|
commit e6a24206e53adfab3b150c498f05e32ad6b6c83c
Author: Simon Xu <xu.simon@oracle.com> Date: Thu Jan 15 15:39:35 2015 +0800 test_1_to_1_accept_close: also expect EACCES when accept on an established socket When SELinux is set to enforcing, errno may be EACCES when accept on an established socket. Signed-off-by: Simon Xu <xu.simon@oracle.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 45c59f29a46bed3d024c4962bacbbb08e3806b07 Author: Michele Baldessari <michele@acksyn.org> Date: Thu Oct 23 23:46:35 2014 +0200 Fix mailing list address Signed-off-by: Michele Baldessari <michele@acksyn.org> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 7aa589acd6f74d76f49f99609cd608047ff1c567 Author: Michele Baldessari <michele@acksyn.org> Date: Thu Oct 23 23:46:34 2014 +0200 Fix usage help for sctp_test Signed-off-by: Michele Baldessari <michele@acksyn.org> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 080c3be6d9371c0bfbd56b404807c27ced7c6cb5 Author: Michele Baldessari <michele@acksyn.org> Date: Thu Oct 23 23:46:33 2014 +0200 Fix recieved->received typos Signed-off-by: Michele Baldessari <michele@acksyn.org> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit d4635c8d664130de676a1285000c4a46f8217cc1 Author: Wang Weidong <wangweidong1@huawei.com> Date: Fri May 23 11:14:33 2014 +0800 lksctp-tools: make bind_test can do while disable IPV6 When disable IPv6, the bind_test can't work for IPv4. So add a variable 'domain' for support working for IPv4. Signed-off-by: Wang Weidong <wangweidong1@huawei.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 043ac46eb52a02ff25ee71dfc2d4d81a2520ebcf Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Feb 18 11:42:49 2014 +0100 Release lksctp-tools-1.0.16 Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit faf774528fc967eaadfc923554b6911ec8105aac Author: Matija Glavinic Pecotic <matija.glavinic-pecotic.ext@nsn.com> Date: Tue Feb 18 11:31:16 2014 +0100 Remove dependency to implementation-specific a_rwnd behavior test_timetolive and test_sctp_sendrecvmsq take advantage of the following lksctp characteristic: - SO_RECVBUF being set prior association is established will control initially advertised rwnd - Modifying SO_RECVBUF later will keep maximum rwnd at the initial value which is defined at the association startup Test cases use this fact in order to initially advertise small rwnd, increase buffer later, but keeping the initial rwnd as maximum one. However, we may at later time increase buffer, and implementation may advertise this to the partner. RFC4960 only forbids decrease of the buffer space: Advertised Receiver Window Credit (a_rwnd): 32 bits (unsigned integer) This value represents the dedicated buffer space, in number of bytes, the sender of the INIT has reserved in association with this window. During the life of the association, this buffer space SHOULD NOT be lessened (i.e., dedicated buffers taken away from this association); however, an endpoint MAY change the value of a_rwnd it sends in SACK chunks. Having in mind that rwnd reflects free space in receiver's buffer, implementation may advertise new value, and therefore test case should not rely on above defined behavior. There is no impact on functional behavior of test cases due to this change. This change was done in course of '[PATCH v2] net: sctp: Fix a_rwnd/rwnd management to reflect real state of the receiver's buffer', as with this patch behavior is changed from described to rwnd reflecting real buffer state. Signed-off-by: Matija Glavinic Pecotic <matija.glavinic-pecotic.ext@nsn.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 0600c8968cc2dea04cbf13ec739216e2939d08fe Author: Markos Chandras <markos.chandras@imgtec.com> Date: Thu Dec 12 12:11:15 2013 +0000 lib/connectx: Fix build for prefixed architectures We need to take into consideration the __USER_LABEL_PREFIX__ for architectures that define it when creating symbol aliases. Fixes the build for the Meta(metag) architecture. [...] src/apps/sctp_test.c:801: undefined reference to `_sctp_connectx' [...] [ Improved .symver handling to make it more readable, also x86_64 objdump alias log attached below. -dborkman ] Before: $ objdump --syms ./src/lib/.libs/libsctp.so | grep @ | grep sctp 0000000000000df0 g F .text 0000000000000032 sctp_connectx@ 0000000000000e30 g F .text 0000000000000029 sctp_connectx@VERS_2 0000000000000e60 g F .text 0000000000000103 sctp_connectx@@VERS_3 0000000000000df0 g F .text 0000000000000032 sctp_connectx@VERS_1 After: $ objdump --syms ./src/lib/.libs/libsctp.so | grep @ | grep sctp 0000000000000df0 g F .text 0000000000000032 sctp_connectx@ 0000000000000e30 g F .text 0000000000000029 sctp_connectx@VERS_2 0000000000000e60 g F .text 0000000000000103 sctp_connectx@@VERS_3 0000000000000df0 g F .text 0000000000000032 sctp_connectx@VERS_1 Signed-off-by: Markos Chandras <markos.chandras@imgtec.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit c43ead6daa0b9ae5763d6cf3d21357a789651417 Author: Simon Xu <xu.simon@oracle.com> Date: Tue Dec 10 16:19:22 2013 +0800 lksctp: Avoid using file descriptor 0 to get ENOTSOCK On some systems, errno will not be ENOTSOCK when operating on file descriptor 0 in an ssh session: test_1_to_1_accept_close.c 2 TBROK : accept with invalid socketerror:-1, errno:22 test_1_to_1_addrs.c 2 TBROK : sctp_getladdrs with invalid socket error:-1, errno:95 test_1_to_1_connect.c 2 TBROK : connect with invalid socket error:-1, errno:22 test_1_to_1_recvfrom.c 2 TBROK : recvfrom with invalid socket count:0, errno:9 test_1_to_1_recvmsg.c 2 TBROK : recvmsg with invalid socket count:0, errno:9 test_1_to_1_send.c 2 TBROK : send with invalid socket count:15, errno:9 test_1_to_1_sendmsg.c 2 TBROK : sendmsg with invalid socket count:-1, errno:106 test_1_to_1_shutdown.c 2 TBROK : shutdown with an invalid socket error:0, errno:9 test_1_to_1_socket_bind_listen.c 6 TBROK : bind() with invalid socket descriptor error:-1, errno:22 test_1_to_1_sockopt.c 2 TBROK : setsockopt with an invalid socket error:-1, errno:95 test_getname.c 9 TBROK : getsockname on an invalid socket error:0 errno:9 This patch gets a non-socket file descriptor by creating a temporary file to replace file descriptor 0. Signed-off-by: Simon Xu <xu.simon@oracle.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit b7ad1afc16cb0c5da793554ed59d8f16842473e3 Author: Wang Weidong <wangweidong1@huawei.com> Date: Mon Dec 2 13:55:56 2013 +0100 tests: fix test_1_to_1_sockopt sock buff test case When I do the test_1_to_1_sockopt, I got a Segmentation fault: DUMP_CORE test_1_to_1_sockopt.c: 343. I used the kernel is 3.12. The problem is that the SOCK_MIN_SNDBUF is (TCP_SKB_MIN_TRUESIZE * 2) while TCP_SKB_MIN_TRUESIZE is 2048 + SKB_DATA_ALIGN(sizeof(struct sk_buff)) The SKB_DATA_ALIGN I test is 256, so I make the sndbuf_val_set to 2304. Signed-off-by: Wang Weidong <wangweidong1@huawei.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 630ce8e1b21c8fbc0dd191e130288ed233059507 Author: Daniel Borkmann <dborkman@redhat.com> Date: Wed Oct 30 16:22:45 2013 +0100 sctp_status: fix usage of max_stream parameter Hangbin reported the following error: remote:addr=127.0.0.1, port=distinct, family=2 local:addr=127.0.0.1, port=distinct32, family=2 Starting tests... socket(SOCK_STREAM, IPPROTO_SCTP) -> sk=3 bind(sk=3, [a:127.0.0.1,p:distinct32]) -- attempt 1/10 connect(sk=3) Client: Sending packets.(1/10) sendmsg(sk=3, assoc=0) 17768 bytes. SNDRCV(stream=1 flags=0x1 ppid=846930886) sendmsg(sk=3, assoc=0) 6250 bytes. SNDRCV(stream=2 flags=0x1 ppid=1714636915) sendmsg(sk=3, assoc=0) 23634 bytes. SNDRCV(stream=3 flags=0x1 ppid=424238335) sendmsg(sk=3, assoc=0) 5195 bytes. SNDRCV(stream=4 flags=0x1 ppid=1649760492) sendmsg(sk=3, assoc=0) 7978 bytes. SNDRCV(stream=5 flags=0x1 ppid=1189641421) sendmsg(sk=3, assoc=0) 22715 bytes. SNDRCV(stream=6 flags=0x1 ppid=1350490027) sendmsg(sk=3, assoc=0) 16883 bytes. SNDRCV(stream=7 flags=0x1 ppid=1102520059) sendmsg(sk=3, assoc=0) 10724 bytes. SNDRCV(stream=8 flags=0x1 ppid=1967513926) sendmsg(sk=3, assoc=0) 125 bytes. SNDRCV(stream=9 flags=0x1 ppid=1540383426) sendmsg(sk=3, assoc=0) 2133 bytes. SNDRCV(stream=10 flags=0x1 ppid=1303455736) *** sendmsg: Invalid argument *** The problem is that we use stream numbers that are larger than the default ones. So in the case the user requests a particular maximum stream number, tell that to the kernel so that we won't end up with an -EINVAL when sending. Reported-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 05c6d334a0ca4f3a489ab0031daa37ba4796a208 Author: Hangbin Liu <liuhangbin@gmail.com> Date: Sun Oct 27 10:56:23 2013 +0800 sctp_status: server side should also receive msg_cnt times As client side send msg_cnt times for each repeat. Server will exit too early if we only receive one time for each repeat and leave client keep sent, which makes client could not exit normally. Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 121b86050d044987c8fb6b199cbcf732ccc82764 Author: Hangbin Liu <liuhangbin@gmail.com> Date: Mon Oct 28 17:42:31 2013 +0800 sctp_status: add missed parameter description and remove unused option Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 54e5063c15e503492fdceb6f9473f504b324c883 Author: Hangbin Liu <liuhangbin@gmail.com> Date: Sun Oct 27 10:55:49 2013 +0800 test_1_to_1_sockopt: we have changed the minimum value to 2048 Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 23e0481e8f24e0c050bc0049fdf06c64fb15d424 Author: Hangbin Liu <liuhangbin@gmail.com> Date: Fri Jul 12 15:01:25 2013 +0800 Move bzero to memset Use memset as bzero has been deprecated Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit de18ef5b1f878835769173ce1edc8d019c67c51b Author: Hangbin Liu <liuhangbin@gmail.com> Date: Wed Jul 10 18:13:29 2013 +0800 Remove duplicate sys/errno.h As sys/errno.h only include errno.h, remove or change it to errno.h Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit cd5dbfa35a175511bb21b1bae797ed6e6312c8c8 Author: Daniel Borkmann <dborkman@redhat.com> Date: Sun May 12 11:06:13 2013 +0200 Release lksctp-tools-1.0.15 Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit e9a8682a1d251829e055b23884babe05529ae4b3 Author: Nicolas Dichtel <nicolas.dichtel@6wind.com> Date: Fri Apr 26 14:17:15 2013 +0200 do not returns an error if AM_SILENT_RULES does not exist On my host: ./bootstrap + libtoolize --force --copy libtoolize: putting auxiliary files in AC_CONFIG_AUX_DIR, `bin'. libtoolize: copying file `bin/ltmain.sh' libtoolize: putting macros in AC_CONFIG_MACRO_DIR, `m4'. libtoolize: copying file `m4/libtool.m4' libtoolize: copying file `m4/ltoptions.m4' libtoolize: copying file `m4/ltsugar.m4' libtoolize: copying file `m4/ltversion.m4' libtoolize: copying file `m4/lt~obsolete.m4' + aclocal configure.ac:23: warning: macro `AM_SILENT_RULES' not found in library + autoheader + automake --foreign --add-missing --copy + autoconf configure.ac:23: error: possibly undefined macro: AM_SILENT_RULES If this token and others are legitimate, please use m4_pattern_allow. See the Autoconf documentation. Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit d9d906453927dd7fb3b6e52e05cb2b62a13b2b7d Author: Daniel Borkmann <dborkman@redhat.com> Date: Fri Apr 19 14:40:46 2013 +0200 sctp_status: fix printstatus output sctp_status has a pretty broken implementation, i.e. in printstatus. At least fix the stringification and check against status without hardcoding stuff. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 9eb4c271c3b561d0141adb1e31c3ffc3725a1630 Author: Daniel Borkmann <dborkman@redhat.com> Date: Thu Apr 11 15:57:18 2013 +0200 withsctp: fix quoted command line arguments withsctp breaks quoted command line arguments. Trivial example: $ withsctp echo "a b c" a b c This is really bad with sshd: $ withsctp /usr/sbin/sshd -D -e -o "Port 4560" command-line line 0: missing port number. After the fix, sshd works as expected: withsctp /usr/sbin/sshd -D -e -o "Port 4560" Server listening on 0.0.0.0 port 4560. Server listening on :: port 4560. Reported-by: Mikko Markus Torni <mikko.markus.torni@iki.fi> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 31d23431133e62d20d759af54fa8aa00406bfc85 Author: Daniel Borkmann <dborkman@redhat.com> Date: Fri Apr 5 12:05:23 2013 +0200 Release lksctp-tools-1.0.14 Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 41251182e9f228835b226eb581ea1d8a7d27b613 Author: Neil Horman <nhorman@tuxdriver.com> Date: Tue Apr 2 15:07:59 2013 -0400 sctp: Add new spinfo state values to enumeration A while back the kernel added two enumeration values to sctp_spinfo_state. Unfortunately, those were never added to the user space header file, and as a result we get inconsistent behavior when reading spinfo state. This patch adds the missing fields in, and corrects the behavior. Signed-off-by: Neil Horman <nhorman@tuxdriver.com> CC: William Reich <reich@ulticom.com> CC: Daniel Borkmann <dborkman@redhat.com> CC: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 5275d4820c087434b30319f0b4fc0e24aba624f2 Author: Daniel Borkmann <dborkman@redhat.com> Date: Wed Mar 6 17:49:57 2013 +0100 test_sctp_sendrecvmsg: fix bug in msg_flag passing It must be ensured that this is zeroed out before passing to recvmsg. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 19f561b36be254801ffc6cb6a4f57874c6544d64 Author: Hangbin Liu <liuhangbin@gmail.com> Date: Tue Feb 26 10:03:59 2013 +0800 test_1_to_1_threads: remove unused variable for t_recv and t_send I think this is enough, we can't give NULL to pthread_create as we need to make sure there is one thread(I use the first one) only run t_send in case of test huang for ever. Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit a2765e638f4f0b5c49b836927c4d670c197b58a6 Author: Hangbin Liu <liuhangbin@gmail.com> Date: Tue Feb 26 10:03:58 2013 +0800 test_1_to_1_threads: fixup pthread hung by giving an array instead of a pointer As we give a pointer arg to pthread_create, some thread's cnt will be set to 1 when we run into next loop, which may cause all the threads blocked by t_recv. If we still want to use a pointer to avoid compile warning, then an array should be a good choice. Here I'd like to analyse the logic of function relay(). As Daniel said, this test's purpose is communications across different threads, which would like: Thread 1: send(sock_client), exit Thread 2: recv(sock_server), send(sock_client), exit ... Thread X-1: recv(sock_server), send(sock_client), exit Thread X: recv(sock_server), exit But this may cause test huang as we don't know the order of threads. e.g. Thread 1: send(sock_client), exit Thread 2: recv(sock_server), send(sock_client), exit Thread X: recv(sock_server), exit Thread 3: recv(sock_server), send(sock_client), hung ... Thread X-1: recv(sock_server), send(sock_client), hung If we receive first and put send at last. It may also cause hung like: Thread X: send(sock_client), exit Thread X-1: recv(sock_server), send(sock_client), exit Thread 1: recv(sock_server), exit Thread X-2: recv(sock_server), send(sock_client), hung ... Thread 2: recv(sock_server), send(sock_client), hung So, keep relay()'s logic like the following should be the best choice: Thread 1: send(sock_client), exit Thread 2: recv(sock_server), send(sock_client), exit ... Thread X-1: recv(sock_server), send(sock_client), exit Thread X: recv(sock_server), send(sock_client), exit Then there will be no hung and we can keep communications across different threads. Thus, we also don't need if (id == THREADS -1) now. Signed-off-by: Hangbin Liu <liuhangbin@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 2266bb3ea3f7f334947c5382b6a9db75a1a2e1b8 Author: Diego Elio Pettenò <flameeyes@flameeyes.eu> Date: Sat Feb 23 23:10:27 2013 -0800 build: fix out-of-tree build of withsctp. Signed-off-by: Diego Elio Pettenò <flameeyes@flameeyes.eu> commit 02df1af3734225481ec5247e9b9c0a40eb0a7994 Author: Daniel Borkmann <dborkman@redhat.com> Date: Thu Jan 24 14:00:04 2013 +0100 build: only use AM_CPPFLAGS instead of a mixture of AM_CFLAGS and AM_CPPFLAGS Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 27437ba73170937035f561e3bec6d65a7ad66f86 Author: Sébastien Luttringer <seblu@seblu.net> Date: Thu Jan 24 12:41:49 2013 +0100 build: Modernize autotools Allow compilation with last version of autotools - libtool 2.4.2 - m4 1.4.16 - autoconf 2.69 - automake 1.13.1 - make 3.82 Signed-off-by: Sébastien Luttringer <seblu@seblu.net> create mode 100644 configure.ac delete mode 100644 configure.in commit aa9787aab00b0857494ca8a1f1b9c63862a72838 Author: Daniel Borkmann <dborkman@redhat.com> Date: Wed Jan 23 15:27:55 2013 +0100 Release lksctp-tools-1.0.13 Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 1c9fe66640e43464985696287b0498150c75c90d Author: Daniel Borkmann <dborkman@redhat.com> Date: Wed Jan 23 11:55:16 2013 +0100 libsctputil: libtool: do not use -version-info for static, internal util lib Fixes: CCLD libsctputil.la libtool: link: warning: `-version-info/-version-number' is ignored for convenience libraries Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 398f377fbce6cf7c994d525ab0471cb5bc8d832c Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 16:42:44 2013 +0100 sctp_darn: fix __u64 format string issues on ppc64 / x86_64 On 64-bit powerpc, __u64 is defined to be unsigned long rather than unsigned long long. This causes compiler warnings every time we print a __u64 value with %PRIu64. Since we cannot change or define our own u64 type in this case, because it is within a structiure from the kernel, we just cast it to uint64_t. This behaviour can be reduced down to the following program: #include <stdio.h> #include <inttypes.h> #include <netinet/sctp.h> int main(void) { __u64 foo = 1; printf("%" PRIu64 "\n", foo); printf("%" PRIu64 "\n", (uint64_t) foo); printf("__u64:%zu versus uint64_t:%zu\n", sizeof(__u64), sizeof(uint64_t)); return 0; } On x86_64: $ gcc foo.c -Wall -O2 foo.c: In function ‘main’: foo.c:9:2: warning: format ‘%lu’ expects argument of type ‘long unsigned int’, but argument 2 has type ‘__u64’ [-Wformat] $ ./a.out [...] __u64:8 versus uint64_t:8 On ppc64: $ gcc -Wall -O2 foo.c $ ./a.out [...] __u64:8 versus uint64_t:8 So simply cast it in this case to uint64_t. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 57911d27b7a5b694aa7ad1ed6b3d8547fcf2d93e Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 15:23:22 2013 +0100 apps, func_tests: adapt cflags for older architectures Quoting RFC3493: One simple addition to the sockets API that can help application writers is the "struct sockaddr_storage". This data structure can simplify writing code that is portable across multiple address families and platforms. This data structure is designed with the following goals. - Large enough to accommodate all supported protocol-specific address structures. - Aligned at an appropriate boundary so that pointers to it can be cast as pointers to protocol specific address structures and used to access the fields of those structures without alignment problems. The sockaddr_storage structure contains field ss_family which is of type sa_family_t. When a sockaddr_storage structure is cast to a sockaddr structure, the ss_family field of the sockaddr_storage structure maps onto the sa_family field of the sockaddr structure. When a sockaddr_storage structure is cast as a protocol specific address structure, the ss_family field maps onto a field of that structure that is of type sa_family_t and that identifies the protocol's address family. So then don't let the compiler bitch about this and use the same flag to work around this as the Linux kernel does. Here's what GCC version is affected: gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-3) --> complains gcc (GCC) 4.7.2 20120921 (Red Hat 4.7.2-2) --> doesn't complain Also, we add -O2 for additional data flow analysis that otherwise we wouldn't have. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 296e5aba4798d8dbf2d949bdebbf70496c312c1e Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 14:04:34 2013 +0100 apps: sctp_darn: fix format string warning on some archs like ppc64 Fixes: sctp_darn.c: In function ‘get_assocstats_func’: sctp_darn.c:2261: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2262: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2263: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2264: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2265: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2266: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2267: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2268: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2269: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2270: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2271: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2272: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2273: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2274: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ sctp_darn.c:2275: warning: format ‘%llu’ expects type ‘long long unsigned int’, but argument 2 has type ‘__u64’ We use the generic PRIu64 type instead for printing. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 2b606fcd434b993293cb334f23bf1b6d006f24b1 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 13:28:08 2013 +0100 test_1_to_1_sockopt: fix deprecated SO_RCVBUF/SO_SNDBUF testcase Actually, it is not guaranteed, that the rcv/snd buffer size, we send is really set by the kernel. For small values, the SOCK_MIN_SNDBUF and SOCK_MIN_RCVBUF sizes in include/net/sock.h are used instead. This was the sole reason why this test, whose implementation was in 2005, failed. Just change it to something bigger is not critical, and we're good to go. $ ./test_1_to_1_sockopt test_1_to_1_sockopt.c 1 PASS : setsockopt() with a bad socket descriptor - EBADF test_1_to_1_sockopt.c 2 PASS : setsockopt() with an invalid socket - ENOTSOCK test_1_to_1_sockopt.c 3 PASS : setsockopt() with an invalid level - ENOPROTOOPT test_1_to_1_sockopt.c 4 PASS : setsockopt() with invalid option buffer - EFAULT test_1_to_1_sockopt.c 5 PASS : setsockopt() with invalid option name - EOPNOTSUPP test_1_to_1_sockopt.c 6 PASS : getsockopt() with a bad socket descriptor - EBADF test_1_to_1_sockopt.c 7 PASS : getsockopt() with an invalid socket - ENOTSOCK test_1_to_1_sockopt.c 8 PASS : getsockopt() with invalid option buffer - EFAULT test_1_to_1_sockopt.c 9 PASS : getsockopt() with invalid option name - EOPNOTSUPP test_1_to_1_sockopt.c 10 PASS : getsockopt() SCTP_INITMSG - SUCCESS test_1_to_1_sockopt.c 11 PASS : setsockopt() SCTP_INITMSG - SUCCESS test_1_to_1_sockopt.c 12 PASS : setsockopt() SO_LINGER - SUCCESS test_1_to_1_sockopt.c 13 PASS : getsockopt() SO_LINGER - SUCCESS test_1_to_1_sockopt.c 14 PASS : getsockopt() SO_RCVBUF - SUCCESS test_1_to_1_sockopt.c 15 PASS : getsockopt() SCTP_STATUS - SUCCESS test_1_to_1_sockopt.c 16 PASS : setsockopt() SO_RCVBUF - SUCCESS test_1_to_1_sockopt.c 17 PASS : setsockopt() SO_SNDBUF - SUCCESS test_1_to_1_sockopt.c 18 PASS : getsockopt() SO_SNDBUF - SUCCESS test_1_to_1_sockopt.c 19 PASS : getsockopt() SCTP_PRIMARY_ADDR - SUCCESS test_1_to_1_sockopt.c 20 PASS : setsockopt() SCTP_PRIMARY_ADDR - SUCCESS test_1_to_1_sockopt.c 21 PASS : getsockopt() SCTP_ASSOCINFO - SUCCESS test_1_to_1_sockopt.c 22 PASS : setsockopt() SCTP_ASSOCINFO - SUCCESS Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 5e2d80cc1fd8b08446f36729652bfa9b5040a732 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 11:48:55 2013 +0100 apps: peel: no reason to keep unused/assigned-only variables Fixes: peel_server.c: In function ‘my_sctpReadInput’: peel_server.c:417:8: warning: variable ‘disped’ set but not used [-Wunused-but-set-variable] peel_server.c:416:26: warning: variable ‘s_info’ set but not used [-Wunused-but-set-variable] peel_server.c: In function ‘poll_fd’: peel_server.c:459:10: warning: variable ‘notdone’ set but not used [-Wunused-but-set-variable] peel_client.c: In function ‘my_sctpReadInput’: peel_client.c:433:11: warning: variable ‘disped’ set but not used [-Wunused-but-set-variable] peel_client.c:432:26: warning: variable ‘s_info’ set but not used [-Wunused-but-set-variable] Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 7e07485fb23d53036bdc8990bcde45678cb8b341 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 11:38:03 2013 +0100 nagle_snd: remove variable with assignment-only Fixes: nagle_snd.c: In function ‘main’: nagle_snd.c:96:13: warning: variable ‘bytes_sent’ set but not used [-Wunused-but-set-variable] It is not needed, since test_sendmsg would fail anyway if the return value matches not the expected one within the wrapper function. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit f009608eba8455e28e81c2c54091d0b3984a9922 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 11:31:42 2013 +0100 test_autoclose: remove unneeded variable Fixes: test_autoclose.c: In function ‘main’: test_autoclose.c:68:13: warning: variable ‘bytes_sent’ set but not used [-Wunused-but-set-variable] We don't need it here since otherwise the test will fail anyway. $ ./test_autoclose test_autoclose.c 0 INFO : Waiting for the associations to close automatically in 5 secs test_autoclose.c 1 PASS : Autoclose of associations Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 6f6d952d883b468bb49eee895242499c876bc799 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 11:23:25 2013 +0100 func_tests: remove unused associds all over the place Fixes: test_fragments.c: In function ‘main’: test_fragments.c:92:25: warning: variable ‘associd2’ set but not used [-Wunused-but-set-variable] test_fragments.c:92:15: warning: variable ‘associd1’ set but not used [-Wunused-but-set-variable] test_peeloff.c: In function ‘main’: test_peeloff.c:65:32: warning: variable ‘clt_associd’ set but not used [-Wunused-but-set-variable] test_connect.c: In function ‘main’: test_connect.c:61:57: warning: variable ‘clt_associd2’ set but not used [-Wunused-but-set-variable] test_connect.c:61:43: warning: variable ‘clt_associd1’ set but not used [-Wunused-but-set-variable] test_connect.c:61:29: warning: variable ‘svr_associd2’ set but not used [-Wunused-but-set-variable] test_timetolive.c: In function ‘main’: test_timetolive.c:107:25: warning: variable ‘associd2’ set but not used [-Wunused-but-set-variable] test_sctp_sendrecvmsg.c: In function ‘main’: test_sctp_sendrecvmsg.c:91:25: warning: variable ‘associd2’ set but not used [-Wunused-but-set-variable] However, ifdef out some examples that might be useful for application developers to see, but not for the compiler. $ ./test_fragments test_fragments.c 1 PASS : setsockopt(SCTP_DISABLE_FRAGMENTS) test_fragments.c 2 PASS : getsockopt(SCTP_DISABLE_FRAGMENTS) test_fragments.c 3 PASS : Send a message that exceeds assoc frag point test_fragments.c 0 INFO : Sent 1353 byte message test_fragments.c 0 INFO : Received 1353 byte message test_fragments.c 0 INFO : Sent 2000 byte message test_fragments.c 0 INFO : Received 2000 byte message test_fragments.c 0 INFO : Sent 5000 byte message test_fragments.c 0 INFO : Received 5000 byte message test_fragments.c 0 INFO : Sent 10000 byte message test_fragments.c 0 INFO : Received 10000 byte message test_fragments.c 0 INFO : Sent 20000 byte message test_fragments.c 0 INFO : Received 20000 byte message test_fragments.c 0 INFO : Sent 32768 byte message test_fragments.c 0 INFO : Received 32768 byte message test_fragments.c 4 PASS : Send/Receive fragmented messages $ ./test_peeloff test_peeloff.c 1 PASS : sctp_peeloff test_peeloff.c 2 PASS : listen on a peeled off socket test_peeloff.c 3 PASS : sctp_peeloff on a peeled off socket test_peeloff.c 4 PASS : Receive msgs on peeled off sockets test_peeloff.c 5 PASS : Send msgs on peeled off sockets test_peeloff.c 6 PASS : peeled off socket cannot initialize a new assoc $ ./test_connect test_connect.c 1 PASS : non-blocking connect test_connect.c 2 PASS : connect on a socket to create an assoc that is already established test_connect.c 3 PASS : blocking connect test_connect.c 4 PASS : connect on a peeled off socket test_connect.c 5 PASS : connect to create an assoc that matches a peeled off assoc $ ./test_timetolive test_timetolive.c 0 INFO : Creating fillmsg of size 3087 test_timetolive.c 1 PASS : Send a message with timeout test_timetolive.c 2 PASS : Send a message with no timeout test_timetolive.c 3 PASS : Send a fragmented message with timeout test_timetolive.c 0 INFO : ** SLEEPING for 3 seconds ** test_timetolive.c 4 PASS : Receive message with no timeout test_timetolive.c 5 PASS : Receive SEND_FAILED for message with timeout test_timetolive.c 6 PASS : Receive SEND_FAILED for fragmented message with timeout $ ./test_sctp_sendrecvmsg test_sctp_sendrecvmsg.c 1 PASS : sctp_sendmsg test_sctp_sendrecvmsg.c 2 PASS : sctp_recvmsg SCTP_COMM_UP notification test_sctp_sendrecvmsg.c 3 PASS : sctp_recvmsg data test_sctp_sendrecvmsg.c 0 INFO : creating a fillmsg of size 3087 test_sctp_sendrecvmsg.c 4 PASS : sctp_sendmsg with ttl test_sctp_sendrecvmsg.c 5 PASS : sctp_sendmsg with zero ttl test_sctp_sendrecvmsg.c 6 PASS : sctp_sendmsg fragmented msg with ttl test_sctp_sendrecvmsg.c 0 INFO : ** SLEEPING for 3 seconds ** test_sctp_sendrecvmsg.c 7 PASS : sctp_recvmsg msg with zero ttl test_sctp_sendrecvmsg.c 8 PASS : sctp_recvmsg SEND_FAILED for message with ttl test_sctp_sendrecvmsg.c 9 PASS : sctp_recvmsg SEND_FAILED for fragmented message with ttl test_sctp_sendrecvmsg.c 10 PASS : sctp_send Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit f3f92d60b1299f43e9d0f36f7f311f85d9e13b2f Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 11:11:49 2013 +0100 test_1_to_1_recv: minor: fix compiler warning We don't need both, since we test afterwards anyway; if it would fail, also the test would not pass. $ ./test_1_to_1_recvfrom test_1_to_1_recvfrom.c 1 PASS : recvfrom() with a bad socket descriptor - EBADF test_1_to_1_recvfrom.c 2 PASS : recvfrom() with invalid socket - ENOTSOCK test_1_to_1_recvfrom.c 3 PASS : recvfrom() with invalid message ptr - EFAULT test_1_to_1_recvfrom.c 4 PASS : recvfrom() on listening socket - ENOTCONN test_1_to_1_recvfrom.c 5 PASS : recvfrom() on a socket that has received shutdown - EOF test_1_to_1_recvfrom.c 6 PASS : recvfrom() on a socket with pending message that has sent shutdown - SUCCESS test_1_to_1_recvfrom.c 7 PASS : recvfrom() on a socket with no pending messages and has sent shutdown - ENOTCONN $ ./test_1_to_1_recvmsg test_1_to_1_recvmsg.c 1 PASS : recvmsg() with a bad socket descriptor - EBADF test_1_to_1_recvmsg.c 2 PASS : recvmsg() with invalid socket - ENOTSOCK test_1_to_1_recvmsg.c 3 PASS : recvmsg() with invalid iovec ptr - EFAULT test_1_to_1_recvmsg.c 4 PASS : recvmsg() with invalid msghdr ptr - EFAULT test_1_to_1_recvmsg.c 5 PASS : recvmsg() on listening socket - ENOTCONN test_1_to_1_recvmsg.c 6 PASS : recvmsg() on a socket that has received shutdown - EOF test_1_to_1_recvmsg.c 7 PASS : recvmsg() on a socket with pending message that has sent shutdown - SUCCESS test_1_to_1_recvmsg.c 8 PASS : recvmsg() on a socket with no pending messages and has sent shutdown - ENOTCONN Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 6c8a3849d600d5f689ace594efb53aef4832bbf6 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 11:06:25 2013 +0100 test_1_to_1_nonblock: remove unused iov and malloced send buffer Fixes: test_1_to_1_nonblock.c: In function ‘main’: test_1_to_1_nonblock.c:75:13: warning: variable ‘count’ set but not used [-Wunused-but-set-variable] test_1_to_1_nonblock.c:72:22: warning: variable ‘iov’ set but not used [-Wunused-but-set-variable] Also remove count, we probe for received messages anyways. $ ./test_1_to_1_nonblock test_1_to_1_nonblock.c 1 PASS : non-blocking accept() - EAGAIN test_1_to_1_nonblock.c 2 PASS : non-blocking connect() - EINPROGRESS test_1_to_1_nonblock.c 3 PASS : accept() after a non-blocking connect - SUCCESS test_1_to_1_nonblock.c 4 PASS : non-blocking recvmsg() - EAGAIN test_1_to_1_nonblock.c 5 PASS : non-blocking recvmsg() when data is available - SUCCESS Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 6d02e67dad7cfdab43bc046a66b56ca8c7d11e6f Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 10:55:18 2013 +0100 test_1_to_1_addrs: remove unused socket, iov and malloced buffer Fixes: test_1_to_1_addrs.c: In function ‘main’: test_1_to_1_addrs.c:86:22: warning: variable ‘iov’ set but not used [-Wunused-but-set-variable] test_1_to_1_addrs.c:82:6: warning: variable ‘sk’ set but not used [-Wunused-but-set-variable] $ ./test_1_to_1_addrs test_1_to_1_addrs.c 1 PASS : sctp_getladdrs() with a bad socket descriptor - EBADF test_1_to_1_addrs.c 2 PASS : sctp_getladdrs() with invalid socket - ENOTSOCK test_1_to_1_addrs.c 3 PASS : sctp_getladdrs() with socket of different protocol - EOPNOTSUPP test_1_to_1_addrs.c 4 PASS : sctp_getladdrs() - SUCCESS test_1_to_1_addrs.c 5 PASS : sctp_freeladdrs() - SUCCESS test_1_to_1_addrs.c 6 PASS : sctp_getpaddrs() with a bad socket descriptor - EBADF test_1_to_1_addrs.c 7 PASS : sctp_getpaddrs() with invalid socket - ENOTSOCK test_1_to_1_addrs.c 8 PASS : sctp_getpaddrs() with socket of different protocol - EOPNOTSUPP test_1_to_1_addrs.c 9 PASS : sctp_getpaddrs() - SUCCESS test_1_to_1_addrs.c 10 PASS : sctp_freepaddrs() - SUCCESS Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 2e5e1046396e6a3a0be02a3d7289c2655f4ed0a9 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 10:49:25 2013 +0100 test_1_to_1_threads: fixup pthread init and exit Fixes: test_1_to_1_threads.c: In function ‘t_send’: test_1_to_1_threads.c:93:15: warning: variable ‘buffer_snd’ set but not used [-Wunused-but-set-variable] test_1_to_1_threads.c: In function ‘main’: test_1_to_1_threads.c:177:22: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] Also, don't use pthread_exit() as a return in main(), but rather in the thread handler instead of return! $ ./test_1_to_1_threads test_1_to_1_threads.c 1 PASS : send and receive data across multiple threads - SUCCESS Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 916a159bb2311f8d7ebce4482f271094a0dad39b Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 10:38:06 2013 +0100 test_1_to_1_sendmsg: remove set-only iov variable with malloced buffers Fixes: test_1_to_1_sendmsg.c: In function ‘main’: test_1_to_1_sendmsg.c:79:22: warning: variable ‘iov’ set but not used [-Wunused-but-set-variable] Then, we also do not need allocated 'buffer' and 'buffer_snd' anymore. $ ./test_1_to_1_sendmsg test_1_to_1_sendmsg.c 1 PASS : sendmsg() with a bad socket descriptor - EBADF test_1_to_1_sendmsg.c 2 PASS : sendmsg() with invalid socket - ENOTSOCK test_1_to_1_sendmsg.c 3 PASS : sendmsg() on a listening socket - EPIPE test_1_to_1_sendmsg.c 4 PASS : sendmsg() with invalid iovec ptr - EFAULT test_1_to_1_sendmsg.c 5 PASS : sendmsg() with invalid iovec length - EINVAL test_1_to_1_sendmsg.c 6 PASS : sendmsg() with invalid msghdr ptr - EFAULT test_1_to_1_sendmsg.c 7 PASS : sendmsg() with invalid sinfo flags - EINVAL test_1_to_1_sendmsg.c 8 PASS : sendmsg() with SCTP_EOF flag - EINVAL test_1_to_1_sendmsg.c 9 PASS : sendmsg() with SCTP_ABORT flag - EINVAL test_1_to_1_sendmsg.c 10 PASS : sendmsg() on a closed association - EPIPE test_1_to_1_sendmsg.c 11 PASS : sendmsg() from accept socket to client - SUCCESS test_1_to_1_sendmsg.c 12 PASS : sendmsg() from client to server - SUCCESS test_1_to_1_sendmsg.c 13 PASS : sendmsg() from unconnected clt to server - SUCCESS test_1_to_1_sendmsg.c 14 PASS : sendmsg() on a SHUT_RD socket - SUCCESS Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 79ccf2e451cfc6c4aed300a11bf105bd9b26b57a Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 10:29:53 2013 +0100 func_tests: drop more af_family occurences Again, af_family is set but not used. Instead pf_class is the one in charge. Fixes: test_fragments.c:89:16: warning: variable ‘af_family’ set but not used [-Wunused-but-set-variable] test_inaddr_any.c:82:16: warning: variable ‘af_family’ set but not used [-Wunused-but-set-variable] test_sockopt.c:81:16: warning: variable ‘af_family’ set but not used [-Wunused-but-set-variable] test_timetolive.c:104:16: warning: variable ‘af_family’ set but not used [-Wunused-but-set-variable] test_sctp_sendrecvmsg.c:80:16: warning: variable ‘af_family’ set but not used [-Wunused-but-set-variable] Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 046a2c0102ef683c1d8e2a6719037d2b9d16739d Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 10:24:32 2013 +0100 sctp_darn: remove never read peer_prim_len variable Fixes: sctp_darn.c: In function ‘peer_primary_func’: sctp_darn.c:1963:6: warning: variable ‘peer_prim_len’ set but not used [-Wunused-but-set-variable] Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit b8e5a8cff4300dd14bf7685ceb114df0fd775952 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 10:12:49 2013 +0100 apps: nagle: remove unused af_family variable The pf_class variable is being used instead, so just remove the other one. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 4248a1ab470470277f5853d290ccd795f3a6e8fd Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 09:53:10 2013 +0100 apps: fix format string warnings due to size_t usage Fix all warnings of type: warning: format ‘%d’ expects argument of type ‘int’, but argument 3 has type ‘size_t’ [-Wformat]. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit c67ed7fd2a243d06f60febc8ab41489b50b924b8 Author: Daniel Borkmann <dborkman@redhat.com> Date: Tue Jan 22 09:44:02 2013 +0100 docs: update current maintainers of lksctp-tools While at it, also credit all contributors that appear in the Git log. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 00c7705991e299c51ad7eee22bafa8586964b525 Author: Fan Du <fan.du@windriver.com> Date: Tue Jan 15 09:52:03 2013 +0800 sctp_xconnect: memory leak when malloc big buffer CLIENT repeatly call process_ready_sockets, which malloc without free, so sctp_xconnect exit unexpectly. v1->v2: Declare this buffer statically suggested by Neil Horman. Signed-off-by: Fan Du <fan.du@windriver.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 2a52eaaf164a162b32fc82e2b0deeca9b7d46d60 Author: Daniel Borkmann <dborkman@redhat.com> Date: Fri Dec 21 13:31:18 2012 +0100 test_fragments: increase message size since it succeeds We get the following error: test_fragments.c 3 BROK : Send a message that exceeds assoc frag point error:30000 errno:0 This means, it just succeeds, so increase the message length! Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 09300466d016765fd9c4e8144175feb23f395307 Author: Daniel Borkmann <dborkman@redhat.com> Date: Thu Dec 20 16:12:06 2012 +0100 autotools: make build with silent rules Just let autotools shut up, so that we can actually see useful gcc warnings and the like. Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit bc01beed0397982659e710d469ccbd0343298f6c Author: Vlad Yasevich <vyasevic@redhat.com> Date: Tue Jan 8 10:58:57 2013 -0500 Release lksctp-tools-1.0.12 Signed-off-by: Vlad Yasevich <vyasevich@gmail.com> commit 28b9e3c3111daf0999675a681518decbb4ff46f8 Author: Michael Tuexen <tuexen@fh-muenster.de> Date: Tue Nov 6 08:16:26 2012 -0500 lksctp-tools: Update sctp.h with info for DTLS Add definitions needed for DTSL/SCTP. Signed-off-by: Michael Tuexen <tuexen@fh-muenster.de> Signed-off-by: Vlad Yasevich <vyasevich@gmail.com> commit c5336c5eb01e72683ccaf44a0879ffa4ce3a69bf Author: Daniel Borkmann <dborkman@redhat.com> Date: Mon Dec 17 17:32:04 2012 +0100 sctp_send: fix msg_control data corruption The byte array outcmsg is allocated on the stack within the if-block that test for a valif sctp_sndrcvinfo structure. There, it is assigned to outmsg.msg_control, which is later on after leaving the if-block passed to sendmsg. With this minimal example, the following is happening: int main(void) { int fd; struct sctp_sndrcvinfo sndinfo; fd = socket(AF_INET, SOCK_SEQPACKET, IPPROTO_SCTP); assert(fd > 0); sctp_send(fd, "bla", strlen("bla") + 1, &sndinfo, 0); return 0; } strace ./a.out before this patch: sendmsg(3, {msg_name(0)=NULL, msg_iov(1)=[{"bla\0", 4}], msg_controllen=48, {cmsg_len=3364590592, cmsg_level=0x4003e8 /* SOL_??? */, cmsg_type=, ...}, msg_flags=0}, 0) --> cmsg_len corrupted strace ./a.out after this patch: sendmsg(3, {msg_name(0)=NULL, msg_iov(1)=[{"bla\0", 4}], msg_controllen=48, {cmsg_len=48, cmsg_level=0x84 /* SOL_??? */, cmsg_type=, ...}, msg_flags=0}, 0) This is basically the case since 2005, introduced in the commit 91239acf ("Add sctp_send() API support and testcases"). However, probably this changed due to a different compiler behaviour / optimization (?), since it was not visible / affected by older Linux versions. Cc: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> commit 3fde8eb195690695b67bfca868c14a4a6983fb80 Author: Michele Baldessari <michele@acksyn.org> Date: Tue Dec 4 17:02:57 2012 +0100 Implement SCTP_GET_ASSOC_STATS Implement SCTP_GET_ASSOC_STATS to retrieve per association statistics. This API is not present in rfc 6458, but is modeled after the same call in OpenSolaris. V2: - Do not include SCTP_GET_ASSOC_STATS in the official sctp_opt_info() API Signed-off: Michele Baldessari <michele@acksyn.org> commit c7f44173f204bff1a1913849e3802c791c7f7c44 Author: Wei Yongjun <yjwei@cn.fujitsu.com> Date: Fri Feb 26 11:19:32 2010 +0800 sctp_test: add -O option for specifying live time of messages This patch add -O option to sctp_test for specifying live time of messages. Use for PRSCTP test. Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit ae28087b6f9ca62cf818933b539fdc10ceb89bbd Author: Wei Yongjun <yjwei@cn.fujitsu.com> Date: Fri Feb 26 11:18:28 2010 +0800 sctp_test: add -B option and -C option for specifying addresses This patch add -B option and -C option for specifying addresses. Copy from sctp_dran. -B option add the specified address(es) as additional bind addresses of the local socket. Multiple addresses can be specified by using this argument multiple times. For example, '-B 10.0.0.1 -B 20.0.0.2'. -C option use the specified address(es) for connection to the peer socket. Multiple addresses can be specified by using this argument multiple times. For example, '-C 10.0.0.1 -C 20.0.0.2'. This option is incompatible with the -h option. Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 37f513c82095b94aeff4a42dc339b3ae591ea492 Author: Wei Yongjun <yjwei@cn.fujitsu.com> Date: Fri Feb 26 11:15:31 2010 +0800 sctp_darn: add inter command heartbeat for user initiated heartbeat Add inter command heartbeat to sctp_darn, which is used for request a user initiated heartbeat to be made immediately. Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 3e107907f11145a63e6faed9831410b8cba98957 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Nov 12 11:00:47 2009 -0500 api: Remove the sctp option enumerator The sctp_optname enumerator is actually getting in the way as we try to deprecate and remove really old options, because removing items from the enum breaks binary compatibility with the kernel. As a result, we do not gain any benefits from the enumerator, but in fact get obfuscation of option values. Now mirror the kernel change and turn the enum into a defined list with specific values. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 6c00ed9653870e5108286f3f9663c5ed1140c00e Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Nov 12 10:59:51 2009 -0500 api: Remove the old/obsolete getaddrs interfaces. The new kernel has removed the 5 year old interfaces for the 64-bit unsafe get_[l|p]addr interfaces. Remove them from the user library as well. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit ecdc6c0d3c3e40843916427923a45a93d748c1c6 Author: Andrei Pelinescu-Onciul <andrei@iptel.org> Date: Thu Nov 12 16:03:39 2009 +0100 api: SCTP_DELAYED_SACK typo fix According to the api draft, the socket option for setting the delayed sack timer should be called SCTP_DELAYED_SACK and not SCTP_DELAYED_ACK. Added a new define for SCTP_DELAYED_SACK (SCTP_DELAYED_ACK is not removed to maintain compatibility with existing applications). Signed-off-by: Andrei Pelinescu-Onciul <andrei@iptel.org> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 5e7c83ad0b352791e175f626b35bc81eed8b805e Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Nov 12 10:38:44 2009 -0500 API: Add SCTP_SACK_IMMEDIATELY defintion. commit 57dff214cf21e36b10a86d7b04b0b7ab88401878 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Nov 12 10:38:34 2009 -0500 lib: Fix the new connectx api to prefent SEGFAULTS The new version connectx() that supports non-blocking sockets fails the regression tests due to additional allocations that it performes. It can also segfault as it attempts to copy addresses from the user provided buffer into a newly allocated buffer. To prevent that, we change the api to remove allocations and pass the user-allocated address buffer unchanged to the kernel. The api uses sctp_getadds_old structure, whete assoc_id is used to return the association id, the addr_num is used to convey the length of the address buffer to the kernel and addrs points to the user-allocated addrs array the the library now passed on as is. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 2e0968e36e2520a0c043e53f4f9a468f4d1cd4a4 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Oct 23 20:43:39 2009 -0400 lib: Support non-blocking sctp_connectx() calls 1: I decided to keep the previous "new" sctp_connectx() implementation, because a: it costs nothing, and b: some very recent programs may rely on the broken behaviour. 2: The new-new sctp_connectx() tries the new method first. If it fails due to ENOPROTOOPT (old kernel) then it falls back to the old method first, but only if it is safe to do so (blocking socket, or assoc_id not returned) 3: (note of users of this feature): If a non-blocking connect fails quickly then a new call to sctp_connectx() may return an identical association id becaue the ID has been freed up already. Additional slight code clean-up. Signed-off-by: Ivan Skytte Jørgensen <isj-sctp@i1.dk> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 7f518dde33a97b9182a28947356a0558eb39e128 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Oct 23 20:43:39 2009 -0400 Release 2.6.28-1.0.11 commit 768bb51405408d15fcb4626f8b18cedd80ef8d04 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Oct 23 20:43:39 2009 -0400 func_test: fix compiler warnings in functional tests. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit bafe2eacbfbba550667e8115a556dad8b6c8ed4d Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Oct 23 20:43:36 2009 -0400 lib: Fix compiler warning for connectx.c Do not discard the const modifier. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit f8955b286554446679e75d283b96fe3ead00e64b Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Oct 23 16:18:34 2009 -0400 apps: Fix compiler warnings in sctp_darn Code was calling fprintf() without a format, but with an argument to evaluate. That's really pointless as it doesn't produce anything and generates a compiler warning. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 41ac5624533d126b791feae79ed3df6f772057a4 Author: Sam Miller <samjmill@us.ibm.com> Date: Fri Jun 5 15:31:54 2009 -0500 tools: fix parallel build warning Makefiles in src/apps and src/func_tests use 'make' instead of $(MAKE) which causes failures in parallel builds. Signed-off-by: Sam Miller <samjmill@us.ibm.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit d4d1be386f1c49eae348a5008e2a19073678a097 Author: Michael Biebl <biebl@debian.org> Date: Thu Apr 16 18:27:07 2009 +0200 lib: fix make distcheck for lksctp-tools lksctp-tools-1.0.10 currently fails "make distcheck". Problem is, that src/lib/Version.map is not correctly referenced in src/lib/Makefile.am leading to problems when builddir != srcdir. From: Michael Biebl <biebl@debian.org> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 326aff394cd48f1c674e43a59b0869e8f2285639 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Mar 27 14:58:37 2009 -0400 api: Update the sctp.h header to match the kernel. The updates include new socket options and fixed spelling of structure members (with backward compatiblity hack). Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit d80ea89fe0e827c2d25e702737c1227b0d80b95c Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Mar 27 14:23:11 2009 -0400 Release 1.0.9 Features: * Updated sctp_connectx() implementation * New tools to support nettest project http://networktest.sourceforge.net/ * Frame tests have been moved to a separate package. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit ffd2a31476f2c553b6ab680ec80df77d052a7c6f Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Mar 27 14:19:35 2009 -0400 lib: Update Makefile.am to include the versions file Since the library now has versioned symbols, we need to make sure that Versions.map is part of the distribution. The easiest way to acheave this is to include it into the _SOURCES list. Signed-off-by: Vlad Yaseivich <vladislav.yasevich@hp.com> commit 88897c36e3f82f78d5d22597636b8ae128b91114 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Mar 19 16:10:12 2009 -0400 Move frame tests to a separate repository Kernel developement has been acceleriting not just in SCTP, but in other areas of the kernel. Since the frame tests depend on a large amount of kernel code, they have been very fragile over the last year or so. At times, they break due to a change in a completely unrelated subsystem (like vm). As such, I am moving them to a separate repository so that they do not hold back changes to SCTP library and user applications. The new repository is available at: git://git.kernel.org/pub/scm/linux/kernel/git/vxy/lksctp-test.git This new repository is aimed at kernel developers who wish to extend the SCTP kernel implementation. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> delete mode 100644 test/Makefile delete mode 100644 test/ft_frame_CACC.c delete mode 100644 test/ft_frame_abort.c delete mode 100644 test/ft_frame_addip_bindx.c delete mode 100644 test/ft_frame_addip_set_prim.c delete mode 100644 test/ft_frame_addr_cookie.c delete mode 100644 test/ft_frame_addrover.c delete mode 100644 test/ft_frame_assoc_abort.c delete mode 100644 test/ft_frame_assoc_shutdown.c delete mode 100644 test/ft_frame_autoclose.c delete mode 100644 test/ft_frame_basic.c delete mode 100644 test/ft_frame_basic_nat.c delete mode 100644 test/ft_frame_bigdata.c delete mode 100644 test/ft_frame_biggap.c delete mode 100644 test/ft_frame_bind.c delete mode 100644 test/ft_frame_bindx.c delete mode 100644 test/ft_frame_congestion.c delete mode 100644 test/ft_frame_connect.c delete mode 100644 test/ft_frame_connectx.c delete mode 100644 test/ft_frame_connectx_collision1.c delete mode 100644 test/ft_frame_connectx_collision2.c delete mode 100644 test/ft_frame_connectx_collision3.c delete mode 100644 test/ft_frame_connectx_t1_timer.c delete mode 100644 test/ft_frame_cookie_ports.c delete mode 100644 test/ft_frame_crossinit.c delete mode 100644 test/ft_frame_delip_bindx.c delete mode 100644 test/ft_frame_dftsndparam.c delete mode 100644 test/ft_frame_ecn_basic.c delete mode 100644 test/ft_frame_ecn_negotiate.c delete mode 100644 test/ft_frame_evenpackets.c delete mode 100644 test/ft_frame_fast_rtx.c delete mode 100644 test/ft_frame_frag_interleave.c delete mode 100644 test/ft_frame_fragments.c delete mode 100644 test/ft_frame_getfreeaddrs.c delete mode 100644 test/ft_frame_hbACK.c delete mode 100644 test/ft_frame_hb_rtx.c delete mode 100644 test/ft_frame_heartbeat.c delete mode 100644 test/ft_frame_init1.c delete mode 100644 test/ft_frame_init2.c delete mode 100644 test/ft_frame_init4.c delete mode 100644 test/ft_frame_init5.c delete mode 100644 test/ft_frame_init_ack.c delete mode 100644 test/ft_frame_init_bomb.c delete mode 100644 test/ft_frame_init_streams.c delete mode 100644 test/ft_frame_init_timer.c delete mode 100644 test/ft_frame_invalid_length.c delete mode 100644 test/ft_frame_kamikaze.c delete mode 100644 test/ft_frame_local_addr.c delete mode 100644 test/ft_frame_lost_cookie_ack.c delete mode 100644 test/ft_frame_lost_cwr.c delete mode 100644 test/ft_frame_lostlink.c delete mode 100644 test/ft_frame_lostpacket.c delete mode 100644 test/ft_frame_lostpackets.c delete mode 100644 test/ft_frame_maxseg.c delete mode 100644 test/ft_frame_missing_cookie.c delete mode 100644 test/ft_frame_multihoming.c delete mode 100644 test/ft_frame_no_data.c delete mode 100644 test/ft_frame_nobind.c delete mode 100644 test/ft_frame_ootb.c delete mode 100644 test/ft_frame_part_delivery.c delete mode 100644 test/ft_frame_peeloff.c delete mode 100644 test/ft_frame_pmtu.c delete mode 100644 test/ft_frame_port.c delete mode 100644 test/ft_frame_prsctp.c delete mode 100644 test/ft_frame_recvmsg.c delete mode 100644 test/ft_frame_renege.c delete mode 100644 test/ft_frame_restart.c delete mode 100755 test/ft_frame_retran_path.c delete mode 100644 test/ft_frame_rto.c delete mode 100644 test/ft_frame_rto_sockopt.c delete mode 100644 test/ft_frame_rtx_gapack.c delete mode 100644 test/ft_frame_rtx_order.c delete mode 100644 test/ft_frame_rtx_timer.c delete mode 100644 test/ft_frame_rwnd_receiver.c delete mode 100644 test/ft_frame_rwnd_sender.c delete mode 100644 test/ft_frame_sack.c delete mode 100644 test/ft_frame_send_failed.c delete mode 100644 test/ft_frame_shutdown.c delete mode 100644 test/ft_frame_sndrcvinfo.c delete mode 100644 test/ft_frame_sockopt.c delete mode 100644 test/ft_frame_source_addr.c delete mode 100644 test/ft_frame_stale_cookie.c delete mode 100644 test/ft_frame_t2_shutdown_timer.c delete mode 100644 test/ft_frame_t4_timer.c delete mode 100644 test/ft_frame_tcp_style.c delete mode 100644 test/ft_frame_tsnmap_rollover.c delete mode 100644 test/ft_frame_tx_bug1.c delete mode 100644 test/ft_frame_unkchunk.c delete mode 100644 test/ft_frame_unkparam.c delete mode 100644 test/ft_frame_v4_mapped_v6.c delete mode 100644 test/ft_frame_v6only.c delete mode 100644 test/ft_kernel_sysctl.c delete mode 100644 test/funtest.h delete mode 100644 test/funutil.c delete mode 100644 test/linux/include/asm/uaccess.h delete mode 100644 test/linux/include/linux/rcupdate.h delete mode 100644 test/test_cmsgs.c delete mode 100644 test/test_command.c delete mode 100644 test/test_frame.h delete mode 100644 test/test_inqueue.c delete mode 100644 test/test_kernel.c delete mode 100644 test/test_kernel.h delete mode 100644 test/test_packet.c delete mode 100644 test/test_sn_compare.c delete mode 100644 test/test_tsnmap.c delete mode 100644 test/test_ulpevent.c delete mode 100644 test/test_ulpqueue.c commit a102b7c49869b2da85d3d849d3a72517614d38e7 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Oct 9 16:05:40 2008 -0400 test: Update tsnmap test to use new implementation. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit c7f626ada53cce61f04deb391c0aaee5ac5ab476 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Oct 9 15:59:31 2008 -0400 frame-test: Add a test for COOKIE-ECHO rtx on alternate transport. Durring connectx, it should be possible to retransmit cookie-echo on alternate transports. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 3ac4de23dd04a247b2a9d966f440cefead349cd3 Author: Wei Yongjun <yjwei@cn.fujitsu.com> Date: Mon Jan 12 17:29:53 2009 +0800 lksctp-tools: Add sctp_status for congestion control test Add sctp_status for congestion control test. Based on sctp_test. It used by SCTP conformance test suite. Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> create mode 100644 src/apps/sctp_status.c commit dc8b33a46ead93dc5abd94a684f985dacca08e24 Author: Wei Yongjun <yjwei@cn.fujitsu.com> Date: Mon Jan 12 17:18:55 2009 +0800 lksctp-tools: Add -T option to sctp_test Add -T option to allow sctp_test use SOCK_STREAM tcp-style sockets. Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit c647541439045751bc711f090ca20a50b9c8ede8 Author: Wei Yongjun <yjwei@cn.fujitsu.com> Date: Fri Jan 9 13:58:07 2009 +0800 lksctp-tools: Fix -e option of sctp_darn If sctp_darn -e option is used, sctp_darn will send back different message to which it received. This is because test_print_message() will change the received message when print out it. Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 93fd7448317419594472045663ff13034c31efa9 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Wed Jan 14 14:30:20 2009 -0500 [lib] Fix building of static libraries When running "./configure --disable-shared --enable- static" the build fails with the following error: gcc -g -Wall -Wstrict-prototypes -Wimplicit-function- declaration -g -O2 -o nagle_snd nagle_snd.o ../../src/lib/.libs/libsctp.a ../../src/t estlib/.libs/libsctputil.a ../../src/testlib/.libs/libsctputil.a(sctputil.o): In function `test_peer_addr': /home/roubert/src/lksctp/lksctp-tools- 1.0.6/src/testlib/sctputil.c:376: undefined reference to `sctp_getpaddrs' /home/roubert/src/lksctp/lksctp-tools- 1.0.6/src/testlib/sctputil.c:382: undefined reference to `sctp_freepaddrs' /home/roubert/src/lksctp/lksctp-tools- 1.0.6/src/testlib/sctputil.c:416: undefined reference to `sctp_freepaddrs' /home/roubert/src/lksctp/lksctp-tools- 1.0.6/src/testlib/sctputil.c:400: undefined reference to `sctp_freepaddrs' collect2: ld returned 1 exit status make[3]: *** [nagle_snd] Error 1 Changing the order of libsctp.a and libsctputil.a fixes this. The attached patch updates src/apps/Makefile.am and src/func_tests/Makefile.am to do that. Patch submitted by Fredrik Roubert <roubert@sf.net>. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 7b7c0cba8d6a669d3604c80982be8c4310904703 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Jan 13 17:07:04 2009 -0500 [lib]: Fix compile warning in addrs.c Include <string.h> to fix compile warning. Patch submitted by Marcus Meissner. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 3bc6c79e12f350b17c2ff1f53471bb811df24452 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Nov 6 12:26:26 2008 -0500 func_test: Fix the rest of the tests/apps that use sctp_connectx. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 96468b71fc06ca98002a11a3ceb49199c4265a67 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Oct 9 16:31:37 2008 -0400 frame test: Update to v2.6.27-rc9 Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 8399a2da90e1bf75c38f1420403f9e2a55cf14b3 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Aug 7 13:33:57 2008 -0400 frame_test: provide a working test for fast retransmissions. We've broken fast retransmits a few times recently and we really needed a test for regression purposes just like that. The existing test was completely bogus in it's expected behavior and the way it was written. So this re-writes the test to make sure that different fast retransmission scenarios are covered and that we correctly implement fast recovery algorithm. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit e8dddd2f39559c07895ad6855dd711958883283c Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Wed Aug 6 15:17:50 2008 -0400 frame_test: Add test for SO_REUSEADDR and IPV6_V6ONLY functionality. ft_frame_bind.c will test different binding scenarios with and without SO_REUSEADDR. This is to make sure that we don't break the conflict detection scemantics during binding. ft_frame_v6only.c will test IPV6_V6ONLY socket option support in SCTP. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> create mode 100644 test/ft_frame_bind.c create mode 100644 test/ft_frame_v6only.c commit 73ffbc7cdbaf19c41220672bd6b847e36ae9c385 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Wed Aug 6 14:42:56 2008 -0400 frame_test: Make frame tests work with 2.6.27-rc2 kernel. Changes were to update checksum computations and account for additional bundling of packets that we didn't do before. Some tests expected packets where we are now sending one. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 5b9eb80bed2090cbfcc3eaa82bcaf53f73b19c89 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Jun 10 09:29:51 2008 -0400 [FUNC_TEST]: Convert sctp_connectx() functional test to new API. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 84db487cfd8322fc39e77a9454afd63ed877b4f0 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Jun 10 09:28:57 2008 -0400 [API] Update sctp_connectx() to conform to spec. The prototype and implementation of sctp_connectx() has changed to include an additional 'id' parameter that is supposed to contain association id of the newly created association. The internal implementation is that the old socket option has changed it's name, but not it's value. A new option has also been added with the SCTP_CONNECTX name. To preserve the ABI, symbol versioning has been added. As a result, applications do not need to recompile. However, if the application is recompiled, it will need to update it's API since there is no to preserve it in the header file. A simple way to fix the apps is to add a NULL arguemnt at the end of sctp_connectx() call. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> create mode 100644 src/lib/Versions.map commit 762dd1ec42bdfea04c8e928767445a79f405083f Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jun 6 22:52:44 2008 -0400 frame_test: Update test_connectx() to account for new API. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 7df37b14da018e83a82e85201f6b10d91c4da472 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Oct 2 10:32:47 2008 -0400 api: let sctp_recvmsg honor passed in flags It should be possible to pass MSG_ flags to sctp_recvmsg and have it honor those flags. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 7c7a1cf0faed35fba3b89cc2f05348b9239610d3 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Aug 5 15:42:15 2008 -0400 Release 1.0.9 Frame tests require kernel version 2.6.26. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit db41d66f76e4d0013eb2dc49263267033a701a18 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Aug 5 15:40:38 2008 -0400 doc: Add the latest SCTP RFCs. rfc4820 - SCTP PAD chunk rfc4895 - SCTP Authentication Extensions rfc4960 - Revised SCTP base specification rfc5061 - Add-IP rfc5062 - SCTP Threats Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> create mode 100644 doc/rfc4820.txt create mode 100644 doc/rfc4895.txt create mode 100644 doc/rfc4960.txt create mode 100644 doc/rfc5061.txt create mode 100644 doc/rfc5062.txt commit 2bd370a343602a5b42302e447b2003c0666a867d Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Wed Jun 4 14:30:12 2008 -0400 frame_test: Make the kernel frame tests work with 2.6.26 commit fc7c7b275cafa9dced33aea90e1b368b5c275a97 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Wed Jun 4 10:46:15 2008 -0400 [FRAME TESTS] Make frame tests work with 2.6.25. One caviate: 2.6.25 introduced some changes to ECN markings that I didn't catch before and as such IPv6 ecn tests fail. I have a patch pending that should correct this issue. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 06b79cbe620364d0b4ce8ad73366f35f07162229 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Feb 29 17:20:19 2008 -0500 [API]: AUTH API updates User side implementation of the AUTH API. This exports all the needed pieces for usersapce. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 675fcb440f5303f9881e22fa5fc35312f02f6260 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Feb 26 14:35:16 2008 -0500 [APP]: Fix withsctp Makefile so that it builds commit 89bc3c864639e305e436c496098d8978ce7d6c65 Author: Michael Biebl <biebl@teco.edu> Date: Tue Feb 26 20:06:47 2008 +0100 [APP] Process withsctp.in at build time This patch creates withsctp at build time, rather than configure time. This has two advantages: 1.) It works with unexpanded dir variables Currently if you run "./configure && make", withsctp will contain LIBDIR=${exec_prefix}/lib/lksctp-tools BINDIR=${exec_prefix}/bin This obviously doesn't work. The patch fixes that. 2.) make install prefix=/foo works as expected. It's a recommended practice (by the gnu autotools) to expand directory variables at build time, rather than configure time for the above reasons. commit 46e79257603f670834334f1158137c9a47c776fd Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Feb 5 15:48:10 2008 -0500 Release 1.0.8 Frame tests require kernel version 2.6.24. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit fef441da0338af177216a592a910fd369dd06781 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Feb 1 13:26:43 2008 -0500 [FRAME-TEST]: Update tests for 2.6.24 kernel. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit ad4acb71ea2edd25461359098362c42af393529c Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Feb 1 13:26:06 2008 -0500 [FRAME-TEST] Update hb_rtx test to wait for an extra HB. HB are special in that they are not retransmissions. So, we need to be able to send a total for max_retrans HBs and timeout on all of them to terminate the association. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit a483cc18418ee1acafb365b6764215187c99ffbc Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Jan 31 14:06:22 2008 -0500 [ALL]: Do not claim that this is a reference implementation. This implementation has never really fit the description of a "reference" implementation. The first user space implementation was probably a reference implementation. Not even BSD implementation triys to claim this and we are being way too presumptuos. Signed-off-by: VladYasevich <vladislav.yasevich@hp.com> commit 4570d5dde103d015cc757a39ec781e53a233bd8d Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Mon Jan 28 13:55:06 2008 -0500 [LIB]: Copy correct length in sctp_getaddrs When performing the memmove trick in sctp_getaddrs, do not adjust the length. This is needed because the kernel has always returned back just the length of the packed address array, excluding the sctp_getaddrs structure. Thanks to Thomas Dreibholz <dreibh@iem.uni-due.de> for tracking down and fixing this. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit c34d7fb961225102969398d8258500c7792501ab Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jan 11 16:09:45 2008 -0500 [FAME_TEST]: Build with 2.6.24-rc7. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit f191e11e36a17dc660345b9d7d2842be3ce69b3b Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Mon Dec 10 16:40:32 2007 -0500 [Func Tests]: Fix TCP tests to correctly close accepted sockets. Two tests accept connection, but never close the socket. They subsequently try to bind() to the same port and fail. Once the socket is correctly closed, the tests work. Also, add a sleep while running tests so that we know everything got flushed from the kernel. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 9226213056678f2a19cfdd1fc9617a92335ece39 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Tue Nov 6 17:01:35 2007 -0500 [PATCH]: Make tests work with 2.6.24 rc2 kernel This involved added the network namespace initialization, calling init_Internet() in the right place in all the tests and enabling addip_noauth for any add-ip tests. We also no longer need a private copy proc_fs.h. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> delete mode 100644 test/linux/include/linux/proc_fs.h commit 1db21a7bbaf2ad7a0ced7b89e51969c19ed2ba00 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Mon Nov 5 10:38:04 2007 -0500 [PATCH] Update frame tests to work with 2.6.23 kernel. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 69457dc36bf2647e0fcf4fe9512b6ef60117cd0b Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Jul 12 12:24:14 2007 -0400 Release 1.0.7 Frame tests require kernel version 2.6.22. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 0389261f46b316a61dc786ab91b73f7eea96b4ba Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Jul 13 10:25:34 2007 -0700 [frame tests] Add ip6_dst_ops structure Also allow NF_CONTACK to be a module. commit 40f9a71ddf7c99ec5c8921a4817f85e28720c78b Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Jul 12 12:22:51 2007 -0400 Update ChangeLog. commit 55458118f872822053a478c0c1930d9aaa598468 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu Jul 12 12:16:36 2007 -0400 [frame tests] Add sock_lock_nested to frame tests. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit cee1e37bde561c82baebd7ae5dfb4e2537f13f04 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Wed Jun 27 15:25:07 2007 -0400 [PATCH] Fix possible memory leak in sctp_getaddrs() Found by Paolo Galtieri <pgaltieri@mvista.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 333297b66d042facc30e9c726f1c1307e0cd4afb Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jun 22 12:02:26 2007 -0400 [PATHC] Fix sctp_assoc_abort functional test case The SCTP_COMM_LOST notification that was a result of an abort now contains the additional parameter that includes the error from the abort chunk. Need to account for that when checking the received notification. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 56e895e60d2dc757e77b522dfd669303adde5073 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jun 22 11:20:11 2007 -0400 [Frame tests] Add dst_ops structure to test_kernel The recent path mtu bug fix in the kernel introduced a derenference through the dst_ops which we didn't have before. We now need this here. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 884ca8ce67fa31aaefe3e7254ae5ee3d0a7e3ac9 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jun 22 11:18:32 2007 -0400 [Frame test]: Define num_cpus to fix compilation on SMP. commit 03040519eb6e2399ae14c4ae693a6a9a4a7efd69 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jun 8 10:28:28 2007 -0400 [Frame tests] Fix IPv6 compilation problem. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit ec456585f637f4ccfdfc1a33d6990606f1349222 Author: Vlad Yasevich <vlad@galen.zko.hp.com> Date: Wed Jan 10 15:24:47 2007 -0500 Add a test for fragment interleave option Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> create mode 100644 test/ft_frame_frag_interleave.c commit a9691cecc1545fcf467f7fea57d82ccb50c42e22 Author: Vlad Yasevich <vlad@galen.zko.hp.com> Date: Wed Jan 10 15:30:42 2007 -0500 Add a test of SCTP_PARRIAL_DELIVERY_POINT option. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit dca409dec1abc31a7aabb1fe7b04aa9b8ddf8ced Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu May 10 15:03:09 2007 -0400 [SCTP] Update socket API definitons. Add the following socket options to userspace include: SCTP_FRAGMENT_INTERLEAVE SCTP_PARTIAL_DELIVERY_POINT SCTP_MAX_BURST Also, update the flag definitions to SCTP_PEER_ADDR_PARAMs and structure definitions. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 4c69391a0efff97bc3aaac6020bf63468e65f17e Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu May 10 13:53:08 2007 -0400 [PATCH] Correct the use for SCTP_PEER_ADDR_PARAMS API. The api uses flags to specify if specific values should be used or ignored. Frametests need to set these flags. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 623239032e6b784737834eb78fdf51f3dd6f9248 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Thu May 10 13:38:38 2007 -0400 [PATCH] Update frame tests to build with 2.6.21-rc7+ kernel Convert to using netdev_for_each() for device interation as well as remove the use of skb->nh and skb->h variables. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> commit 7c9963a12df8cd8c9cce744dd02013d98bdcefa5 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Apr 26 15:07:41 2007 -0700 ChangeLog update commit 93ae1232a97acbff84cfaa6c5c9ad78bd2cb803c Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Apr 26 15:02:36 2007 -0700 Add stubs for _read_lock_bh() and _read_unlock_bh() commit 0891ab2ccb6bec9e26f1e986189eff64cc1b154b Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Feb 28 12:04:47 2007 -0800 frametests update to 2.6.21-rc2 commit f75da31712c99013adbeb523154452cc7d21a30e Author: Sridhar Samudrala <sri@us.ibm.com> Date: Mon Feb 5 14:48:54 2007 -0800 framtests update to 2.6.20 & ChangeLog update. commit 307f2b429f5d3f6a4849d4d653811c1f970c674f Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jan 26 15:19:15 2007 -0800 [Frame Tests]: Update frame tests for new SACK+SHUTDOWN handling New handling in which SHUTDOWN is sent first followed by a SACK (when in SHUTDOWN-SENT state) caused this test to fail because the order of the chunks changed. We now detect new order. commit 7860d6594a4a2e7dec9e4facc9e21baff0f7eb4b Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jan 12 11:11:53 2007 -0800 [Frame Tests]: Update frame tests for new unexpected INIT-ACKs This monster of test (ft_frame_connectx_collision3.c) was the only one that was failing. The problem was that it was holding back and re-inserting INIT-ACK at such a time that neither association nor the endpoint could be found in the hashes. This result in a proper OOTB ABORT on which the test died. This fix is to 'listen' on the endpoint so that it can be found in the hash. commit 2c500815e1db7fd59e4346213cc1435df5b2d286 Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jan 12 11:10:21 2007 -0800 [Frame Tests]: Fix timers using SCTP_RTO_* variables SCTP_RTO_* variables used to be in jiffies, but were converted to milliseconds (a while ago). This patch fixes the usage in the frametests. commit dfbbc9927fa680de4fa6aa44d7dabbfb51413a0d Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Fri Jan 12 10:43:01 2007 -0800 Build v6 tests only if IPV6 is enabled in the kernel config. commit ccb1963617d14b2af9e68cf0e146064dde0e5d84 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Dec 28 16:31:57 2006 -0800 Add support for sctp_getaddrlen() API. commit b0986da72b1dd333450ae15337bb8cdae16a1479 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Dec 28 15:55:26 2006 -0800 Add SCTP_CONTEXT socket option. commit aae0b9236f4ba64ecc301fec84996ca60be72c2e Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Dec 28 15:52:47 2006 -0800 Fix typo adaption -> adaptation as per the latest API draft. commit 6cf017319c62c2129cce32bef4b899f8c95366dd Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Dec 15 19:15:38 2006 -0800 Changelog update. commit a20fce5a3cb7e4d8270be9b42a7e0c6ec7b05cec Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Dec 15 19:05:37 2006 -0800 update frametests to work with 2.6.20-rc1 commit ea425754b0edbe2ae453f8464b0a3f7341d8c91e Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Dec 13 10:51:31 2006 -0800 Update sockets API to draft14 in doc directory delete mode 100644 doc/draft-ietf-tsvwg-sctpsocket-13.txt create mode 100644 doc/draft-ietf-tsvwg-sctpsocket-14.txt commit 571bc65a79b735a50c4833a75d031cdbcc16ece1 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Dec 6 12:17:28 2006 -0800 update frametests to 2.6.19+ (handle endian annotation fixes) commit a5a630940c849e59752e601dbc72df721501a8bb Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Dec 6 12:16:45 2006 -0800 Adding RFC4460 to doc create mode 100644 doc/rfc4460.txt commit 73a2d128a15e6473e1de51f56f1d0ca2ea60dfb9 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Nov 29 11:02:47 2006 -0800 frametests update to 2.6.19 commit 541c47e161553a936ba4a93a50a4d6622aef8e70 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Oct 5 10:23:45 2006 -0700 Changelog update commit f8b4068125beabbc058e5428ef527cea7ea5496b Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Oct 5 10:00:27 2006 -0700 Fixes to frametests to sync up with kernel changes to enable Nagle by default and including sk_buff overhead while updating the peer's rwnd. commit 176851f96c4e52edccf190389ed1d872fca298c2 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Sep 26 10:01:48 2006 -0700 frametests update to 2.6.18+ commit 3aeca711f2fb8318d571428985e9b33a3d16f49b Author: Sridhar Samudrala <sri@us.ibm.com> Date: Sat Sep 23 23:21:53 2006 -0700 Changelog update commit 641e51e246c96477d871b8a86d5725d60e1e0e80 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Mon Aug 28 14:13:19 2006 -0700 fix sctp_sendmsg() calls in ft_frame_assoc_abort.c commit f99433eaf94c19546bde2f9afd0bd9ae9a4ecef2 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Jul 19 11:02:24 2006 -0700 Fix bug in the calculation of path verification heartbeat timers. commit 21409ef317df7f7fbee0fb408d5128f9327e0f72 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Jul 19 11:01:28 2006 -0700 Add stubs for kmap_atomic, kunmap_atomic, local_bh_enable, local_bh_disable when CONFIG_HIGHMEM is set. commit 80e879d15f6a5d3fbfa0cebb0469957357d024be Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jul 18 17:23:49 2006 -0700 Changelog update. commit 434b686d987309c5918999abc6e0bc66e448b133 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jul 18 17:17:56 2006 -0700 frametests update to 2.6.18-rc2 commit 7129adf882991deba2a9bd869a2c78fbb49440e2 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jul 18 17:17:20 2006 -0700 Add testcases for path verification via heartbeat and fix the tests to do the path verification. commit a3846f34f1daeae2c8eb4f332806dbf15b4f8997 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Jul 6 15:14:07 2006 -0700 update to 2.6.18-rc1 commit 52ff53648550ced1e2ea01863b00e429c25bb98e Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jun 20 16:21:46 2006 -0700 doc updates delete mode 100644 doc/draft-blanton-dsack-use-02.txt delete mode 100644 doc/draft-coene-sctp-multihome-04.txt delete mode 100644 doc/draft-hss-conformance-test-sctp-00.txt delete mode 100644 doc/draft-ietf-sigtran-sctp-mib-10.txt delete mode 100644 doc/draft-ietf-sigtran-signalling-over-sctp-applic-09.txt delete mode 100644 doc/draft-ietf-sigtran-srwnd-sctp-00.txt delete mode 100644 doc/draft-ietf-tsvwg-addip-sctp-14.txt create mode 100644 doc/draft-ietf-tsvwg-addip-sctp-15.txt delete mode 100644 doc/draft-ietf-tsvwg-sctpimpguide-16.txt delete mode 100644 doc/draft-ietf-tsvwg-sctpsocket-12.txt create mode 100644 doc/draft-ietf-tsvwg-sctpsocket-13.txt delete mode 100644 doc/draft-iyengar-sctp-cacc-01.txt delete mode 100644 doc/draft-price-rohc-epic-sctp-00.txt delete mode 100644 doc/draft-riegel-tuexen-mobile-sctp-03.txt delete mode 100644 doc/draft-sjkoh-mobile-sctp-handover-00.txt delete mode 100644 doc/draft-sjkoh-mobile-sctp-mobileip-02.txt delete mode 100644 doc/draft-sjkoh-sctp-mobility-02.txt delete mode 100644 doc/draft-stewart-rddp-sctp-02.txt delete mode 100644 doc/draft-stewart-tsvwg-sctpipv4-00.txt delete mode 100644 doc/draft-tuexen-tsvwg-tls-over-sctp-00.txt delete mode 100644 doc/draft-west-sctp-epic-01.txt create mode 100644 doc/rfc3873.txt commit 6537931b52c400fbfdb04432552b1c3b5fe2d784 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jun 20 16:21:27 2006 -0700 update ___pskb_trim() in test_kernel.c to match linux 2.6.17+ commit 2f7d6aa7d5516f4a2a60396a9443ecf821418008 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jun 20 11:28:28 2006 -0700 Changelog update. commit a893282202a42e4575b10e0fb0f2fbaa35162208 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jun 20 11:27:36 2006 -0700 Initialize rt->rt_flags in test_kernel.c:ip_route_output_key commit ebfc5bf65b8ab8aa1e73bc7caa566b82d9a6afd7 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue May 9 14:30:39 2006 -0700 update to 2.6.16+ commit da93eef44e90b382e582d94f59880196e4d5685f Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri May 5 15:53:16 2006 -0700 Fix the type of 'len' argument in sctp_recvmsg() man page. commit dcf53cb1674b09c6389734824fbcf3b3877fc079 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri May 5 15:51:53 2006 -0700 Update a couple of drafts in the doc directory. delete mode 100644 doc/draft-ietf-tsvwg-addip-sctp-13.txt create mode 100644 doc/draft-ietf-tsvwg-addip-sctp-14.txt delete mode 100644 doc/draft-ietf-tsvwg-sctpsocket-11.txt create mode 100644 doc/draft-ietf-tsvwg-sctpsocket-12.txt commit d472cef604c9e24133892e3e244de92888bd1cf2 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Feb 3 10:58:37 2006 -0800 lksctp-tools 1.0.6 release commit be2c27d57bb063645c4717e7b99a1ced62efe0f3 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Feb 1 14:48:47 2006 -0800 Fix ft_frame_congestion to match the change in tsn_missing_reports to 3. commit b9db385c1393944ced6fc346640b4d662f78d26e Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Mon Jan 30 11:25:03 2006 -0800 Fix ft_frame_init_timer to correctly count init retransmissions. commit 9641a0ceaf336ce9553515e82274747734142bcc Author: Vlad Yasevich <vladislav.yasevich@hp.com> Date: Mon Jan 30 11:12:04 2006 -0800 New frametest to make sure that heartbeats are transmitted a total of max_retrans times. create mode 100644 test/ft_frame_hb_rtx.c commit fc0b5cfbbe056023cae3cdc3adbf2b5c29aa061a Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Jan 27 16:45:57 2006 -0800 Changelog update commit 0e9e0b91e61f6dfbb0f06bf69275c9f268c261da Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Jan 27 16:45:36 2006 -0800 Add -D KBUILD_BASENAME to CFLAGS to avoid build errors with latest git tree. commit 29a08fd74b8f5a33d98bff4716115dd99f3298b6 Author: Frank Filz <ffilz@us.ibm.com> Date: Tue Jan 17 16:22:09 2006 -0800 Add testcases for updated SCTP_PEER_ADDR_PARAMS and the new SCTP_DELAYED_ACK_TIME socket option. commit d5829c9ffe963ee1a7778961a63f924b66eb1a2b Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Jan 12 17:26:20 2006 -0800 update to 2.6.15+ commit 1aff1d738fa9f3cf0708771195ff6211d9686b1f Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Jan 3 15:58:39 2006 -0800 lksctp-tools 1.0.5 release commit 4c46df1a7d7c14f1b6d06d286585619e2cf23dca Author: Sridhar Samudrala <sri@us.ibm.com> Date: Sun Jan 1 22:24:16 2006 -0800 minor updates to a couple of sctp man pages. commit 5b057eaaea489125eb4987b590fd3bcec8b093f7 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Sun Jan 1 22:22:35 2006 -0800 Fix TST_TOTAL values in some of the functional tests. commit 87f38123645c8678b53edd6cfe6b7425e28ee706 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Sun Jan 1 22:20:15 2006 -0800 update addip draft to the latest version in the doc directory. delete mode 100644 doc/draft-ietf-tsvwg-addip-sctp-12.txt create mode 100644 doc/draft-ietf-tsvwg-addip-sctp-13.txt commit a6315ed75b486617644fd550107c463ca5246dc2 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Dec 20 10:46:02 2005 -0800 Remove SCTP_PEER_ADDR_PARAMS tests from test_sockopt.c that are based on older version of the SCTP sockets API draft. commit bc5d93841bad933ae125b3c6408c76d7f93fc312 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Sat Nov 12 18:47:44 2005 -0800 Fix to build frametests with NETFILTER enabled. commit 14ad23514f57fc1ffe43eef143354866b869e718 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Nov 11 11:03:05 2005 -0800 changelog update. commit 4d46cbaf2afe869d398c7477b30db914a4e47b9d Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Nov 11 10:52:26 2005 -0800 Fix ipv6_addr_type and kmem_cache_create to match with mainline changes. commit f5e75d4d2918b8284d002cd388b329e12aec8d0e Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Nov 10 16:42:11 2005 -0800 fixes to address ulprcvbuf accounting changes commit 48f9b782ce28fccc857e91a1347a6a171542d409 Author: Neil Horman <nhorman@tuxdriver.com> Date: Wed Nov 2 23:27:50 2005 -0800 Fix functional tests to work with recieve buffer accounting changes. commit f1110b56e4b25ad0da4e027d46f5d0174506150b Author: Sridhar Samudrala <sri@us.ibm.com> Date: Wed Nov 2 23:23:47 2005 -0800 update to 2.6.14+ commit 5d029d2cb487b893bb761eca97472dba404731cf Author: Ivan Skytte Jorgensen <isj-sctp@i1.dk> Date: Mon Oct 31 11:55:53 2005 -0800 Rename SCTP specific control message flags to use SCTP_ prefix rather than MSG_ prefix as per the latest sctp sockets API draft. commit 8ad34d370a0eba32d5f5301f936f474dc6f969ac Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Fri Oct 28 11:04:18 2005 -0700 lksctp-tools 1.0.4 release commit 52a91b9ed5c0cf7cf565e2d0d61d36f395e476b5 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Fri Oct 28 11:02:50 2005 -0700 update sctpimpguide to the latest draft 16 in the doc directory. delete mode 100644 doc/draft-ietf-tsvwg-sctpimpguide-15.txt create mode 100644 doc/draft-ietf-tsvwg-sctpimpguide-16.txt commit 49d57a9a48c8b9b774a15b7a817b8204437f09a7 Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Thu Oct 27 16:42:45 2005 -0700 Add the missing file sctp_send.3 create mode 100644 man/sctp_send.3 commit cd8af66288f0b0bf216c0d2514467f637c0e89f2 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Tue Oct 25 14:19:39 2005 -0700 Add manpage for sctp_send() commit 91239acfc91f0bbd9e28ba84049a40c747c391db Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Fri Oct 21 10:42:09 2005 -0700 Add sctp_send() API support and testcases commit 8287f40463dc719569eafea7cbbf10e227813b3d Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Mon Oct 17 14:29:05 2005 -0700 Add preprocessor constants for PR-SCTP, ADDIP and CANSET_PRIMARY commit c50a9115989ae34d865877a74d265f60fa8990a3 Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Fri Oct 14 10:58:29 2005 -0700 update to 2.6.14 commit c69e853435f9cc1a0bb6cde1187eee52922751f1 Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Mon Oct 10 16:54:42 2005 -0700 Update rfcs/drafts in docs directory. delete mode 100644 doc/draft-ietf-tsvwg-prsctp-03.txt delete mode 100644 doc/draft-ietf-tsvwg-sctpimpguide-14.txt create mode 100644 doc/draft-ietf-tsvwg-sctpimpguide-15.txt delete mode 100644 doc/draft-ietf-tsvwg-sctpsocket-10.txt create mode 100644 doc/draft-ietf-tsvwg-sctpsocket-11.txt commit 06cefae87a9998bdcec2b6c0e033c7030915ee90 Author: Sridhar Samudrala <sri@us.ibm.com> Date: Thu Oct 6 10:52:20 2005 -0700 Fix ft_frame_getfreeaddrs() to use the old socket options. commit 95677cb2fea4e90e44193ec3433b7d71de8b63fb Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Thu Oct 6 10:51:07 2005 -0700 Fix alignment of structures used with SCTP socket options to allow 32-bit apps work with 64-bit kernels. commit d1b6966e95a4e244db8488e4b9c1f20c4e84f844 Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Thu Oct 6 10:49:38 2005 -0700 Fix sctp_get{pl}addrs() API to work with 32-bit apps on 64-bit kernels. commit bc385c08bb15de607f89729fc3b53d212cbcc94d Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Mon Sep 26 20:53:47 2005 -0700 Ignore MSG_CMSG_COMPAT flag leaked to 32-bit userspace with 64-bit kernels <= 2.6.13. commit c30614a260c615c2351965601dbffcebed053bed Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Mon Sep 26 15:49:32 2005 -0700 Add testcase to validate reception of SCTP_SHUTDOWN notification on 1-1 style sockets. commit d94aca5ac587dfffe80ca3aa5ca4de5442ca3881 Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Thu Sep 1 14:34:22 2005 -0700 lksctp-tools 1.0.3 release delete mode 100644 doc/.cvsignore delete mode 100644 src/.cvsignore delete mode 100644 src/include/.cvsignore delete mode 100644 src/include/netinet/.cvsignore commit 64541346063bd6f5f75b7a2a10ad1293e2dd126b Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Tue Aug 30 14:40:45 2005 -0700 Fix a couple of frame tests to do a recvmsg() before close(). commit 626c34ef557a8498a484159121db808007bc301e Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Mon Aug 1 23:16:16 2005 -0700 Fix gcc 4.0 warnings. commit 47479393f2227b24dc484a63ed7f19662fe4f55e Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Tue Jul 19 11:02:49 2005 -0700 update to 2.6.13-rc3 commit 6d30fc0c6c2696f73ad2e7b6e3db708280f9cec8 Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Fri Jul 1 11:02:26 2005 -0700 update to 2.6.13-rc1 commit 19b55a2af145fe769d76f7518eacabc5ff0e9dae Author: Sridhar Samudrala <sridhar@w-sridhar2.beaverton.ibm.com> Date: Fri Jun 24 15:43:21 2005 -0700 Initial commit last bk lksctp-tools create mode 100644 AUTHORS create mode 100644 COPYING create mode 100644 COPYING.lib create mode 100644 ChangeLog create mode 100644 INSTALL create mode 100644 Makefile.am create mode 100644 Makefile.dirs create mode 100644 Makefile.rules create mode 100644 Make ... |