OpenBSD src changes summary for 2016-02-21 to 2016-02-28 inclusive ==================================================================
distrib/macppc distrib/miniroot distrib/sets games/caesar games/hangman games/worms gnu gnu/usr.bin/perl lib/libc regress/lib regress/sys regress/usr.bin sbin/fdisk sbin/ifconfig sbin/pdisk share/man sys/arch/alpha/alpha sys/arch/alpha/include sys/arch/amd64/amd64 sys/arch/amd64/include sys/arch/arm/arm sys/arch/arm/include sys/arch/hppa/hppa sys/arch/hppa/include sys/arch/hppa64/hppa64 sys/arch/hppa64/include sys/arch/i386/conf sys/arch/i386/i386 sys/arch/i386/include sys/arch/landisk/landisk sys/arch/macppc/macppc sys/arch/mips64/mips64 sys/arch/sgi/sgi sys/arch/sh/include sys/arch/sh/sh sys/arch/socppc/socppc sys/arch/sparc/include sys/arch/sparc/sparc sys/arch/sparc64/dev sys/arch/sparc64/include sys/arch/sparc64/sparc64 sys/arch/vax/include sys/arch/vax/vax sys/arch/zaurus/stand/zboot sys/conf sys/ddb sys/dev/pci sys/dev/pckbc sys/dev/usb sys/isofs/cd9660 sys/kern sys/miscfs/fuse sys/net sys/netinet sys/netinet6 sys/ntfs sys/sys sys/ufs/ext2fs sys/ufs/ffs sys/ufs/ufs usr.bin/kdump usr.bin/patch usr.bin/rdist usr.bin/skeyinit usr.bin/ssh usr.sbin/eigrpd usr.sbin/lpr usr.sbin/pkg_add usr.sbin/rtadvd usr.sbin/smtpd usr.sbin/tcpdump usr.sbin/tftp-proxy == distrib =========================================================== 01/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib macppc ~ ramdisk/install.md > In the great pdisk churn we lost a pdisk pecularity in disk name > handling. "pdisk -l /dev/wd0c" used to work as well as "pdisk -l > /dev/rwd0c". The new, more fdisk-like, code does not accept the > former. > Adjust the install scripts to eliminate the use of "/dev/wd0c" > constructs and, like fdisk, simply use 'wd0'. > Fixes macppc installs to HFS partitioned disks. > Discovered and diff by rpe@. > ok rpe@ deraadt@ (krw@) miniroot ~ install.sub > In the great pdisk churn we lost a pdisk pecularity in disk name > handling. "pdisk -l /dev/wd0c" used to work as well as "pdisk -l > /dev/rwd0c". The new, more fdisk-like, code does not accept the > former. > Adjust the install scripts to eliminate the use of "/dev/wd0c" > constructs and, like fdisk, simply use 'wd0'. > Fixes macppc installs to HFS partitioned disks. > Discovered and diff by rpe@. > ok rpe@ deraadt@ (krw@) ~ install.sub > Shorter version that does the same. > Change description to better match the purpose of this function. > OK halex@ (rpe@) ~ install.sub > Cleanup the list of usernames that are rejected during installation. > - sendmail, named, popa3d are not part of base anymore and their ports > versions use _<username>. > - _tftp_proxy and _ftp_proxy are used instead of the proxy user. > OK ajacoutot@ (rpe@) sets ~ lists/man/mi > sync (sthen@) == games ============================================================= 02/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/games caesar ~ caesar.c > - Convert atoi(3) to strtonum(3) > - Remove lint-style comment > - Remove usage() since errors are now more informative (the usage is still > available in the manpage) > With this diff I made it accept 0 as rotation whereas before it didn't, but > alas if you use 0 then it defeats the whole purpose of the game. > Initial diff sent by Peter Williams <peterbw at gmail.com>, tweaked by me > and > several hints and OK by tb@. I'm in desperate need of a coffee, thank you > tb@ > for making me notice that! (mestre@) hangman ~ main.c > - Remove -h and -? from getopt(3), they weren't documented on manpage > anyway > - Replace hardcoded string "hangman" by getprogname(3) > OK tb@ (mestre@) ~ main.c > Add back undocumented -h switch defaulting to usage(). > pjanzen@ correctly pointed out that a majority of games do > this, so no need to deviate here. (tb@) worms ~ worms.c > Untangle a mess of nested ternary operators. > ok mestre@ (tb@) == gnu =============================================================== 03/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/gnu gnu ~ gcc/gcc/config/alpha/alpha.c > In alpha_expand_block_clear(), cope with the offset being negative; > this can happen due to the frame layout change introduced in order to > support the stack protector. Fix from miod. > Bug originally observed by jca and condensed to a 3-liner by myself, > basically local [] arrays being initialized with shorter strings. > (deraadt@) ~ gcc/gcc/config/alpha/alpha.c > In emit_insxl() force the first operand of the insbl or inswl pattern > into a register. Fixes an ICE when building Mesa with __sync builtins. > From Roger Sayle in gcc svn rev 121779 in Feb 2007 before the license > change. > Tested by miod and matthieu. (jsg@) usr.bin/perl ~ perl.c ~ vms/vms.c > Cope with ambiguous environments > from Ricardo Signes > ok deraadt@ (afresh1@) == lib =============================================================== 04/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib libc ~ sys/sendsyslog.2 > Document sendsyslog2(2) in man page. > from Rafael Neves; input jmc@; OK deraadt@ (bluhm@) ~ asr/asr.c > Avoid a possible double-free if the "search" keyword is used multiple > times. > ok jca@ gilles@ (eric@) ~ stdlib/malloc.c > refactor option letter parsing into a subfunction, to increase clarity > about which options are turned on/off by 's' and 'S' > ok tedu (deraadt@) ~ gen/fnmatch.c > Fix negation of POSIX character classes; passed new regress test. > OK tb@ (millert@) ~ locale/mbtowc.c > If an incomplete character is passed to mbtowc(3), set errno to EILSEQ. > This is unambiguously required both by POSIX and by our own manual page. > It also makes a lot of sense because having a function that can > fail and that sets errno in some failure modes but does not set > errno in other failure modes would be a terrible idea. Such a > function would be ridiculously complicated to use. To detect the > reason for failure, you would have to: > - save errno > - reset errno to zero > - call the function > - inspect the return value to detect failure > - inspect errno to decide about the reason for failure > - if errno is zero, restore the saved errno > That is completely unreasonable, in particular for a seemingly innocous > function like mbtowc(3). Next to no programmer would get that right in > any real-world program. > Note that this bug is very widespread, it also affects NetBSD, > FreeBSD, Solaris 11, and glibc. I will also send a note around to > the other systems. > There may be fallout from programs using the interface incorrectly. > If you run into any, please report that to me. > OK millert@ (schwarze@) ~ locale/mbtowc.3 > Fix STANDARDS: use the appropriate .St argument and add the information > that setting errno is required by POSIX, but not by ISO C. (schwarze@) ~ sys/truncate.2 > fruncate() returns EFBIG when length exceeds the max file size. > ok mpi@ (natano@) == regress =========================================================== 05/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress lib ~ libc/fnmatch/fnm_test.in > Add a test for negated POSIX characer classes. (millert@) sys ~ ffs/run > Repair the namegen function to make ffs regress useful again. > openssl md5 adds a "(stdin)= " prefix to the output, but all we need is > the actual md5 sum itself. This caused most of the ffs regress tests to > fail. > ok mpi@ (natano@) usr.bin ~ patch/Makefile + patch/t18.diff + patch/t18.in + patch/t18.out > Properly handle ed-files which fully replace input file content. This > misbehaviour is triggered with 'c' and 'i' commands on empty buffers. > Spotted and fixed by Martin Natano <natano at natano dot net>. > ok millert@ (tobias@) == sbin ============================================================== 06/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sbin fdisk ~ fdisk.8 > A disk name of /dev/wd0c doesn't work. Correct verbiage and example. > Discovered by rpe@. Tweaked by schwarze@. > ok rpe@ schwarze@ jmc@ deraadt@ (krw@) ifconfig ~ ifconfig.8 > Add a missing Ns macro (tim@) pdisk ~ dump.c > Fix display of pdisk partition sizes by casting the 32-bit number > of blocks to (long long) before multiplying by 512 and passing the > value to fmt_scaled(). > Discovered and fix tested by rpe@. > ok rpe@ deraadt@ (krw@) ~ pdisk.8 > A disk name of /dev/wd0c doesn't work. Correct verbiage and example. > Discovered by rpe@. Tweaked by schwarze@. > ok rpe@ schwarze@ jmc@ deraadt@ (krw@) == share ============================================================= 07/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/share man ~ man8/man8.amd64/boot_amd64.8 ~ man8/man8.i386/boot_i386.8 > Xr biosboot and describe more precisely how this versatile program > will eventually load the first-stage bootstrap. From Michal Mazurek. > tweaks and ok jmc@ (tb@) ~ man4/man4.amd64/apm.4 ~ man4/man4.i386/apm.4 ~ man4/man4.loongson/apm.4 ~ man4/man4.macppc/apm.4 ~ man4/man4.sparc/tctrl.4 ~ man4/man4.zaurus/apm.4 > Add documentation for the APM_IOC_HIBERNATE ioctl. > tweaks and ok jmc@ (natano@) ~ man4/options.4 ~ man8/Makefile > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) ~ man4/wscons.4 > remove mention of Linux and FreeBSD binary emulation; pointed out by jmc@ > (naddy@) ~ man4/uhid.4 > Remove USB_SET_IMMED ioctl and corresponding code and manpage bits. > Nothing uses it. > suggested by and ok mpi@ (stefan@) ~ man8/autoinstall.8 > Tweak info about using subdirectories. > feedback from and OK jmc@ (rpe@) == sys =============================================================== 08/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys arch/alpha/alpha ~ pmap.c > Move #include <sys/mutex.h> from pmap.c to pmap.h, like every other archs > using struct mutex in pmap.h do. Fixes net-snmp on alpha. > ok deraadt@ (landry@) ~ clock.c > Make alpha and sh 2038-safe. Now that we have 64-bit time_t, we don't > need to ignore the hardware clock when it reports a year > 2037. Irony > noted. > both ok millert, alpha ok deraadt (mmcc@) ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) arch/alpha/include ~ pmap.h > Move #include <sys/mutex.h> from pmap.c to pmap.h, like every other archs > using struct mutex in pmap.h do. Fixes net-snmp on alpha. > ok deraadt@ (landry@) ~ pmap.h > put mutex.h lower (deraadt@) ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/amd64/amd64 ~ vmm.c > Plug memleak: make sure that the MSR bitmap is freed in the VCPU init > error path. > ok mlarkin@, deraadt@ (stefan@) ~ vmm.c > Fix previous: Need to free vc_msr_bitmap_va, not > vc_msr_bitmap_pa. (stefan@) ~ locore.S > SYMTAB_SPACE is no longer used (last used with a.out ddb) (mlarkin@) ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_trace.c > Rename and move x86 calllframe definitions in <machine/frame.h> to use > it in MI code. > ok mlarkin@, visa@ (mpi@) ~ db_interface.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/amd64/include ~ frame.h > Rename and move x86 calllframe definitions in <machine/frame.h> to use > it in MI code. > ok mlarkin@, visa@ (mpi@) ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/arm/arm ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_interface.c ~ fault.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/arm/include ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/hppa/hppa ~ disksubr.c > Restore validity checks for the disklabels read from disk. They > were suppressed in 2011 to address a problem that has since been > solved in a more permanent way. Without the validity check proper > disklabel discovery relied on a particular I/O failure, and the I/O > failure was 'solved' by the introduction of readdisksector(). > Fixes problem reading CDROM disklabels discovered by mlarkin@ while > trying to install hppa. > Compile check/typo discovery and ok deraadt@. (krw@) ~ db_interface.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/hppa/include ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/hppa64/hppa64 ~ disksubr.c > Restore validity checks for the disklabels read from disk. They > were suppressed in 2011 to address a problem that has since been > solved in a more permanent way. Without the validity check proper > disklabel discovery relied on a particular I/O failure, and the I/O > failure was 'solved' by the introduction of readdisksector(). > Fixes problem reading CDROM disklabels discovered by mlarkin@ while > trying to install hppa. > Compile check/typo discovery and ok deraadt@. (krw@) ~ db_interface.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) ~ intr.c > Remove a couple malloc casts. No functional change. (mmcc@) arch/hppa64/include ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/i386/conf ~ GENERIC ~ files.i386 > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) arch/i386/i386 ~ locore.s > SYMTAB_SPACE is no longer used (last used with a.out ddb) (mlarkin@) ~ locore.s > pctr.h is not used here, so no need to include it. (mlarkin@) ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_trace.c > Rename and move x86 calllframe definitions in <machine/frame.h> to use > it in MI code. > ok mlarkin@, visa@ (mpi@) ~ db_interface.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) ~ genassym.cf ~ locore.s ~ trap.c > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) arch/i386/include ~ frame.h > Rename and move x86 calllframe definitions in <machine/frame.h> to use > it in MI code. > ok mlarkin@, visa@ (mpi@) ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/landisk/landisk ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) arch/macppc/macppc ~ disksubr.c > Restore validity checks for the disklabels read from disk. They > were suppressed in 2011 to address a problem that has since been > solved in a more permanent way. Without the validity check proper > disklabel discovery relied on a particular I/O failure, and the I/O > failure was 'solved' by the introduction of readdisksector(). > Fixes problem reading CDROM disklabels discovered by mlarkin@ while > trying to install hppa. > Compile check/typo discovery and ok deraadt@. (krw@) ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_interface.c ~ locore.S > Rename ddb_trap_glue() to db_trap_glue(). > Let's keep the 'db_' prefix for all ddb(4) functions that should not be > instrumented. (mpi@) arch/mips64/mips64 ~ db_machdep.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/sgi/sgi ~ disksubr.c > Restore validity checks for the disklabels read from disk. They > were suppressed in 2011 to address a problem that has since been > solved in a more permanent way. Without the validity check proper > disklabel discovery relied on a particular I/O failure, and the I/O > failure was 'solved' by the introduction of readdisksector(). > Fixes problem reading CDROM disklabels discovered by mlarkin@ while > trying to install hppa. > Compile check/typo discovery and ok deraadt@. (krw@) ~ ip27_machdep.c ~ ip30_machdep.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/sh/include ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/sh/sh ~ clock.c > Make alpha and sh 2038-safe. Now that we have 64-bit time_t, we don't > need to ignore the hardware clock when it reports a year > 2037. Irony > noted. > both ok millert, alpha ok deraadt (mmcc@) ~ db_interface.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/socppc/socppc ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_interface.c ~ locore.S > Rename ddb_trap_glue() to db_trap_glue(). > Let's keep the 'db_' prefix for all ddb(4) functions that should not be > instrumented. (mpi@) ~ fdt.c > Remove useless code, from Patrick Wildt. (mpi@) ~ fdt.c > Convert FDT values from big endian to host endian. > This is a no-op on socppc but will allows ARM ports (little-endian) > to make use of this code. > From Patrick Wildt. (mpi@) arch/sparc/include ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/sparc/sparc ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_interface.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/sparc64/dev ~ vnet.c > Prevent memory leak when the ldc gets reset. (kettenis@) arch/sparc64/include ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/sparc64/sparc64 ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_interface.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/vax/include ~ db_machdep.h > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/vax/vax ~ conf.c > Remove stale RAIDframe entries from chrtoblktbl. > While there truncate the tables to the minimum required size; chrtoblk() > and blktochr() are designed to handle a table shorter than cdevsw. > "Looks good to me" deraadt@ (natano@) ~ db_machdep.c ~ trap.c > Rename kdb_trap() into db_ktrap(). > The goal is to include it in the list of functions that must not be > instrumented. All ddb(8) functions should be in this list and have > their names start with 'db_'. > ok visa@, deraadt@ (mpi@) arch/zaurus/stand/zboot ~ compat_linux.h > fix a typoed ioctl name; ok millert@ deraadt@ (naddy@) conf ~ newvers.sh > we are now hacking on 5.9-current (deraadt@) ~ GENERIC > activate POOL_DEBUG during -current; miod (deraadt@) ~ files > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) ddb ~ db_hangman.c ~ db_sym.c > Remove support for multiple symbol tables in ddb(4), we only have one. > ok mlarkin@, visa@ (mpi@) ~ db_elf.c ~ db_hangman.c ~ db_sym.c ~ db_sym.h > Remove the "symbol table" argument of the functions dealing with > symbols. > There's no reason to ask the programmer to supply a specific symbol > table since the kernel only have one. > ok visa@, guenther@ (mpi@) ~ db_sym.c > Remove leftovers from mutliple symbol tables support. (mpi@) dev/pci ~ drm/i915/dvo_ch7017.c ~ drm/i915/dvo_ch7xxx.c ~ drm/i915/dvo_ivch.c ~ drm/i915/dvo_ns2501.c ~ drm/i915/dvo_sil164.c ~ drm/i915/dvo_tfp410.c > Fix the build with DRMDEBUG. Committing now instead of after unlock to > make it easier for people running release to diagnose drm problems. > ok deraadt@ (jsg@) dev/pckbc ~ pms.c > pakets -> packets (mmcc@) dev/usb ~ uhid.c ~ usb.h > Remove USB_SET_IMMED ioctl and corresponding code and manpage bits. > Nothing uses it. > suggested by and ok mpi@ (stefan@) isofs/cd9660 ~ cd9660_vfsops.c > Move mnt_maxsymlink from struct mount to struct ufsmount. > The concept of differentiating between "short" and "long" symlinks is > specific to ufs/, so it shouldn't creep into the generic fs layer. > Inspired by a similar commit to NetBSD. > While there replace all references to mnt_maxsymlinklen in ufs/ext2fs > with EXT2_MAXSYMLINKLEN, which is the constant max short symlink len for > ext2fs. This allows to get rid of some (mnt_maxsymlinklen == 0) checks > there, which is always false for ext2fs. > input and ok stefan@ > ok millert@ (natano@) kern ~ uipc_mbuf.c > provide m_dup_pkt() for doing fast deep mbuf copies with a specified > alignment > if a physical interface receives a multicast/broadcast packet and > has carp interfaces on it, that packet needs to be copied for > reception by each of those carp interfaces. > previously it was using m_copym2, but that doesn't respect the > alignment of the source packet. this meant the ip header in the > copies were aligned incorrectly for the network stack, which breaks > strict alignment archs. > m_dup_pkt lets carp specify that the payload needs an ETHER_ALIGN > adjustment, so the ip header inside will be aligned correctly. > reported and tested by anthony eden who hit this on armv7 > i reproduced the problem on sparc64 and verified the fix on amd64 > and sparc64 > ok mpi@ mikeb@ deraadt@ (dlg@) ~ exec_conf.c ~ exec_elf.c > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) miscfs/fuse ~ fuse_vfsops.c > fuse doesn't support NFS; prevent a uvm fault caused by claiming otherwise > ok stefan@ (natano@) ~ fuse_vfsops.c > Replace a usage of ROOTINO ((ufsino_t)2 with FUSE_ROOTINO ((ino_t)1). > The FUSE_ROOTINO define is used everywhere else in fuse, so the mismatch > caused a file handle for the filesystem root to be falsely rejected with > ESTALE. > ok stefan@ (natano@) net ~ route.c ~ route.h ~ rtable.c > Fix ECMP routing by passing the correct destination address to the > hash routine. > Bug reported and fix analysed by Jean-Daniel Dupas <jddupas AT xooloo DOT > net> > ok deraadt@ (mpi@) ~ route.c > L2 entries are always in the first table of a routing domain, this fixes > a regression introduced during 5.7 and 5.8 as reported by Jean-Daniel > Dupas on misc@ > diff provided by mpi, ok claudio (deraadt@) ~ if.c > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) netinet ~ ip_carp.c > provide m_dup_pkt() for doing fast deep mbuf copies with a specified > alignment > if a physical interface receives a multicast/broadcast packet and > has carp interfaces on it, that packet needs to be copied for > reception by each of those carp interfaces. > previously it was using m_copym2, but that doesn't respect the > alignment of the source packet. this meant the ip header in the > copies were aligned incorrectly for the network stack, which breaks > strict alignment archs. > m_dup_pkt lets carp specify that the payload needs an ETHER_ALIGN > adjustment, so the ip header inside will be aligned correctly. > reported and tested by anthony eden who hit this on armv7 > i reproduced the problem on sparc64 and verified the fix on amd64 > and sparc64 > ok mpi@ mikeb@ deraadt@ (dlg@) ~ ipsec_output.c > When IPsec UDP encapsulation is used for IPv6, the stack should > construct an IPv6 packet instead of an IPv4. > Diff from Patrick Wildt <patrick at blueri ! se> with input from > bluhm@; ok mpi, bluhm (mikeb@) netinet6 ~ in6.c ~ in6_var.h > Remove SIOCSIFALIFETIME_IN6 ioctl, as NetBSD did. > As described in NetBSD kern/35897 PR, the parameters > this ioctl needs overlay each other in a union. The ioctl > cannot have worked properly. > Discovered while discussing overflow checks with mmcc@ and mpi@ > The checks were part of the removed code. > ok deraadt@ (stefan@) ntfs ~ ntfs_vfsops.c > Move mnt_maxsymlink from struct mount to struct ufsmount. > The concept of differentiating between "short" and "long" symlinks is > specific to ufs/, so it shouldn't creep into the generic fs layer. > Inspired by a similar commit to NetBSD. > While there replace all references to mnt_maxsymlinklen in ufs/ext2fs > with EXT2_MAXSYMLINKLEN, which is the constant max short symlink len for > ext2fs. This allows to get rid of some (mnt_maxsymlinklen == 0) checks > there, which is always false for ext2fs. > input and ok stefan@ > ok millert@ (natano@) sys ~ mbuf.h > provide m_dup_pkt() for doing fast deep mbuf copies with a specified > alignment > if a physical interface receives a multicast/broadcast packet and > has carp interfaces on it, that packet needs to be copied for > reception by each of those carp interfaces. > previously it was using m_copym2, but that doesn't respect the > alignment of the source packet. this meant the ip header in the > copies were aligned incorrectly for the network stack, which breaks > strict alignment archs. > m_dup_pkt lets carp specify that the payload needs an ETHER_ALIGN > adjustment, so the ip header inside will be aligned correctly. > reported and tested by anthony eden who hit this on armv7 > i reproduced the problem on sparc64 and verified the fix on amd64 > and sparc64 > ok mpi@ mikeb@ deraadt@ (dlg@) ~ socketvar.h > corrects a misleading comment about SS_DNS (and adjust tab) > SS_DNS is now created by using SOCK_DNS with socket(2). dnssocket(2) was > removed since October 28. (semarie@) ~ conf.h > Move misplaced cdev_ipmi_init() back into within #ifdef _KERNEL. > Pointed out by natano@. (uebayasi@) ~ mount.h > Convert mnt_maxsymlinklen to unsigned. > This allows to remove some truncating casts in symlink handling code. > Also, validate fs_maxsymlinklen in the superblock at mount time and on > fsck to make sure we don't use bogus data. > discussion & ok millert@, stefan@ (natano@) ~ mount.h > Move mnt_maxsymlink from struct mount to struct ufsmount. > The concept of differentiating between "short" and "long" symlinks is > specific to ufs/, so it shouldn't creep into the generic fs layer. > Inspired by a similar commit to NetBSD. > While there replace all references to mnt_maxsymlinklen in ufs/ext2fs > with EXT2_MAXSYMLINKLEN, which is the constant max short symlink len for > ext2fs. This allows to get rid of some (mnt_maxsymlinklen == 0) checks > there, which is always false for ext2fs. > input and ok stefan@ > ok millert@ (natano@) ~ ioctl.h > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) ufs/ext2fs ~ ext2fs_readwrite.c ~ ext2fs_vnops.c > Convert mnt_maxsymlinklen to unsigned. > This allows to remove some truncating casts in symlink handling code. > Also, validate fs_maxsymlinklen in the superblock at mount time and on > fsck to make sure we don't use bogus data. > discussion & ok millert@, stefan@ (natano@) ~ ext2fs_inode.c ~ ext2fs_readwrite.c ~ ext2fs_vfsops.c ~ ext2fs_vnops.c > Move mnt_maxsymlink from struct mount to struct ufsmount. > The concept of differentiating between "short" and "long" symlinks is > specific to ufs/, so it shouldn't creep into the generic fs layer. > Inspired by a similar commit to NetBSD. > While there replace all references to mnt_maxsymlinklen in ufs/ext2fs > with EXT2_MAXSYMLINKLEN, which is the constant max short symlink len for > ext2fs. This allows to get rid of some (mnt_maxsymlinklen == 0) checks > there, which is always false for ext2fs. > input and ok stefan@ > ok millert@ (natano@) ufs/ffs ~ ffs_vfsops.c ~ ffs_vnops.c > Convert mnt_maxsymlinklen to unsigned. > This allows to remove some truncating casts in symlink handling code. > Also, validate fs_maxsymlinklen in the superblock at mount time and on > fsck to make sure we don't use bogus data. > discussion & ok millert@, stefan@ (natano@) ~ ffs_inode.c ~ ffs_vfsops.c ~ ffs_vnops.c > Move mnt_maxsymlink from struct mount to struct ufsmount. > The concept of differentiating between "short" and "long" symlinks is > specific to ufs/, so it shouldn't creep into the generic fs layer. > Inspired by a similar commit to NetBSD. > While there replace all references to mnt_maxsymlinklen in ufs/ext2fs > with EXT2_MAXSYMLINKLEN, which is the constant max short symlink len for > ext2fs. This allows to get rid of some (mnt_maxsymlinklen == 0) checks > there, which is always false for ext2fs. > input and ok stefan@ > ok millert@ (natano@) ufs/ufs ~ ufs_dirhash.c ~ ufs_lookup.c ~ ufs_vnops.c ~ ufsmount.h > Move mnt_maxsymlink from struct mount to struct ufsmount. > The concept of differentiating between "short" and "long" symlinks is > specific to ufs/, so it shouldn't creep into the generic fs layer. > Inspired by a similar commit to NetBSD. > While there replace all references to mnt_maxsymlinklen in ufs/ext2fs > with EXT2_MAXSYMLINKLEN, which is the constant max short symlink len for > ext2fs. This allows to get rid of some (mnt_maxsymlinklen == 0) checks > there, which is always false for ext2fs. > input and ok stefan@ > ok millert@ (natano@) == usr.bin =========================================================== 09/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin kdump ~ kdump.1 > there is no emulation "bsd", our default is "native"; ok deraadt@ millert@ > (naddy@) ~ Makefile > Support for running Linux binaries under emulation is going away. > Remove "option COMPAT_LINUX" and everything directly tied to it from the > kernel and the corresponding man page documentation. > ok visa@ guenther@ (naddy@) ~ kdump.1 ~ kdump.c > drop the support for Linux emulation; ok guenther@ visa@ (naddy@) patch ~ ed.c > Properly handle ed-files which fully replace input file content. This > misbehaviour is triggered with 'c' and 'i' commands on empty buffers. > Spotted and fixed by Martin Natano <natano at natano dot net>. > ok millert@ (tobias@) rdist ~ rdist.1 > remove extra word; from edgar pettijohn (jmc@) skeyinit ~ skeyinit.c > Fix skeyinit -E by hoisting argument parsing and the call to enable_db() > above the calls to pledge(): /etc/skey wants to have its sticky bit. > ok deraadt@ (tb@) ssh ~ sshconnect2.c > fix spurious error message when incorrect passphrase entered for > keys; reported by espie@ ok deraadt@ (djm@) ~ version.h > openssh-7.2 (djm@) ~ dh.c > rearrange DH public value tests to be a little more clear > rearrange DH private value generation to explain rationale more > clearly and include an extra sanity check. > ok deraadt (djm@) == usr.sbin ========================================================== 10/10 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin eigrpd ~ kroute.c > Filter our RTM_GET messages which are not from us. > Pulled from ospfd. Original author: claudio@ (renato@) ~ tlv.c > Make eigrpd work against newer IOS routers. > Different versions of IOS can use a different number of bytes to encode > the same IPv6 prefix inside route TLVs. This sucks but we have to deal > with it. Instead of calculating the number of bytes based on the value > of the prefixlen field, let's get this number by subtracting the size > of all other fields from the total size of the TLV. It works because > all the other fields have a fixed length. > For reference, the EIGRP draft says that length of the prefix field > should be obtained according to this function: > ((Bit Count - 1) / 8) + 1 > But older IOS versions use this for IPv6 (obtained through reverse > engineering): > ((Bit Count == 128) ? 16 : ((Bit Count / 8) + 1)) > Now, the new IOS-XR apparently uses the first formula for both IPv4 and > IPv6. With this patch, eigrpd will work against both older and newer > versions of IOS. (renato@) ~ packet.c > Add scope id only for unicast IPv6 packets. (renato@) ~ eigrpd.c ~ eigrpd.h ~ eigrpe.h ~ interface.c ~ kroute.c ~ packet.c ~ parse.y ~ rde_dual.c ~ tlv.c ~ util.c > Several minor tweaks. (renato@) ~ kroute.c ~ neighbor.c ~ rde_dual.c ~ util.c > Extend eigrp_addrcmp() and use it in more places. > The idea is to remove unnecessary code duplication throughout the code. > (renato@) ~ eigrpd.h ~ kroute.c ~ packet.c ~ parse.y ~ tlv.c ~ util.c > Introduce the bad_addr() family of functions. > These functions improve code reusability as there's no more need to > check for all possible "bad" addresses in several different places. > Besides that, this patch introduce additional checks in the code. (renato@) ~ eigrpd.c ~ eigrpe.c ~ interface.c ~ kroute.c ~ log.c ~ neighbor.c ~ packet.c ~ parse.y ~ rde.c ~ rde_dual.c ~ rtp.c ~ tlv.c > Copy structs by assignment instead of memcpy. > Copying by straight assignment is shorter, easier to read and has a > higher level of abstraction. We'll only avoid it when copying from an > unaligned source (e.g., network buffers). (renato@) ~ tlv.c > Simplify the handling of the sequence tlv. > The C standard says: "A pointer to a union object, suitably converted, > points to each of its members". > This means that we can use the same code to process both v4 and v6 > addresses. (renato@) ~ eigrpe.c ~ eigrpe.h ~ interface.c ~ kroute.c ~ parse.y ~ rde_dual.c > Move some functions around and rename a few variables and functions. > The sooner we do these changes the better. This rearrangement will make > the code easier to read. (renato@) ~ rde_dual.c > Skip redistributed routes when updating the FIB. (renato@) ~ eigrpd.c > I just stumbled on this and noticed that eigrpd requires inet and rpath in > pledge(2) call to make "reload", > without that, daemon will crash on SIGHUP signal.... > as discurssed/OK renato@ > commit it now, deraadt@ (gsoares@) lpr ~ lpd/lpdchar.c > Fix ghastly whitespace. From Chris Bennett (tb@) pkg_add ~ OpenBSD/PackageRepository.pm > sanitize environment thru a whitelist. Only pass what's relevant for ftp > and > ftp clones. > Give "decent" values to your normal environment variables, so that commands > won't be surprised. > This fixes the spurious warnings from ftp:// url reported by Jiri B. > Much feedback from deraadt@, zhuk@, sthen@ (espie@) rtadvd ~ config.c ~ rtadvd.h > Kill some useless members of struct rainfo; ok florian@ (jca@) ~ config.c > Remove #if'0ed code that uses a deprecated ioctl; ok florian@ (jca@) ~ rtadvd.c ~ rtadvd.h > GC remnants of the Router Renumbering code; ok florian@ (jca@) ~ rtadvd.c > Unused variable. (jca@) ~ rtadvd.c > sys/signal.h -> signal.h (jca@) smtpd ~ smtpd.h > bump version (gilles@) ~ parse.y > the default address family for a listener was set too late and would > override the value provided by the user if any. this commit moves the > initialization earlier to avoid this. > issue reported by and fix ok jturner@ (gilles@) tcpdump ~ print-802_11.c > Make tcpdump show details of beacon country elements in verbose mode only. > ok sthen@ (stsp@) tftp-proxy ~ tftp-proxy.c > satisfy whiny compiler on m88k, noticed by aoyama, solution from gsoares > (deraadt@) =============================================================================== _______________________________________________ owc mailing list [email protected] http://www.squish.net/mailman/listinfo/owc
