OpenBSD src changes summary for 2015-09-13 to 2015-09-20 inclusive ==================================================================
bin/chmod bin/ed bin/ksh bin/pax distrib/alpha distrib/notes distrib/sets distrib/sgi etc/etc.sgi/Makefile.inc etc/netstart etc/rc etc/systrace/usr_sbin_lpd games/fortune gnu gnu/usr.bin/cc include/Makefile include/arpa/inet.h include/rpc/auth.h include/rpc/svc.h include/rpc/svc_auth.h include/tib.h lib/csu lib/libarch lib/libc lib/libcrypto lib/libedit lib/libkeynote lib/libm lib/libpthread lib/libsndio lib/libssl lib/libtls lib/libutil libexec/ld.so regress/bin regress/lib regress/sys regress/usr.bin sbin/ping6 sbin/swapctl share/man sys/arch/alpha/alpha sys/arch/alpha/conf sys/arch/alpha/include sys/arch/amd64/amd64 sys/arch/amd64/conf sys/arch/amd64/include sys/arch/amd64/stand sys/arch/amd64/stand/boot sys/arch/amd64/stand/cdboot sys/arch/amd64/stand/libsa sys/arch/amd64/stand/pxeboot sys/arch/arm/arm sys/arch/armish/include sys/arch/armv7/conf sys/arch/armv7/include sys/arch/hppa/gsc sys/arch/hppa/hppa sys/arch/hppa/include sys/arch/i386/conf sys/arch/i386/i386 sys/arch/i386/include sys/arch/i386/stand sys/arch/i386/stand/boot sys/arch/i386/stand/cdboot sys/arch/i386/stand/libsa sys/arch/i386/stand/pxeboot sys/arch/loongson/dev sys/arch/loongson/include sys/arch/m88k/m88k sys/arch/macppc/conf sys/arch/macppc/macppc sys/arch/macppc/stand sys/arch/mips64/mips64 sys/arch/octeon/dev sys/arch/octeon/include sys/arch/powerpc/include sys/arch/powerpc/powerpc sys/arch/sgi/dev sys/arch/sgi/hpc sys/arch/sgi/include sys/arch/sgi/localbus sys/arch/sgi/sgi sys/arch/sparc/dev sys/arch/sparc/include sys/arch/sparc/sparc sys/arch/sparc/stand/common sys/arch/sparc64/dev sys/arch/sparc64/include sys/arch/sparc64/sparc64 sys/arch/sparc64/stand/ofwboot sys/arch/vax/mbus sys/arch/vax/qbus sys/arch/vax/vax sys/arch/vax/vsa sys/arch/zaurus/include sys/arch/zaurus/stand/zboot sys/compat/linux sys/ddb sys/dev/ic sys/dev/pci sys/dev/sun sys/dev/usb sys/dev/wsfont sys/kern sys/lib/libsa sys/net sys/net80211 sys/netinet sys/netinet6 sys/netmpls sys/sys sys/uvm usr.bin/aucat usr.bin/awk usr.bin/calendar usr.bin/chpass usr.bin/col usr.bin/doas usr.bin/head usr.bin/hexdump usr.bin/kdump usr.bin/m4 usr.bin/mandoc usr.bin/mg usr.bin/nc usr.bin/nm usr.bin/openssl usr.bin/passwd usr.bin/ssh usr.bin/tail usr.bin/tmux usr.bin/vi usr.sbin/amd usr.sbin/bgpctl usr.sbin/bgpd usr.sbin/dhcpd usr.sbin/ldpctl usr.sbin/npppd usr.sbin/ospf6ctl usr.sbin/ospfctl usr.sbin/pkg_add usr.sbin/portmap usr.sbin/pppd usr.sbin/relayd usr.sbin/ripctl usr.sbin/rmt usr.sbin/syslogd usr.sbin/sysmerge usr.sbin/tcpdump usr.sbin/traceroute == bin =============================================================== 01/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/bin chmod ~ chown.8 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) ed ~ ed.1 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) ksh ~ ksh_limval.h ~ shf.c > zap unused FP support. from Michael McConville. ok nicm (tedu@) ~ tty.c > move tty_close up a bit so we can use it. > combine two equiv if blocks. > from Michael McConville (tedu@) - INSTALL > directions for using autoconf cannot be relevant. > from Michael McConville (tedu@) - ksh_limval.h ~ PROJECTS ~ sh.h ~ shf.c ~ var.c > Apply the ancient art of tedu to ksh_limval.h. ok guenther (nicm@) - BUG-REPORTS - ChangeLog - ChangeLog.0 - IAFA-PACKAGE - NEWS > these files are a combination of old, lies, and old lies. zap. > ok halex (tedu@) ~ c_ksh.c ~ exec.c ~ proto.h ~ sh.h ~ table.h ~ var.c > Replace Tflag typedef with just int; ok tedu (nicm@) ~ config.h > SIZEOF_INT is no longer needed. (nicm@) ~ ksh.1 ~ sh.1 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) ~ c_ksh.c ~ jobs.c ~ misc.c > stop abusing null (the variable) for an empty string in output formats. > from Michael McConville > ok nicm (tedu@) ~ c_ksh.c ~ c_sh.c ~ c_test.c ~ c_ulimit.c ~ edit.c ~ eval.c ~ exec.c ~ expr.c ~ history.c ~ io.c ~ jobs.c ~ lex.c ~ main.c ~ misc.c ~ path.c ~ shf.c ~ syn.c ~ trap.c ~ tree.c ~ var.c ~ vi.c > correct spelling of NULL from (char *)0. from Michael McConville. > ok md5 (tedu@) ~ shf.c > Expand the one use of POP_INT() macro into if()s and fix some errors > (now that llnum is long long, int needs to be sign extended on all > platforms, not just when sizeof(int) < sizeof(long); and sign extend > %ld, %li and %i as well as %d. Also simplify the code for %p since > pointers are always sizeof (long). > ok tedu (nicm@) ~ c_sh.c ~ edit.c ~ eval.c ~ exec.c ~ expand.h ~ expr.c ~ history.c ~ io.c ~ jobs.c ~ lex.c ~ mail.c ~ main.c ~ misc.c ~ shf.c ~ syn.c ~ table.c ~ tree.c ~ var.c ~ vi.c > Remove unnecessary casts, from Michael McConville. No binary change. > (nicm@) ~ history.c ~ jobs.c > More (foo *)0 -> NULL, from Michael McConville. No binary change. (nicm@) ~ c_ksh.c ~ c_sh.c ~ c_test.c ~ edit.c ~ emacs.c ~ eval.c ~ exec.c ~ expr.c ~ io.c ~ lex.c ~ main.c ~ misc.c ~ shf.c ~ syn.c ~ table.c ~ trap.c ~ var.c ~ vi.c > Last of the (thing *)0 -> NULL, from Michael McConville. No binary > change. (nicm@) pax ~ tar.1 > Describe tar's -p option correctly > noted by trondd(at)kagu-tsuchi.com > ok deraadt@ (guenther@) == distrib =========================================================== 02/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib alpha ~ floppy/Makefile ~ floppyB/Makefile ~ floppyC/Makefile > Remove bogus .PATH statement. (miod@) ~ common/Makefile.inc ~ inst-common/Makefile.inc > Add the kernel in its various setup stages (bsd, bsd.gz, bsd.strip...) to > CLEANFILES and make sure the `clean' target actually removes CLEANFILES. > (miod@) ~ miniroot/Makefile > Create the filesystem image with the same size as its `c' slice from > disktab. (miod@) notes ~ sgi/contents ~ sgi/hardware ~ sgi/prep ~ sgi/whatis > More IP26 bits. (miod@) sets ~ lists/base/md.alpha ~ lists/base/md.amd64 ~ lists/base/md.armish ~ lists/base/md.armv7 ~ lists/base/md.aviion ~ lists/base/md.hppa ~ lists/base/md.hppa64 ~ lists/base/md.i386 ~ lists/base/md.landisk ~ lists/base/md.loongson ~ lists/base/md.luna88k ~ lists/base/md.macppc ~ lists/base/md.octeon ~ lists/base/md.sgi ~ lists/base/md.socppc ~ lists/base/md.sparc ~ lists/base/md.sparc64 ~ lists/base/md.zaurus > sync (deraadt@) ~ lists/etc/mi ~ lists/base/mi > Move /etc/ssl/cert.pem from base to etc. Previously this was updated every > upgrade when untarring base sets; now it is done when sysmerge is run; if > there are no local changes the update is seamless, otherwise with local > changes (adding your own CAs, or removing ones which you don't want to > trust) sysmerge will ask to diff as usual. ok miod deraadt (sthen@) ~ lists/base/md.alpha ~ lists/base/md.amd64 ~ lists/base/md.armish ~ lists/base/md.armv7 ~ lists/base/md.aviion ~ lists/base/md.hppa ~ lists/base/md.hppa64 ~ lists/base/md.i386 ~ lists/base/md.landisk ~ lists/base/md.loongson ~ lists/base/md.luna88k ~ lists/base/md.macppc ~ lists/base/md.octeon ~ lists/base/md.sgi ~ lists/base/md.socppc ~ lists/base/md.sparc ~ lists/base/md.sparc64 ~ lists/base/md.vax ~ lists/base/md.zaurus ~ lists/comp/mi > sync (deraadt@) ~ lists/base/md.hppa64 ~ lists/base/md.i386 ~ lists/base/md.landisk ~ lists/base/md.loongson ~ lists/base/md.luna88k ~ lists/base/md.macppc ~ lists/base/md.octeon ~ lists/base/md.sgi ~ lists/base/md.socppc ~ lists/base/md.sparc ~ lists/base/md.sparc64 ~ lists/base/md.zaurus > sync (deraadt@) ~ lists/base/md.alpha ~ lists/base/md.amd64 ~ lists/base/md.armish ~ lists/base/md.armv7 ~ lists/base/md.aviion ~ lists/base/md.hppa ~ lists/base/md.hppa64 ~ lists/base/md.i386 ~ lists/base/md.landisk ~ lists/base/md.loongson ~ lists/base/md.luna88k ~ lists/base/md.macppc ~ lists/base/md.octeon ~ lists/base/md.sgi ~ lists/base/md.socppc ~ lists/base/md.sparc ~ lists/base/md.sparc64 ~ lists/base/md.zaurus > sync (sthen@) ~ lists/comp/mi > sync (sthen@) ~ lists/comp/mi > sync (miod@) ~ lists/base/md.loongson ~ lists/base/md.octeon ~ lists/base/md.sgi ~ lists/comp/md.loongson ~ lists/comp/md.octeon ~ lists/comp/md.sgi ~ lists/comp/mi > sync (miod@) sgi ~ ramdisk/Makefile > Enable IP26 builds. (miod@) == etc =============================================================== 03/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/etc etc.sgi/Makefile.inc ~ etc.sgi/Makefile.inc > Enable IP26 builds. (miod@) netstart ~ netstart > only print the "IPv6 autoconf" line if there are interfaces to configure > feedback/ok rpe (sthen@) rc ~ rc > Remove setup_X_sockets(): > The creation of Unix sockets directories in /tmp for X happens right > after pruning /tmp. So the whole dance of checking for their > existence, ownership or permissions is not necessary. It's safe to > just create them with the right permissions if X is installed. > Changes to do_fsck(): > Remove the _flags variable and pass flags to fsck directly with "$@". > Feedback and OK halex@ > OK krw@ on a similar diff (rpe@) ~ rc > - FOO=bar; export FOO -> export FOO=bar > - run domainname only with a non-empty /etc/defaultdomain file > - Make single-user if-block more intuitive, which also matches > better what the comment actually says > OK halex@, krw@ on a similar diff (rpe@) systrace/usr_sbin_lpd ~ systrace/usr_sbin_lpd > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) == games ============================================================= 04/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/games fortune ~ strfile/strfile.c > big blocks need braces for support (tedu@) == gnu =============================================================== 05/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/gnu gnu ~ gcc/libstdc++-v3/src/debug.cc > Make the famous _S_debug_messages warning go away, by hidining that symbol. > This symbol isn't used anywhere outside libstdc++, thus no bump. > Upstream initially went the samy way, but then implemented a different fix, > which don't work for us. Eventually we should move to whitelisting the list > of symbols exported anyway. > okay miod@, no objections from sthen@; also supported by a few a while ago > (zhuk@) usr.bin/cc ~ cc_tools/Makefile > Enable -msecure-plt by default (on powerpc). > ok pascal@ (kettenis@) == include =========================================================== 06/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/include Makefile ~ Makefile > Prefer the standardized <poll.h> over <sys/poll.h> (guenther@) ~ Makefile > Revert: finger slip (guenther@) arpa/inet.h ~ arpa/inet.h > Prefer the standardized <poll.h> over <sys/poll.h> (guenther@) ~ arpa/inet.h > Revert: finger slip (guenther@) rpc/auth.h ~ rpc/auth.h > authdes_create() was never implemented (guenther@) rpc/svc.h ~ rpc/svc.h > Wrap <rpc/*.h> so that calls go direct and the symbols are all weak. > Hide __xprt_register() and _authenticate(); truncate <rpc/svc_auth.h> > ok deraadt@ (guenther@) ~ rpc/svc.h > Prefer the standardized <poll.h> over <sys/poll.h> (guenther@) rpc/svc_auth.h ~ rpc/svc_auth.h > Wrap <rpc/*.h> so that calls go direct and the symbols are all weak. > Hide __xprt_register() and _authenticate(); truncate <rpc/svc_auth.h> > ok deraadt@ (guenther@) tib.h + tib.h > Prefer the standardized <poll.h> over <sys/poll.h> (guenther@) == lib =============================================================== 07/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib csu ~ boot.h > Make sure we set PROT_EXEC on the GOT for BSS-PLT binaries on powerpc. > ok guenther@ (kettenis@) ~ powerpc/md_init.h > We don't need to calculate the load offset. This also means we don't need > to calculate the GOT address anymore. > ok guenther@ (who spotted the latter bit) (kettenis@) libarch ~ arm/arm_drain_writebuf.2 ~ mips64/get_fpc_csr.3 > in the SYNOPSIS, make void function arguments explicit (schwarze@) libc ~ hidden/signal.h ~ compat-43/sigcompat.c ~ arch/alpha/gen/setjmp.S ~ arch/amd64/gen/setjmp.S ~ arch/amd64/gen/sigsetjmp.S ~ arch/arm/gen/setjmp.S ~ arch/hppa/gen/setjmp.S ~ arch/hppa64/gen/setjmp.S ~ arch/i386/gen/setjmp.S ~ arch/i386/gen/sigsetjmp.S ~ arch/m88k/gen/setjmp.S ~ arch/m88k/gen/sigsetjmp.S ~ arch/mips64/gen/setjmp.S ~ arch/powerpc/gen/setjmp.S ~ arch/vax/gen/setjmp.S > Do provide hidden _libc_* aliases for sig{block,setmask} and use them in > the ASM *setjmp implementations. > Skip the PLT when calling them on amd64 (other archs to do this after > testing) > ok miod@ (guenther@) ~ arch/amd64/Symbols.list ~ arch/amd64/gdtoa/strtold.c ~ arch/i386/Symbols.list ~ arch/i386/gdtoa/strtold.c ~ arch/mips64/Symbols.list ~ arch/mips64/gdtoa/strtold.c ~ arch/sparc64/Symbols.list ~ arch/sparc64/gdtoa/strtold.c ~ arch/vax/gdtoa/strtof.c ~ crypt/arc4random.c ~ crypt/arc4random_uniform.c ~ gdtoa/gdtoa.h ~ gdtoa/gdtoaimp.h ~ gdtoa/strtod.c ~ gdtoa/strtof.c ~ gdtoa/strtold.c ~ gen/devname.c ~ gen/getcap.c ~ hidden/unistd.h ~ locale/__mb_cur_max.c ~ stdio/mktemp.c ~ stdlib/abort.c ~ stdlib/abs.c ~ stdlib/atoi.c ~ stdlib/div.c ~ stdlib/erand48.c ~ stdlib/exit.c ~ stdlib/getenv.c ~ stdlib/lcong48.c ~ stdlib/malloc.c ~ stdlib/qsort.c ~ stdlib/rand.c ~ stdlib/reallocarray.c ~ stdlib/setenv.c ~ stdlib/srand48.c ~ stdlib/strtol.c ~ stdlib/strtoll.c ~ stdlib/strtonum.c ~ stdlib/strtoul.c ~ stdlib/strtoull.c ~ stdlib/system.c + hidden/stdlib.h > Wrap <stdlib.h> so that calls go direct and the symbols not in the > C standard are all weak. > Apply __{BEGIN,END}_HIDDEN_DECLS to gdtoa{,imp}.h, hiding the > arch-specific __strtorx, __ULtox_D2A, __strtorQ, __ULtoQ_D2A symbols. > (guenther@) - db/man/mpool.3 > mpool(3) is no longer installed; delete it. prodded by jmc@ (guenther@) ~ hidden/md5.h > namespace.h is pulled in via -include, so don't do it here (guenther@) ~ gen/unvis.c ~ gen/vis.c + hidden/vis.h > Wrap <vis.h> so that calls go direct and the symbols are all weak > (guenther@) ~ uuid/uuid_create_nil.c ~ uuid/uuid_is_nil.c + hidden/uuid.h > Wrap <uuid.h> so that calls go direct and the symbols are all weak > (guenther@) ~ Symbols.list ~ gen/ctype_.c ~ gen/isctype.c ~ gen/tolower_.c ~ gen/toupper_.c ~ locale/_wctrans_local.h ~ locale/iswctype.c ~ locale/rune_local.h ~ locale/wctoint.h + hidden/ctype.h + hidden/wctype.h > Wrap <ctype.h> and <wctype.h> so that calls go direct and the symbols not > in standard C are all weak. > Hide several symbols internal to the implementation (guenther@) + hidden/utime.h > Wrap <utime.h> to make utime() a weak symbol (guenther@) ~ gen/getttyent.c + hidden/ttyent.h > Wrap <ttyent.h> so that calls go direct and the symbols are all weak. > (guenther@) ~ termios/tcgetattr.c ~ termios/tcsetattr.c + hidden/termios.h > Wrap <termios.h> so that calls go direct and the symbols are all weak. > (guenther@) ~ gen/getgrent.c + hidden/grp.h > Wrap <grp.h> so that calls go direct and the symbols are all weak. > (guenther@) ~ crypt/bcrypt.c > The number of rounds is just two digits in the salt. We've already > verified that they are there via isdigit() so we can convert from > ASCII to an int without using atoi(). OK guenther@ deraadt@ (millert@) ~ softfloat/eqdf2.c ~ softfloat/eqsf2.c ~ softfloat/fpgetmask.c ~ softfloat/fpgetround.c ~ softfloat/fpgetsticky.c ~ softfloat/fpsetmask.c ~ softfloat/fpsetround.c ~ softfloat/fpsetsticky.c ~ softfloat/gedf2.c ~ softfloat/gesf2.c ~ softfloat/gtdf2.c ~ softfloat/gtsf2.c ~ softfloat/ledf2.c ~ softfloat/lesf2.c ~ softfloat/ltdf2.c ~ softfloat/ltsf2.c ~ softfloat/nedf2.c ~ softfloat/negdf2.c ~ softfloat/negsf2.c ~ softfloat/nesf2.c ~ softfloat/softfloat.c ~ softfloat/timesoftfloat.c ~ softfloat/unorddf2.c ~ softfloat/unordsf2.c ~ arch/arm/softfloat/Makefile.inc ~ arch/sh/softfloat/Makefile.inc > Uses angle brackets to include softfloat.h, and add the proper directory to > CFLAGS. > ok guenther@ (miod@) ~ softfloat/softfloat-specialize.h + hidden/softfloat.h > Make softfloat's internal float_raise also available under a hidden, > internal > name, so that local references to it do not need to go through the plt. > (this routine is used by libm and thus can't be made internal to libc) > help and ok guenther@ (miod@) ~ hidden/stdlib.h ~ stdlib/seed48.c > Overshot w/PROTO_DEPRECATED: seed48_deterministic() is used internally > (guenther@) ~ Symbols.list ~ crypt/bcrypt.c ~ crypt/cryptutil.c ~ gen/getpwent.c ~ gen/pw_dup.c > Wrap <pwd.h> so that calls go direct and the symbols are all weak. > Hide bcrypt_autorounds(), prefixing with an underbar for static builds. > (guenther@) + hidden/pwd.h > Wrap <pwd.h> so that calls go direct and the symbols are all weak. > Hide bcrypt_autorounds(), prefixing with an underbar for static builds. > (guenther@) ~ Symbols.list ~ rpc/auth_none.c ~ rpc/auth_unix.c ~ rpc/authunix_prot.c ~ rpc/clnt_perror.c ~ rpc/clnt_tcp.c ~ rpc/clnt_udp.c ~ rpc/get_myaddress.c ~ rpc/getrpcent.c ~ rpc/pmap_clnt.c ~ rpc/pmap_getport.c ~ rpc/pmap_prot.c ~ rpc/pmap_prot2.c ~ rpc/pmap_rmt.c ~ rpc/rpc_callmsg.c ~ rpc/rpc_prot.c ~ rpc/svc.c ~ rpc/svc_auth.c ~ rpc/svc_auth_unix.c ~ rpc/svc_tcp.c ~ rpc/svc_udp.c ~ rpc/xdr.c ~ rpc/xdr_array.c ~ rpc/xdr_mem.c ~ rpc/xdr_rec.c ~ rpc/xdr_reference.c + hidden/rpc/auth.h + hidden/rpc/auth_unix.h + hidden/rpc/clnt.h + hidden/rpc/pmap_clnt.h + hidden/rpc/pmap_prot.h + hidden/rpc/pmap_rmt.h + hidden/rpc/rpc.h + hidden/rpc/rpc_msg.h + hidden/rpc/svc.h + hidden/rpc/svc_auth.h + hidden/rpc/xdr.h > Wrap <rpc/*.h> so that calls go direct and the symbols are all weak. > Hide __xprt_register() and _authenticate(); truncate <rpc/svc_auth.h> > ok deraadt@ (guenther@) ~ arch/i386/stdlib/abs.S ~ arch/i386/stdlib/div.S ~ arch/i386/stdlib/labs.S ~ arch/i386/stdlib/ldiv.S > Add hidden _libc_ aliases and symbol sizes > prodded by semarie@ (guenther@) ~ arch/sparc/stdlib/abs.S ~ arch/sparc64/stdlib/abs.S > Add hidden _libc_ aliases and symbol sizes (guenther@) ~ shlib_version > Major bump for symbol removals (guenther@) ~ arch/i386/stdlib/abs.S ~ arch/i386/stdlib/div.S ~ arch/i386/stdlib/labs.S ~ arch/i386/stdlib/ldiv.S ~ arch/sparc64/stdlib/abs.S > Need SYS.h for END_STRONG() (guenther@) ~ Symbols.list ~ gen/Makefile.inc ~ sys/Makefile.inc > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) - gen/sysctl.c > Detritus (guenther@) ~ gen/login_cap.c + hidden/login_cap.h > Wrap <login_cap.h> so that calls go direct and the symbols are all weak > (guenther@) ~ hidden/stdlib.h ~ stdlib/malloc.c > For now, permit overriding of the malloc family, to make emacs happy > (guenther@) + hidden/sys/sysctl.h > Wrap <sys/sysctl.h> so that calls go direct and the symbol is weak > (guenther@) ~ yp/yp_bind.c ~ yp/yp_first.c ~ yp/yp_get_default_domain.c ~ yp/ypmatch_cache.c ~ yp/ypprot_err.c + hidden/rpcsvc/ypclnt.h > Wrap <rpcsvc/ypclnt.h> so that calls go direct and the symbols are all weak > (guenther@) ~ gen/fstab.c + hidden/fstab.h > Wrap <fstab.h> so that calls go direct and the symbols are all weak > (Do you sense a theme?) (guenther@) ~ net/inet_addr.c ~ net/inet_network.c ~ net/inet_ntop.c ~ net/inet_pton.c ~ net/res_comp.c + hidden/arpa/inet.h + hidden/arpa/nameser.h > Wrap <arpa/inet.h> and <arpa/nameser.h> so that calls go direct and the > symbols without underbar prefix are all weak (guenther@) ~ asr/getaddrinfo.c ~ asr/gethostnamadr.c ~ asr/getnameinfo.c ~ asr/getrrsetbyname.c ~ hidden/netdb.h ~ net/freeaddrinfo.c ~ net/gai_strerror.c ~ net/getproto.c ~ net/getprotoent.c ~ net/getprotoname.c ~ net/getservbyname.c ~ net/getservbyport.c ~ net/getservent.c ~ net/herror.c > Finish wrapping <netdb.h> so that calls go direct and the symbols are all > weak (guenther@) ~ hidden/termios.h > Use PROTO_DEPRECATED() make tcdrain() weak in libc, to guarantee libpthread > can override it (guenther@) ~ stdlib/system.c > Remove cast of int* to int* (guenther@) ~ hidden/sys/stat.h ~ hidden/sys/mount.h > Make isfdtype() and getmntinfo() weak symbols (guenther@) ~ gen/getnetgrent.c + hidden/netgroup.h > Wrap <netgroup.h> so internal calls go direct and the symbols are all weak > (guenther@) ~ gen/readpassphrase.c + hidden/readpassphrase.h > Wrap <readpassphrase.h> so internal calls go direct and readpassphrase is > weak (guenther@) ~ net/if_indextoname.c ~ net/if_nametoindex.c + hidden/net/if.h + hidden/net/if_dl.h > Wrap <net/if.h> and <net/if_dl.h> so internal calls go direct and all the > symbols are weak (guenther@) ~ net/ethers.c ~ net/getifaddrs.c ~ net/rthdr.c ~ rpc/bindresvport.c + hidden/ifaddrs.h + hidden/netinet/if_ether.h + hidden/netinet/in.h > Wrap <ifaddrs.h>, <netinet/in.h>, and <netinet/if_ether.h> so internal > calls go direct and all the symbols are weak (guenther@) ~ asr/asr.c ~ asr/getaddrinfo_async.c ~ asr/gethostnamadr_async.c ~ asr/getnameinfo_async.c ~ asr/getnetnamadr_async.c ~ asr/getrrsetbyname_async.c ~ asr/res_search_async.c ~ asr/res_send_async.c + hidden/asr.h > Wrap <asr.h> so internal calls go direct and all the symbols are weak > (guenther@) ~ sys/semctl.c > KNF (guenther@) + hidden/sys/sem.h > Wrap <sys/sem.h> so internal calls go direct and all the symbols are weak > (guenther@) ~ yp/_yp_check.c ~ yp/yp_bind.c ~ yp/ypinternal.h > Wrap _yp_check() and _yp_unbind() so internal calls go direct and they're > weak (guenther@) ~ locale/localeconv.c ~ locale/setlocale.c + hidden/locale.h > Wrap <locale.h> so internal calls go direct (guenther@) ~ locale/nl_langinfo.c ~ regex/regfree.c + hidden/langinfo.h + hidden/regex.h > Wrap <langinfo.h> and <regexp.h> so internal calls go direct and > the symbols are weak (guenther@) ~ stdio/floatio.h ~ gdtoa/dmisc.c ~ gdtoa/dtoa.c ~ gdtoa/gdtoa.c ~ gdtoa/gdtoa.h ~ gdtoa/hdtoa.c ~ gdtoa/ldtoa.c > Wrap the remaining __*dtoa() functions so that internal calls go direct > (guenther@) ~ arch/vax/gdtoa/hdtoa.c > vax deserves gdtoa love too! (guenther@) ~ time/ctime.3 ~ rpc/rpc.3 ~ stdlib/malloc.3 > use .Va for global variables, and .Vt where the type is included > (schwarze@) ~ stdlib/drand48.c ~ stdlib/erand48.c ~ stdlib/rand48.h > Only two of the *rand48.c files need <math.h>, so just #include it in them > (guenther@) ~ stdio/getwc.3 > in the SYNOPSIS, make void function arguments explicit (schwarze@) ~ gmon/gmon.c > Wrap moncontrol() so that internal calls go direct and it's a weak symbol > (guenther@) ~ gen/fpgetmask.3 ~ hash/SipHash24.3 > Remove useless quoting from .Fo and .Fn function names, to prevent > development of a cargo cult in case people look at existing files > for examples. This achieves a consistent .Fo and .Fn quoting style > across the whole tree. (schwarze@) ~ gen/auth_subr.c ~ gen/authenticate.c ~ gen/fts.c ~ gen/getcap.c ~ gen/getgrent.c ~ gen/getmntinfo.c ~ gen/getnetgrent.c ~ gen/getpwent.c ~ gen/getusershell.c ~ gen/glob.c ~ gen/login_cap.c > remove null check before free. from Michael McConville > ok semarie (tedu@) ~ regex/re_format.7 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) ~ hidden/ctype.h ~ hidden/errno.h ~ hidden/getopt.h ~ hidden/signal.h ~ hidden/stdio.h ~ hidden/stdlib.h ~ hidden/time.h ~ gen/errlist.c ~ gen/siglist.c ~ gen/signame.c ~ gen/tolower_.c ~ gen/toupper_.c ~ gen/ctype_.c ~ stdlib/getopt_long.c ~ locale/__mb_cur_max.c ~ time/localtime.c > Don't wrap initialized variables: binutils appears to be mishandling them > on arm and m88k > problems with optind observed by jsg@ (guenther@) ~ asr/getaddrinfo_async.c > remove unused static function (eric@) ~ asr/asr.c > use _PATH_RESCONF directly (eric@) ~ asr/asr.c ~ asr/getaddrinfo_async.c ~ asr/gethostnamadr_async.c ~ asr/getnameinfo_async.c ~ asr/getnetnamadr_async.c ~ asr/getrrsetbyname_async.c ~ asr/res_search_async.c ~ asr/res_send_async.c > remove bogus includes of err.h (eric@) libcrypto ~ crypto/shlib_version > Add a new interface, OPENSSL_cpu_caps(), to return the currently running > cpu's specific hardware capabilities users of libcrypto might be interested > in, as an integer value. This deprecates the existing OPENSSL_ia32cap() > macro and the OPENSSL_ia32cap_loc() function (which returns the pointer so > that you can mess with stuff you shouldn't mess with). > Interpreting the value returned by OPENSSL_cpu_caps() is, of course, > machine-dependent. > Minor version bump for libcrypto. > ok beck@ jsing@ (miod@) ~ cert.pem > Add Certplus CA root certificate: > C=FR, O=Certplus, CN=Class 2 Primary CA > req by beck@, ok miod@ beck@ (sthen@) ~ crypto/Makefile > Nuke openssl/e_os2.h, since nothing should be using it. > ok deraadt@ "hurray! finally!" miod@ "Yay!" sthen@ (jsing@) ~ crypto/Makefile > Merge ech_ossl.c into ech_key.c - not much point having one file with a > four line function and a tonne of license text. > ok beck@ (jsing@) ~ crypto/Makefile > Remove SHA-0 support. > SHA-0 was withdrawn shortly after publication 20 years ago and replaced > with SHA-1. This will require a major crank. > ok bcook@, jsing@ (doug@) ~ crypto/Makefile ~ man/Makefile > Remove MD4 support from LibreSSL. > MD4 should have been removed a long time ago. Also, RFC 6150 moved it to > historic in 2011. Rides the major crank from removing SHA-0. > Discussed with many including beck@, millert@, djm@, sthen@ > ok jsing@, input + ok bcook@ (doug@) ~ crypto/shlib_version > Crank major version due to removal of SHA-0 and MD4 from libcrypto. (doug@) ~ crypto/Makefile ~ man/Makefile > Temporarily revive MD4 for MS CHAP support. (doug@) ~ man/Makefile > Remove more EVP_sha() SHA-0 references. (bcook@) libedit ~ editline.3 > in the SYNOPSIS, make void function arguments explicit (schwarze@) libkeynote ~ keynote.4 ~ keynote.5 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) libm ~ man/lgamma.3 > use .Va for global variables, and .Vt where the type is included > (schwarze@) ~ Makefile > Move s_signgam.c to PURE_SRCS, so that it also gets built on vax. (miod@) libpthread ~ man/pthread_once.3 > use .Va for global variables, and .Vt where the type is included > (schwarze@) ~ man/getc_unlocked.3 > in the SYNOPSIS, make void function arguments explicit (schwarze@) libsndio ~ mio_open.3 ~ sio_open.3 > Remove useless quoting from .Fo and .Fn function names, to prevent > development of a cargo cult in case people look at existing files > for examples. This achieves a consistent .Fo and .Fn quoting style > across the whole tree. (schwarze@) libssl ~ src/ssl/ssl_ciph.c > If we have hardware acceleration for AES, prefer AES as a symmetric cipher > over CHACHA20. Otherwise, prefer CHACHA20 with AES second. > ok beck@ miod@ (jsing@) ~ src/ssl/d1_srvr.c ~ src/ssl/s3_srvr.c > The *_accept() functions increment in_handshake at the start of the > function, > then decrement it and call a callback on exit from the function. As such, > these functions should not return in the middle, otherwise in_handshake is > never decremented and the callback never called. > ok beck@ "with many sighs" miod@ (jsing@) ~ src/crypto/cryptlib.c ~ src/crypto/crypto.h > Add a new interface, OPENSSL_cpu_caps(), to return the currently running > cpu's specific hardware capabilities users of libcrypto might be interested > in, as an integer value. This deprecates the existing OPENSSL_ia32cap() > macro and the OPENSSL_ia32cap_loc() function (which returns the pointer so > that you can mess with stuff you shouldn't mess with). > Interpreting the value returned by OPENSSL_cpu_caps() is, of course, > machine-dependent. > Minor version bump for libcrypto. > ok beck@ jsing@ (miod@) ~ ssl/shlib_version > Mechanical minor bump to follow libcrypto. (miod@) ~ src/crypto/ecdh/ecdh.h ~ src/crypto/ecdh/ech_err.c ~ src/crypto/ecdh/ech_key.c ~ src/crypto/ecdh/ech_lib.c ~ src/crypto/ecdh/ech_locl.h ~ src/crypto/ecdh/ech_ossl.c > Lob a style(9) grenade in here. (jsing@) ~ src/crypto/ecdh/ecdh.h ~ src/crypto/ecdh/ech_err.c ~ src/crypto/ecdh/ech_ossl.c > Check ECDH output buffer length and avoid truncation. > Currently, if you call ECDH_compute_key() it will silently truncate the > resulting key if the output buffer is less than the key size. Instead, > detect this condition and return an error. If the buffer provided is larger > than the key length, zero the remainder. > ok beck@ miod@ "+ shivers" (jsing@) ~ src/crypto/ecdh/ecdh.h ~ src/crypto/ecdh/ech_lib.c ~ src/crypto/ecdh/ech_ossl.c > Provide ECDH_size(). > "jajaja" miod@ (jsing@) - src/e_os2.h > Nuke openssl/e_os2.h, since nothing should be using it. > ok deraadt@ "hurray! finally!" miod@ "Yay!" sthen@ (jsing@) - src/crypto/ecdh/ech_ossl.c ~ src/crypto/ecdh/ech_key.c > Merge ech_ossl.c into ech_key.c - not much point having one file with a > four line function and a tonne of license text. > ok beck@ (jsing@) ~ src/ssl/ssl_ciph.c > Switch to miod's shiny new OPENSSL_cpu_caps() and we can now also enable > the AES acceleration checking for i386. > ok beck@ miod@ (jsing@) ~ src/ssl/s3_clnt.c ~ src/ssl/s3_srvr.c > Use ECDH_size() instead of rolling our own. > ok beck@ (jsing@) ~ src/ssl/s3_clnt.c > Stop generating private keys in a network buffer. > The current client key exchange code generates DH and ECDH keys into the > same buffer that we use to send data to the network - stop doing this and > malloc() a new buffer, which we explicit_bzero() and free() on return. > This also benefits from ASLR and means that the keys are no longer > generated in a well known location. > ok beck@ (jsing@) ~ src/crypto/ecdh/ech_key.c > Only check for key truncation if no KDF function is being used. > ok beck@ miod@ (jsing@) ~ src/crypto/x509/x_all.c > Reorder functions for readability/consistency. (jsing@) ~ src/crypto/bn/bn_print.c > BN does support negative-zero -- BN_print() sets the standard here. > BN_bn2hex() had a 1-byte overflow when creating "-0\0". Reported to > me a while back by unknown person -- did not have enough experience > to push this through then. > advice from jsing, ok miod (deraadt@) ~ src/crypto/bn/bn_print.c > Handle negative-zero in BN_bn2dec() too, just like in BN_print(). > ok miod (deraadt@) ~ src/crypto/crypto.h ~ src/crypto/cryptlib.c > Since a major bump of libcrypto is coming, remove OPENSSL_ia32cap and > OPENSSL_ia32cap_loc; nothing in ports uses them besides embedded copies of > OpenSSL. This opens the `all hell gets loose' window. (miod@) - src/crypto/evp/m_sha.c - src/crypto/sha/sha_dgst.c - src/crypto/sha/sha_one.c ~ src/crypto/opensslfeatures.h ~ src/crypto/evp/c_all.c ~ src/crypto/evp/evp.h ~ src/crypto/sha/sha.h ~ src/crypto/sha/sha1dgst.c ~ src/crypto/sha/sha_locl.h ~ src/doc/crypto/EVP_DigestInit.pod > Remove SHA-0 support. > SHA-0 was withdrawn shortly after publication 20 years ago and replaced > with SHA-1. This will require a major crank. > ok bcook@, jsing@ (doug@) - src/crypto/evp/m_md4.c - src/crypto/md4/md4.h - src/crypto/md4/md4_dgst.c - src/crypto/md4/md4_locl.h - src/crypto/md4/md4_one.c ~ src/crypto/opensslfeatures.h ~ src/crypto/evp/c_all.c ~ src/crypto/evp/evp.h ~ src/doc/standards.txt ~ src/doc/apps/dgst.pod ~ src/doc/apps/openssl.pod ~ src/doc/apps/speed.pod ~ src/doc/apps/ts.pod ~ src/doc/crypto/MD5.pod ~ src/doc/crypto/crypto.pod > Remove MD4 support from LibreSSL. > MD4 should have been removed a long time ago. Also, RFC 6150 moved it to > historic in 2011. Rides the major crank from removing SHA-0. > Discussed with many including beck@, millert@, djm@, sthen@ > ok jsing@, input + ok bcook@ (doug@) ~ ssl/shlib_version > Crank major version due to removal of SHA-0 and MD4 from libcrypto. (doug@) ~ src/crypto/opensslfeatures.h ~ src/crypto/evp/c_all.c ~ src/crypto/evp/evp.h ~ src/doc/standards.txt ~ src/doc/apps/dgst.pod ~ src/doc/apps/openssl.pod ~ src/doc/apps/speed.pod ~ src/doc/apps/ts.pod ~ src/doc/crypto/MD5.pod ~ src/doc/crypto/crypto.pod + src/crypto/evp/m_md4.c + src/crypto/md4/md4.h + src/crypto/md4/md4_dgst.c + src/crypto/md4/md4_locl.h + src/crypto/md4/md4_one.c > Temporarily revive MD4 for MS CHAP support. (doug@) ~ src/doc/ssl/SSL_set_session.3 > Remove useless quoting from .Fo and .Fn function names, to prevent > development of a cargo cult in case people look at existing files > for examples. This achieves a consistent .Fo and .Fn quoting style > across the whole tree. (schwarze@) ~ src/doc/ssl/SSL_CTX_get_ex_new_index.3 ~ src/doc/ssl/SSL_CTX_set_session_id_context.3 > fix formatting by adding the required quotes to .Fa in the SYNOPSIS > (schwarze@) ~ src/doc/ssl/SSL_alert_type_string.3 > delete bogus trailing .Ns from SYNOPSIS .Ft macros (schwarze@) ~ src/crypto/x509/x509_vfy.c ~ src/crypto/x509/x509_vfy.h > Add support for disabling certificate and CRL validity checking. > Loosely based on changes in OpenSSL. > ok beck@ (jsing@) ~ src/crypto/crypto.h > include stdint.h for uint64_t > noted by Bernard Spil (bcook@) ~ src/doc/crypto/EVP_DigestInit.pod > Re-add missing comma from SHA-0 removal which breaks mlinks generation. > Worked out by bcook@ (sthen@) ~ src/doc/crypto/EVP_DigestInit.pod > Remove more EVP_sha() SHA-0 references. (bcook@) ~ src/crypto/bn/bn_print.c > Revert bn_print.c:r1.25 ("handle negative-zero in BN_bn2dec() too") for > now, it has a NULL deref. Segfault reported by Mikolaj Kucharski, ok bcook > (sthen@) ~ src/crypto/ecdh/ech_key.c > avoid void * pointer arithmetic > ok miod@ (bcook@) libtls ~ tls.h ~ tls_conninfo.c ~ tls_init.3 ~ tls_internal.h > add visibility of ciper and connection version strings > ok jsing@ (beck@) ~ shlib_version > Mechanical minor bump to follow libcrypto. (miod@) ~ tls.c > Don't leak conninfo - spotted by marko kreen. > ok jsing@ (beck@) ~ tls.c > work around the stupid semantics of SSL_read and SSL_write to make sure > we can indicate an EOF properly on tls_read and tls_write > ok jsing@ (beck@) ~ shlib_version > Crank major version due to removal of SHA-0 and MD4 from libcrypto. (doug@) ~ tls_init.3 > add missing function return types (schwarze@) ~ tls.c > Return an error if tls_handshake() or tls_close() is called on a context > for which they are not valid operations. > ok beck@ (jsing@) ~ tls.c ~ tls_internal.h > Expose EOF without close-notify via tls_close(). > Make tls_read(3)/tls_write(3) follow read(2)/write(2) like semantics and > return 0 on EOF with and without close-notify. However, if we saw an EOF > from the underlying file descriptors without getting a close-notify, save > this and make it visible when tls_close(3) is called. This keeps the > semantics we want, but makes it possible to detect truncation at higher > layers, if necessary. > ok beck@ guenther@ (jsing@) ~ Makefile ~ tls_init.3 > some conn_version and conn_cipher bits; (jmc@) ~ tls_init.3 > Remove useless quoting from .Fo and .Fn function names, to prevent > development of a cargo cult in case people look at existing files > for examples. This achieves a consistent .Fo and .Fn quoting style > across the whole tree. (schwarze@) ~ Makefile ~ tls.c ~ tls.h ~ tls_config.c ~ tls_init.3 ~ tls_internal.h > Provide tls_config_insecure_noverifytime() in order to be able to disable > certificate validity checking. > ok beck@ (jsing@) ~ tls_init.3 > tweak previous; (jmc@) libutil ~ pw_init.3 ~ pw_lock.3 > in the SYNOPSIS, make void function arguments explicit (schwarze@) ~ fparseln.3 > Remove useless quoting from .Fo and .Fn function names, to prevent > development of a cargo cult in case people look at existing files > for examples. This achieves a consistent .Fo and .Fn quoting style > across the whole tree. (schwarze@) == libexec =========================================================== 08/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/libexec ld.so ~ alpha/ldasm.S ~ amd64/ldasm.S ~ arm/ldasm.S ~ hppa/ldasm.S ~ i386/ldasm.S ~ m88k/ldasm.S ~ mips64/ldasm.S ~ powerpc/ldasm.S ~ sh/ldasm.S ~ sparc/ldasm.S ~ sparc64/ldasm.S > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) ~ powerpc/ldasm.S > Make bootstrapping code compatible with the Secure-PLT ABI. > ok guenther@ (kettenis@) ~ powerpc/archdep.h > Remove PROT_EXEC bit from GOT_PERMS. ld.so doesn't need it anymore, and > rcrt0.o will add it for BSS-PLT binaries if necessary. > ok guenther@ (kettenis@) ~ dlfcn.c ~ resolve.h ~ alpha/ldasm.S ~ alpha/syscall.h ~ amd64/ldasm.S ~ amd64/syscall.h ~ arm/ldasm.S ~ arm/syscall.h ~ hppa/ldasm.S ~ hppa/syscall.h ~ i386/ldasm.S ~ i386/syscall.h ~ m88k/ldasm.S ~ m88k/syscall.h ~ mips64/ldasm.S ~ mips64/syscall.h ~ powerpc/ldasm.S ~ powerpc/syscall.h ~ sh/ldasm.S ~ sh/syscall.h ~ sparc/ldasm.S ~ sparc/syscall.h ~ sparc64/ldasm.S ~ sparc64/rtld_machine.c ~ sparc64/syscall.h > kbind has eliminated the need for and use of the bind lock. Delete it, the > the callback, and the sigprocmask stub. > Keep around the DL_SETBINDLCK case until libpthread stops using it. > discussed with miod@ at l2k15 > ok kettenis@ (guenther@) == regress =========================================================== 09/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress bin ~ systrace/id/id.policy ~ systrace/symlink/ln.policy > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) lib ~ libcrypto/ecdh/ecdhtest.c > style(9) grenade. (jsing@) ~ libcrypto/ecdh/ecdhtest.c > Wrap long lines. (jsing@) ~ libcrypto/engine/enginetest.c ~ libssl/ssl/ssltest.c > remove e_os2.h includes > ok jsing@ (bcook@) ~ libcrypto/ecdh/ecdhtest.c > Use ECDH_size() and do some other clean up. (jsing@) ~ libcrypto/ecdh/ecdhtest.c > Remove explicit NULL checks before *_free() calls and tidy some code. > (jsing@) ~ libssl/client/clienttest.c > Revise regression test so that it works correctly with AES taking priority > if hardware acceleration is available. (jsing@) - libcrypto/sha/Makefile - libcrypto/sha/shatest.c ~ libcrypto/Makefile ~ libcrypto/sha1/sha1test.c > Remove SHA-0 support. > SHA-0 was withdrawn shortly after publication 20 years ago and replaced > with SHA-1. This will require a major crank. > ok bcook@, jsing@ (doug@) ~ libcrypto/ec/ectest.c > Another style(9) grenade. (jsing@) - libcrypto/md4/Makefile - libcrypto/md4/md4test.c ~ libcrypto/Makefile > Remove MD4 support from LibreSSL. > MD4 should have been removed a long time ago. Also, RFC 6150 moved it to > historic in 2011. Rides the major crank from removing SHA-0. > Discussed with many including beck@, millert@, djm@, sthen@ > ok jsing@, input + ok bcook@ (doug@) ~ libcrypto/Makefile + libcrypto/md4/Makefile + libcrypto/md4/md4test.c > Temporarily revive MD4 for MS CHAP support. (doug@) ~ libc/strerror/strerror_test.c > Test the exported strsignal() API, not the (now hidden) internal API > noted by daniel@ (guenther@) ~ libpthread/Makefile + libpthread/cancel_wait/Makefile + libpthread/cancel_wait/cancel_wait.c > Add regress for cancellation of wait/waitpid/wait3/wait4 (guenther@) ~ libpthread/cancel2/cancel2.c > Check pselect, poll, and ppoll too (guenther@) ~ libpthread/cancel_wait/cancel_wait.c > init a variable to avoid spurious failure (guenther@) sys ~ kern/tame/generic/main.c > remove check for LD_BIND_NOW as kbind(2) is available on almost (all ?) > arch > where LD_BIND_NOW was in use. (semarie@) usr.bin ~ sdiff/Iflag12.out ~ sdiff/Iflag21.out ~ sdiff/Ilflag12.out ~ sdiff/Ilflag21.out ~ sdiff/file1 ~ sdiff/file2 ~ sdiff/lflag.out ~ sdiff/same.out ~ sdiff/short.out ~ sdiff/tabs.out ~ sdiff/tabs1 ~ sdiff/tabs2 ~ sdiff/wflag.out > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) ~ sdiff/tabs.out > Update expected output to reflect changed sort order (guenther@) - mandoc/mdoc/Vt/child.out_lint ~ mandoc/mdoc/Vt/Makefile > Remove the warning about children of .Vt blocks because actually, > .Vt type global_variable No = Dv defined_constant ; > is the best way to specify in the SYNOPSIS how a global variable > is initialized in the rare case where that matters. > Issue noticed by jmc@. (schwarze@) ~ openssl/Makefile ~ openssl/openssl.cnf > Zap RANDFILE. (lteo@) ~ tsort/Makefile > Repair. (daniel@) == sbin ============================================================== 10/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sbin ping6 ~ ping6.c > Some cleanup from dhill AT mindcry.org, thanks! > - remove unused defines > - use socklen_t for getnameinfo > no object change (florian@) swapctl ~ swapctl.8 > fix markup of the -t argument (schwarze@) ~ swapctl.c > update spacing in usage(); reminded by jmc@ (schwarze@) == share ============================================================= 11/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/share man ~ man9/srpl_rc_init.9 > forgot to prototope SRPL_RC_INITIALIZER (dlg@) ~ man5/bsd.port.mk.5 > Mark IGNORE_IS_FATAL and IGNORE_SILENT as user settings explicitly. > This makes syncing list of such variables with portcheck code easier. > (zhuk@) ~ man9/Makefile + man9/intr_barrier.9 > intr_barrier(9) (kettenis@) ~ man9/intr_barrier.9 > vallue -> value; (jmc@) ~ man9/bio_register.9 ~ man9/ieee80211.9 > add missing function return types (schwarze@) ~ man9/SipHash24.9 ~ man9/aml_evalnode.9 ~ man9/bufq_init.9 ~ man9/fork1.9 ~ man9/hardclock.9 ~ man9/if_rxr_init.9 ~ man9/log.9 ~ man9/ml_init.9 ~ man9/mq_init.9 ~ man9/physio.9 ~ man9/printf.9 ~ man9/psignal.9 ~ man9/sensor_attach.9 ~ man9/srp_enter.9 ~ man9/task_add.9 ~ man9/timeout.9 ~ man9/tsleep.9 ~ man9/vwaitforio.9 > Remove useless quoting from .Fo and .Fn function names, to prevent > development of a cargo cult in case people look at existing files > for examples. This achieves a consistent .Fo and .Fn quoting style > across the whole tree. (schwarze@) ~ man9/ieee80211_node.9 > delete stray .Ft line (schwarze@) ~ man4/bktr.4 ~ man4/bridge.4 ~ man4/gpio.4 ~ man4/hotplug.4 ~ man4/pipex.4 ~ man4/pppx.4 ~ man4/radio.4 ~ man4/tun.4 ~ man4/usb.4 > convert .Fd #include to .In (schwarze@) ~ man5/pf.conf.5 ~ man5/spamd.conf.5 ~ man7/mdoc.7 ~ man8/release.8 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) ~ man5/bsd.port.mk.5 > use .Sq rather than manual `...' and fix two typos; > from Michael Reed <m dot reed at mykolab dot com> (schwarze@) ~ man9/srpl_rc_init.9 ~ man9/Makefile > implement SRPL_INSERT_AFTER_LOCKED. > i thought id committed this at l2k15. sorry for the delay. (dlg@) ~ man9/srpl_rc_init.9 > tweak previous; (jmc@) ~ man9/rtalloc.9 > Document that rtfree(9) accepts NULL. > From Michael McConville via jmc@ (mpi@) ~ man4/man4.sgi/intro.4 > More IP26 bits. (miod@) ~ man5/bsd.port.mk.5 > fix an Xr to pkg_create(1); > from michael reed (jmc@) == sys =============================================================== 12/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys arch/alpha/alpha ~ interrupt.c > intr_barrier(9) for alpha. (kettenis@) arch/alpha/conf ~ RAMDISK ~ RAMDISKB ~ RAMDISKBIG ~ RAMDISKC > No need for wsmux on these installation media, which lack USB support. > (miod@) arch/alpha/include ~ intr.h > intr_barrier(9) for alpha. (kettenis@) arch/amd64/amd64 ~ intr.c > Introduce intr_barrier(4), an interface that guarantees that an interrupt > handler that was running has finished. > ok miod@, guenther@ (both for the equivalent sparc64 diff) (kettenis@) arch/amd64/conf ~ RAMDISK_CD > remove delta (deraadt@) arch/amd64/include ~ intr.h > Introduce intr_barrier(4), an interface that guarantees that an interrupt > handler that was running has finished. > ok miod@, guenther@ (both for the equivalent sparc64 diff) (kettenis@) arch/amd64/stand ~ Makefile.inc > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/amd64/stand/boot ~ conf.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/amd64/stand/cdboot ~ conf.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/amd64/stand/libsa ~ cmd_i386.c ~ dev_i386.c ~ machdep.c ~ memprobe.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) ~ biosdev.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/amd64/stand/pxeboot ~ conf.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/arm/arm ~ cpu.c > intr_barrier(9) for arm > ok kettenis@ (jsg@) arch/armish/include ~ intr.h > intr_barrier(9) for arm > ok kettenis@ (jsg@) arch/armv7/conf ~ RAMDISK > remove delta (deraadt@) arch/armv7/include ~ intr.h > intr_barrier(9) for arm > ok kettenis@ (jsg@) arch/hppa/gsc ~ if_ie_gsc.c > Fix more ifmedia64 fallout in the kernel. It's hiding everywhere... (stsp@) arch/hppa/hppa ~ mainbus.c > Adjust printf for the fact that device locators are now long. (kettenis@) ~ intr.c > intr_barrier(9) for hppa. (kettenis@) ~ mutex.c > Put memory barriers in the appropriate place. > Discssed with geunther@, tested by landry@ (kettenis@) arch/hppa/include ~ intr.h > intr_barrier(9) for hppa. (kettenis@) arch/i386/conf ~ RAMDISK_CD > remove delta (deraadt@) arch/i386/i386 ~ machdep.c > intr_barrier(9) for i386. (kettenis@) ~ bios.c > Fix uninitialized variable found by Maxime Villard's Brainy Code Scanner. > ok uebayasi@, mlarkin@ (daniel@) arch/i386/include ~ intr.h > intr_barrier(9) for i386. (kettenis@) arch/i386/stand ~ Makefile.inc > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/i386/stand/boot - crt0.c ~ Makefile ~ conf.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/i386/stand/cdboot ~ conf.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/i386/stand/libsa - unixsys.S ~ cmd_i386.c ~ dev_i386.c ~ exec_i386.c ~ machdep.c ~ memprobe.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) ~ biosdev.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/i386/stand/pxeboot ~ conf.c > Remove support for building the boot blocks with DEBUGFLAGS=-D_TEST, which > is > supposed to create a userland binary in order to test non-boot related > functionality. This feature has been bitrotting in a non-compiling state > for years, and causes a too-many-ifdefs disease now that there are > intrusive > EFI changes. > No functional change. (miod@) arch/loongson/dev ~ bonito.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/loongson/include ~ intr.h > intr_barrier(9) for loongson, octeon and sgi. (kettenis@) arch/m88k/m88k ~ trap.c > <sys/systm.h> must be before <sys/syscall_mi.h> now. (miod@) arch/macppc/conf ~ RAMDISK > remove delta (deraadt@) arch/macppc/macppc ~ openprom.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/macppc/stand ~ alloc.c > unify free(NULL,size) behaviour by allowing passing NULL > ok millert@ jasper@ (semarie@) arch/mips64/mips64 ~ interrupt.c > intr_barrier(9) for loongson, octeon and sgi. (kettenis@) ~ cache_tfp.c > Reset the `pending I$ flushes' flag in tfp_InvalidateICache() if > invalidating > the whole I$. (miod@) ~ exception_tfp.S > Correctly compute the userland pte index in a pte page in the userland tlb > miss > handler; from Naruaki Etomi (nullnilaki on gmail), thanks! (miod@) arch/octeon/dev ~ cn30xxgmxvar.h > Fix more ifmedia64 fallout in the kernel. It's hiding everywhere... (stsp@) arch/octeon/include ~ intr.h > intr_barrier(9) for loongson, octeon and sgi. (kettenis@) arch/powerpc/include ~ intr.h > intr_barrier(9) for macppc and socppc. (kettenis@) arch/powerpc/powerpc ~ intr.c > intr_barrier(9) for macppc and socppc. (kettenis@) arch/sgi/dev ~ if_iec.c > Fix an uninitialized variable found by Maxime Villard's Brainy. > While here, fix the size parameter of bus_dmamem_unmap() in > iec_alloc_physical(). > ok miod@ (visa@) arch/sgi/hpc ~ if_sq.c > Fix more ifmedia64 fallout in the kernel. It's hiding everywhere... (stsp@) ~ hpc.c ~ hpcvar.h ~ if_sq.c ~ if_sqvar.h > Go back to the previous approach when managing individual HPC DMA > descriptors: > provide again an optional storage for a copy of the descriptor in the > `sync' > (fetch) function, and use the returned address afterwards. > On IP22 systems (in the broader sense of the term, thus IP20/IP22/IP24), > descriptors will remain in uncached memory and no local copies need to be > made. > On IP28 systems, descriptors will remain in cached memory (so as to avoid > switching to `slow mode'), but a local copy will be performed with the > necessary > cache eviction work, so that speculative code execution on R10000 will not > touch the real descriptor. > With this in place, all the explicit descriptor cache operations in if_sq, > some of them being redundant or operating on the wrong number of > descriptors, can be removed, with the HPC DMA wrappers taking care of doing > the right thing. > Tested on IP22 and IP28. IP26 still unhappy but no worse than before. > (miod@) arch/sgi/include ~ intr.h > intr_barrier(9) for loongson, octeon and sgi. (kettenis@) arch/sgi/localbus ~ tcc.c > Invoke tfp_IOSyncDCache() from tcc_IOSyncDCache() as well, so that > invalidate > operations are also carried on the internal write-through cache. (miod@) arch/sgi/sgi ~ l1.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/sparc/dev ~ dmavar.h ~ qe.c > Fix more ifmedia64 fallout in the kernel. It's hiding everywhere... (stsp@) arch/sparc/include ~ vmparam.h > Drop the workarounds allowing SMALL_KERNEL (i.e. installation media) to be > booted from old (< 2.10) bootblocks. (miod@) arch/sparc/sparc ~ autoconf.c ~ machdep.c ~ pmap.c > Drop the workarounds allowing SMALL_KERNEL (i.e. installation media) to be > booted from old (< 2.10) bootblocks. (miod@) ~ clock.c ~ openprom.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/sparc/stand/common ~ dvma.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/sparc64/dev ~ vdsp.c > Fix overlapping memcpy found by Maxime Villard's Brainy Code Scanner. > (kettenis@) ~ fhc.c ~ sbus.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/sparc64/include ~ intr.h > Introduce intr_barrier(4), an interface that guarantees that an interrupt > handler that was running has finished. > ok miod@, guenther@ (kettenis@) arch/sparc64/sparc64 ~ intr.c > Introduce intr_barrier(4), an interface that guarantees that an interrupt > handler that was running has finished. > ok miod@, guenther@ (kettenis@) ~ autoconf.c ~ openprom.c > trivial "if(x) free(x)" replacement by "free(x)" > ok miod@ (semarie@) arch/sparc64/stand/ofwboot ~ alloc.c > unify free(NULL,size) behaviour by allowing passing NULL > ok millert@ jasper@ (semarie@) arch/vax/mbus ~ fwio.c > Locators are long now. (miod@) arch/vax/qbus ~ uba.c > Locators are long now. (miod@) arch/vax/vax ~ trap.c > Needs <sys/systm.h> before <sys/syscall_mi.h> now. (miod@) arch/vax/vsa ~ vsbus.c > Locators are long now. (miod@) arch/zaurus/include ~ intr.h > intr_barrier(9) for arm > ok kettenis@ (jsg@) arch/zaurus/stand/zboot ~ alloc.c > unify free(NULL,size) behaviour by allowing passing NULL > ok millert@ jasper@ (semarie@) ~ unixdev.c > Use the local unixdev.h file rather than the outdated libsa one. (miod@) compat/linux ~ linux_misc.c > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) ddb ~ db_trap.c > On show the ps/o output and ddb.html blurb once, so that we don't get them > after each "mach ddbcpu N" > ok miod@ deraadt@ (guenther@) dev/ic ~ qlw.c > Do not define QLW_DEBUG if SMALL_KERNEL, and turn qlw_dump_iocb() and > qlw_dump_iocb_segs() into empty macros if QLW_DEBUG is not defined (if only > to make this compile, then). (miod@) ~ i82596.c ~ i82596var.h > More ifmedia64 fallout. (miod@) dev/pci ~ if_tl.c > Fix more ifmedia64 fallout in the kernel. It's hiding everywhere... (stsp@) ~ if_vmx.c > make vmx(4) interrupts mpsafe. > the vmx rx path is only touched in the interrupt handler, so it is > already guaranteed to be accessed by only one cpu at a time. > the tx path has been massaged so the the producer is only touched > by the start routine, and the consumer is only touched by the > interrupt path, and can therefore be run concurrently. the only > interlock is a count of the free descriptors. > if txintr clears IFF_OACTIVE, it takes the kernel lock before running > the start routine. > other interrupts, eg, link state handling, take the kernel lock. (dlg@) ~ if_em.c ~ if_em.h > Avoid using a mutex in the rx completion path. Instead rely on > intr_barrier(9) to avoid having the interrupt handler touch the rx data > structures while we're brining down the interface. This actually reverts > many of the changes in rev. 1.300. > ok mikeb@ (kettenis@) ~ if_vmx.c > need to keep bpf in the tx path. got a bit ahead of myself there... > noticed by brad (dlg@) ~ if_vmx.c > brad points out i need bpf_mtap_ether to reconstruct vlan headers (dlg@) dev/sun ~ z8530ms.c > Remove stray splx(9) call. Found by Maxime Villard's Brainy Code Scanner. > From miod@. (kettenis@) dev/usb ~ ugen.c > Fix kpanic I found the hard way while using pcsc-lite > From Grant Czajkowski <czajkow2 AT illinois DOT edu> > OK mpi@ (dcoppa@) ~ usbdevs > add fingerprint reader as found the x250 as well as another similar > device found while figuring out the chipset. > ok jsg@ (jasper@) ~ usbdevs.h ~ usbdevs_data.h > regen (jasper@) ~ if_urtwn.c > Remove unused variables, from Michael McConville. > ok stsp@ (mpi@) dev/wsfont ~ wsfont.c > Put the 12x22 font on alpha and macppc installation kernels, the 8x16 font > is really fugly on large displays (tga, vgafb). It's only 10KB more data > before compression, which fits in our floppies. (miod@) kern ~ kern_sched.c > Introduce sched_barrier(9), an interface that acts as a scheduler barrier > in > the sense that it guarantees that the specified CPU went through the > scheduler. This also guarantees that interrupt handlers running on that > CPU > will have finished when sched_barrier() returns. > ok miod@, guenther@ (kettenis@) ~ subr_disk.c > Rename readgptlabel() to spoofgptlabel() because that's what we > really want it to do. Handle all the actual disklabel reading in > readdoslabel(). Makes the code much simpler to understand. > ok deraadt@ (krw@) ~ subr_disk.c > Move prototype for spoofgptlabel() from disklabel.h to subr_disk.c. > It's a helper function for readdoslabel(). Not something called > outside of subr_disk.c. (krw@) ~ syscalls.master ~ kern_ktrace.c ~ kern_sysctl.c ~ kern_tame.c > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) ~ init_sysent.c ~ syscalls.c > sync (guenther@) ~ subr_disk.c > Unbeak boot from GPT partition. Reported by Henrik Friedrichsen. > ok krw (yasuoka@) ~ kern_tame.c > In the internal conversion of _TM_* to TAME_*, some bits were lost for > "dns" and "cmsg" in the `tamereq' array. > Restore the previous behaviour. > ok guenther@ (semarie@) ~ kern_sched.c > Short circuit if we're running on the CPU that we want to sync with. Fixes > suspend on machines with em(4) now that it uses intr_barrier(9). > ok krw@ (kettenis@) lib/libsa ~ alloc.c > unify free(NULL,size) behaviour by allowing passing NULL > ok millert@ jasper@ (semarie@) - unixdev.c - unixdev.h ~ Makefile > Remove orphaned files. (miod@) net ~ if_vlan.c ~ if_vlan_var.h > back out mpsafe vlans while i chase down a panic when operations > are done in a certain order. > panic found by naddy@ (dlg@) ~ if_vlan.c ~ if_vlan_var.h > reintroduce mpsafe vlan_input. > the last commit was missing SRPL_INITs of the vlan_tagh buckets, > which was in a different tree to the one i committed from. > apologise to naddy@ again (dlg@) ~ if.c ~ if_var.h > Run the interface watchdog timer routine as a task such that we have > process > context. > ok mpi@, claudio@ (kettenis@) ~ if_vlan.c ~ if_vlan_var.h > replace hand rolled reference counting with refcnts. (dlg@) ~ if.c > replace hand rolled refcounts and sleep_setup/finish with refcnts and > refcnt_finalize. (dlg@) ~ if_ethersubr.c > queue revarps to softnet so we can defer processing to a context with > the kernel lock. > "do it" claudio@ mpi@ (dlg@) ~ bpf.c ~ if.c ~ if_pppoe.c ~ if_var.h ~ pf.c ~ pipex.c ~ ppp_tty.c > There's no point in abstracting ifp->if_output() as long as pf_test() > needs to see lo0 in the output path. > ok claudio@ (mpi@) ~ if.c > Fix double rtfree(9) in non-INET6 kernel. > With and ok claudio@ (mpi@) ~ if_dl.h > link_addr() is gone > ok mpi@ (guenther@) ~ if_ethersubr.c > instead of leaking rarp packets, break from the switch statement > so packets can get enqueued for the arp netisr to process. > fixes jsgs nfs mountroot problem. (dlg@) net80211 ~ ieee80211_pae_output.c > Make code in ieee80211_send_4way_msg3() more readable and make sure a local > variable is initialized to NULL in the WPA case. No functional change. > Pointed out by Maxime Villard's brainy code scanner. > ok claudio@ zhuk@ (stsp@) netinet ~ if_ether.c ~ if_ether.h > queue revarps to softnet so we can defer processing to a context with > the kernel lock. > "do it" claudio@ mpi@ (dlg@) ~ ip_output.c > Get the default loopback interface pointer just after doing a route > lookup to ensure pf_test() is called with the same interface in the > input annd output path for local traffic. > Fix a regression reported by Heiko Zimmermann on bugs@, thanks! > ok mikeb@, claudio@ (mpi@) ~ igmp.c > if_put() after if_get(). > ok dlg@ (mpi@) ~ if_ether.c ~ ip_output.c > There's no point in abstracting ifp->if_output() as long as pf_test() > needs to see lo0 in the output path. > ok claudio@ (mpi@) ~ if_ether.c > Fix uninitalized use of variable. > Issue _17 found by Maxime Villard Brainy scanner. > OK bluhm@ (claudio@) netinet6 ~ ip6_output.c > Get the default loopback interface pointer just after doing a route > lookup to ensure pf_test() is called with the same interface in the > input annd output path for local traffic. > Fix a regression reported by Heiko Zimmermann on bugs@, thanks! > ok mikeb@, claudio@ (mpi@) ~ ip6_mroute.c ~ nd6.c > There's no point in abstracting ifp->if_output() as long as pf_test() > needs to see lo0 in the output path. > ok claudio@ (mpi@) ~ icmp6.c ~ in6_src.c ~ nd6.c ~ nd6_nbr.c ~ nd6_rtr.c > Do not manually decrement rt's refcounter in nd6_lookup() and let the > callers rtfree(9) it. > Inputs and ok bluhm@ (mpi@) netmpls ~ mpls_input.c ~ mpls_output.c > There's no point in abstracting ifp->if_output() as long as pf_test() > needs to see lo0 in the output path. > ok claudio@ (mpi@) sys ~ refcnt.h > tweak ordering slightly (dlg@) ~ sched.h > Introduce sched_barrier(9), an interface that acts as a scheduler barrier > in > the sense that it guarantees that the specified CPU went through the > scheduler. This also guarantees that interrupt handlers running on that > CPU > will have finished when sched_barrier() returns. > ok miod@, guenther@ (kettenis@) ~ disklabel.h > Rename readgptlabel() to spoofgptlabel() because that's what we > really want it to do. Handle all the actual disklabel reading in > readdoslabel(). Makes the code much simpler to understand. > ok deraadt@ (krw@) ~ disklabel.h > Move prototype for spoofgptlabel() from disklabel.h to subr_disk.c. > It's a helper function for readdoslabel(). Not something called > outside of subr_disk.c. (krw@) ~ disklabel.h > Remove unused and incorrect defines GPT_PARTSPERSEC and GPT_SECOFFSET. > (krw@) ~ syscall.h ~ syscallargs.h > sync (guenther@) ~ srp.h > implement SRPL_INSERT_AFTER_LOCKED. > i thought id committed this at l2k15. sorry for the delay. (dlg@) uvm ~ uvm_glue.c > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) ~ uvm_km.c > Back out rev. 1.125. This bit was left behind (intentionally?) when the > remainder of that commit was backed out. However,clearing the PQ_AOBJ bit > here is definitely wrong. > Our pagedaemon uses two separate lists to keep track of inactive > pages. It uses PQ_SWAPBACKED, which really is both PQ_ANON and > PQ_AOBJ to keep track of which inactive queue a page is sitting on. > So if you twiddle PQ_AOBJ (or PQ_ANON) for an inactive page, a > subsequent uvm_pagefree(9) will remove the page from the wrong queue! > This usually goes unnoticed, but if the page happens to be the last > one on the queue, the queues get corrupted. The damage quickly > spreads to the free page queues and almost certainly results in the > uvm_pmr_size_RB_REMOVE_COLOR() faults that people have seen > sporadically since the spring of this year. > ok visa@, beck@, krw@, guenther@ (kettenis@) == usr.bin =========================================================== 13/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin aucat ~ aucat.1 > typo resultung in a markup error (schwarze@) awk ~ awk.1 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) calendar ~ calendar.1 > for argument placeholders for file names use .Ar, not .Pa (schwarze@) chpass ~ Makefile > Supply DEF_WEAK() macro to suppress warnings during compilation via > reach-around into libc > noted by daniel@ (guenther@) col ~ col.1 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) doas ~ doas.c > doas doesn't need any files to be passed in. closefrom STDERR+1. ok benno > (tedu@) head ~ head.1 > avoid the useless idiom Fl Ns Ar, use Fl Ar (schwarze@) hexdump ~ od.1 ~ odsyntax.c > Fix two bugs (at least 13 year old) of the [+]offset[.][Bb] XSI offset > syntax: > Make it work with decimal offsets and with standard input. > Also fix the documentation; it showed the argument in the wrong place. > According to XSI (and our code), it must come *after* the filename, > and it can be used with at most one single filename. > Remove it from the SYNOPSIS; it only provides a subset of -j, > is less standard and contains lots of traps (i counted 7 distinct traps). > Instead, mention it as an alternative below -j where it belongs. > Also fix some markup quirks (that's what originally made me find this > mess). > Feedback and OK sobrado@ jmc@. (schwarze@) kdump ~ kdump.c > Rename __sysctl syscall to just sysctl, as the userland wrapper is no > longer > necessary > ok deraadt@ jsing@ (guenther@) m4 ~ m4.1 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) mandoc ~ mdoc_term.c > The .Dv macro actually forces normal font. (schwarze@) ~ mandoc.1 ~ mandoc.h ~ mdoc_validate.c ~ read.c > Remove the warning about children of .Vt blocks because actually, > .Vt type global_variable No = Dv defined_constant ; > is the best way to specify in the SYNOPSIS how a global variable > is initialized in the rare case where that matters. > Issue noticed by jmc@. (schwarze@) ~ mandocdb.c > for portability, use (char *)NULL in execlp(3) as discussed on tech@ > (didn't blow up anywhere yet, but better safe than sorry) (schwarze@) mg ~ dired.c > Calculate the correct line number when opening in dired mode. > ok florian@ (lum@) nc ~ netcat.c > display negotiated TLS version and cipher suite in verbose mode. > ok jsing@ (beck@) nm ~ nm.1 > fix markup of the -t argument (schwarze@) openssl ~ openssl.1 > tweak STANDARDS; (jmc@) ~ openssl.1 > document extra algorithms available with openssl speed command > ok jmc@ (bcook@) ~ apps.c ~ apps.h ~ apps_posix.c ~ openssl.c > Factor out setup_up / destroy_ui functions. > This pulls out and renames setup_ui/destroy_ui so we have something that > can be replaced as-needed, moving the the console setup code for Windows > to app_win.c in -portable, instead of needing a local patch to enable > binary > console mode > ui_read/write are also simplified. (bcook@) ~ openssl.1 > nc(1) seems worth an Xr in SEE ALSO now; (jmc@) ~ openssl.1 ~ openssl.c ~ req.c ~ speed.c ~ ts.c > Remove MD4 support from LibreSSL. > MD4 should have been removed a long time ago. Also, RFC 6150 moved it to > historic in 2011. Rides the major crank from removing SHA-0. > Discussed with many including beck@, millert@, djm@, sthen@ > ok jsing@, input + ok bcook@ (doug@) ~ openssl.1 ~ openssl.c ~ req.c ~ speed.c ~ ts.c > Temporarily revive MD4 for MS CHAP support. (doug@) ~ speed.c > Pack the algorithm numbers, to avoid printing a useless (null) 0 0 0 0 > line in the summary. (miod@) passwd ~ Makefile > Supply DEF_WEAK() macro to suppress warnings during compilation via > reach-around into libc > noted by daniel@ (guenther@) ssh ~ ssh-add.c > When adding keys to the agent, don't ignore the comment of keys for which > the > user is prompted for a passphrase. > Tweak and OK djm@ (tim@) ~ authfile.c ~ ssh-add.c ~ ssherr.c ~ sshkey.c ~ sshkey.h > - Fix error message: passphrase needs to be at least 5 characters, not 4. > - Remove unused function argument. > - Remove two unnecessary variables. > OK djm@ (tim@) tail ~ tail.1 > avoid the useless idiom Fl Ns Ar, use Fl Ar (schwarze@) tmux ~ cmd-wait-for.c > Set woken flag when flushing so that the channel is freed, while here > use the same loop construct for both loops. (nicm@) ~ cmd-copy-mode.c ~ tmux.1 ~ tmux.h ~ window-copy.c > Add copy-mode -e to exit copy mode when scrolling off the bottom, useful > for quick view of history, from Cam Hutchison. (nicm@) ~ cmd-refresh-client.c ~ format.c ~ server-client.c ~ status.c ~ tmux.h > Make refresh-client force update of jobs, from Sina Siadat. (nicm@) ~ cmd-select-pane.c ~ style.c ~ tmux.h ~ window.c > When the active pane changes, redraw panes if the style has > changed. From Cam Hutchison. (nicm@) ~ tmux.h ~ window-copy.c > Style nit, int for flags not u_int. (nicm@) ~ log.c ~ tmux.c > Move tzset() from log_open to main. (nicm@) ~ cmd-set-buffer.c > Should add buffer if no -b. (nicm@) ~ cmd-find.c ~ cmd-set-option.c ~ paste.c ~ screen-write.c ~ window-clock.c > Remove some extra blank lines. (nicm@) ~ cmd-if-shell.c ~ cmd-queue.c ~ cmd-run-shell.c ~ server-client.c ~ tmux.h > Rename cmd_q dead flag to a general flags bitmask (will be more flags > later). (nicm@) ~ cmd-attach-session.c > Hoist some common code out of both branches of an if/else. (nicm@) ~ cmd-list-panes.c > A few minor style nits. (nicm@) ~ cmd-queue.c > Log when cmdq_continue is called. (nicm@) ~ cmd-kill-session.c > Give some variables less silly names. (nicm@) ~ cmd-break-pane.c > Redraw both src and dst sessions in break-pane. (nicm@) ~ layout.c > -l should apply to the new not the old pane with -b, from "MadMaverick9" > on GitHub. (nicm@) vi ~ docs/USD.doc/vi.man/vi.1 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) == usr.sbin ========================================================== 14/14 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin amd ~ amd/rpc_fwd.c > More FNDELAY/FIONBIO --> SOCK_NONBLOCK conversions > ok millert@ (guenther@) bgpctl ~ bgpctl.c > match a few if_media uint64_t conversions; ok miod (deraadt@) ~ bgpctl.c > Another ifmedia64 fix, this time for bgpd's ift2ifm(). > And rename struct kif's media_type to if_type to avoid confusion with > ifmedia's media type. Same rename in other daemons will follow. > ok claudio henning benno (stsp@) bgpd ~ pfkey.c > explicit_bzero() from Michael McConville, thanks! > OK claudio (florian@) ~ bgpd.h ~ kroute.c > Another ifmedia64 fix, this time for bgpd's ift2ifm(). > And rename struct kif's media_type to if_type to avoid confusion with > ifmedia's media type. Same rename in other daemons will follow. > ok claudio henning benno (stsp@) ~ rde.c > Use correct function name in error message (claudio@) dhcpd ~ dhcp-options.5 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) ldpctl ~ ldpctl.c > match a few if_media uint64_t conversions; ok miod (deraadt@) npppd ~ npppd/npppd_subr.c ~ pppoe/pppoe_session.c > Set the index of the interface in handroller link_addr(3) like code. > This is now necessary on OpenBSD because our routing interface no longer > use ifunit() to find the corresponding interface. > ok yasuoka@, dlg@ (mpi@) ospf6ctl ~ ospf6ctl.c > match a few if_media uint64_t conversions; ok miod (deraadt@) ospfctl ~ ospfctl.c > match a few if_media uint64_t conversions; ok miod (deraadt@) pkg_add ~ pkg_create.1 > avoid useless .Ns before punctuation after in-line macros (schwarze@) portmap ~ portmap.c > Rename the routines used for packing/unpacking rmtcall RPCs so they don't > conflict with those in libc. They're already static, but best to be clear > that they're different in functionality, supporting operations that libc's > don't need or have. > with deraadt@ (guenther@) ~ portmap.c > More FNDELAY/FIONBIO --> SOCK_NONBLOCK conversions > ok millert@ (guenther@) pppd ~ pppd.8 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) relayd ~ ca.c ~ config.c ~ parse.y ~ pfe.c ~ relay.c ~ relay_http.c ~ relayd.c ~ relayd.h ~ ssl.c TAGGED OPENBSD_5_7 > maintainance diff for relayd > MFC the following changes > - Missing free(3) in error path (ssl.c,v 1.29) > - fix a memory leak. (pfe.c,v 1.80) > - allocate se_log evbuffer before loging errors with relay_close() > (relay.c,v 1.192) > - fix a file descriptor leak in http protocol handling > (relay.c,v 1.193 and relay_http.c,v 1.44) > - Fix obvious problems with relayd config reload > (ca.c,v 1.13; config.c,v 1.25; parse.y,v 1.204; > relayd.c,v 1.139; relayd.h,v 1.209) > - http protocol: you cannot append to the previous key-value before > line three of a request (relay_http.c,v 1.45) > - fix a crash / use after free (relay.c,v 1.194; relay_http.c,v 1.46) > - fix a non safe use of TAILQ_FOREACH with TAILQ_REMOVE > (relay_http.c,v 1.47) > - Plug a memory leak by simplifying kv_free() (relayd.c,v 1.141) > - Fix memory leak in error case (relay_http.c,v 1.48) > - track the connection state of a session and stops doing double > opens in certain situations (relay.c,v 1.195; relay_http.c,v 1.49; > relayd.h,v 1.210) > - coding style (relay.c,v 1.196; relay_http.c,v 1.50; relayd.h,v 1.212) > ok claudio@, sthen@ and feedback tedu@ (benno@) ripctl ~ ripctl.c > match a few if_media uint64_t conversions; ok miod (deraadt@) rmt ~ rmt.8 ~ rmt.c > Add a set of flags to rmt(8) to make it run in a restricted mode, > designed to work with rdump(8) to remote disk. > -d <directory> confines rmt to operate within a single directory. > -r enforces read-only mode. > -w enforces write-only mode. > This is quite usable with public ssh key setup, e.g. having the following > in .ssh/authorized/keys: > command="/etc/rmt -rd > /backups/machine.example.conf",no-agent-forwarding,... ssh-rsa AAAAB3... > ok semarie@ (halex@) syslogd ~ evbuffer_tls.c > Misuse of libevent in TLS read and write might have caused strange > event loss in syslogd. The ev_read and ev_write event structures > were used for both EV_READ and EV_WRITE events intermixedly. The > correct way is to use each event for its read and write purpose and > instead switch the handler. Then libevent is no longer confused. > When doing TLS read and a TLS_WANT_POLLOUT occures, call event_set() > for the write event to change the callback to read. As there might > be an event scheduled on the event, call event_del() before > event_set(). After a successful TLS read, change back the write > event. Add analog code for TLS write and handshake. > OK beck@ (bluhm@) sysmerge ~ sysmerge.sh > More dead code. (ajacoutot@) ~ sysmerge.sh > Tweak warning/error output a bit. > Check for examples changes at the end. (ajacoutot@) ~ sysmerge.8 ~ sysmerge.sh > Simplify: we don't need a work directory anymore, temproot is enough. > (ajacoutot@) ~ sysmerge.sh > In case of a hard error, remove all checksums except for examplessum which > is special. Otherwise, we may miss comparing files at next run; we may miss > some automation on the next run but it's safer this way. (ajacoutot@) tcpdump ~ tcpdump.8 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) traceroute ~ traceroute.8 > Avoid .Ns right after .Pf, it's pointless. > In some cases, do additional cleanup in the immediate vicinity. (schwarze@) =============================================================================== _______________________________________________ owc mailing list [email protected] http://www.squish.net/mailman/listinfo/owc
