OpenBSD src changes summary for 2015-11-19 ==========================================
distrib/sets lib/libkeynote libexec/login_passwd libexec/login_reject libexec/login_token regress/lib sbin/fdisk sbin/growfs sbin/iked sbin/wsconsctl share/man sys/arch/octeon/dev sys/dev/pci sys/kern sys/netinet sys/sys usr.bin/cvs usr.bin/less usr.bin/lex usr.bin/locale usr.bin/locate usr.bin/mg usr.bin/newsyslog usr.bin/ssh usr.bin/tail usr.bin/telnet usr.bin/tmux usr.bin/vi usr.sbin/config usr.sbin/hotplugd usr.sbin/httpd usr.sbin/ntpd usr.sbin/rarpd usr.sbin/relayd usr.sbin/sensorsd == distrib =========================================================== 01/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib sets ~ lists/base/mi ~ lists/comp/mi > sync (deraadt@) ~ lists/comp/mi > sync (deraadt@) == lib =============================================================== 02/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib libkeynote ~ auxil.c ~ keynote-keygen.c ~ keynote-sign.c ~ keynote-sigver.c ~ keynote-verify.c ~ keynote.l ~ keynote.y ~ parse_assertion.c > Remove a ton of ugly, needless casts for NULL, calloc(), and strdup(). > No binary change. (mmcc@) ~ keynote-verify.c > Remove a sixteen-year-old testing macro. > "yes." deraadt@ (mmcc@) ~ keynote-sign.c > Fix typo in previous commit (guenther@) ~ keynote.l ~ keynote-ver.l > lexer fixes to work with new flex. from Serguey Parkhomovsky (tedu@) == libexec =========================================================== 03/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/libexec login_passwd ~ login_passwd.c > Call syslog() if login_* pledge fails; OK deraadt@ (millert@) login_reject ~ login_reject.c > Call syslog() if login_* pledge fails; OK deraadt@ (millert@) login_token ~ login_token.c > Call syslog() if login_* pledge fails; OK deraadt@ (millert@) == regress =========================================================== 04/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress lib ~ libpthread/socket/1/socket1.c > hint: compile before commit (deraadt@) == sbin ============================================================== 05/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sbin fdisk ~ cmd.c ~ disk.c ~ gpt.c ~ mbr.c ~ misc.c ~ part.c ~ user.c > Reduce overburden of unnecessary #include's. Note that param.h is > included only for DEV_BSIZE. Use INT64_MAX instead of LLONG_MAX > so stdint.h is enough. General tidying up of #include sections. > ok GCC (krw@) ~ fdisk.c > Oops. Missed a file. > Reduce overburden of unnecessary #include's. Note that param.h is > included only for DEV_BSIZE. Use INT64_MAX instead of LLONG_MAX > so stdint.h is enough. General tidying up of #include sections. > ok GCC (krw@) ~ cmd.c > When an existing partition is modified in LBA mode, ensure that the > partition > table is marked dirty so that it gets written when "quit" is issued. > This commit contains the MBR part of the fix only. The GPT code needs a few > adjustments before the GPT part can be fixed properly. > OK krw@ (tim@) growfs ~ growfs.c > calloc((size_t)1, ...) -> calloc(1, ...) > No binary change (for real this time). (mmcc@) ~ growfs.c > Don't bother casting memset() argument to void*. No binary change. (mmcc@) ~ growfs.c > Remove some weird comments containing only the name of the next function. > (mmcc@) iked ~ util.c > Simplify all instances of get_string() and get_data() using malloc() and > strndup(). > ok millert@ (mmcc@) wsconsctl ~ map_scan.l > lexer fixes to work with new flex. from Serguey Parkhomovsky (tedu@) == share ============================================================= 06/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/share man ~ man3/Makefile ~ man3/queue.3 > Add SIMPLEQ_CONCAT and TAILQ_CONCAT for moving one queue onto the end > of another one. Adapted from FreeBSD. OK jmc@ dlg@ nicm@ (millert@) == sys =============================================================== 07/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys arch/octeon/dev ~ cn30xxfpa.c > Remove an unused bus space mapping and a few other leftovers. Compact > the resulting code a little. (visa@) dev/pci ~ if_myx.c > get rid of sc_tx_free and the atomic ops on it in myx_start and myx_txeof. > myx_start calculates the free space by reading the consumer index > and doing some maths, which lets us avoid the interlocked cpu ops. (dlg@) kern ~ uipc_syscalls.c > corrects leaks refs to files introduced by my previous commit for > pledge_socket. > reported by Mateusz Guzik with a diff. > this one is a slightly modified version. > ok deraadt@ (semarie@) ~ kern_task.c > dont try and wakeup other threads to handle pending work when we > know there's only one thread in the taskq. wakeups are much more > expensive than a simple compare. > from haesbart (dlg@) netinet ~ ip_output.c > Grab the KERNEL_LOCK around ip_mforward(), in preparation for unlocking > ip_output(). > Note that ipmforwarding is not enabled by default. (mpi@) sys ~ queue.h > Add SIMPLEQ_CONCAT and TAILQ_CONCAT for moving one queue onto the end > of another one. Adapted from FreeBSD. OK jmc@ dlg@ nicm@ (millert@) ~ queue.h > The CIRCLEQ_* macros have been deprecated and removed from queue.3 > over a year ago. It's now time to remove them from queue.h. > OK krw@ guenther@ deraadt@ (millert@) == usr.bin =========================================================== 08/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin cvs ~ cvs.c > opencvs can use pledge "stdio rpath wpath cpath fattr proc exec". > I have not looked into adding stricter pledges later in the code. > (deraadt@) less ~ lesskey.c > if you're going to call err(), need to use <err.h> (tedu@) lex - VERSION ~ COPYING ~ FlexLexer.h ~ Makefile ~ ccl.c ~ config.h ~ dfa.c ~ ecs.c ~ flex.skl ~ flexdef.h ~ gen.c ~ initscan.c ~ libmain.c ~ libyywrap.c ~ main.c ~ misc.c ~ mkskel.sh ~ nfa.c ~ parse.y ~ scan.l ~ sym.c ~ tblcmp.c ~ version.h ~ yylex.c + ChangeLog + NEWS + ONEWS + README + buf.c + filter.c + flexint.h + gettext.h + initparse.c + initparse.h + initskel.c + options.c + options.h + regex.c + scanflags.c + scanopt.c + scanopt.h + tables.c + tables.h + tables_shared.c + tables_shared.h > Update flex from ancient 2.5.4 to the recent 2.5.39. > This work was all done by Serguey Parkhomovsky. Thanks. > Some changes from upstream: > * Removed autotools cruft/localization/texinfo manual/etc > * Kept the old manpage, as the new manpage is content-free > * Used safe string handling functions and fixed several compiler warnings > * pledge(2). Flex 2.5.39 now forks/execs its filter chains and needs proc > exec in addition to what was previously pledged > * Removed register keyword from all variable declarations > * renamed parse.c, parse.h, scan.c, skel.c with init prefix so compiling > flex outside of obj by accident wouldn't clobber the bootstrap files > * Minor fixes (spelling, accessing buf[strlen(buf) - 1] for zero-length > strings in initscan.c/scan.l, etc) that were already in our tree > This is a huge change, so it's going in the tree code bomb style. > I'm not excited about the growth in complexity (like now running m4 > to _post_ process lexers) but hopefully this will be a one time update > and we will now take "ownership" of the code. (tedu@) ~ Makefile > add back flex++.1 link noticed by sthen (tedu@) ~ buf.c ~ filter.c ~ initskel.c ~ options.c ~ options.h ~ regex.c ~ scanflags.c ~ scanopt.c ~ scanopt.h ~ tables.c ~ tables.h ~ tables_shared.h > we don't keep vim modelines in files (tedu@) ~ filter.c > repair some of the preposterously damaged indentation (tedu@) ~ main.c > typically indentation decreases after a block is closed (tedu@) ~ buf.c ~ filter.c ~ gen.c ~ libmain.c ~ libyywrap.c ~ main.c ~ misc.c ~ nfa.c > orbital strike from moonbase knf (tedu@) ~ ccl.c ~ tblcmp.c > mechanical knf (tedu@) ~ flexdef.h ~ main.c > remove VMS and MSDOS support (tedu@) ~ flexdef.h ~ misc.c ~ tblcmp.c > it should be safe to assume the presence of memset these days instead of > implementing a local version. (tedu@) ~ buf.c ~ dfa.c ~ filter.c ~ flexdef.h ~ gen.c ~ initscan.c ~ main.c ~ misc.c ~ regex.c ~ scan.l ~ scanflags.c ~ sym.c > flex_alloc and flex_free are nothing more than malloc and free, so replace > them with the real functions so as to not trick people into thinking they > are special (tedu@) ~ buf.c ~ filter.c ~ gen.c > it is not necessary to cast the result of malloc/calloc. > also replace some 0s with the modern concept of NULL (tedu@) ~ ecs.c ~ flexdef.h ~ initscan.c ~ main.c ~ misc.c ~ scan.l ~ scanopt.c ~ sym.c ~ tblcmp.c > Replace Char (defined as unsigned char) with u_char. > ok tedu@ (mmcc@) ~ buf.c > split out some variable initialization to be easier (tedu@) - gettext.h ~ flexdef.h > remove the LGPL gettext.h since we aren't using it anyway. > spotted by nicm (tedu@) ~ Makefile > make sure parse.c depends on initparse.c (tedu@) ~ config.h ~ flexdef.h ~ initparse.c > burn down all the gratutious alloca() compatability that autofuck injected > to ensure nobody uses alloca by accident. (tedu@) locale ~ locale.c > pledge "stdio rpath". Might not need "rpath", but ... (deraadt@) locate ~ locate/locate.c > Sort and simplify includes, remove a couple needless extern decls. > From Michael Reed. ok tedu@ (mmcc@) mg ~ main.c > "tty proc exec", not "proc exec tty" (bentley@) newsyslog ~ newsyslog.c > backout pledge completely for now, newsyslog needs chown() which is > restricted (sthen@) ssh ~ ssh-keygen.c ~ sshkey.c ~ sshkey.h > move the certificate validity formatting code to sshkey.[ch] (djm@) ~ sshconnect.c > print host certificate contents at debug level (djm@) ~ ssh-keygen.c > trailing whitespace (djm@) ~ ssh.c ~ sshconnect.c > ban ConnectionAttempts=0, it makes no sense and would cause > ssh_connect_direct() to print an uninitialised stack variable; > bz#2500 reported by dvw AT phas.ubc.ca (djm@) tail ~ extern.h ~ forward.c ~ misc.c ~ read.c ~ reverse.c ~ tail.c > another try to allow tailing multiple files. maybe it works? > commit now to allow people to test. > from Martijn van Duren (tedu@) telnet ~ main.c > how did the pledge "dns" get forgotten?!?! wow.. (deraadt@) tmux ~ alerts.c > The activity flag could already be set, so queue the callback always (if > not already queued) rather than only if the flag is being added. Fixes a > problem reported by tim@ (nicm@) ~ server-client.c ~ tmux.h > Only assume pasting with at least two characters, reduces problems for > people who can type ^B c very fast, or who are using tmux inside > something else that buffers. (nicm@) vi - ex/ex_cscope.c ~ build/Makefile ~ common/common.h ~ common/exf.c ~ common/msg.c ~ common/screen.h ~ common/search.c ~ docs/help ~ docs/USD.doc/vi.man/vi.1 ~ ex/ex.h ~ ex/ex_cmd.c ~ ex/ex_display.c ~ ex/ex_init.c ~ ex/ex_subst.c ~ ex/ex_tag.c ~ ex/tag.h ~ include/ex_def.h ~ include/ex_extern.h > Remove cscope support in vi. > It makes no sense to keep support for a non-base tool in base, especially > for a feature that few if any people use. (bentley@) ~ common/main.c > "tty proc exec", not "proc exec tty" (bentley@) == usr.sbin ========================================================== 09/09 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin config ~ scan.l > lexer fixes to work with new flex. from Serguey Parkhomovsky (tedu@) hotplugd ~ hotplugd.c > pledge "stdio rpath proc exec" for entire lifetime. rpath must remain > even late, because of access() before doing wasteful fork+execs. (deraadt@) httpd ~ httpd.c > Simplify all instances of get_string() and get_data() using malloc() and > strndup(). > ok millert@ (mmcc@) ntpd ~ constraint.c > Simplify all instances of get_string() and get_data() using malloc() and > strndup(). > ok millert@ (mmcc@) rarpd ~ arptab.c ~ rarpd.c > hoist the SOCK_RAW setup to the top before pledge > issues discovered by sebastia > comments from semarie (deraadt@) relayd ~ relayd.c > Simplify all instances of get_string() and get_data() using malloc() and > strndup(). > ok millert@ (mmcc@) sensorsd ~ sensorsd.c > KNF (deraadt@) ~ sensorsd.c > pledge "stdio rpath proc exec" (deraadt@) =============================================================================== _______________________________________________ odc mailing list [email protected] http://www.squish.net/mailman/listinfo/odc
