OpenBSD src changes summary for 2017-08-04 ==========================================
gnu lib/libcompiler_rt lib/libexpat regress/libexec sbin/dhclient sbin/slaacd share/man sys/dev/fdt sys/dev/pci sys/net sys/net80211 sys/netinet sys/netinet6 usr.sbin/dhcpd usr.sbin/pkg_add usr.sbin/smtpd == gnu =============================================================== 01/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/gnu gnu - llvm/.gitignore > remove llvm's .gitignore file; upstream have separate repos for clang, lld, > etc but as they're directly in our main tree, so the ignore file causes > problems for people using cvs->git conversions. kettenis suggested removing > the file rather than just commenting-out the relevant repos, patrick > agrees. (sthen@) ~ llvm/tools/clang/include/clang/Basic/Builtins.def > Backport https://reviews.llvm.org/D29778. Fixes a bug where clang was > optimizing code around an lgamma(3) call a bit too much. Fixes the > lib/libm/lgamma regress test. > ok patrick@, millert@ (kettenis@) == lib =============================================================== 02/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib libcompiler_rt ~ Makefile > Enable atomic support functions. > ok mlarkin@ (kettenis@) libexpat - README ~ Changes ~ expat_config.h ~ lib/expat.h ~ lib/siphash.h ~ lib/xmlparse.c ~ lib/xmlrole.c ~ lib/xmltok.c ~ lib/xmltok_impl.c ~ tests/chardata.c ~ tests/chardata.h ~ tests/minicheck.c ~ tests/runtests.c + README.md > Update libexpat to 2.2.3. Only few changes affect OpenBSD. > OK deraadt@ (bluhm@) == regress =========================================================== 03/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress libexec ~ ld.so/constructor/libaa/aa.C ~ ld.so/constructor/libab/Makefile > Use volatile to keep clang from optimizing things too much. Force linking > with the c++ command instead of cc by setting CC=${CXX}. This is a hack > and a better solution is under development. Linking with c++ is necessary > to make the shared libraries depend up libc++.so such that the iostream > objects are constructed before we use them. > ok bluhm@ (kettenis@) ~ ld.so/init-env/libaa/Makefile > Use ${CC}=${CXX} here as well. > ok bluhm@ (kettenis@) ~ ld.so/initfirst/test2/libaa/aa.C ~ ld.so/initfirst/test2/libab/ab.C ~ ld.so/initfirst/test2/libac/ac.C ~ ld.so/initfirst/test2/libad/ad.C ~ ld.so/initfirst/test2/libae/ae.C > Use C stdio instead of C++ iostream as we cannot use the latter in > constructors for an "initfirst" object as the iostream objects have > not been constructed yet. > ok bluhm@ (kettenis@) == sbin ============================================================== 04/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sbin dhclient ~ kroute.c > Do not attempt to parse zero'd input buffer after > read() on routing socket fails. Just bail. (krw@) slaacd ~ engine.c > When free(3)'ing an address_proposal we need to stop the libevent > timer. > Resulting use-after-free found the hard way by me. > Introduce free_address_proposal() to make sure we always do this. > While here also stop the interface timer when deleting an interface. > (florian@) ~ engine.c > No need for LIST_FOREACH_SAFE, we want to remove exactly one interface > and then stop the iteration. get_slaacd_iface_by_id() is perfectly > fine for this; we also gain one level of indentation. > While here LIST_REMOVE the currently deleted dfr_proposal. (florian@) == share ============================================================= 05/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/share man ~ man4/Makefile ~ man4/dwge.4 + man4/dwmmc.4 > Add dwmmc(4) and update dwge(4). (kettenis@) == sys =============================================================== 06/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys dev/fdt ~ dwmmc.c > Fix defenition of reset bits. (kettenis@) dev/pci ~ pcidevs > add ASIX AX99100 device ID > ok by deraadt@ (uaa@) ~ pcidevs.h ~ pcidevs_data.h > regen (uaa@) ~ pucdata.c > add ASIX AX99100 support > ok by deraadt@ (uaa@) net ~ if.c > When we disable INET6 on an interface that also removes the autoconf6 > flag. Notify userland about this via the route socket, otherwise > slaacd gets confused about the state the interface is in. > OK on a previous version sthen > input & OK bluhm (florian@) net80211 ~ ieee80211_proto.c > Show net80211 state transitions in desmg if 'ifconfig debug' was used. > This information is needed in bug reports. > Convert the invalid state transitions from panic() to a printf() which is > also guarded by ifconfig debug. There are many races exposed by these > panics > which should all be fixed. But that will surely take some time, and the > panics have now served their purpose. Thanks to everyone who reported > these panics being triggered, your help is appreciated. (stsp@) ~ ieee80211_input.c > Remove ieee80211_input_print() which printed information about received > frames to dmesg, if debug mode was enabled with ifconfig. > This debug output was much too verbose and not actually useful for > debugging. > tcpdump -y IEEE802_11_RADIO will show the same information. > ok sthen@ (stsp@) ~ ieee80211_node.c > Compile a debug printf in ieee80211_match_bss() by default, previously > guarded by the IEEE80211_DEBUG preprocessor flag. This shows one line > per detected AP after a scan, and indicates which APs are considered > candidates for association. > Shorten the output a bit to fit into 80 columns more likely. > ok sthen@ (stsp@) netinet ~ ip_input.c > We do have SO_TIMESTAMP since some time and there is other code in the > kernel that uses it without the #ifdef guard. > OK bluhm (florian@) ~ in_pcb.c > The in_pcbhashlookup() in in_pcbconnect() enforces that the 4 tupel > of src/dst ip/port is unique for TCP. But if the socket is not > bound, the automatic bind by connect happens after the check. If > the socket has the SO_REUSEADDR flag, in_pcbbind() may select an > existing local port. Then we had two colliding TCP PCBs. This > resulted in a packet storm of ACK packets on loopback. The softnet > task was constantly holding the netlock and has a high priority, > so the system hung. > Do the in_pcbhashlookup() again after in_pcbbind(). This creates > sporadic "connect: Address already in use" errors instead of a hang. > bug report and testing Olivier Antoine; OK mpi@ (bluhm@) netinet6 ~ ip6_input.c > We do have SO_TIMESTAMP since some time and there is other code in the > kernel that uses it without the #ifdef guard. > OK bluhm (florian@) ~ in6_pcb.c > The in_pcbhashlookup() in in_pcbconnect() enforces that the 4 tupel > of src/dst ip/port is unique for TCP. But if the socket is not > bound, the automatic bind by connect happens after the check. If > the socket has the SO_REUSEADDR flag, in_pcbbind() may select an > existing local port. Then we had two colliding TCP PCBs. This > resulted in a packet storm of ACK packets on loopback. The softnet > task was constantly holding the netlock and has a high priority, > so the system hung. > Do the in_pcbhashlookup() again after in_pcbbind(). This creates > sporadic "connect: Address already in use" errors instead of a hang. > bug report and testing Olivier Antoine; OK mpi@ (bluhm@) ~ nd6.c > Since we have TAILQ_FOREACH_SAFE in nd6_purge() it is not necessary > that nd6_free() advances the next pointer itself. Make nd6_free() > a void function and remove unused parameter gc. > OK florian@ millert@ (bluhm@) == usr.sbin ========================================================== 07/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin dhcpd ~ dhcpd.h > Remove unused structs and defines, from Edgar Pettijohn. ok jca@ (rob@) pkg_add ~ OpenBSD/PackageRepository.pm ~ OpenBSD/PackageRepository/HTTP.pm > Adjust the html index-parsing regex to cope with anchor tags that aren't > just '<a href="foo">'; several mirrors now include a title attribute which > pkg_add can't cope with. > Diff from Taylor Stearns to handle attributes following the href; tweaked > by me to also handle attribute before the href. > ok zhuk@ (sthen@) ~ OpenBSD/PkgInfo.pm > Only call find_pkg when we are using other options with -Q. > This resolves the speed issues Michael Reed <[email protected]> > reported that were introduced by my changes in 1.45. > OK landry@ (abieber@) smtpd ~ smtpd.h > the PURGE_EVERYTHING flag used to purge config bits was inaccurate > ok eric@ (gilles@) =============================================================================== _______________________________________________ odc mailing list [email protected] http://www.squish.net/mailman/listinfo/odc
