OpenBSD src changes summary for 2016-08-25 ==========================================
bin/dd bin/pax lib/libc regress/sys regress/usr.sbin share/man sys/arch/arm/arm sys/arch/arm/include sys/isofs/udf sys/kern sys/netinet6 sys/sys usr.bin/grep usr.bin/openssl usr.bin/ssh usr.bin/tmux usr.sbin/inetd usr.sbin/pkg_add usr.sbin/rtadvd usr.sbin/switchd == bin =============================================================== 01/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/bin dd ~ misc.c > remove lint comments (tedu@) pax ~ ar_subs.c ~ cache.c ~ cache.h ~ extern.h ~ gen_subs.c ~ options.c ~ tar.c > Replace name_{uid,gid}() with the libc routines user_from_uid() and > group_from_gid(). Eliminate some superfluous strncpy() calls. > ok millert@ (guenther@) == lib =============================================================== 02/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib libc ~ stdio/fgetln.c > After read errors, fgetln(3) sometimes succeeded (returning non-NULL) > and failed (setting errno and ferror(3)) both at the same time. > That's a bad idea in general, and here in particular since > returning partial lines was neither reliable (sometimes, you > got NULL anyway) nor predictable (almost always, the line would > be truncated long before the actual read error). > Instead, on read failure, fail properly and always return NULL. > Issue found in a discussion with Andrey Chernov <ache at freebsd dot org> > who finally agreed to move FreeBSD into the same direction. > The fix is joint work with and OK by millert@. (schwarze@) == regress =========================================================== 03/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress sys ~ uvm/mmap_fixed/mmap_fixed.c > do not request RWX mappings, RW will do (deraadt@) ~ kern/sosplice/error/args-idle-EINVAL.pl ~ kern/sosplice/error/args-max-EINVAL.pl > Fix sosplice tests on 32 bit systems by putting the correct number > of bytes into the time_t Perl pack template. Also fix error messages. > (bluhm@) usr.sbin ~ relayd/Proc.pm ~ relayd/Server.pm ~ relayd/args-http-slow-consumer.pl ~ relayd/relayd.pl ~ relayd/remote.pl > Make relayd test slow-consumer more reliable. Set SO_SNDTIMEO > sockopt correctly with Perl pack on i386. Make it possible to grep > in each others client and server logfile. Client does not simply > sleep but waits for short write at server. (bluhm@) ~ relayd/args-http-slow-consumer.pl > Set relayd socket buffer size to reasonable value to make test pass > in different environments. (bluhm@) == share ============================================================= 04/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/share man ~ man4/pfsync.4 > Fix pfsync(4)'s carp examples: as of ip_carp.c r1.245, carpdev must be > specified. From Bryan Stenson. (sthen@) ~ man5/port-modules.5 > Amend after clean-up of fortran.port.mk (dcoppa@) == sys =============================================================== 05/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys arch/arm/arm ~ cpufunc.c > Enable the UWXN bit in the SCTRL register when available. This should > prevent the kernel from accidentally executing userland pages that are > writable. > ok jsg@, patrick@ (kettenis@) arch/arm/include ~ armreg.h > Enable the UWXN bit in the SCTRL register when available. This should > prevent the kernel from accidentally executing userland pages that are > writable. > ok jsg@, patrick@ (kettenis@) isofs/udf ~ udf_vfsops.c > pool_setipl for udf > ok phessler@ krw@ (dlg@) kern ~ kern_descrip.c ~ kern_event.c ~ kern_proc.c ~ kern_resource.c ~ kern_sig.c > pool_setipl > ok kettenis@ (dlg@) ~ vfs_cache.c ~ vfs_init.c ~ vfs_lockf.c ~ vfs_subr.c > pool_setipl > ok kettenis@ (dlg@) ~ uipc_socket.c > Completely revert the M_WAIT change on the cluster allocation and > bring back the behaviour of rev 1.72. Although allocating small > mbufs when allocating an mbuf cluster fails seems suboptimal, this > should not be changed as a side effect when introducing m_getuio(). > OK claudio@ (bluhm@) ~ uipc_socket.c > Spliced TCP sockets become faster when the output part is running > as its own task thread. This is inspired by userland copy where a > process also has to go through the scheduler. This gives the socket > buffer a chance to be filled up and tcp_output() is called less > often and with bigger chunks. > When two kernel tasks share all the workload, the current scheduler > implementation will hang userland processes on single cpu machines. > As a workaround put a yield() into the splicing thread after each > task execution. This reduces the number of calls of tcp_output() > even more. > OK tedu@ mpi@ (bluhm@) netinet6 ~ ip6_output.c > Simplify ip6_getpmtu() to use a 'struct rtentry *' instead of two > 'struct route_in6 *'. > This is another little step towards deprecating 'struct route{,_in6}' > Inputs from and ok bluhm@ (mpi@) ~ nd6_rtr.c > Remove a check to expire received prefix. > This check does not make sense since November 2000 when IPv6 autoconf > address deletion has been made independent from prefix lifetimes. > Fix a case when a route was added to the table but the corresponding > address was not, leaving v6 unusable. > Found the hardway by and ok sthen@ (mpi@) sys ~ socketvar.h > Spliced TCP sockets become faster when the output part is running > as its own task thread. This is inspired by userland copy where a > process also has to go through the scheduler. This gives the socket > buffer a chance to be filled up and tcp_output() is called less > often and with bigger chunks. > When two kernel tasks share all the workload, the current scheduler > implementation will hang userland processes on single cpu machines. > As a workaround put a yield() into the splicing thread after each > task execution. This reduces the number of calls of tcp_output() > even more. > OK tedu@ mpi@ (bluhm@) == usr.bin =========================================================== 06/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin grep ~ util.c > when using -o, we may restart a match in the middle of the line. > set NOTBOL so that anchored patterns don't match. > from a patch by Daniël de Kok in freebsd bug 201650 > ok martijn (tedu@) openssl ~ openssl.1 > shorten the pkeyutl text; > help/ok guenther (jmc@) ssh ~ readconf.c > remove superfluous NOTREACHED comment (djm@) ~ servconf.c > add a sIgnore opcode that silently ignores options and use it to > suppress noisy deprecation warnings for the Protocol directive. > req henning, ok markus (djm@) tmux ~ cmd-display-message.c > Do not crash if display-message used without a client, issue reported by > Serge Aleynikov, fix from Thomas Adam. (nicm@) == usr.sbin ========================================================== 07/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin inetd ~ inetd.c > remove lint comments (tedu@) pkg_add ~ pkg_check.8 ~ OpenBSD/PkgCheck.pm > make the full filesystem check optional, by popular demand. > doubling -q removes the plist thorough check. (espie@) ~ pkg_check.8 ~ OpenBSD/PkgCheck.pm > -F before -f (as -I before -i currently is...); (jmc@) rtadvd ~ rtadvd.conf.5 > Fix rtadvd.conf(5) example for pinfoflags, it accepts numbers not strings. > ok phessler (sthen@) switchd ~ ofp.h > goda@ contributed most of the OpenFlow 1.3.5 parts in ofp.h (for the > "still-unreleased" switch(4) driver), so it is only fair to add his > copyright. This header file will eventually move to net/ofp.h. (reyk@) ===============================================================================
_______________________________________________ odc mailing list [email protected] http://www.squish.net/mailman/listinfo/odc
