3 o Initial Snow Leopard port. Thanks to various people contributing patches in
4 the bugtracker. The new official version can only be compiled on Snow
5 Leopard but has been tested to work on all Tiger, Leopard and Snow Leopard
6 o Clean up unused locking code and switch to rwlocks even for simple mutexes,
7 which avoids a symbol incompatibility for Tiger and Leopard.
8 o Clean up compilation flags in the Makefiles.
11 o Adapt the former Leopard package to also be installable on Tiger systems.
12 This obsoletes the Tiger version, both Leopard and Tiger are now supported
16 o Protect the selwakeup() call by the lock. This fixes incorrect select()
17 behaviour, thanks to Roland Wendelin for reporting this.
18 o Fix tuntap_mbuf_queue::size initialization
19 o Use a proper wait condition for synchronization when detaching the network
20 interface. The old code would crash if the if_detached() handler was called
21 from a different thread than unregister_interface().
24 o Work around an issue in the Darwin kernel. When unregistering an interface,
25 addresses are not properly removed from the interface. This leads to
26 crashes and other problems when reusing the interface. Introduce an ugly
27 hack that tries to remove all interface addresses when shutting the
29 o Fix a small mbuf leak that could occur when the output queue was full.
30 Thanks to Oleg Dolgov for reporting this.
33 o Fix paths in the startup item postflight scripts
34 o Check if_ioctl arguments more defensively after a report of a panic after
38 o I have done a complete rework of the installer package generation. The
39 package is now edited in PackageMaker. The distribution package can still
40 be built from the commandline though.
41 o Fix incorrect permission & ownership of the installed files.
44 o Fix the permissions of the postflight scripts. Installer packages should work
46 o Drop the kmod and kmodc++ in the linker command, they seem to be unneeded
50 o Initial Leopard port, it's basically the latest Tiger version with some
51 Leopard-related fixes and s/Tiger/Leopard/g
52 o I have switched to a proper version management system (git) and could
53 remove some of the CVS hacks subsequently.
54 o The installation packages have been reworked a bit, they now install into
55 /System/Extensions and /System/StartupItems directly by using
59 o This version is not stable, it may crash, sorry.
60 o Universal binaries that run on ppc and intel macs.
61 o Adds tap MAC address randomization
63 o Better multicast support
64 o mDNSResponder workaround, so that the tap interfaces should get picked up
65 now. Note that we are fixing ifconfig/kernel behaviour here.
66 o All tapX and tunX devices are visible in /dev at all times, network
67 interfaces still created dynamically, though.
68 o Startup items moved to /Library/StartupItems
71 o Initial Tiger port. We now have KPI-style interfaces. I guess the Tiger
72 version is little slower than the Panther version because of all the
73 wrapping and hiding in the kernel.
74 o The kernel extensions moved to /Library/Extensions. That is the place where
75 non-Apple kexts are supposed to live.
78 o I added support in tun for AF prepending like some BSDs do. Thanks to Dennis
79 kSchneider for mailing the initial patch. You can also set the value of
81 o I finally found that major bug causing crashes (especially on multiprocessor
82 machines). It also caused a memory leak (lost mbufs), and might have caused
83 performance/througput/data-loss problems. Everyone is recommended to upgrade.
86 o I rewrote the common part concerning the tun and tap initialization and
87 cleanup. This should make the code more maintainable (less duplication).
88 o The devices now reinitialize to the state they were started in when they
89 are closed by an application. This concerns IP addresses for example.
90 o I changed the package building system to use PackageMaker.app in batch
91 mode. The packages also check for version 10.3 now, so nobody should be
92 able to install tun/tap on 10.2 using installer packages. Furthermore I
93 have sprinkled some warnings telling you not to use tun/tap on SMP machines
94 over the installation process ;-)
95 o Some minor locking fixes.
98 o Jamie Wood reported that the packet queue in the driver could be considered
99 empty even if there were packets in it. This was probably caused by a
100 synchronization problem that should be fixed now. People encountering
101 timeouts etc. should try the new version.
102 o I finally implemented support for changing the interface MTU. The driver
103 enforces the MTU when writing packets to the character device now. However,
104 packets coming from the kernel are not checked.
107 o Marcello Teodori told me that the tun driver wasn't working with openvpn.
108 The problem was the fcntl call, fixed that. Should work now. Thanks
110 o changed the tun driver not to prepend the address family field before each
111 and every packet (which is the behaviour of OpenBSD). As there is currently
112 only IPv4 and IPv6 support there is no problem with the standard tun
113 approach used on other OSes. This should make the driver much more
115 o Did a script and makefile support so that the installer packages can now be
116 built from the command prompt. Unfortunately this might break things
117 someday as I am not using the 'official' way to build the packages
118 o Cleaned up installer packages a little.
121 o initial version put online
122 o basic tun/tap support, tap working with qemu