summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2015-11-28Back out previous commit. Not emitting DT_RPATH isn't helpful if our ld.soMark Kettenis
doesn't grok DT_RUNPATH yet.
2015-11-28Use SOCK_NONBLOCK in relayd as well.Reyk Floeter
OK benno@
2015-11-28Rework gem_start to check that there's enough space in the ring beforeJonathan Matthew
dequeueing a packet, then drop the packet if it can't be sent, rather than using two phase dequeue then commit. Adjust sc_tx_cnt using atomic operations, take the kernel lock before calling gem_start or gem_init from the interrupt handler, and use an interrupt barrier when taking the interface down. With all this done, we can mark the interrupt handler for pci (but not sbus) attached gem(4) as mpsafe. mpi@ wrote this initially, I fixed bugs to keep sparc64 happy and kept it up to date with ifq changes. ok mpi@ dlg@
2015-11-28do not leak temp file if there is no known_hosts fileTheo de Raadt
from craig leres, ok djm
2015-11-28don't include port number in tcpip-forward replies for requestsDamien Miller
that don't allocate a port; bz#2509 diagnosed by Ron Frederick ok markus
2015-11-28Oops, left the pledge "tty" active in the previous commitedTheo de Raadt
pointed out by Chris Hettrick
2015-11-28chflags -l flag went away. Noticed this myself 15 minutesTheo de Raadt
before Chris Hettrick did.
2015-11-28add missing "log" bitsGleydson Soares
spotted out by Luiz Gustavo Costa < luizgustavo at mundounix.com.br >, Thanks! OK jmc@ jsg@
2015-11-28fix exit status on pledge(2) error, where it should be >1Gleydson Soares
OK millert@ deraadt@
2015-11-27Imporve error messages for the imsg handler code. OK sthen@Claudio Jeker
2015-11-27add getpw to pledge. rpath would normally suffice, but there's some doubleTed Unangst
checking code in snapshots, and it serves as a useful annotation. from Carlin Bingham
2015-11-27after reading a too long line, restart at the beginning of the buffer soTed Unangst
we don't keep writing past the end. (the perils of trying to recover from parse errors.) noticed by Jan Schreiber
2015-11-27Two additional ioctls for pledge("disklabel"), needed by installboot.Jeremie Courreges-Anglas
Committing on behalf of tb@, problem reported by Rolf Sommerhalder on misc@.
2015-11-27the rm code in mv inherited pledge from rm. it shoud be safe, but thereTed Unangst
are cases when it is too strict. after pledge() we continue to run some of the cp code which requires a very broad set of permissions. remove pledge entirely for now, pending a complete analysis of needs.
2015-11-27spacesTheo de Raadt
2015-11-27After evaluating patches from Ricardo Mestre, it became obvious thatSebastian Benoit
route needs pledge("stdio rpath dns") in all cases. Looks best to me deraadt@
2015-11-27Keep "struct vxlan_softc" private to prevent pulling more headers whenMartin Pieuchot
<net/if_vxlan.h> is included.
2015-11-27Don't panic in pmap_enter() if we deplete the pool of pv entries and theMark Kettenis
PMAP_CANFAIL flag is set. Return ENOMEM instead.
2015-11-27Don't report a bus conflict for bridges that are left (partly) unconfigured byMark Kettenis
the system firmware.
2015-11-27Do not set a limit on the length of commands when printing them.Nicholas Marriott
2015-11-27Keep lo(4) definitions inside if_loop.cMartin Pieuchot
2015-11-27Use strlcpy to copy strings in struct mount, not memmove. AvoidsTodd C. Miller
reading past the end of a buffer. Problem found and OK jsg@
2015-11-27Document that routing table heads are never freed as suggested by dlg@Martin Pieuchot
and kill rtable_put() because we're not going to use it. The overhead of keeping a "struct art_root/radix_node_head" around is very small compared to the added complexity needed to reference count such structures.
2015-11-27Protect the growth of the routing table arrays used by rtable_get()Martin Pieuchot
with SRPs. This is a simplified version of the dynamically sizeable array of pointers used by if_get() because routing table heads are never freed. ok dlg@
2015-11-27Attach the iPhone 6 as ugen(4), from Laurent GUALDI.Martin Pieuchot
2015-11-27regenMartin Pieuchot
2015-11-27iPhone 6, from Laurent GUALDI.Martin Pieuchot
2015-11-27Fix after removal of the setgid bit:tb
Add pledge "stdio rpath wpath cpath tty". Move score file to $HOME, add $USER as high score entry. Maintain hall of fame of 10 high scores, make snake wink if it ate you while you were eligible for eternal fame. Based on an initial diff by Ricardo Mestre and with helpful input from tedu@. ok beck@
2015-11-27Change the command line syntax to a "command + getopts" style, eg.Reyk Floeter
vmmctl start "myvm" -m 512M -i 1 -d disk.img -k /bsd Networking people like me are used to the human-readable CLI-style syntax that we have in some of our OpenBSD *ctl tools, including bgpctl, other routing daemons, or relayctl. But this didn't work for vmmctl, so change it to something a bit less human-readable but more BSD- and POSIX-ish. After all, the command line is only intended for simple configuration, more details will go into the configuration file, so we can avoid getopt flag conflicts or scary things like the getsubopt(3) style. OK krw@ jasper@, with input from kettenis@ jmc@ deraadt@ mlarkin@ etc.
2015-11-27correct logic for a IEEE80211_MODE_11N testJonathan Gray
ok stsp@
2015-11-27syncTheo de Raadt
2015-11-27yppasswd went awayTheo de Raadt
2015-11-27Remove three NULL-checks before free(). ok millert@mmcc
2015-11-27syncTheo de Raadt
2015-11-27pledge "stdio rpath wpath cpath fattr tty proc exec" except for theTheo de Raadt
-p option (which sadly has insane semantics...) ok semarie dtucker
2015-11-26Use the backchannel for all error messages instead of syslog(3).Todd C. Miller
OK deraadt@ beck@
2015-11-26document MODPY_SETUP_ARGSJasper Lievisse Adriaanse
2015-11-26Avoid setting mode to GOP if the mode is unchanged. Also don't panicYASUOKA Masahiko
if the setting mode is failed. reported and tested by Joe Gidi
2015-11-26Fix "Bad value" error seen when building Mesa 11.0.6 which uses -Bsymbolic.Mark Kettenis
Inspired by https://sourceware.org/ml/binutils/2010-08/msg00333.html, but expressed differently so there are no GPLv3 issues.
2015-11-26passwd no longer has an -l flag; ok deraadtJason McIntyre
2015-11-26remove rpc.yppasswdd bits;Jason McIntyre
2015-11-26pledge in login_radius(8).YASUOKA Masahiko
ok deraadt
2015-11-26rpc.yppasswdd goes to the bitbucket.Theo de Raadt
(The story is that some people do use YP still as a libc -> non-file lookup mechanism; some of them with standard ypbind/ypserv, but more of them with ypbind/ypldap. That however does not justify these cruddy tools dealing with passwords, which are more likely to contain problems)
2015-11-26Apply pledge.Theo de Raadt
The people I asked to review this did not get back to me, so we will test this a different way.
2015-11-26Delete YP password related code. As a result, these can also beTheo de Raadt
pledged. Keep an eye out for regressions, because they could be uncomfortable. ok beck semarie
2015-11-26Delete YP lookup code; user base of 0.Theo de Raadt
2015-11-26syncTheo de Raadt
2015-11-26Disable by default the *.emerg block; ok millert beckTheo de Raadt
2015-11-26Fix copy-paste error in scorefile nametb
From Ricardo Mestre
2015-11-26Add checks for getenv(), snprintf() and getlogin().tb
From Ricardo Mestre (up to a small tweak). Thanks!