1 OpenAFS Release Notes - Version 1.6.0pre2
2 _________________________________________________________________
3 All systems: Minor bugfixes.
4 _________________________________________________________________
6 ADDITIONAL CHANGES IN 1.6.0PRE2
10 - Documentation updates.
12 - Don't stop Rx keepalives after an ackall is received, avoiding
13 spurious connection timeouts. (128848)
15 - Don't retry Rx calls on channels returning busy errors. (128671)
17 - vos will not die with a double free error at command completion.
19 - Properly enable Rx connection hard timeouts.
21 - Initialize rx_multi lock before use.
23 - Avoid spurious crashes when initializing in "backup" client.
27 - Check for /afs existance before starting, unless -nomount is specified.
29 - Avoid a potential panic when using /afs/.:mount syntax.
31 - Avoid a panic in memcache mode due to missing CellItems file.
35 - Attempt to recovery more quickly from timed out volume release
38 - Auditing now properly byte order swaps IP addresses when printing.
40 - vos split now has improved error handling.
42 - Many changes to again support Windows fileservers.
44 - During volume removal, data removal speed improved.
46 - Improve CPU utilization during volume attaching by DAFS.
48 - In salvager check-only mode, avoid potentially fixing a vnode.
50 - Fix support for large (greater than 2gb) volume special files.
52 - Salvager will not crash if multiple or bad volume link tables
55 - Avoid erroneous full dump by remembering which sites were out of date
56 at the start of the release.
60 - Remove support for "Giant" lock as we no longer need to use it.
62 - Don't sleep with AFS GLOCK.
64 - Properly enable 64 bit long long support.
66 - Restore support for FreeBSD 7 (128612)
68 - Fix locking issues at shutdown.
72 - support through kernel 2.6.38.
74 - RedHat packaging now properly supports RHEL6.
76 - Use rx_Readv in cache bypass to improve performance.
78 - Properly handle 0-length replies during cache bypass operations.
80 - Properly handle non-contiguous readpage cache bypass operations.
82 - Do proper locking when transitioning to or from cache bypass.
84 - Avoid extra runs of vcache freeing routine. (128756)
88 - Check for unloaded kernel extensions when decoding AFS panics.
90 - Properly handle setpag errors. PAGs are not supported.
92 - Disable "get tokens at login" in prefs pane if AD authentication
97 - support through OpenBSD 4.8.
101 - Fix support for Solaris pre-10.
105 - afs_config will not longer set the Tray Icon State
106 in the registry if the checkbox is not present in
109 - AFS Explorer Shell Extension now works from folder
110 backgrounds. Overlays for mount points and symlinks
111 are present in the dll, but are not registered at present
114 - Do not use RankServerInterval registry value as the value for
115 PerformanceTuningInterval.
117 - When the data version of a mountpoint or symlink changes,
118 the target string in the cm_scache_t object must be cleared.
120 - "fs checkservers" now includes vldb servers in the output
121 and only lists multi-homed servers once. A multi-homed
122 server that has at least one up interface is no longer
123 considered to be down.
125 - When asynchronously storing dirty data buffers to the
126 file server ensure that (a) the cm_scache_t object and
127 the cm_buf_t object are for the same File ID so that
128 locking and signalling work properly; and (b) if the
129 FID no longer exists on the file server, do not panic,
130 just discard the buffer.
132 - When processing VNOVOL, VMOVED and VOFFLINE errors perform
133 server comparisons by UUID or address and not simply by
134 cm_server_t pointer. Otherwise, server failover may not
137 - Do not preserve status information for cm_scache_t objects
138 when the issuing server is multi-homed.
140 - Giving up all callbacks when shutting down or suspending
141 the machine is now significantly faster due to the use
142 of an rx_multi implementation. (This functionality is
143 still off by default and must be activated by a registry
146 - Race conditions were possible when updating the state
147 of the cm_volume_t flags and when moving the volumes
148 within the least recently used list.
150 - Ensure that the lanahelper library does not perform a
151 NCBRESET of each lan adapter when enumerating the
152 current network bindings. Correcting this permits OpenAFS
153 to work on Windows 7 when the network adapter settings
156 - Fix creation of mount points and symlinks as \\AFS\xxxx
162 - vos now properly deals with matching sites when servers are multihomed.
166 - Servers now marked down when GetCapabilities returns error.
168 - In-use vcache count is now properly tracked.
170 All server platforms:
172 - Fix ptserver supergroups support on 64 bit platforms.
174 - Demand attach salvaging doesn't use freed volume pointers.
176 - Properly hold host lock during host enumeration in fileserver.
180 - Fix socket termination on shutdown.
182 - Support for 7.2, 7.3, 7.4 and 8.2 included.
184 - References to vcaches are no longer leaked during root or reclaim.
188 - Define llseek handler to avoid ESPIPE error in 2.6.37.
190 - Mount interface replaces get_sb (new for 2.6.37, not yet required).
192 - RedHat init script allows deferring for a new binary restart.
194 - DEFINE_MUTEX replaces DECLARE_MUTEX for 2.6.37.
198 - Correct return value from setpag syscall.
202 - Bug fixes for issues introduced previously in 1.5 series.
206 - Switch to ioctl() syscall replacement for Solaris 11 since syscall 65