| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2002-02-12 | Fix my total blackout of the brain. | Artur Grabowski | |
| I rearrange half of this function to properly handle unlocking in one error case, then I forget to handle that error case. I even had a regression test written for this! In other words: fix accept(2) to unlock when it runs out of fds. | |||
| 2002-02-12 | Fix all users of getsock to use FREF/FRELE properly. | Artur Grabowski | |
| 2002-02-11 | In {send,recv}msg allow zero sized message. We can have control data. | Artur Grabowski | |
| 2002-02-08 | - Rename FILE_{,UN}USE to FREF and FRELE. USE is a bad verb and we don't have | Artur Grabowski | |
| the same semantics as NetBSD anyway, so it's good to avoid name collissions. - Always fdremove before freeing the file, not the other way around. - falloc FREFs the file. - have FILE_SET_MATURE FRELE the file (It feels like a good ortogonality to falloc FREFing the file). - Use closef as much as possible instead of ffree in error paths of falloc:ing functions. closef is much more careful with the fd and can deal with the fd being forcibly closed by dup2. Also try to avoid manually calling *fo_close when closef can do that for us (this makes some error paths mroe complicated (sys_socketpair and sys_pipe), but others become simpler (sys_open)). | |||
| 2001-12-27 | makes a new accept() socket have the nonblock behavior. | Federico G. Schwindt | |
| slightly modified version from the original diff by theo, based on comments by niels@ and deraadt@; both ok'ed. | |||
| 2001-12-02 | release the right descriptors when pipe fails; bug report from | Niels Provos | |
| marcodsl@swbell.net | |||
| 2001-11-28 | avoid "thundering herd" problem in accept by waking just one process. | Niels Provos | |
| based on freebsd. okay art@ markus@ | |||
| 2001-11-27 | change socket connection queues to use TAILQ_ | Niels Provos | |
| from NetBSD: Wed Jan 7 23:47:08 1998 UTC by thorpej Make insertion and removal of sockets from the partial and incoming connections queues O(C) rather than O(N). | |||
| 2001-10-26 | - every new fd created by falloc() is marked as larval and should not be used | Artur Grabowski | |
| any anyone. Every caller of falloc matures the fd when it's usable. - Since every lookup in the fd table must now check this flag and all of them do the same thing, move all the necessary checks into a function - fd_getfile. | |||
| 2001-09-20 | occured->occurred | Mike Pechkin | |
| idea from deraadt@ via NetBSD millert@ ok | |||
| 2001-09-17 | Don't treated return length of recvfrom() as a size_t... it's a socklen_t. | Jason Wright | |
| 2001-06-26 | implement djb's getpeereid(2), to allow local-domain servers to determine ↵ | Dug Song | |
| client credentials. mostly from superscript.com. deraadt@ ok | |||
| 2001-06-22 | KNF | Theo de Raadt | |
| 2001-05-16 | No need to check M_WAIT/M_WAITOK malloc return values. (art@ ok) | Hakan Olsson | |
| 2001-05-14 | gc OLD_PIPE. | Artur Grabowski | |
| 2001-02-19 | Indentation nit. | Artur Grabowski | |
| 2001-02-09 | return error from soaccept() upwards. part of accept(2) fix against RST | Jun-ichiro itojun Hagino | |
| right after handshake. | |||
| 2000-11-16 | support kernel event queues, from FreeBSD by Jonathan Lemon, | Niels Provos | |
| okay art@, millert@ | |||
| 2000-11-10 | Change the ktrace interface functions from taking the trace vnode to taking the | Artur Grabowski | |
| traced proc. The vnode is in the proc and all functions need the proc. | |||
| 2000-10-12 | allow buflen > MLEN for !SO_NAME case (like ancillary data, necessary for | Jun-ichiro itojun Hagino | |
| IPv6 advanced API). sync with netbsd. | |||
| 2000-09-27 | replace MALLOC/FREE w/ malloc/free for non-constant-sized memory ↵ | Michael Shalayeff | |
| allocations; art@ ok | |||
| 2000-01-17 | oops, another syscall pipe() botch; karls@inet.no | Theo de Raadt | |
| 1999-12-08 | bring in KAME IPv6 code, dated 19991208. | Jun-ichiro itojun Hagino | |
| replaces NRL IPv6 layer. reuses NRL pcb layer. no IPsec-on-v6 support. see sys/netinet6/{TODO,IMPLEMENTATION} for more details. GENERIC configuration should work fine as before. GENERIC.v6 works fine as well, but you'll need KAME userland tools to play with IPv6 (will be bringed into soon). | |||
| 1999-07-13 | introduce fdremove() to mark a file descriptor as unused. fdremove makes | Niels Provos | |
| sure that the fd_freefile hints stay in sync, otherwise free file descriptors might not be overlooked by fdalloc(); ok millert@ | |||
| 1999-06-15 | do not mangle retval, makes return value sane; problem found by millert | Theo de Raadt | |
| 1999-06-11 | oops | Niels Provos | |
| 1999-06-08 | copy int sized objects, not register_t sized objects in pipe() | Theo de Raadt | |
| 1999-06-08 | better fd leak prevention | Theo de Raadt | |
| 1999-06-07 | need seperate sys_pipe() versions, for pipeclose() or soclose() calls | Theo de Raadt | |
| 1999-06-07 | replacement pipe() system call; copies data into place inside kernel, so | Theo de Raadt | |
| that EFAULT return value is possible | |||
| 1999-06-07 | do not leak file descriptors if copyout() fails | Theo de Raadt | |
| 1999-06-07 | store NULL in fd_ofiles | Theo de Raadt | |
| 1999-02-15 | sendto(2) takes a socklen_t as well | Todd C. Miller | |
| 1999-02-15 | use socklen_t | Todd C. Miller | |
| 1999-02-14 | Use IOV_MAX, not the deprecated UIO_MAXIOV | Todd C. Miller | |
| 1999-02-14 | Don't do range checking on namelen unless name is non-NULL, otherwise | Todd C. Miller | |
| namelen is uninitialized. | |||
| 1999-02-14 | readv/writev with iov_len == 0 is legal (was already ok in uipc_syscalls.c). | Todd C. Miller | |
| Make the check more readable by comparing against SSIZE_MAX instead of checking for wraparound. This is safe because SSIZE_MAX * 2 <= SIZE_T_MAX. Fixes recno problems in the db routines exposed by a perl test. | |||
| 1999-02-11 | avoid insane lengths for various syscalls | Theo de Raadt | |
| 1998-09-14 | allow the recvfrom (fd, NULL, 0, MSG_PEEK, ...) again, pr 584. | Niels Provos | |
| 1998-08-05 | return EMSGSIZE, not EINVAL is msg_iovlen <= 0 as per XPG 4.2 | Todd C. Miller | |
| When comparing against UIO_SMALLIOV/UIO_MAXIOV check for >, not >= | |||
| 1998-08-05 | delete bogus casts of msg_iovlen to u_int since msg_iovlen is already a u_int | Todd C. Miller | |
| 1998-07-28 | For read, write, send, sendto, recv, recvfrom return EINVAL is size arg > ↵ | Todd C. Miller | |
| SSIZE_MAX. For readv, writev, sendmsg, recvmsg return EINVAL if sum of the over iov_len values overflows an ssize_t. Based on what XPG 4.2 says (though XPG is not entirely consistent here). Basically since the return value is ssize_t allowing size > SSIZE_MAX is bad since people who (incorrectly mind you) check the return value for < 0 instead of == -1 will have bad things happen to them. | |||
| 1998-07-28 | Don't assume uio_resid can go negative when detecting oflow | Todd C. Miller | |
| 1998-07-28 | Return EINVAL when msg_iovlen or iovcnt <= 0; Make uio_resid unsigned ↵ | Todd C. Miller | |
| (size_t) and don't return EINVAL if it is < 0 in sys_{read,write}. Remove check for uio_resid < 0 uiomove() now that uio_resid is unsigned and brack remaining panics with #ifdef DIAGNOSTIC. vn_rdwr() must now take a size_t * as its 9th argument so change that and clean up uses of vn_rdwr(). Fixes 549 + more | |||
| 1998-02-08 | Updates to match type changes in syscall tables | Thorsten Lockert | |
| 1997-12-16 | evaluate length before doing compat stuff which might overwrite the | Niels Provos | |
| length field we are going to evaluate. From FreeBSD, originally submitted by: pb@fasterix.freenix.org (Pierre Beyssac) | |||
| 1997-06-05 | delete cannot happen code | Theo de Raadt | |
| 1996-08-27 | New fast pipe(2) from freebsd without fancy vm stuff. | shawn | |
| The old pipes can be used with the "OLD_PIPE" config option. | |||
| 1996-05-18 | socketpair() is not a hack in libc like pipe() | Theo de Raadt | |
| 1996-03-03 | From NetBSD: 960217 merge | Niklas Hallqvist | |
 |
index : src | |
| OpenBSD base system |
| summaryrefslogtreecommitdiff |