file name pattern matching. Question marks can match nothing
if found at the end of a component.
Since 1.4.0 rc6: 1.4.0 rc7 released (8 Oct 2005)
* Kerberos 5 ticket lifetimes should adhere to KFW Leash lifetimes
* The MSI installer now has a new Product ID for each build. This
is to allow for automated upgrades. The Upgrade ID is remaining
constant.
* Volumes that were cached and then subsequently were salvaged,
moved or otherwise taken temporarily offline (as opposed to being
busy) would become unreachable. A distinction is now made between
a volume being offline and a server being down.
Since 1.4.0 rc5: 1.4.0 rc6 released (2 Oct 2005)
* Fixed several AFS Shell Extension issues:
- Checking Server Status would only work if a specific cell was
specified.
- If servers were down, the Server Down dialog would write to
invalid memory and cause explorer.exe to crash
- Default Push Buttons were not specified for many dialogs
- The extension would be unloaded by the Explorer Shell and
never be reloaded under some circumstances.
* Fixed the AFS System Tray menu to disappear if another Window
was selected before a menu item was selected
* Added volume owner and group information to "fs examine" output
* Added fs and registry support for enabling/disabling/clearing
rx statistics gathering.
Since 1.4.0 rc4: 1.4.0 rc5 released (25 Sep 2005)
* New Product Code GUIDs issues for 1.4 and 1.5 branches. Separate
GUIDs for each platform.
* Documentation directory structure and default installation rules
for MSI
* New OpenAFS for Windows HTML Release Notes installed
as part of the documentation set.
* Add thread id to rx debugging messages
* When all servers are offline, return a bad network path error
immediately. This is necessary to allow Explorer to do the
right thing and not hang. The background daemon thread will
check the down servers every 30 seconds to see if they are
back up.
Since 1.4.0 rc3: 1.4.0 rc4 released (14 Sep 2005)
* The AFSCache file has been modified to store both the
serial number of the volume on which the file resides
and the SID of the local machine. If either change,
a new UUID is generated.
* add "fs flushall" command that forces all data buffers
to be flushed.
Since 1.4.0 rc2: 1.4.0 rc3 released (4 Sep 2005)
* add support for '/' instead of '\' in pioctl() calls
* Apply AFS Client Admins group protect to AFS Shell Extension
* Add support for \\afs\<name> to most AFS Shell Extension
commands except the symlink methods.
* when installing the Microsoft Loopback Adapter, enable
MS Client for Networks and prevent an install failure
by not calling CoInitialize twice in the same thread.
* reload cell vldb values from the CellServDB every two
hours in case it changes
* When updating cell information from DNS, be sure to set
a new timeout.
* Add support to allow use of \\AFS\<foo> where <foo> is
either a mount point or symlink. As <foo> is normally
treated as a share name, we transform it into \\AFS\all\<foo>
for processing.
* Init 'code' to prevent false errors when integrated logon
is disabled and the service is not running
Since 1.4.0 rc1: 1.4.0 rc2 released (28 Aug 2005)
* Removed trace log messages that were hampering performance.
* Fixed a deadlock that was being triggered by editing Word
documents stored within AFS with WinWord 2003.
* Bit 3 of the TraceOption registry value is now used to set the
default for "fs trace" in the non-Debug builds. The new default
is off for release builds and on for debug builds.
* Bit 2 of the TraceOption registry value can be used to turn on
real-time output of debug log entries to the Windows Debug
Output monitor interface. This data can be viewed with tools
such as Sysinternal's DbgView.
Since 1.3.87: 1.4.0 rc1 released (17 Aug 2005)
* When the cache manager reports ALLBUSY or ALLOFFLINE for
a resource we should not be returning status codes such
as Network Busy or Remote Host Down because these errors
will imply that the AFS Client Service is Busy or Down
and that is simply not the case. Instead we will return
Bad Network Path as the path is temporarily not available.
Instead of returning Sharing Paused when there is a
timeout error, return Timeout. Once again, the AFS Client
Service is not paused.
* afscreds.exe would display an Obtain Creds dialog when
the expired credentials reminder was triggered even
if there was no network path to the KDC. This is prevented
by adding KDC probe logic to the reminder thread.
* afscreds.exe would display expired tokens no differently
than unexpired ones. This would make it difficult for the
user to distiguish when the tokens were expired. For the
English build added a new resource string "(expired) that
is displayed instead of the expiration time.
Since 1.3.86:
* "fs wscell" when executed in freelance mode will return the
name of the cell configured in the registry. The root.afs
volume is not loaded from this cell, but it is used for the
default for aklog and integrated login.
* "fs mkmount, fs rmmount, symlink make, symlink remove" will
no longer work on \\AFS\all when freelance is being used
unless the user is a member of the "AFS Client Admins" Windows
Security group.
* some more checks performed during persistent cache validation.
Since 1.3.85:
* Optimize calls to cm_CheckServer from cm_Analyze. Only check
down servers for the current cell not all cells.
* If the scache LRU list becomes corrupted, the AFS Client Service
might enter an infinite loop at startup. Detect the corruption
and discard the cache contents in this case.
* Fixed the Explorer Shell's remove mountpoint functionality.
* Fixed a deadlock caused by the holder of cm_aclLock attempting to
obtain a mutex lock on a cm_scache_t object whose lock is held by
a thread attempting to obtain the cm_aclLock.
Since 1.3.84:
* Really fixed DNS AFSDB queries to ensure that "csail" cannot be
misinterpretted as "csail.mit.edu" when the resolver
is configured to append ".mit.edu" to failed DNS queries.
* Added a new registry key, "LogoffPreserveTokens" (see registry.txt),
that can be used to force the preservation of user tokens upon logout.
* Update the NSIS install scripts to use NSIS 2.07. This release adds
recommended installation categories: AFS Client, AFS Administrator,
AFS Server, AFS Development Kit. Each category includes a different
default subset of the OpenAFS components.
The OpenAFS logo is now associated with the NSIS uninstall entry
in the Add/Remove Programs control panel.
* The user name associated with AFS tokens when obtained with
integrated login, the afs systray tool, or aklog will always
include the full Kerberos 5 user principal name regardless of
whether or not the cell is local to the realm.
* Modify integrated login so that it does not enter an infinite
loop if the service is not set to auto start.
* Added asetkey.exe used to set a Kerberos 5 key for use by the
AFS server daemons
* Added uninstall.exe to Wix installation
* More modifications to algorithms used to wake sleeping threads.
Since 1.3.83:
* Changes to the algorithms used to wake threads when they are
put to sleep because of conflicts over buffer operations.
Since 1.3.82:
* The OpenAFS integrated logon obtains Kerberos 5 tickets as part of
the process of producing AFS tokens. The tokens are stored within
the AFS Client Service but the Kerberos 5 tickets have been discarded.
New functionality has been added to temporarily copy the tickets to
a file ccache during the short window between the NPLogonNotify and
WinLogon Logon Event. The Logon Event handler starts a process as
user, afscpcc.exe, that copies the tickets into the default ccache.
The file ccache is then destroyed by the event handler.
* During synchronization operations on stat cache entries, if threads
are forced to sleep it was possible that the threads would not be
woken under heavy load.
* Enforce the delete on close flag specified by the NTCreateX()
operation.
* A race condition in the rx library was fixed that resulted in
a deadlock in rx_NewCall when the RX_CONN_MAKECALL_WAITING flag
was cleared when threads were still waiting to use the
connection. Also, fixed a potential case in which the
connection could be destroyed while threads were waiting to use it.
* The DNLC freelist has been seen to become corrupted with still
active entries being stored on the freelist. Changes were made
to perform a better job of cleaning entries before placing them
onto the freelist; marking them with the correct magic value when
purging the list; and allowing the list to be purged on startup
if the cache validation fails.
* Windows Crash Reporting does not get triggered for afsd_service.exe
because it provides its own top level exception handler. This patch
places an explicit call to ReportCrash() on platforms which support
it. (XP and above.)
If you configure Crash Reporting via AD Policy to capture crash
reports within the domain, then you will receive this output.
* Generate MiniDumps independent of the Windows Crash Reporting.
(a) If an exception occurs,a minidump will be generated at the path
%TEMP%\afsd.dmp. The type of minidump defaults to Normal but
can be set to any other type via use of the
registry.
[HKLM\SOFTWARE\OpenAFS\Client] MiniDumpType (DWORD)
(b) MiniDumps can be generated at any time using the "fs minidump"
command. This allows you to generate a minidump without
stopping the process or even requiring the use of a debugger
on the system.
* an RX connection which reports bad ticket now treats it the same way
the expired ticket is treated. The ticket is removed from the
connection and a new anonymous connection is established.
* fs memdump now requires that the caller be in the AFS Client Admins
group
* additional debugging information is provided in cm_Analyze when the
error code is ignored.
* Fix a regression introduced into the pioctl() client code which
will cause a crash in winlogon.exe, explorer.exe and fs.exe if
MIT KFW is not installed. (1.3.82a)
* Remove AFS Gateway option from OpenAFS Control Panel (1.3.82a)
Since 1.3.81:
* Fix a race condition in cm_EndCallbackGrantingCall() that
could leave 'nrevp' pointing to freed memory. In the process,
optimize the processing to ensure that the applications monitoring
the status of the stat cache entry are only notified at most
once.
* Conditionalize access to the Explorer Shell AFS->Mount Point
->Remove menu item. Disable the item if the selected item(s)
are not mount points.
* Activate AFS RPC (RX) Free Packet Queues stored in Thread
Local Storage. This eliminates thread contention on the
global RX free packet queue. This should improve performance
on multi-CPU systems.
* Fix the IP address filtering code to properly load/unload
the IPHelper DLL so that Windows XP/2003 and beyond systems
do not have to manually probe the registry.
* CleanupACLEnt() was not being called consistently with
the cm_scache_t object referred to by the back pointer
mutex locked. This could in very rare conditions lead to
an invalid memory access.
* Added a script command to the msi installer "afs_replace"
which can be used to replace one version of OpenAFS with
another without requiring an uninstall and reinstall if
the installed msi is no longer in its original location.
* Apparently the problem with multi-domain forests with cross-
realm trusts to non-Windows realms was not entirely solved.
The authentication to the AFS SMB service failed because
the wrong name was being used. Using ASU as an example,
the authentication was being performed with the name
"QAAD\user" (an account in the forest root) and not
"user@ASU.EDU (the MIT Kerberos principal used to login with)
The solution was to add an additional dependency on KFW
in order or to be able to easily obtain the client principal
name stored in the MSLSA ccache TGT. This information is
used in two locations:
- the pioctl() function
- a new WinLogon Event Handler for the "logon" event.
The pioctl function will now be able to use the correct
name when calling WNetAddConnection2() and the "logon"
event handler will now be able to call WNetAddConnection2().
The hope is that the "logon" event handler will be called
before the profile is loaded but I have not guarrantee
that will happen.
Since 1.3.80:
* Fixed a locking error in cm_TryBulkStat() which had the
potential of deadlocking the system for the length of time
it takes to perform a bulk status fetch operation.
* Replaced time conversion code (UnixTime <-> FILETIME) to
be completely arithmetic instead of relying on a bizarre
algorithm involving a variety of C RTL time functions.
This has the side effect that UnixTime and FILETIME which
are both stored in UTC are interpretted as UTC throughout
the year. Windows will apply the same localization to AFS
as it does to NTFS. Applications which rely on the ability
to sync files between the two file systems will no longer
see the timestamps of files in AFS change an hour relative
to the files stored in NTFS or Windows based backup devices.
* Fixed a invalid memory access under a bizarre circumstance.
Windows will allow a physical mass media device to be
installed via Plug N Play to the system and will assign it
the lowest available drive letter. This is true even if the
drive letter is currently assigned in the user session to
a network device via NET USE (or its equivalent.) When this
happens, queries sent to the CIFS server will contain invalid
data. This invalid data was not being caught by the AFS
Client Service and was resulting in a crash when Freelance
mode was in use.
* Fixed a reference count error when registering callbacks
on a stat cache object if the callback was already registered
* Add a case to cm_Analyze() in order to handle the case where
cm_GetConnByMServers() returns CM_ERROR_NOSUCHVOLUME because
the server list for the volume is empty. In this case, force
an update of the volume info and retry.
* Insert a missing cm_EndCallbackGrantingCall() which could
result in threads waiting for a callback to complete to never
be woken up.
* In the persistent cache, there is a maximum number of volume
entries. Allow volume entries to be reused if the maximum
number have been allocated and their reference count is zero.
* If we already have a dead virtual connection object, don't
ignore the fact that additional dead vc objects must be taken
care of.
* Removed a deadlock condition introduced in the summer of 2004.
Do not hold mx locks around calls to RX functions including
rx_NewCall(). That is what reference counts on the rx_connection
object are for.
* Fixed an initialization error in afslogon.dll which could
result in random behavior including a failure to terminate
the NPLogon function. (1.3.80b)
* Fixed an error preventing the use of SMB authentication
on Windows 2000. (1.3.80a)
Since 1.3.77:
* Updated CellServDB to Public CellServDB 16 Mar 2005
* Fixed DNS AFSDB queries to ensure that "csail" cannot be
misinterpretted as "csail.mit.edu" when the resolver
is configured to append ".mit.edu" to failed DNS queries.
* Fixed another case in which the client would replace connections
to the host when it was not necessary. In this case the
situation would occur if cryptall was on and the connections
were unauthenticated due to lack of an appropriate token.
* OpenAFS for Windows has failed to work at sites which are
utilizing a cross-realm trust between an MIT/Heimdal realm
and a multi-domain Windows forest when the workstation being
accessed is not located in the root domain. This is caused
by a bug in the workstation which was triggered after the
introduction of Windows 2003 Server. When the bug is triggered,
the workstation attempts to authenticate users to afsd_service.exe
by contacting the Domain Controller instead of using the
LSA loopback authentication mechanism.
One of the reasons this bug occurs is because the workstation
does not have a reliable way of knowing that the service whose
netbios name is "AFS" is located on the workstation. This will
be fixed starting in Longhorn Beta 1 by Microsoft. The
"BackConnectionHostNames" registry value will be used to
indicate that the authentications to that service name should
be performed using the loopback authentication mechanism.
In the meantime, when Logon Caching is enabled, we can force
afsd_service.exe to authenticate using the logon cache before
contacting the Domain Controller. This will work with both
password and smart card based logons.
* The allDown logic in cm_ConnByMServers() was wrong. The allDown
flag should not be cleared if a volume's server reference is
marked as "offline". In the case where all of the volume's
servers are either "down" or the volumes are "offline", we want
cm_Analyze() to process the condition as CM_ERROR_ALLOFFLINE
instead of as CM_ERROR_TIMEDOUT. In fact, CM_ERROR_TIMEDOUT
should never occur in practice.
In the case of CM_ERROR_ALLOFFLINE, cm_Analyze() will sleep for
5 seconds, clear the server down and volume busy flags, and
then force an update from the VLDB. This allows the client to
update the location of a volume if the reason for it being
marked offline is because it is being moved. Calls to
cm_ConnByMServers() will be retried either until success or
the RDRtimeout period is reached.
* Correct the Power Management code behavior in response to
Standby, Suspend, and Shutdown events. Instead of flushing the
buffers associated with the mounted SMB submounts, simply write
all buffers which are marked dirty.
* Added support for "TheseCells" to afslogon.dll. "TheseCells"
provides a list of cells other than the default cell for which
tokens should be obtained using the default Kerberos principal.
This functionality is logon domain specific and is only available
when using KFW for authentication.
* Fixed FindNext Invalid Handle error caused by over agressive
attempts at garbage collecting dirSearch entries when the
dirSearch ID wraps from the maximum value to 1.
* Add support for registry defined server preferences for VLDB and
File servers. See registry.txt for details.
* Increased default cache size to 96MB and default number of cache
entries to 10,000.
* Fixed refCount leaks related to directory lookups and pioctl calls.
* Callbacks revoked during a race condition with an attempt to
obtain the same callback no longer result in an inaccessible
stat cache entry.
* New command line tool:
afsdacl : Set or reset the DACL to allow starting or stopping
the afsd service by any ordinary user.
Usage : afsdacl [-set | -reset] [-show]
-set : Sets the DACL
-reset : Reset the DACL
-show : Show current DACL (SDSF)
* IP addresses are no longer queried once at startup. Instead IP
addresses are obtained as needed. Loopback adapter addresses are
no longer published to the server.
* Pay attention to the MIT KFW registry configuration for automatic
importation of MSLSA credentials. (SOFTWARE\MIT\Leash32,MsLsaImport)
* Fix (once again) case-sensitive comparisons which was apparently
broken in 1.3.72.
* Activate support for vos listvol -format
* Population of the "AFS Client Admins" group failed on non-English
versions of Windows because the name of the "Administrators" group
is localized. Now we lookup the name of the group by using the
Administrators Group Alias Relative ID.
* Multiple cell token acquisition within afscreds.exe was broken in the
case of Kerberos 5 cross realm authentication. Instead of contacting
the KDC associated with the cell's realm, afscreds would obtain a
token from the KDC of the user's realm. This would result in
"invalid kvno errors" while authenticating to the AFS servers.
Unauthenticated access would work. The symptoms would vary based upon
whether or not the VLDB servers had been contacted using unauthenticated
connections prior to the user obtaining tokens.
* The list of ACL entries was becoming corrupted. It appears as if
an ACL when it expires was not being appropriately cleaned up. In
fact, it was left in the list of ACLs associated with the scache
entry.
* Changed the default @sys name list to "x86_win32 i386_w2k i386_nt40"
for 32-bit x86 systems. The default for itanium will be "ia64_win64"
and "amd64_win64" for amd 64-bit processors.
* When executing executables, Windows provides a mechanisms by which
the normal search for DLLs can be bypassed. If a file foo.exe is
being executed and there exists a file or directory "foo.exe.local"
(the contents are ignored if a file), then Windows will search for
DLLs first in the "foo.exe.local" directory and second in the
directory in which the "foo.exe" file is located.
Previous releases of OAFW would improperly return
STATUS_NOT_A_DIRECTORY instead of the expected
STATUS_OBJECT_PATH_NOT_FOUND. This would cause the Windows SMB client
to terminate the search for the DLL causing the execution of the
application to fail.
* Rework the reference counting of smb_vc_t objects. The references
stored in the waiting locks were not counted. This could result in
an assertion if the reference count drops to zero.
* "fs wscell" returns "Freelance.Local.Root" for the workstation cell
when running in freelance mode.
* Added support for persistent caching of file contents, stat entries,
volume data, cell data, and name lookup data. When the registry
value, "NonPersistentCaching", is not defined or set to 0 the file
specified by "CachePath" becomes a persistent cache file. The
size of the cache file is computed at run time. It has a maximum
size of approximately 1.2GB. The contents of the file will be
validated according to the rules specified by the "ValidateCache"
registry value.
* Added support for UUIDs. UUIDs are kept for as long as the
cache file is intact.
* Added cmdebug.exe and support cache manager debugging callback
interfaces in afsd_service.exe.
Usage: cmdebug -servers <server machine> [-port <IP port>] [-long]
[-addrs] [-cache] [-help]
Where: -long print all info
-addrs print only host interfaces
-cache print only cache configuration
* Symlinks to \\AFS\[all\]... will now be treated the same as
symlinks to /afs/.... However, please use /afs/... as otherwise
the symlinks won't work on Unix.
* Correct a problem with local CellID allocation for cells whose
server list is obtained via DNS instead of a CellServDB file. If
the DNS information expires the CellID assigned to the entry will
be changed. This causes all of the cm_scache_t objects which refer
to the old cellID number to become useless. Attempts to access
files or directories with cm_scache_t objects using the old cellID
will fail since the server list cannot be obtained.
* Correct deadlock condition in cm_EndCallbackGrantingCall which can
be triggered if the volume referenced by the cm_scache_t object is
offline.
* The AFS Service needs to respond to SERVICE_ACCEPT_SHUTDOWN messages
in addition to SERVICE_ACCEPT_STOP.
* Move RPC shutdown until after the SMB and RX shutdown procedures
complete. Block until RPC shutdown is complete.
* Modify afslogon.dll (integrated logon) to wait for service if its
state is START_PENDING. If the timeout period occurs, reset to
the retry interval and not the sleep interval.
* When renewing the server list for a cell obtained via DNS AFSDB
records, the cm_cell_t entry must be removed from the list of all
cells. Otherwise, the list of cells will be corrupted.
* In the dcache and scache modules, use the cm_scache_t dataVersion
instead of the cm_fakeDirVersion.
* Synchronize fs.c with the unix version.
* The variable used to determine whether a file or virtual memory
mapped cache is used was not properly initialized to a default
value. If the registry setting "NonPersistentCaching" was not
set, the choice would be random. Properly initialized to be
"file".
* The memory mapped view was never unmapped before closing the file
at service shutdown. This is now properly cleaned up.
* Default location of Cache file is now %TEMP%\AFSCache
Since 1.3.76:
* A bug affecting new installations of 1.3.75/76 would result in
the creating of incorrect mountpoints in the freelance root.afs
volume for the default cell.
If "fs \\afs\all\<cellname>" lists a volume name of "root" instead
of "root.cell", you have been affected by the bug. To correct the
problem, execute the following commands:
fs rmmount \\afs\all\<cellname>
fs rmmount \\afs\all\.<cellname>
fs mkmount \\afs\all\<cellname> root.cell <cellname>
fs mkmount \\afs\all\.<cellname> root.cell <cellname> -rw
Since 1.3.75:
* A bug has been fixed in the auto-generation of Freelance root.afs
symlinks which produced random entries in the \\AFS\all directory.
* Support has been added for multi-homed servers
Since 1.3.74:
* Added a new registry value, "StoreAnsiFilenames", which can be used
to force the use of ANSI character sets instead of OEM Code Pages.
This feature is useful when users require the ability to create
filenames with 8-bit characters and need to access the files from
both Latin-1 based Unix systems as well as from Windows.
Activation of this feature will prevent access to files stored with
8-bit OEM characters.
* Shutdown all SMB threads in a synchronized manner when stopping the
service.
* There is currently a maximum cache size of 1.3GB. The limit is imposed
by the largest contiguous block of unused memory within the 2GB process
space which can be assigned to the memory mapped file. Unfortunately,
when the executable digital signature verification code is activated
Windows sees fit to further segment the process memory which in turn
reduces the size of the maximum cache file to less then 800MB. If
larger cache sizes are desired, a new registry value should be set:
HKLM\SOFTWARE\OpenAFS\Client (DWORD) "VerifyServiceSignature" = 0x0
Setting this value will disable the runtime verification of digital
signatures on afsd_service.exe and the afs dlls which it loads. It
will not disable the the version number check on those same files.
The signature verification is not a security messure and is only meant
to enhance the ability to afsd_service.exe to detect potential
destablizing mixtures of DLLs from incompatible distributions.
Added code to auto-disable the signature verification check if
the desired cache size is greater then 700MB.
* Windows' WinTrustVerify(WIN_SPUB_ACTION_PUBLISHED_SOFTWARE) is
used to verify the validity of the afsd_service.exe binary
as well as each of the AFS DLLs loaded by the service. Not only
must the digital signature be valid but the signatures of the
DLL must be signed by the same entity as the service.
* Implement new functions: cm_freelanceMountPointExists and
cm_freelanceSymlinkExists. Use them along with other validity
checks in cm_freelanceAddMount and cm_freelanceAddSymlink to
ensure that name collisions do not occur and that empty strings
are not valid file names.
A symlink may not have a name which would resolve to a valid
cell name. Doing so would prevent access to the cell.
* Add missing cm_HoldSCacheNoLock call to Freelance mount point
re-initialization code. The reference counts of the fake root.afs
volume scache object(s) would become invalid when the mount point
or symlink lists were altered.
* Add registry entries to provide mappings from the afsdsbmt.ini
to the new locations for applications which count on the use
of the old Profile file APIs. These apps are likely to fail
if the user does not have administrator privileges and the
registry is locked down.
* The afs_config.exe submounts dialog had two errors.
First, attempts to remove entries failed because the registry
key was being opened without KEY_WRITE privileges.
Second, when editing a submount entry, changing the name
would add a new key and leave the original one in place.
Now the original submount will be removed if its name is
changed.
* In recent months there have been several incidents in which
users have experienced problems starting or accessing
afsd_service.exe and after significant effort has been spent
it has turned out that they have two versions of AFS on the
machine or an inconsistent set of DLLs.
Code has now been added to afsd_service.exe which will walk
the list of modules loaded by afsd_service.exe and validate
that the version of the AFS DLLs matches the version of the
afsd_service.exe executable. If they do not match the service
will not start.
* When Freelance mode is enabled and there is no registry
key HKLM\SOFTWARE\OpenAFS\Client\Freelance, afsd_service.exe
will attempt to import the afs_freelance.ini file contents.
If the file does not exist, it was creating a dummy file
with a r/o and r/w entry for the default cell and then
importing those values.
This process has been changed. The temporary file is no
longer created. Also, both the OpenAFS Client install
directory as well as %WINDIR% are checked for previous
afs_freelance.ini files.
* Added support for VL_GetEntryByNameN(). Still need to add
support for VL_GetEntryByNameU() for multi-homed support.
* Fix a deadlock situation in afscreds.exe when canceling an
auto-generated Obtain Tokens dialog
Since 1.3.72/73:
* Fix the locking of objects during Directory Searches in the
SMB/CIFS server. The failure to properly lock the reference
counts was resulting in the premature freeing of smb_dirSearch_t
objects while they were still in use by the SMB/CIFS client.
This does not solve the "Invalid Handle" problem.
* Fix Find Cell By Name pioctl call to return a valid cell
name for the Freelance fake root.afs volume.
"Freelance.Local.Root".
* Fix the Explorer Shell Extension Symlinks->Add operation.
The dialog template was missing and the link destination
string was too short.
* Add support for symlinks to Freelance root.afs volume
Stored at HKLM\SOFTWARE\OpenAFS\Client\Freelance\Symlinks
<number> = "<linkname>:<relative-path>."
Use symlink.exe to create, list, or remove
* Remove the fallback to the use of KFW's KRB4 library when obtaining
tokens. We never obtain KRB4 tickets.
* Fix AFS Client Configuration Control Panel to support new SysName
protocol.
* Fix a bug in afsd_service.exe which could result in the SysName
not being read from the registry.
Since 1.3.71:
* Add code to block the issuance of AFS tokens by aklog.exe or
afscreds.exe when the Kerberos 5 principal name contains a dot.
* Modify the IsAdmin() function to always treat the local SYSTEM
account as an AFS client administrator. Affects fs.exe and
afs_config.exe.
* Modify the internal handling of Quota Exceeded errors
* Upgrade all reference count fields in the Windows cache manager
and the osi library to use unsigned long instead of signed short.
A similar fix has been applied to the afs rpc (rx) library.
* fix the Windows cache manager to prevent it from replacing the
rx_connection object associated with the cm_conn_t object on each
and every operation if "fs crypt" was set. This explains the
dramatic performance difference when crypt is used vs clear.
The problem: 'cryptall', a boolean flag indicating whether or not
"fs crypt" is set, was being compared to the rx_connection
cryptlevel which is either rxkad_clear:0 or rxkad_crypt:2.
1 != 2 and therefore the rx_connection was always destroyed
and replaced on each and every operation.
Lock the cm_conn_t object around every call to RXAFS_xxxx functions.
It is not safe for the cm_conn_t object to not be locked because
rx_DestroyConnection might be called from another thread if:
- the user's tokens have changed or expired
- the crypt mode has changed
This fix appears to have also taken care of the problems associated
with Overlapped Writes resulting in Delayed Write errors.
* fix NSIS installer's AdminGroup.exe to properly create and
remove groups when given -create or -remove. The string comparison
test was wrong.
* fs sysname now accepts a list of sysname values
* added a new registry value HKLM\SOFTWARE\OpenAFS\Client "IoctlDebug"
DWORD which when set to a non-zero value will cause error message
text to be output to stderr from the pioctl() routine. Useful in
debugging failures of fs.exe, tokens.exe, etc.
* added a test to the power management code to only perform a
flush operation if there is at least one network adapter which
is not a loopback adapter.
* Fix bug in loading of registry value HKLM\SOFTWARE\OpenAFS\Client
"EnableKFW". This value will not be read if the key
HKCU\SOFTWARE\OpenAFS\Client exists; even if the "EnableKFW"
value under that key does not.
* provide mechanisms to force the use of krb524d for Kerberos 5
ticket to AFS token conversion. For afslogon.dll and afscreds.exe
there is a new registry value "Use524" and for aklog.exe a new
command line parameter "-m".
* Fix the pattern matching algorithm to properly match patterns
ending with a '*'.
* smb_ReceiveCoreRename() was factored to produce smb_Rename()
which is used by both the original function and the new
smb_ReceiveNTRename(). smb_ReceiveNTRename() supports the
creation of HardLinks in addition to Renaming. smb_Link()
is a new function which creates HardLinks via cm_Link().
cm_Link() is a new vnodeops function which creates links
using RXAFS_Link().
smb_ReceiveNTRename() does not support the File Copy and
Move Cluster Information operations described in its interface.
ReceiveNTRename is under documented in CIFS-TR-1p00_FINAL.pdf.
* When opening files via symlinks, we should follow the symlinks
until we reach the actual file stat cache entry. The stat cache
entry of the file should then be stored in the FID instead of
stat scache entry of the symlink.
* return bad operation errors for all unimplemented functions
even if we do not know the functions exist.
* Log bad packets and unknown operation packets to the trace log
* Map CM_ERROR_BADOP to STATUS_NOT_SUPPORTED instead of
0xC09820FF
* Update list of known CIFS operations to include all those listed
in CIFS-TR-1p00_FINAL.pdf.
* Modify the handling of HKLM\SOFTWARE\OpenAFS\Client\Submounts
to support the REG_EXPAND_SZ type.
Since 1.3.70:
* A new Windows authorization group "AFS Client Admins" is now
created and populated with the members of the "Administrators"
group. The group is used to determine which accounts on the
machine may be used to modify the AFS Client Configuration via
the UI and command line tools. afs_config.exe, fs.exe,
* Modify the WinLogon Logoff Event Handler to query NT4 domain
controllers for the remote profile path if Active Directory
services are not available.
* Fix aklog.exe to not add the AFS ID to the username
* PTS registration of new users to foreign cells has been added to
afscreds.exe
* The cm_Daemon thread is used to perform checks for
down servers, up servers, volumes, callback expirations,
lock maintenance and token expiration. Due to a gaff in
larger integer division the thread never performed any
work. Instead the current time computation would always
be less then the trigger times. This had an adverse affect
on the client's ability to maintain communication with servers,
keep volumes up to date, and flush user tokens and acls
when they have expired. This was broken when the 1.3 branch
was modified to support VC7 which no longer included
largeint.lib
* An initialization problem with the Freelance code was
detected while fixing the callbackRequest. The cm_rootSCachep
object is obtained during afsd_InitDaemons() but the callback
information is incomplete. The callback information will not
be obtained until cm_MergeStatus is called from within
cm_GetCallback. Unfortunately, cm_SyncOp did not properly
test for the conditions under which the callback information
must be obtained.
* Reports have been filed indicating that callbacks were
being lost. An examination of the code indicated that the
cm_server_t objects were not being properly reference
counted by the cm_scache_t and cm_callbackRequest_t objects.
In particular, the cm_server_t objects may have been freed
from beneath the cm_conn_t objects.
All of the reference counting is now done via the functions:
cm_GetServer
cm_GetServerNoLock
cm_PutServer
cm_PutServerNoLock
this improves the ability to track the referrals.
Each cm_BeginCallbackGranting Call now allocates a reference
to the cm_server_t. The cm_EndCallbackGrantingCall either
frees the reference or transfers it to the cm_scache_t
cbServerp field. These are then appropriately tracked
through the cm_Analyze call.
* Ensure that the dnlc hash table is the same size as the
dir name hash table (as per original author's note).
Increase the dnlc CM_AFSNCNAMESIZE to a multiple of 8
for compatibility with 64-bit systems.
* fix smb_ApplyV3DirListPatches to properly apply the hidden
attribute to dotfiles when the infoLevel < 0x101 and
cm_SyncOp has failed.
* Fix the Freelance registry initialization code. There
was a possibility that some systems could end up with
garbage in the registry during a clean install.
Since 1.3.66:
* file and directory names beginning with "." will now be given the
hidden attribute when the volume access is anonymous. this matches
the behavior when the volume access is via an authenticated user.
* Added a change monitor to the HKLM\SOFTWARE\OpenAFS\Client\Freelance
key. When a change occurs mark the root.afs data as invalid and
for it to be reloaded on the next access. This allows administrators
to modify the mount point list without restarting the service.
The freelance client used to provide a fake modification time for
the root.afs volume data and its mount points of 7/09/2001 14:24 EDT.
Added code to extract the last modification time of the Freelance
registry key and use that instead. The time now represents the
most recent mount point change.
* PTS registration of new users to foreign cells has been added to
aklog.exe
* Additional Cache Control and Credential Manager options have been
added to the WiX installer. See deployment guide for details.
* The CachePath setting is now optionally a REG_EXPAND_SZ type
* The WiX installer has been upgraded. Version 2.0.1927.1 is now
required.
* The loopback installation code may have had a problem updating the
%ETC%\HOSTS file which could have resulted in a premature failure.
Work around code has been added for the case where the file cannot
be deleted.
* The default max chunksize was increased from 15 (32K) to 17 (128K)
because Windows sends 64K blocks when using overlapped writes.
* The default number of server threads was increased from 4 to 25 to
better handle overlapped writes.
* The "AfscredsShortcutParams" registry value was not being properly
loaded by afscreds.exe. Therefore, the default value was always being
used instead of the value set by the installer.
* Windows XP provides downgrade attack detection to prevent an attacker
from being able to force the use of NTLM simply by disrupting
communication with the KDC. This attack cannot exist between the
Windows CIFS client and the AFS Client Service. Therefore, when a
downgrade has been detected the afs pioctl library will force the
establishment of a new CIFS connection using NTLM.
* A locking error was discovered surrounding all references to volume
server lists within the cm_cell.c source file.
* The logged into Windows username was incorrect on Terminal Server
machines.
* A new registry value "NonPersistentCaching" was added to the service
parameters key. When set to a non-zero value, the afs cache is stored
in the Windows paging file. There are two limitations to choosing
this option:
1. when persistent caching is implemented it won't work with
this flag set since there will be nothing to persist.
2. with this flag set the initial paging allocation cannot be
changed while the service is running
* An initialization bug was discovered in aklog.exe which affected users
who have a domain name for their afs servers which could not be mapped
to a realm
Since 1.3.65:
* afs_config.exe now validates cell names against DNS in addition
to the CellServDB file.
* In order to allow the freelance client to connect to a volume with ID
equal to 1 on the default cell we changed the fake root.afs volume ID
once again. This time we choose 0xFFFFFFFF. In addition, we change
the cell ID of the fake root.afs volume from 1 to 0xFFFFFFFF as well.
It will now be impossible for a volume ID to match that of another
cell unless the client is connected to 0xFFFFFFFD cells. That should
be enough room for growth.
* Fix "fs mkmount" command to work with UNC paths and when
started from non-AFS drives. It is now possible to create a mount
point in the freelance fake root.afs volume with the command
fs mkmount \\AFS\all\<directory-name> <volume-name> <cellname>
For example,
fs mkmount \\AFS\all\openafs.org root.cell openafs.org
fs mkmount \\AFS\all\.openafs.org root.cell openafs.org -rw
* The algorithm used to re-attempt access to the servers associated with
a volume has been altered to properly address the case in which all
servers have been marked down. The previous algorithm did not reset
the server's down flags so the servers were never actually retried.
This caused a problem with active volumes if the network connectivity
was lost as could be the case with a network cable removal, wireless
drop, or laptop hibernation. With the fix volume access is restored
almost instantenously when network connectivity becomes available.
* Support for SMB/CIFS browsing has been added to the AFS Client Service
SMB server. It is now possible to use "NET VIEW \\AFS" to obtain a
listing of AFS submounts and freelance mount points. Support for
NETSHAREENUM, NETSHAREGETINFO, NETSERVERENUM2, NETSERVERGETINFO
significantly enhances the behavior of AFS volumes within the Explorer
Shell. For instance, "AFS" now shows up as server in the Explorer
with each submount or freelance mount point visible as a share.
The right click menu in each folder now works with full functionality
on a consistent basis.
* The network provider can be configured to have different behavior
depending on the domain that the user logs into. These settings are
only relevant when using integrated login. A domain refers to an
Active Directory (AD) domain, a trusted Kerberos (non-AD) realm or the
local machine (i.e. local account logins). The domain name that is
used for selecting the domain would be the domain that is passed into
the NPLogonNotify function of the network provider. (see registry.txt
for details)
* Added a new registry value [HKCU\SOFTWARE\OpenAFS\Client]
"Authentication Cell" which may be used to specify a default
authentication cell for afscreds.exe which is different from
the default cell for the AFS Client Service daemon.
* Added a Logoff WinLogon Event Notification function to afslogon.dll.
afslogon.dll moved to %WINDIR%\System32\.
New registry entries added to register the dll for Winlogon events.
The logoff event will now force a call to ktc_ForgetAllTokens()
using the context of the user being logged off as long as the
user's profile is not loaded from within AFS. If the profile
was loaded from AFS we can't release the tokens since the Logoff
event is triggered prior to the profile being written back to
the its source location. This is now performed in an XP SP2
safe manner.
* Windows XP SP2 Internet Connection Firewall interoperability
has been added.
* The %WINDIR%\afsdsbmt.ini contains four sections:
Submounts, Drive Mappings, Active Maps and CSC Policies.
The Submounts and CSC policies are now stored in the registry under
[HKLM\SOFTWARE\OpenAFS\Client\Submounts]
[HKLM\SOFTWARE\OpenAFS\Client\CSCPolicy]
The Drive Mappings and Active Maps are stored in the registry under
[HKCU\SOFTWARE\OpenAFS\Client\Mappings]
[HKCU\SOFTWARE\OpenAFS\Client\Active Maps]
There is no automatic migration of this data as it would be impossible
to consistently migrate data to user profiles which may not be active
when the machine is updated.
* The %WINDIR%\afs_freelance.ini contains lists of mountpoints for the
fake root.afs volume. For the same reasons as for the cellservdb file,
this information should not be in %WINDIR%. This information is now
kept under the registry key
[HKLM\SOFTWARE\OpenAFS\Client\Freelance]
The data from the afs_freelance.ini file will be automatically
migrated to the registry on first execution of afsd_service.exe
* Keeping the CellServDB file in the location %WINDIR%\afsdcell.ini is
troublesome for several reasons. One, it is confusing for those who
expect the file to be named "CellServDB" instead of "afsdcell.ini".
Two, this file is not a Windows Profile formatted file. Three,
applications should not be reading or writing to %WINDIR%. It causes
problems for Windows Terminal Server.
The new location of CellServDB will be the OpenAFS Client install
directory which is by default C:\Program Files\OpenAFS\Client and can
be determined by querying the registry for
[HKLM\SOFTWARE\TransarcCorporation\AFS Client\CurrentVersion]PathName
The existing afsdcell.ini will be migrated by the NSIS installer.
The Wix installer must still be updated to do the same.
* Change NSIS installer to use DNS by default; to remove Integrated Logon
High Security mode; and to add Terminal Services compatibility registry
entries to allow the OpenAFS tools to find the afsdcell.ini and other
configuration files in %WINDIR%.
* Add support for authenticated SMB connections. This will remove
the need for high security mode in most situations. Both NTLM
and Extended Security (GSS SPNEGO) modes are supported. Effectively,
only NTLM can be used even though Kerberos is now supported. The
reason is that it is not possible to construct a service principal
which is unique to each individual machine.
SMB Extended Auth does not work on XP SP2 unless one of two registry
modifications are made:
(1) To disable the check for matching host names on loopback connections
set this key. This does not require a reboot:
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa]
"DisableLoopbackCheck"=dword:00000001
(2) To add the AFS SMB/CIFS service name to an approved list. This
does require a reboot:
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa\MSV1_0]
"BackConnectionHostNames"=multi-sz "AFS" "MACHINE-AFS"
afsd_service.exe will automatically add the current Netbios Name
to the BackConnectionHostNames list and then temporarily disable
the loopback check for one cycle of startup/shutdown of the service.
We assume most folks do not start/stop without a reboot so this
will be adequate in most cases.
* Fix security hole in afslogon.dll which allowed passwords to be
sent in clear text to the KDC in a misformed principal name.
* Fix cm_GetCell() to properly handle expired dns entries
without crashing
* If Freelance mode is active and the afs_freelance.ini
file does not exist, do not create an empty file.
Instead create a file containing ro and rw mountpoints
to the default cell using the standard conventions.
* Modify the Freelance support to handle the ability
to create rw mount points in the fake root.afs volume.
* Changed the RPC mechanism used for token setting from
named pipes to local. Use of named pipes can be restored
by setting the environment variable AFS_RPC_PROTSEQ to
"ncacn_np".
Named pipes were required when a Windows 9x system was
using a NT system in gateway mode which is incompatible
with our use of local loopback adapters.
* In afscreds.exe, if a username of the form user@REALM is
specified and no password is specified, do not perform a
kinit operation. Only perform the aklog functionality.
* Add a new registry value which allows the number of processors
on which afsd_service.exe executes to be restricted. Valid
values are 1..numOfProcessors
HKLM\SYSTEM\CurrentControlSet\Services\TransarcAfsDaemon\Parameters
(DWORD) MaxCPUs
Since 1.3.64:
* A second MSI based installer option is now available.
* Fixed Kerberos 5 kinit functionality in afscreds.exe to properly
request tickets for user/instance@REALM instead of just user@REALM
* Modify the Power Management Notify routine to wait for the Hard Dead
timeout period instead of a fixed 19 seconds. With the longer timeout
periods Hibernation and Standby could never succeed when network
connectivity is not available.
* The following fs.exe commands are now restricted to Administrator:
- checkservers with a non-zero timer value
- setcachesize
- newcell
- sysname with a new sysname list
- exportafs
- setcell
- setserverprefs
- storebehind
- setcrypt
- cscpolicy
- trace
setting the default sysname for a machine should be done via the
registry and not via "fs sysname".
* NSIS installer adds options to install Debugging Symbols
and the Microsoft Loopback Adapter; the user is now also
given the ability to select the afscreds.exe startup options.
* Build system modified to generate symbols for FREE (aka RELEASE)
builds as well as CHECKED (aka DEBUG) builds
* Sites which have a volume ID of 0x20000001 assigned to their
root.cell volumes have been experiencing problems with accessing
the root.cell volume of their cell when Freelance mode has been
active. This was because 0x20000001 was assigned to the fake
root.afs volume created by freelance. The fake volume id is
now set to 0x00000001 to prevent conflicts.
* The timeout logic in the AFS Client Service has been wrong
for sometime. It is based on two different assumptions.
First, the SMB client timeout is a fix value as was the case
with OS/2 Lan Manager. This assumption is incorrect. The
SMB timeout in Windows is a dynamic value computed based upon
a fixed minimum timeout to which is added time based upon the
size of the request and the performance characteristics of
the connection. Second, it is the responsibility of the
SMB Server to enforce the timeout requirements of the client.
This is untrue. The SMB Server cannot be expected to know
the requirements of the client. More importantly, if the
SMB server uses the SMB client timeout as a value to restrict
its behavior as an RX client, the performance characteristics
of the local SMB session would be used to prematurely terminate
WAN connections with significantly different performance
characteristics.
The timeout logic has therefore been modified in the following
manner:
. the Lan Manager Workstation (SMB) Session Timeout is used only
as a basis for configuring the Connection Dead Timeout
and Hard Dead Timeout values. The Connection Dead Timeout
must be at least 15 seconds longer than the SMB Timeout
and the Hard Dead Timeout must be at least double the
Connection Dead Timeout.
. New registry entries have been added to allow the Connection
Dead Timeout and Hard Dead Timeout values independent of the
Lan Manager Workstation Session Timeout
. The test to enforce the SMB Client Timeout has been removed.
One of the side-effects of removing the enforcement of the SMB
Client Timeout is that regardless of whether or not the SMB client
is available to receive the response (and how would the SMB server
know) the RX protocol response can be used to update the AFS
Client Service state for ready access by future SMB client
requests.
This should be the end of the "Server paused or restarting messages"
* Add support for arbitrary UNC paths to the pioctl() support.
This enables the fs commands as well as the AFS Shell Extension
to work correctly when UNC paths are being used.
* Fix afscreds.exe (by updating afskfw.lib) to search for cells via
DNS if the cell configuration cannot be determined via CellServDB
* Add debug info to test whether CM_BUF_WAITING or CM_SCACHE_WAITING
are ever set more than once at a time
* Fix the management of lists of cm_cell_t structures when using
DNS to lookup cell information. The previous code would fail to
reuse the same cellID for a cell if DNS was used more than once
for a given cell name. When the ttl expired, a single cm_cell_t
could be inserted into the cm_allCellsp list more than once
producing a loop. In addition, the vlServerp list belonging to
the cell was not freed resulting in improper refCounting of the
servers.
* Add DNS support to cm_IoctlNewCell() which previous only examined
the CellServDB file
* Add cm_FreeServer() function and call it from cm_FreeServerList()
to properly garbage collect cm_server_t objects
* Add numVCs variable to smb.c to track the number of smb_vc_t
objects created and use it to initialize the vcID field which
previously was set to 0 in all objects resulting in FindByID
collisions.
* Fixed DNS lookups to work consistently throughout the OpenAFS
product instead of just from within the afsd_service.exe
* Added a runtime check to ensure that AFS Client Service SMB
Server is accepting connections before attempting to mount
global drives.
* Read IP addresses for volume servers out of the CellServDB
file if gethostbyname() on the hostname fails.
* Fix getcellconfig() to populate both the Hostnames as well
as IP addresses when loading cell data via DNS
* Increase the Connection Dead Time to 50 from 20 seconds
Increase the Hard Dead Time to 120 from 40 seconds
(matches the Unix values)
* Fixed an assertion validating the number of allocated NCBs
* Fixed the build environment to consistently build for
Windows 2000 and above. (APPVER = 5.0)
* Fixed rx_debug to properly validate the receipt in incoming
data with select() and recvfrom(). Do not copy data out of
the socket buffer unless success is indicated.
Since 1.3.63:
* afsd_service.exe will now display a message box to the
desktop when it terminates due to an IP Address Change.
* installer no longer deletes AFS Server configuration data
on uninstall
* installer generates a warning dialog if the RPC service
is not properly configured
* installer compressed with lzma instead of bzip2
* afsd_service.exe shutdown crash solved once and for all
* reference counting of smb_vc_t data structures improved
* name space collision of smb_fid_t event objects corrected
* the output of "fs memdump" is now written to
%WINDIR%\TEMP\afsd_alloc.log
* the file TaAfsApp_1033.dll is now properly installed allowing
the User Manager to start
* a new algorithm is used for computing filename pattern matches
* afscreds.exe now accepts user names containing instance
fields.
* Fix the Directory Name Lookup Cache to be case-sensitive.
This is crucial in environments in which a Windows client
is accessing a directory with more than one filename that
differs only by case. If the directory contains "FOO"
and "Foo". You want "DEL Foo" to delete the correct one.
We still have a problem in that "DEL foo" will delete a
random filename. This will be addressed in a future release.
* Fix afscreds.exe -M option (renewMaps) to work when High
Security mode is off. Also, remember to disable the ActiveMap
flag in afsdsbmt.ini when a drive mapping is removed.
* Updates to NSIS installer script. AFS Server configuration
data will not be destroyed on un-install or re-install.
Use a better compression algorithm.
* afslogon.dll now uses KFW to obtain tokens when available
* afslogon.dll when given an all uppercase username will
attempt to authenticate with both the uppercase name
and an all lowercase variation
* DST modification removed. The fix appears to make things
worse after a reboot of the machine.
* fs.exe: added "cscpolicy" which is used to
change client side caching policy for AFS shares
Usage: fs cscpolicy [-share <AFS share>] [-manual] [-programs]
[-documents] [-disable] [-help]
* Several uninitialized variables have been initialized
* It is now possible to obtain tokens using cross realm
Kerberos within afscreds.exe:
cell: dementia.org
user: jaltman@ATHENA.MIT.EDU
password: xxxxxxxx
Will obtain a cross realm ticket for jaltman/DEMENTIA.ORG@ATHENA.MIT.EDU
will will in turn be used to obtain afs@DEMENTIA.ORG.
The resulting token will be stored with the display name
jaltman@ATHENA.MIT.EDU@dementia.org
* aklog.exe has been added to the client
Usage: aklog [-d] [[-cell | -c] cell [-k krb_realm]]
[[-p | -path] pathname]
[-noprdb] [-force]
[-5 | -4]
-d gives debugging information.
krb_realm is the kerberos realm of a cell.
pathname is the name of a directory to which you wish to authenticate.
-noprdb means don't try to determine AFS ID.
-5 or -4 selects whether to use Kerberos V or Kerberos IV.
(default is Kerberos V)
No commandline arguments means authenticate to the local cell.
Since 1.3.62:
* All of the resource files have been restructured to adhere to
a set of rules IBM implemented for loading string resources.
These rules had either been forgotten or were not discovered
by folks working on the OpenAFS sources. The end result was
memory corruption. This is primary item which was preventing
the AFS Server from working.
* Increased the size of the maximum ticket size stored in a token
from 344 bytes to 12,000. Increased the buffers used to convey
messages between the pioctl() caller and the SMB Server from
1000 bytes to 12,512. The code appeared to have been writing
above the top of the stack by quite a few number of bytes.
(The increased ticket size is necessary for the next item.)
* When obtaining AFS Tokens via KFW, krb524 is no longer required.
Instead the raw Kerberos 5 ticket is used in its entirety. This
is extremely important as it allows us to use pure Kerberos 5 KDCs
as the source of the AFS authentication. The use of up to 12,000 byte
tickets will allow tickets produced by all versions of Microsoft
Active Directory to be used.
- create a user account.
- designate it DES only
- disable pre-auth
- specify its UPN to be "afs@realm"
- assign a SPN of "afs/cellname" to the UPN with setspn.exe
* Do not enforce the funky 8dot3 pattern matching rule that the first "."
is special when using long file names. (you must use "*.*" and not "*")
Instead only enforce it when performing 8dot3 searches.
* Fixed the DST problem with creation times being set one hour ahead
* Fixed the problem when using \\afs\cell-alias. For example,
\\afs\uncc instead of \\afs\uncc.edu. Do not a new cell struct
for the alias name; instead simply expand the name. One of the
symptoms of this problem was a loss of acquired tokens.
* Fixed the AFS Shell Extension. The Symbolic Link menu was empty
of strings. (Only English strings provided.)
* Fixed the installer to properly replace in use files.
* Fixed the build system to cleanup generated component version files
* The release build compiled with MSVC 6.0 compiler to avoid the
afsd_service.exe shutdown crash. This does not solve the problem
but simply avoids it for the time being.
Since 1.3.61:
* fix afslogon.dll to not corrupt memory when High Security mode
is not used.
* fix afsd_service.exe to not attempt to restore the stack when
an exception occurs. (not safe in multi-threaded programs)
* fix uninstaller to properly remove the CRT and MFC DLLs
* remove a Message Box from afscreds.exe when getcellconfig()
fails on a kerberos realm which is not a cell
The following is a list of changes to the OpenAFS for Window client
since 1.3.60.
* "fs setserverprefs" will leave afsd service deadlocked
* "vos listaddrs" will core dump
* installer sets the appropriate keys to support Integrated Logon
* installer disables the "Find Lana by Name" functionality as it
was causing headaches for many users
* fix the intermittent crash of the power management thread when
shutting down the AFS Client Service
* optimizes the obtain drive mount list functionality which is
executed every time the mount tab in afscreds.exe and afs_config.exe
are refreshed. (this happens a lot)
* fix the service shutdown logic. add the STOP_PENDING state
and do not accept additional service events after we declare
ourselves STOPPED.
The following is a list of changes to the OpenAFS for Window client
since 1.2.10.
* flexelint was run against the source tree and hundreds (perhaps
thousands) of corrections were applied to ensure prototypes
were in use; types were used consistently; variables were
initialized; unused variables were removed; etc.
* A wide variety of instrumentation was added including the
ability to produce a stack trace from within afsd_service.exe
when it crashes.
* Dynamic configuration of the RDRtimeout value based upon the
LanMan Workstation Session Timeout
* The mount root no longer needs to be called "/afs". This
is now set by a registry value "MountRoot" within the key
HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
* The cell list is now only read out of afsdcell.ini when the
file changes instead of each time a cell is resolved.
* Thread synchronization was added to cm_server.c and ktc_nt.c
* All calls to GlobalAlloc()/GlobalFree() were replaced with
calloc()/free(). The Global functions were needed on Windows 3.x
but have caused a variety of problems on the Win32 platforms.
Avoiding them is highly recommended by several Microsoft
Knowledgebase articles
* Support for Symbolic Links added to the AFS Shell Extension
* Added a registry value "OverlayEnabled" to determine if
Shell Extension Overlays should be enabled.
HKLM\SYSTEM\CurrentControlSet\Services\TransarcAFSDaemon\Parameters
* New Build system to support VC6, VC.NET, VC.NET2003 compilers and
separate trees for checked and free builds. Build system supports
a custom directory src\WINNT\extra which can be used as a grafting
location of organization specific additions to the build tree.
* New installer built using NSIS 2.0.
* Named all kernel objects in order to allow them to be monitored
with tools such as SysInternals' ProcExp.exe.
* Introduced new EventLog framework for AFSD
* Introduced Power Management interface to AFSD for Standby and
Hibernate modes to allow cache to be flushed prior to network
disconnect
* Utilize Win32 DNSQuery API instead of internal routines. This
allows DNS SRV queries to be sent to all current domain name
servers. Not just one specified in an INI file. DNS is now
always activated.
* "NetbiosName" registry value may be used to specify a fixed
Netbios Name such as "AFS" to be used instead of "HOSTNAME-AFS"
when the loopback adapter is in use. If you need to use the
old notation with a loopback adapter installed specify a registry
entry of
"NetbiosName" REG_EXPAND_SZ = "%COMPUTERNAME%-AFS"
* Refactor all modules which depend on LAN Adapter and NetbiosName
determination in a new library: lanahelper.lib. This allows for
consistent behavior throughout the product.
* Move the afsd.log and afsd_init.log files to the directory specified
by the "TEMP" environment variable. This is usually %WINDIR%\TEMP
for services. Added the Date to the log entries.
* New registry value "RxMaxMTU" used to limit the size of the RX
packets sent by the AFS Client Service to the Server. In order
to enable OpenAFS to work across the Cisco IPSec VPN the packet
size must be restricted to 1264 or smaller. The latest NSIS
installer sets a value of 1260 by default.
* New registry value "RxNoJumbo" to disable the use of Jumbo Rx
packets. This is not needed in order to work across the Cisco
VPN but might be needed for other network environments. This
value is not set by the NSIS installer.
* New registry value "HideDotFiles" is used to apply the Hidden
attribute to files whose names begin with a '.'. This value
is set by the NSIS installer.
* New registry value "MaxMpxRequests" allows the maximum number
of multiplexed sessions to be configured at run time. This
value is not set by the NSIS installer. The default value is
50.
* New registry value "MaxVCPerServer" allows the maxmimum number
of VCs per server to be configured at run time. This value is
not set by the NSIS installer. The default value is 100.
* New registry value "AllSubmount" allows the "all" submount to
be disabled by setting its value to 0x00.
* Allow cells names to be valid mount points
\\<netbiosName>\<cellname>
* Store the active state of drive mappings in order for afscreds.exe
to restore them upon startup
* Add exception handling to generate a Stack Trace to the afsd_init.log
file if one happens to occur.
* Add lots of logging to help detect the cause of invalid SMB packets
* Enable Kerberos for Windows to be used to obtain AFS Tokens via
conversion of Kerberos 5 "afs" service tickets. Supports auto-
renewal of expiring tokens as long as afscreds.exe is running.
* New afscreds.exe command line options:
-A = autoinit
-M = renew drive maps
-N = ip address change detection
-Z = unmap drives
* New registry value "EnableKFW" in {HKCU,HKLM}SOFTWARE\OpenAFS\Client
determines whether or not MIT Kerberos for Windows should be used
to obtain tokens via Kerberos 5 tickets.
* New registry value "AfscredsShortcutParams" in
{HKCU,HKLM}SOFTWARE\OpenAFS\Client
determines the command line parameters to be specified when "fixing"
the AFS Shortcut in the user's startup folder.
* The "ShowTrayIcon" registry value has been moved from
HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
{HKCU,HKLM}SOFTWARE\OpenAFS\Client
* The <cell name> registry values used to store the token expiration
reminders have been moved from
HKLM\Software\TransarcCorporation\AFS Client\AfsCreds to
{HKCU,HKLM}SOFTWARE\OpenAFS\Client\Reminders
* Obtain the Logon User Name from the Explorer key when available
* new text document doc\txt\winnotes\registry.txt lists all registry
values used by OpenAFS (excluding the AFS Server)
* BUG: rx_securityClass objects were not properly reference
counted and were never freed.
* BUG: reduce the number of conditions under which CM_ERROR_TIMEOUT
would be generated. The existence of a server does not imply
that it is not down. If all of the servers for a cell are down
return CM_ERROR_NOSUCHVOLUME instead. This prevents the Explorer
Shell from hanging.
* BUG: the directory name lookup cache failed to free the entries
in the cache when the name cache entries cycled. The entries
in the cache would become dereferenced without being freed.
* BUG: fs setserverprefs could be executed without Administrator
privileges
* BUG: the number of allocated NCB objects (100) exceeded the number
which could actually be waited upon by the kernel (64). Any objects
which were utilized above the limit could never have event completions
detected.
* BUG: smb_username_t objects were not being reference counted and
were not properly freed.
* BUG: smb_tid_t objects could under unusual circumstances be freed
before they were no longer referenced.
* BUG: smb_fid_t object pointer were frequently used even when
their value could be NULL. They were not properly released and
therefore they were never freed.
* BUG: smb_packet_t data structures were not completely initialized
upon creation
* BUG: when Rx produces a CM_ERROR_NOIPC error do not return "Access
Denied" because that causes the Explorer Shell to try again until
access is obtained. Instead return "Remote Resources" which allows
the shell to move on and treat the error as transient.
* BUG: when initializing the NCBreturns structure, separate Event objects
were created for each NCB although a single Event object was supposed
to be shared by all.
* BUG: smb_dirSearch_t objects were not being properly referenced counted
or freed.
* BUG: smb_tran2Packet_t objects were not being properly referenced
counted or freed.
* BUG: directory path creation did not handle the case of multiple
directories requiring creation in one attempt
* BUG: SMB requests which required an Extended Response were ignored.
This prevented some files from being written to AFS volumes.
* BUG: character strings were being freed even after they were
inserted into in use data structures
* BUG: inconsistent usernames were used when High Security mode was
enabled. (there is still much to do in this area)
* BUG: pioctl() calls which require out of band RPC operations were
susceptible to race conditions when performed by multiple processes
* BUG: memory allocation and deallocation crossed instances of the
C Runtime Library producing memory leakage and corruption in
afscreds and the client configurator.
Reference in New Issue View Git Blame Copy Permalink