OpenBSD src changes summary for 2016-12-19 ==========================================
distrib/miniroot gnu gnu/lib/libreadline lib/csu lib/libcrypto lib/libcxx lib/libcxxabi regress/usr.bin sys/arch/amd64/amd64 sys/arch/arm64/arm64 sys/arch/arm64/conf sys/dev/pv sys/kern sys/miscfs/fifofs sys/net sys/netinet sys/netinet6 sys/nfs sys/sys usr.bin/ssh usr.sbin/bgpd == distrib =========================================================== 01/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib miniroot ~ install.sub > - no need to export the TERM variable twice > - re-format case-block > - unquote single word default answer > - use 'break' instead of return to leave (same effect here, but > needed later when the function is no function anymore) > - use -q option with kbd to limit output to warnings/errors only > discussed with and OK krw@ tb@ (rpe@) ~ install.sub > Revert r1.934 to unbreak autoinstall (rpe@) == gnu =============================================================== 02/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/gnu gnu ~ gcc/gcc/config/m88k/m88k.md > Convert the gcc 3 scheduling information to the gcc 4 model. (miod@) lib/libreadline ~ histfile.c > A couple more unused variables. > tweak & ok kettenis@ (krw@) == lib =============================================================== 03/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib csu ~ boot.h > A couple more unused variables. > tweak & ok kettenis@ (krw@) libcrypto ~ man/RSA_generate_key.3 > fix typo, from OpenSSL > commit 0b742f93ea7882a447f6523ac56a6f847d9f8e92 > Author: Finn Hakansson <[email protected]> > Date: Thu Dec 15 12:58:19 2016 -0500 (schwarze@) libcxx ~ Makefile > Compile libc++ and libc++abi with clang by default instead of egcc. > This allows our build system to selfhost itself after the initial > clang bootstrap without relying on an external compiler. > with input from and ok kettenis@ (patrick@) libcxxabi ~ Makefile > Compile libc++ and libc++abi with clang by default instead of egcc. > This allows our build system to selfhost itself after the initial > clang bootstrap without relying on an external compiler. > with input from and ok kettenis@ (patrick@) == regress =========================================================== 04/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress usr.bin ~ ssh/allow-deny-users.sh > Use LOGNAME to get current user and fall back to whoami if not set. > Mainly to benefit -portable since some platforms don't have whoami. > (dtucker@) ~ ssh/unittests/utf8/tests.c > remove testcase that depends on exact output and behaviour of > snprintf(..., "%s", NULL) (djm@) ~ ssh/allow-deny-users.sh > use standard /bin/sh equality test; from Mike Frysinger (djm@) == sys =============================================================== 05/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys arch/amd64/amd64 ~ acpi_wakecode.S ~ mptramp.S > Generating mixed 16-bit/32-bit/64-bit code with clang's integrated > assembler is a bit tricky. It supports the .code16, .code32 and > .code64 directives. But it doesn't know about the data16/data32 and > addr16/addr32 instruction prefixes. Instead it tries to determine > those from the instruction opcode. It mostly succeeds, but there are > a couple of corner cases where clang will generate the "addr32" form > where gas generates the "addr16" form in .code16 segments. That > should be no problem (and just waste a couple of bytes), but it makes > comparing the generated code a bit difficult. > Allow the trampoline code to be compiled with both. For clang #define > away the addr32 prefix and avoid using the data32 prefix by using a > mnemonic that explicitly encodes the size of the operand. Add a few > addr32 prefixes in .code16 blocks to reduce the differences between > code generated by clang and gas. > ok patrick@, deraadt@, mlarkin@ (kettenis@) arch/arm64/arm64 ~ sig_machdep.c > Fix an off by one when saving and restoring an array of registers. > ok patrick@ (jsg@) ~ db_trace.c > ansify function declaration; better whitespace (patrick@) ~ pmap.c > remove redundant return statements > ok patrick@ (jsg@) ~ db_trace.c ~ machdep.c > fix uninitialised variable warnings from clang > ok patrick@ (jsg@) ~ pmap.c > Fix debug printf format and arguments. > Found by and input from jsg@ (patrick@) arch/arm64/conf ~ Makefile.arm64 > Adjust Makefile to gain improvements that were made in other archs > already. This means for example dropping -Wno-format or adding -g > by default. > Prompted by jsg@ (patrick@) dev/pv ~ xen.c ~ xenvar.h > Add experimental support for device hot-plugging > We're installing watches on all nodes under "device/" and re-scan > the subtree every time the watch is triggered looking for changes > in the output. Tested with xnf(4) and xbf(4), helpful hints from > Roger Pau Monne, thanks! (mikeb@) ~ if_xnf.c > Implement interface detaching (mikeb@) kern ~ sys_socket.c ~ uipc_socket.c ~ uipc_socket2.c ~ uipc_syscalls.c ~ uipc_usrreq.c > Introduce the NET_LOCK() a rwlock used to serialize accesses to the parts > of the network stack that are not yet ready to be executed in parallel or > where new sleeping points are not possible. > This first pass replace all the entry points leading to ip_output(). This > is done to not introduce new sleeping points when trying to acquire ART's > write lock, needed when a new L2 entry is created via the RT_RESOLVE. > Inputs from and ok bluhm@, ok dlg@ (mpi@) miscfs/fifofs ~ fifo_vnops.c > Introduce the NET_LOCK() a rwlock used to serialize accesses to the parts > of the network stack that are not yet ready to be executed in parallel or > where new sleeping points are not possible. > This first pass replace all the entry points leading to ip_output(). This > is done to not introduce new sleeping points when trying to acquire ART's > write lock, needed when a new L2 entry is created via the RT_RESOLVE. > Inputs from and ok bluhm@, ok dlg@ (mpi@) net ~ if.c ~ if_spppsubr.c ~ raw_usrreq.c ~ route.c ~ rtsock.c > Introduce the NET_LOCK() a rwlock used to serialize accesses to the parts > of the network stack that are not yet ready to be executed in parallel or > where new sleeping points are not possible. > This first pass replace all the entry points leading to ip_output(). This > is done to not introduce new sleeping points when trying to acquire ART's > write lock, needed when a new L2 entry is created via the RT_RESOLVE. > Inputs from and ok bluhm@, ok dlg@ (mpi@) ~ if_pfsync.c > Timer sending packets need to grab the NET_LOCK(). > ok bluhm@ (mpi@) ~ if_bridge.c > Assert that IPL_SOFTNET is needed rather than calling splsoftnet() > recursively. > ok bluhm@ (mpi@) ~ if_gre.c > Sending keep alive ends up in ip_output(), so it needs the NET_LOCK(). > ok bluhm@ (mpi@) ~ pipex.c > Kill useless comment about splsoftnet() which is going away. (mpi@) ~ ppp_tty.c > Stop mentioning splsoftnet() in comments, it's almost history. (mpi@) netinet ~ if_ether.c ~ ip_carp.c ~ ip_divert.c ~ ip_icmp.c ~ ip_input.c ~ ip_ipsp.c ~ ip_output.c ~ raw_ip.c ~ tcp_input.c ~ tcp_timer.c ~ tcp_usrreq.c ~ udp_usrreq.c > Introduce the NET_LOCK() a rwlock used to serialize accesses to the parts > of the network stack that are not yet ready to be executed in parallel or > where new sleeping points are not possible. > This first pass replace all the entry points leading to ip_output(). This > is done to not introduce new sleeping points when trying to acquire ART's > write lock, needed when a new L2 entry is created via the RT_RESOLVE. > Inputs from and ok bluhm@, ok dlg@ (mpi@) ~ igmp.c ~ ip_input.c ~ ip_mroute.c ~ ip_mroute.h ~ ip_output.c ~ ip_var.h ~ raw_ip.c > Extend the multicast sockets and multicast hash table support to multiple > domains. This is one step towards supporting to run more than one multicast > socket in different domains at the same time. > ok mpi@ (rzalamena@) ~ in.c > If rt_ifa_addlocal() in in_ifinit() fails, the address has been > added to the interface address list, but the local route is missing. > This inconsistency could result in a "ifa == rt->rt_ifa" assertion > panic later. So in case of a route add error, remove the interface > address to get a consistent state again. > OK stsp@ mpi@ (bluhm@) ~ in.c > Remove a recursive splsoftnet() in in_ioctl() that already asserts that > it is called at IPL_SOFTNET. (mpi@) ~ ip_mroute.c ~ ip_mroute.h > Kill unused function. > ok mpi@ (rzalamena@) ~ udp_usrreq.c > Remove redundant splsoftnet(). > ok bluhm@ (mpi@) netinet6 ~ icmp6.c ~ ip6_divert.c ~ ip6_input.c ~ nd6.c ~ raw_ip6.c > Introduce the NET_LOCK() a rwlock used to serialize accesses to the parts > of the network stack that are not yet ready to be executed in parallel or > where new sleeping points are not possible. > This first pass replace all the entry points leading to ip_output(). This > is done to not introduce new sleeping points when trying to acquire ART's > write lock, needed when a new L2 entry is created via the RT_RESOLVE. > Inputs from and ok bluhm@, ok dlg@ (mpi@) ~ udp6_output.c > Remove redundant splsoftnet(). > ok bluhm@ (mpi@) nfs ~ nfs_boot.c ~ nfs_socket.c > Introduce the NET_LOCK() a rwlock used to serialize accesses to the parts > of the network stack that are not yet ready to be executed in parallel or > where new sleeping points are not possible. > This first pass replace all the entry points leading to ip_output(). This > is done to not introduce new sleeping points when trying to acquire ART's > write lock, needed when a new L2 entry is created via the RT_RESOLVE. > Inputs from and ok bluhm@, ok dlg@ (mpi@) sys ~ socketvar.h ~ systm.h > Introduce the NET_LOCK() a rwlock used to serialize accesses to the parts > of the network stack that are not yet ready to be executed in parallel or > where new sleeping points are not possible. > This first pass replace all the entry points leading to ip_output(). This > is done to not introduce new sleeping points when trying to acquire ART's > write lock, needed when a new L2 entry is created via the RT_RESOLVE. > Inputs from and ok bluhm@, ok dlg@ (mpi@) == usr.bin =========================================================== 06/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin ssh ~ version.h > openssh-7.4 (djm@) == usr.sbin ========================================================== 07/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin bgpd ~ session.c > In parse_header() not all of the bgp message may be in the buffer yet so > move the MRT msg dump to session_process_msg() after the point the full > message is in the read buffer. > Bug found and fix tested by Ian Bobbitt (claudio@) =============================================================================== _______________________________________________ odc mailing list [email protected] http://www.squish.net/mailman/listinfo/odc
