sr-dev August 2010

sr-dev@lists.kamailio.org

26 participants
346 discussions

git:sr_3.0: io_wait: kqueue: use the entire array during too many errors fallback
by Andrei Pelinescu-Onciul 19 Aug '10

19 Aug '10

Module: sip-router Branch: sr_3.0 Commit: e4d53ae023d3b50702c635beb521e313431cff66 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=e4d53ae… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Thu Jul 8 15:18:52 2010 +0200 io_wait: kqueue: use the entire array during too many errors fallback Minor fix/optimization: if there are too many errors in the changelist and the kevent() call has … [View More]

1 0

git:sr_3.0: io_wait: kqueue: handle ENOENT and more robust error handling
by Andrei Pelinescu-Onciul 19 Aug '10

19 Aug '10

Module: sip-router Branch: sr_3.0 Commit: 4cfd8e1d074e9bb76b20ca3b5722f4a02dd29f15 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=4cfd8e1… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Sat Jun 19 00:44:24 2010 +0200 io_wait: kqueue: handle ENOENT and more robust error handling - handle also ENOENT (along EBADF) when kevent fails due to errors in the changelist. ENOENT can be … [View More]returned in the following valid scenario: fd scheduled for delayed removal from the watched fd list, fd closed (which automatically removes the fd from the kqueue watched list), new opened fd which gets the same number, delayed changes applied (kevent()). - treat all the other kevent errors or EV_ERRORs in a similar way but log them (at BUG() level). - return POLLERR|POLLHUP for EV_EOF with a non-null fflags. (only kqueue, meaning *bsd and darwin are affected by this fix) (cherry picked from commit 8966d0a1787152fa64d1f78321ee539116bd448a) --- io_wait.h | 92 +++++++++++++++++++++++++++++++++++++------------------------ 1 files changed, 56 insertions(+), 36 deletions(-) diff --git a/io_wait.h b/io_wait.h index b110b6b..c28f53d 100644 --- a/io_wait.h +++ b/io_wait.h @@ -259,34 +259,33 @@ static inline int kq_ev_change(io_wait_h* h, int fd, int filter, int flag, again: n=kevent(h->kq_fd, h->kq_changes, h->kq_nchanges, 0, 0, &tspec); if (unlikely(n == -1)){ - if (likely(errno == EBADF)) { + if (unlikely(errno == EINTR)) goto again; + else { + /* for a detailed explanation of what follows see + io_wait_loop_kqueue EV_ERROR case */ + if (unlikely(!(errno == EBADF || errno == ENOENT))) + BUG("kq_ev_change: kevent flush changes failed" + " (unexpected error): %s [%d]\n", + strerror(errno), errno); + /* ignore error even if it's not a EBADF/ENOENT */ /* one of the file descriptors is bad, probably already closed => try to apply changes one-by-one */ for (r = 0; r < h->kq_nchanges; r++) { retry2: n = kevent(h->kq_fd, &h->kq_changes[r], 1, 0, 0, &tspec); if (n==-1) { - if (errno == EBADF) - continue; /* skip over it */ - if (errno == EINTR) + if (unlikely(errno == EINTR)) goto retry2; - LOG(L_ERR, "ERROR: io_watch_add: kevent flush changes" - " failed: %s [%d]\n", - strerror(errno), errno); - /* shift the array */ - memmove(&h->kq_changes[0], &h->kq_changes[r+1], - sizeof(h->kq_changes[0])* - (h->kq_nchanges-r-1)); - h->kq_nchanges-=(r+1); - return -1; + /* for a detailed explanation of what follows see + io_wait_loop_kqueue EV_ERROR case */ + if (unlikely(!(errno == EBADF || errno == ENOENT))) + BUG("kq_ev_change: kevent flush changes failed:" + " (unexpected error) %s [%d] (%d/%d)\n", + strerror(errno), errno, + r, h->kq_nchanges); + continue; /* skip over it */ } } - } else if (errno == EINTR) goto again; - else { - LOG(L_ERR, "ERROR: io_watch_add: kevent flush changes" - " failed: %s [%d]\n", strerror(errno), errno); - h->kq_nchanges=0; /* reset changes array */ - return -1; } } h->kq_nchanges=0; /* changes array is empty */ @@ -1118,17 +1117,18 @@ again: n=kevent(h->kq_fd, h->kq_changes, apply_changes, h->kq_array, h->fd_no, &tspec); if (unlikely(n==-1)){ - if (errno==EINTR) goto again; /* signal, ignore it */ - else if (errno==EBADF) { + if (unlikely(errno==EINTR)) goto again; /* signal, ignore it */ + else { + /* for a detailed explanation of what follows see below + the EV_ERROR case */ + if (unlikely(!(errno==EBADF || errno==ENOENT))) + BUG("io_wait_loop_kqueue: kevent: unexpected error" + " %s [%d]\n", strerror(errno), errno); /* some of the FDs in kq_changes are bad (already closed) and there is not enough space in kq_array to return all of them back */ apply_changes = h->fd_no; goto again; - }else{ - LOG(L_ERR, "ERROR: io_wait_loop_kqueue: kevent:" - " %s [%d]\n", strerror(errno), errno); - goto error; } } /* remove applied changes */ @@ -1148,14 +1148,13 @@ again: r, n, h->kq_array[r].ident, (long)h->kq_array[r].udata, h->kq_array[r].flags); #endif - if (unlikely((h->kq_array[r].flags & EV_ERROR) && - (h->kq_array[r].data == EBADF || - h->kq_array[r].udata == 0))){ + if (unlikely((h->kq_array[r].flags & EV_ERROR) || + h->kq_array[r].udata == 0)){ /* error in changes: we ignore it if it has to do with a bad fd or update==0. It can be caused by trying to remove an already closed fd: race between adding something to the - changes array, close() and applying the changes. - E.g. for ser tcp: tcp_main sends a fd to child fore reading + changes array, close() and applying the changes (EBADF). + E.g. for ser tcp: tcp_main sends a fd to child for reading => deletes it from the watched fds => the changes array will contain an EV_DELETE for it. Before the changes are applied (they are at the end of the main io_wait loop, @@ -1163,6 +1162,16 @@ again: to tcp_main by a sender (send fail) is processed and causes the fd to be closed. When the changes are applied => error for the EV_DELETE attempt of a closed fd. + Something similar can happen when a fd is scheduled + for removal, is close()'ed before being removed and + re-opened(a new sock. get the same fd). When the + watched fd changes will be applied the fd will be valid + (so no EBADF), but it's not already watch => ENOENT. + We report a BUG for the other errors (there's nothing + constructive we can do if we get an error we don't know + how to handle), but apart from that we ignore it in the + idea that it is better apply the rest of the changes, + rather then dropping all of them. */ /* example EV_ERROR for trying to delete a read watched fd, @@ -1176,9 +1185,12 @@ again: udata = 0x0 } */ - if (h->kq_array[r].data != EBADF) - LOG(L_INFO, "INFO: io_wait_loop_kqueue: kevent error on " - "fd %ld: %s [%ld]\n", (long)h->kq_array[r].ident, + if (h->kq_array[r].data != EBADF && + h->kq_array[r].data != ENOENT) + BUG("io_wait_loop_kqueue: kevent unexpected error on " + "fd %ld udata %lx: %s [%ld]\n", + (long)h->kq_array[r].ident, + (long)h->kq_array[r].udata, strerror(h->kq_array[r].data), (long)h->kq_array[r].data); }else{ @@ -1186,20 +1198,28 @@ again: if (likely(h->kq_array[r].filter==EVFILT_READ)){ revents=POLLIN | (((int)!(h->kq_array[r].flags & EV_EOF)-1)&POLLHUP) | - (((int)!(h->kq_array[r].flags & EV_ERROR)-1)&POLLERR); + (((int)!((h->kq_array[r].flags & EV_EOF) && + h->kq_array[r].fflags != 0) - 1)&POLLERR); while(fm->type && (fm->events & revents) && (handle_io(fm, revents, -1)>0) && repeat); }else if (h->kq_array[r].filter==EVFILT_WRITE){ revents=POLLOUT | (((int)!(h->kq_array[r].flags & EV_EOF)-1)&POLLHUP) | - (((int)!(h->kq_array[r].flags & EV_ERROR)-1)&POLLERR); + (((int)!((h->kq_array[r].flags & EV_EOF) && + h->kq_array[r].fflags != 0) - 1)&POLLERR); while(fm->type && (fm->events & revents) && (handle_io(fm, revents, -1)>0) && repeat); + }else{ + BUG("io_wait_loop_kqueue: unknown filter: kqueue: event " + "%d/%d: fd=%d, filter=%d, flags=0x%x, fflags=0x%x," + " data=%lx, udata=%lx\n", + r, n, h->kq_array[r].ident, h->kq_array[r].filter, + h->kq_array[r].flags, h->kq_array[r].fflags, + (long)h->kq_array[r].data, (long)h->kq_array[r].udata); } } } } while(unlikely(orig_changes)); -error: return n; } #endif [View Less]

1 0

git:sr_3.0: io_wait: don't update FD watched status on error
by Andrei Pelinescu-Onciul 19 Aug '10

19 Aug '10

Module: sip-router Branch: sr_3.0 Commit: 311b29d5e73720050f2bc56c36432b2eb45bbe83 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=311b29d… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Fri Jun 18 09:48:21 2010 +0200 io_wait: don't update FD watched status on error If the syscall to change the events or delete a watched FD fails, don't update/delete the FD status in fd_hash. For /… [View More]dev/poll if a change fails when re-adding the FD, delete it from the hash (in the /dev/poll case to change the events a FD is watched for one has to remove it and re-add it with the new events). The syscalls should never fail in an un-handled way, but in the unlikely event that it happens this change will make the code more robust. (cherry picked from commit 2d8cd170ab867ab15296b30f0b784abe1adc1bca) --- io_wait.h | 15 +++++++++------ 1 files changed, 9 insertions(+), 6 deletions(-) diff --git a/io_wait.h b/io_wait.h index 93f1426..01df1e6 100644 --- a/io_wait.h +++ b/io_wait.h @@ -627,7 +627,6 @@ inline static int io_watch_del(io_wait_h* h, int fd, int idx, int flags) goto error; } events=e->events; - unhash_fd_map(e); switch(h->poll_method){ case POLL_POLL: @@ -647,7 +646,6 @@ inline static int io_watch_del(io_wait_h* h, int fd, int idx, int flags) #endif #ifdef HAVE_SIGIO_RT case POLL_SIGIO_RT: - fix_fd_array; /* the O_ASYNC flag must be reset all the time, the fd * can be changed only if O_ASYNC is reset (if not and * the fd is a duplicate, you will get signals from the dup. fd @@ -667,6 +665,7 @@ inline static int io_watch_del(io_wait_h* h, int fd, int idx, int flags) " failed: %s [%d]\n", strerror(errno), errno); goto error; } + fix_fd_array; /* only on success */ break; #endif #ifdef HAVE_EPOLL @@ -737,6 +736,7 @@ again_devpoll: h->poll_method); goto error; } + unhash_fd_map(e); /* only on success */ h->fd_no--; return 0; error: @@ -808,14 +808,14 @@ inline static int io_watch_chg(io_wait_h* h, int fd, short events, int idx ) add_events=events & ~e->events; del_events=e->events & ~events; - e->events=events; switch(h->poll_method){ case POLL_POLL: + fd_array_chg(events #ifdef POLLRDHUP - /* listen to POLLRDHUP by default (if POLLIN) */ - events|=((int)!(events & POLLIN) - 1) & POLLRDHUP; + /* listen to POLLRDHUP by default (if POLLIN) */ + | (((int)!(events & POLLIN) - 1) & POLLRDHUP) #endif /* POLLRDHUP */ - fd_array_chg(events); + ); break; #ifdef HAVE_SELECT case POLL_SELECT: @@ -921,6 +921,8 @@ again_devpoll2: LOG(L_ERR, "ERROR: io_watch_chg: re-adding fd to " "/dev/poll failed: %s [%d]\n", strerror(errno), errno); + /* error re-adding the fd => mark it as removed/unhash */ + unhash_fd_map(e); goto error; } break; @@ -931,6 +933,7 @@ again_devpoll2: h->poll_method); goto error; } + e->events=events; /* only on success */ return 0; error: return -1; [View Less]

1 0

git:sr_3.0: io_wait: fix kqueue and too many errors in changelist
by Andrei Pelinescu-Onciul 19 Aug '10

19 Aug '10

Module: sip-router Branch: sr_3.0 Commit: ef1e1b4af1e96ee46c08400acd05abbc43a51738 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=ef1e1b4… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Thu Jun 17 18:43:14 2010 +0200 io_wait: fix kqueue and too many errors in changelist kevent() tries to return errors in the changelist back in the supplied eventlist array. However if this is not … [View More]large enough, the whole kevent() syscall will fail. Now if kevent() fails with EBADF the call will be retried with a smaller set of changes, until the entire original changelist is applied. Fixes also kq_ev_change() flush mode: on error it will try to apply the changes one-by-one. (this affects only systems that have kqueue: *bsd and darwin) (cherry picked from commit b0bd3201826ee693d6f96c6e336477b9d9db7c32) --- io_wait.h | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++------ 1 files changed, 55 insertions(+), 7 deletions(-) diff --git a/io_wait.h b/io_wait.h index 44ef60c..93f1426 100644 --- a/io_wait.h +++ b/io_wait.h @@ -247,6 +247,7 @@ static inline int kq_ev_change(io_wait_h* h, int fd, int filter, int flag, void* data) { int n; + int r; struct timespec tspec; if (h->kq_nchanges>=h->kq_changes_size){ @@ -257,11 +258,36 @@ static inline int kq_ev_change(io_wait_h* h, int fd, int filter, int flag, tspec.tv_nsec=0; again: n=kevent(h->kq_fd, h->kq_changes, h->kq_nchanges, 0, 0, &tspec); - if (n==-1){ - if (errno==EINTR) goto again; - LOG(L_ERR, "ERROR: io_watch_add: kevent flush changes " + if (unlikely(n == -1)){ + if (likely(errno == EBADF)) { + /* one of the file descriptors is bad, probably already + closed => try to apply changes one-by-one */ + for (r = 0; r < h->kq_nchanges; r++) { +retry2: + n = kevent(h->kq_fd, &h->kq_changes[r], 1, 0, 0, &tspec); + if (n==-1) { + if (errno == EBADF) + continue; /* skip over it */ + if (errno == EINTR) + goto retry2; + LOG(L_ERR, "ERROR: io_watch_add: kevent flush changes" + " failed: %s [%d]\n", + strerror(errno), errno); + /* shift the array */ + memmove(&h->kq_changes[0], &h->kq_changes[r+1], + sizeof(h->kq_changes[0])* + (h->kq_nchanges-r-1)); + h->kq_nchanges-=(r+1); + return -1; + } + } + } else if (errno == EINTR) goto again; + else { + LOG(L_ERR, "ERROR: io_watch_add: kevent flush changes" " failed: %s [%d]\n", strerror(errno), errno); - return -1; + h->kq_nchanges=0; /* reset changes array */ + return -1; + } } h->kq_nchanges=0; /* changes array is empty */ } @@ -1076,22 +1102,43 @@ inline static int io_wait_loop_kqueue(io_wait_h* h, int t, int repeat) int n, r; struct timespec tspec; struct fd_map* fm; + int orig_changes; + int apply_changes; int revents; tspec.tv_sec=t; tspec.tv_nsec=0; + orig_changes=h->kq_nchanges; + apply_changes=orig_changes; + do { again: - n=kevent(h->kq_fd, h->kq_changes, h->kq_nchanges, h->kq_array, + n=kevent(h->kq_fd, h->kq_changes, apply_changes, h->kq_array, h->fd_no, &tspec); if (unlikely(n==-1)){ if (errno==EINTR) goto again; /* signal, ignore it */ - else{ + else if (errno==EBADF) { + /* some of the FDs in kq_changes are bad (already closed) + and there is not enough space in kq_array to return all + of them back */ + apply_changes = h->fd_no; + goto again; + }else{ LOG(L_ERR, "ERROR: io_wait_loop_kqueue: kevent:" " %s [%d]\n", strerror(errno), errno); goto error; } } - h->kq_nchanges=0; /* reset changes array */ + /* remove applied changes */ + h->kq_nchanges -= apply_changes; + if (unlikely(apply_changes < orig_changes)) { + orig_changes -= apply_changes; + memmove(&h->kq_changes[0], &h->kq_changes[apply_changes], + sizeof(h->kq_changes[0])*h->kq_nchanges); + apply_changes = orig_changes<h->fd_no ? orig_changes : h->fd_no; + } else { + orig_changes = 0; + apply_changes = 0; + } for (r=0; r<n; r++){ #ifdef EXTRA_DEBUG DBG("DBG: kqueue: event %d/%d: fd=%d, udata=%lx, flags=0x%x\n", @@ -1148,6 +1195,7 @@ again: } } } + } while(unlikely(orig_changes)); error: return n; } [View Less]

1 0

git:sr_3.0: io_wait: fix: check for EV_ERROR for kqueue()
by Andrei Pelinescu-Onciul 19 Aug '10

19 Aug '10

Module: sip-router Branch: sr_3.0 Commit: ea37f13d610fdab887a256960e4a056dfca45f4a URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=ea37f13… Author: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Committer: Andrei Pelinescu-Onciul <andrei(a)iptel.org> Date: Thu Jun 17 16:15:10 2010 +0200 io_wait: fix: check for EV_ERROR for kqueue() Re-enabled and enhanced the check for EV_ERROR when using kqueue (*bsd). This is needed to workaround errors reported by … [View More]kqueue when trying to delete (EV_DELETE) an already closed FD (this can happen in the tcp code, where we try to avoid applying immediately changes in the set of watched FDs and instead collect them and apply them after all the current kqueue events are processed => in some corner case situations it's possible to try to delete the FD from kqueue after the fd was close()'ed). This fix will ignore EV_ERROR with data == EBADF. All the other errors will result in a POLLERR flag for the callback. It fixes crashes with *bsd under tcp stress tests (lots of very short lived connections). (cherry picked from commit 05c080a53f88babad6729f79015f555f53fdf957) --- io_wait.h | 55 ++++++++++++++++++++++++++++++++++++++++--------------- 1 files changed, 40 insertions(+), 15 deletions(-) diff --git a/io_wait.h b/io_wait.h index 0557c00..44ef60c 100644 --- a/io_wait.h +++ b/io_wait.h @@ -45,6 +45,7 @@ * 2007-11-29 support for write (POLLOUT); added io_watch_chg() (andrei) * 2008-02-04 POLLRDHUP & EPOLLRDHUP support (automatically enabled if POLLIN * is set) (andrei) + * 2010-06-17 re-enabled & enhanced the EV_ERROR for kqueue (andrei) */ @@ -1097,31 +1098,55 @@ again: r, n, h->kq_array[r].ident, (long)h->kq_array[r].udata, h->kq_array[r].flags); #endif -#if 0 - if (unlikely(h->kq_array[r].flags & EV_ERROR)){ - /* error in changes: we ignore it, it can be caused by - trying to remove an already closed fd: race between - adding something to the changes array, close() and - applying the changes */ - LOG(L_INFO, "INFO: io_wait_loop_kqueue: kevent error on " - "fd %ld: %s [%ld]\n", h->kq_array[r].ident, + if (unlikely((h->kq_array[r].flags & EV_ERROR) && + (h->kq_array[r].data == EBADF || + h->kq_array[r].udata == 0))){ + /* error in changes: we ignore it if it has to do with a + bad fd or update==0. It can be caused by trying to remove an + already closed fd: race between adding something to the + changes array, close() and applying the changes. + E.g. for ser tcp: tcp_main sends a fd to child fore reading + => deletes it from the watched fds => the changes array + will contain an EV_DELETE for it. Before the changes + are applied (they are at the end of the main io_wait loop, + after all the fd events were processed), a CON_ERR sent + to tcp_main by a sender (send fail) is processed and causes + the fd to be closed. When the changes are applied => + error for the EV_DELETE attempt of a closed fd. + */ + /* + example EV_ERROR for trying to delete a read watched fd, + that was already closed: + { + ident = 63, [fd] + filter = -1, [EVFILT_READ] + flags = 16384, [EV_ERROR] + fflags = 0, + data = 9, [errno = EBADF] + udata = 0x0 + } + */ + if (h->kq_array[r].data != EBADF) + LOG(L_INFO, "INFO: io_wait_loop_kqueue: kevent error on " + "fd %ld: %s [%ld]\n", (long)h->kq_array[r].ident, strerror(h->kq_array[r].data), (long)h->kq_array[r].data); - }else{ -#endif + }else{ fm=(struct fd_map*)h->kq_array[r].udata; if (likely(h->kq_array[r].filter==EVFILT_READ)){ - revents=POLLIN | - (((int)!(h->kq_array[r].flags & EV_EOF)-1)&POLLHUP); + revents=POLLIN | + (((int)!(h->kq_array[r].flags & EV_EOF)-1)&POLLHUP) | + (((int)!(h->kq_array[r].flags & EV_ERROR)-1)&POLLERR); while(fm->type && (fm->events & revents) && (handle_io(fm, revents, -1)>0) && repeat); }else if (h->kq_array[r].filter==EVFILT_WRITE){ - revents=POLLOUT | - (((int)!(h->kq_array[r].flags & EV_EOF)-1)&POLLHUP); + revents=POLLOUT | + (((int)!(h->kq_array[r].flags & EV_EOF)-1)&POLLHUP) | + (((int)!(h->kq_array[r].flags & EV_ERROR)-1)&POLLERR); while(fm->type && (fm->events & revents) && (handle_io(fm, revents, -1)>0) && repeat); } - /*} */ + } } error: return n; [View Less]

1 0

git:carstenbock/rtpproxy: The Timeout socket must be sent with Update after session confirmation.
by Carsten Bock 19 Aug '10

19 Aug '10

Module: sip-router Branch: carstenbock/rtpproxy Commit: d16d34b1ed18e572705f1c1a8b3113f274e618b3 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=d16d34b… Author: Carsten Bock <lists(a)bock.info> Committer: Carsten Bock <lists(a)bock.info> Date: Thu Aug 19 15:33:30 2010 +0200 The Timeout socket must be sent with Update after session confirmation. --- modules_k/rtpproxy/rtpproxy.c | 10 +++++----- 1 files changed, 5 insertions(+), 5 deletions(-) … [View More]

1 0

preparing kamailio 3.0.3
by Daniel-Constantin Mierla 19 Aug '10

19 Aug '10

Hello, I am planning to release minor version 3.0.3 by beginning of next week. Please commit there what you think it needs to be fixed for this release. Thanks, Daniel -- Daniel-Constantin Mierla http://www.asipto.com/

2 4

New commits on branch sr_3.0
by Andrei Pelinescu-Onciul 19 Aug '10

19 Aug '10

URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=0451e20… Author: Henning Westerholt <henning.westerholt(a)1und1.de> Date: Wed Aug 18 13:45:37 2010 +0200 pdbt: fix two bugs in debian packaging (cherry picked from commit cd71125fcdb434d50f6e7b5d208d260ebc49fca6) URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=7465ebf… Author: Daniel-Constantin Mierla <miconda(a)gmail.com> Date: Thu Aug 12 17:14:38 2010 +0200 … [View More]

1 0

git:carstenbock/rtpproxy: Tiny typo.
by Carsten Bock 18 Aug '10

18 Aug '10

Module: sip-router Branch: carstenbock/rtpproxy Commit: 449cde48834574b069d930842d406ac709083654 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=449cde4… Author: Carsten Bock <lists(a)bock.info> Committer: Carsten Bock <lists(a)bock.info> Date: Wed Aug 18 22:57:37 2010 +0200 Tiny typo. --- modules_k/rtpproxy/rtpproxy.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/modules_k/rtpproxy/rtpproxy.c b/modules_k/rtpproxy/… [View More]

1 0

git:carstenbock/rtpproxy: Merge branch 'master' into rtpproxy
by Carsten Bock 18 Aug '10

18 Aug '10

Module: sip-router Branch: carstenbock/rtpproxy Commit: fdebd2b7a11dfb31d243b433f8fc5a389b6167a8 URL: http://git.sip-router.org/cgi-bin/gitweb.cgi/sip-router/?a=commit;h=fdebd2b… Author: Carsten Bock <lists(a)bock.info> Committer: Carsten Bock <lists(a)bock.info> Date: Wed Aug 18 22:44:53 2010 +0200 Merge branch 'master' into rtpproxy * master: modules*: add k stats. includes where needed kstats: cleanup - move type defs. to separate file modules_k/*: add missing lib/… [View More]

1 0

← Newer
1
...
12
13
14
15
16
17
18
...
35
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

sr-dev August 2010