dll/3rdparty/libtirpc/ChangeLog

   1 2008-11-19  Steve Dickson <steved@redhat.com>
   2
   3     * Version 0.1.10 released.
   4
   5 commit 32ec5931e3debf208972d5146578f08dc113a9b6
   6 Merge: 338af7f... 92cf0dd...
   7 Author: Steve Dickson <steved@redhat.com>
   8 Date:   Mon Nov 17 12:26:22 2008 -0500
   9
  10     Merge branch 'master' of git://git.infradead.org/~steved/libtirpc
  11
  12 commit 92cf0dde310ca341a2f29ff66b19eeb9994a649a
  13 Author: Ian Kent <ikent@redhat.com>
  14 Date:   Tue Oct 28 11:19:07 2008 -0400
  15
  16     Fixed a warings the IPV6 client routines
  17
  18     Signed-off-by: Steve Dickson <steved@redhat.com>
  19
  20 commit 338af7f9f00e096b65a6d823f885c4eeaf1d1f8c
  21 Author: Steve Dickson <steved@redhat.com>
  22 Date:   Mon Oct 27 12:46:54 2008 -0400
  23
  24     __rpc_taddr2uaddr_af() assumes the netbuf to always have a
  25     non-zero data. This is a bad assumption and can lead to a
  26     seg-fault. This patch adds a check for zero length and returns
  27     NULL when found.
  28
  29     Signed-off-by: Steve Dickson <steved@redhat.com>
  30
  31 commit d9a5ae7079d001a9e3b9b384f9153f591a7158bd
  32 Author: Olaf Kirch <okir@suse.de>
  33 Date:   Tue Sep 30 15:10:43 2008 -0400
  34
  35     Fix __rpc_getconfip
  36
  37     __rpc_getconfip is supposed to return the first netconf
  38     entry supporting tcp or udp, respectively. The code will
  39     currently return the *last* entry, plus it will leak
  40     memory when there is more than one such entry.
  41
  42     This patch fixes this issue.
  43
  44     Signed-off-by: Olaf Kirch <okir@suse.de>
  45     Signed-off-by: Steve Dickson <steved@redhat.com>
  46
  47 commit 6c487efe74adb5c29f7bee5bd51b3ebef4968f7d
  48 Author: Olaf Kirch <okir@suse.de>
  49 Date:   Tue Sep 30 15:09:06 2008 -0400
  50
  51     Fix getpeereid
  52
  53     getpeereid fails because it uses an incorrect getsockopt call to obtain
  54     the peer credentials on a AF_LOCAL socket.  This in turn will cause all
  55     RPC services to be registered with rpcbind to show up as having been
  56     registered by "unknown".
  57
  58     This has a serious impact on security - a service owned by "unknown"
  59     can essentially be unregistered (and thus replaced) by anyone.
  60
  61     Signed-off-by: Olaf Kirch <okir@suse.de>
  62     Signed-off-by: Steve Dickson <steved@redhat.com>
  63
  64 commit 851b0f5c6dca22d634603f03f0a5e3e35c6db867
  65 Author: Olaf Kirch <okir@suse.de>
  66 Date:   Tue Sep 30 15:08:07 2008 -0400
  67
  68     svc_getcaller_netbuf macro seems broken
  69
  70     I haven't found any documentation, but the comment in the header
  71     file seems to suggest that svc_getcaller_netbuf should return the
  72     xp_rtaddr netbuf. Returning the address of the socket descripor
  73     seems to be wrong at any rate.
  74
  75     Signed-off-by: Olaf Kirch <okir@suse.de>
  76     Signed-off-by: Steve Dickson <steved@redhat.com>
  77
  78 commit d94b92d5125242ce595c1baf42a1e6d1004b7756
  79 Author: Olaf Kirch <okir@suse.de>
  80 Date:   Tue Sep 30 15:06:54 2008 -0400
  81
  82     Introduce __rpc_set_netbuf helper
  83
  84     The RPC code contains a number of places where a netbuf
  85     is initialized with some data. All the mem_alloc/memcpy
  86     stuff is open-coded. Introduce a helper function and
  87     convert the code.
  88
  89     Signed-off-by: Olaf Kirch <okir@suse.de>
  90     Signed-off-by: Steve Dickson <steved@redhat.com>
  91
  92 commit da5f9861ea3bae59c8eead26d38334721caa9f0a
  93 Author: Olaf Kirch <okir@suse.de>
  94 Date:   Tue Sep 30 15:05:20 2008 -0400
  95
  96     Kill map_ipv4_to_ipv6
  97
  98     After the change to svc_vc.c performed in the previous patch,
  99     this function is no longer needed.
 100
 101     Signed-off-by: Olaf Kirch <okir@suse.de>
 102     Signed-off-by: Steve Dickson <steved@redhat.com>
 103
 104 commit 59c374c4b507aeca957ed0096d98006edf601375
 105 Author: Olaf Kirch <okir@suse.de>
 106 Date:   Tue Sep 30 15:04:17 2008 -0400
 107
 108     Fix xp_raddr handling in svc_fd_create etc
 109
 110     Currently svc_fd_create tries to do some clever tricks
 111     with IPv4/v6 address mapping.
 112
 113     This is broken for several reasons.
 114      1. We don't want IPv4 based transport to look like IPv6
 115         transports. Old applications compiled against tirpc
 116         will expect AF_INET addresses, and are not equipped
 117         to deal with AF_INET6.
 118      2. There's a buffer overflow.
 119                 memcpy(&sin6, &ss, sizeof(ss));
 120         copies a full struct sockaddr to a sockaddr_in6 on
 121         the stack. Unlikely to be exploitable, but I wonder
 122         if this ever worked....
 123
 124     Signed-off-by: Olaf Kirch <okir@suse.de>
 125     Signed-off-by: Steve Dickson <steved@redhat.com>
 126
 127 commit 628788c1cc84c86ee4cb36ee5d4fe8954e90fca5
 128 Author: Steve Dickson <steved@redhat.com>
 129 Date:   Tue Sep 16 11:32:31 2008 -0400
 130
 131     - Fixed version-info in src/Makefile.am to reflect the correct version
 132     - Fixed some of warnings in: src/auth_time.c, src/clnt_dg.c and
 133         src/clnt_raw.c
 134     - Added some #ifdef NOTUSED around some code in src/rpbc_clnt.c
 135       that was not being used...
 136
 137     Signed-off-by: Steve Dickson <steved@redhat.com>
 138
 139 commit 9e7ba0c7a02031294fefadfbca42b3dd5f2d841f
 140 Author: Olaf Kirch <okir@suse.de>
 141 Date:   Tue Sep 16 08:46:29 2008 -0400
 142
 143     Fix for taddr2addr conversion bug of local addresses
 144
 145     When converting af_local socket addresses in taddr2uaddr, an incorrect
 146     sizeof() would result in a truncated path string. As a result,
 147     rpcbind will report the local /var/lib/rpcbind address to clients
 148     as "/v" on a 32bit machine.
 149
 150     Signed-off-by: okir@suse.de
 151     Signed-off-by: Steve Dickson <steved@redhat.com>
 152
 153 commit ea9f048761d0b9a2ab6310bffa07351f0b04d8c5
 154 Author: Olaf Kirch <okir@suse.de>
 155 Date:   Tue Sep 2 12:11:15 2008 -0400
 156
 157     Always make IPv6 sockets V6ONLY
 158
 159     Assume you have a netconfig file looking like this:
 160
 161     udp        tpi_clts      v     inet     udp     -       -
 162     udp6       tpi_clts      v     inet6    udp     -       -
 163     ...
 164
 165     a call to svc_tli_create(... &someaddr, "udp") will fail to create an
 166     IPv6 server socket. The problem is that on Linux, passive IPv6 sockets
 167     will also accept packets/connections from IPv4, and will simply map
 168     the sender's address to an IPv6 mapped IPv4 address. So if you want to
 169     bind both a UDPv4 and UDPv6 socket to the same port, this will fail with
 170     EADDRINUSE.
 171
 172     The way to avoid this behavior is to change the socket to V6ONLY,
 173     which tells the kernel to avoid the autmatic mapping.
 174
 175     The change proposed in the patch below does this. I *think* this is
 176     a good place to do this, as it will also fix applications that do not
 177     use svc_tli_create() - such as rpcbind, which creates the sockets on
 178     its own using __rpc_nconf2fd.
 179
 180     I think this also improves portability, as BSD code assumes BSD
 181     behavior, where this mapping does not occur either.
 182
 183     Signed-off-by: Olaf Kirch <okir@suse.de>
 184     Signed-off-by: Steve Dickson <steved@redhat.com>
 185
 186 commit 95c8f7227e6b15f2e430d7b87dadc95b2acd4a61
 187 Author: Olaf Kirch <okir@suse.de>
 188 Date:   Tue Sep 2 12:09:39 2008 -0400
 189
 190     Fix incorrect sizeof() in __rpc_getbroadifs
 191
 192     __rpc_getbroadifs returns bad broadcast addresses on 32bit
 193     machines because when copying the broadcast addresses, ite
 194     applies the sizeof() operator to a pointer to a sockaddr,
 195     rather than the sockaddr itself.
 196
 197     Signed-off-by: Olaf Kirch <okir@suse.de>
 198     Signed-off-by: Steve Dickson <steved@redhat.com>
 199
 200 2004-10-13  Antoine Fraticelli  <antoine.fraticellie@bull.net>
 201
 202         * Version 0.1 released.
 203
 204 2005-01-07  Gilles Quillard  <Gilles.Quillard@bull.net>
 205
 206         * Version 0.1.5  Fix problems links to the use of Kerberos.