OpenBSD src changes summary for 2016-11-16 ==========================================
distrib/sets gnu/usr.bin/texinfo lib/libpcap regress/sys regress/usr.sbin sys/dev/ic sys/dev/pci sys/dev/pv sys/dev/usb sys/net sys/netinet sys/netinet6 usr.bin/tmux usr.sbin/smtpd usr.sbin/switchd usr.sbin/syspatch usr.sbin/tcpdump usr.sbin/vmd == distrib =========================================================== 01/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/distrib sets ~ lists/base/mi > sync (deraadt@) == gnu =============================================================== 02/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/gnu usr.bin/texinfo ~ Makefile.bsd-wrapper > Set permissions of /usr/share/info/dir explicitly, so they don't depend on > the umask. > ok millert (tb@) == lib =============================================================== 03/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/lib libpcap ~ gencode.c ~ pcap.c ~ shlib_version > Add new DLT_OPENFLOW link-type to allow using tcpdump to debug switch(4), > eg. tcpdump -y openflow -i switch0 > Includes a minor bump for libpcap. > Feedback and OK rzalamena@ (reyk@) == regress =========================================================== 04/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/regress sys ~ net/pf_divert/Makefile ~ net/pf_divert/README > Test multiple combinations of divert connections. The socket cleanup > of the first connection must remove the pf state so that the second > connection can succeed. Disable two raw IP tests that do not work > with the current implementation in the kernel. (bluhm@) usr.sbin ~ switchd/OFP.pm ~ switchd/run.pl > Remove code duplication for receiving/sending OFP packets and simplify > packet-in building. > ok reyk@ (rzalamena@) ~ switchd/OFP.pm ~ switchd/args-packet-jumbo.pm ~ switchd/run.pl > Add support for big reads in perl and fix the jumbo test, also remove > message type specific code from OFP encode. > ok reyk@ (rzalamena@) == sys =============================================================== 05/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys dev/ic ~ re.c ~ rtl81x9reg.h > serialise posts to the txstart register. > without this it is possible for re_txeof to let the chip look at > and complete tx descriptors that re_start has written to the ring > but not accounted for in the software producer index. in this > situation no future interrupts may be generated for packets on the > ring, and tx will stall. > the serialisation is implemented by only doing the writes from tasks > via ifq_serialize. > this is part of a bigger diff that fixes lockups reported by haesbaert. > (dlg@) ~ re.c ~ rtl81x9reg.h > move toward having the a tx packet descriptor per tx ring entry. > previously you could have 64 packets on a ring with up to 1024 > descriptors. this mismatch makes the accounting for free space > complicated and allows for a data race with OACTIVE being set. > it will also let us push more packets. (dlg@) ~ re.c > dont restrict the number of tx descriptors to what can fit on a cpu page. > (dlg@) ~ re.c > allocate dmamaps for every descriptor in the tx ring. > previously there was a limit of 64 packets on the tx ring. soon we > will use as many descriptors as we want to fill the ring. (dlg@) ~ re.c > rework re_start and re_txeof to be more like other drivers. > the important change is to no longer account for packets and > descriptors separately, instead we just deal with available descriptors > on the ring. > this allows us to guarantee that oactive will only be set after > re_start has queued at least one packet, which in turn guarantees > that the chip will interrupt in the future and run re_txeof to clear > oactive. > without this change it was possible for a call of re_start to almost > fill the descriptor ring, followed by a second call that could fail > to post a packet because it had run out of descriptors and would > then set oactive. if re_txeof ran concurrently it could complete > all the previously posted packets without noticing that oactive had > been set and therefore needed to be cleared. in that situation > oactive would be set, but nothing would be set up to clear it > in the immediate future. > this fixes a lockup reported by haesbaert. (dlg@) dev/pci ~ ixgbe_type.h > Add new register definitions from the upstream; no binary change (mikeb@) ~ ixgbe_type.h > Add new structures and new fields; no functional change (mikeb@) + ixgbe_x550.c > Intel X550 specific module; not hooked up to the build yet > Obtained from the Intel ixgbe driver in FreeBSD. (mikeb@) ~ ixgbe.c ~ ixgbe_82598.c ~ ixgbe_82599.c ~ ixgbe_type.h ~ ixgbe_x540.c > Remove unused bits; no functional change (mikeb@) dev/pv ~ if_hvn.c > Always attempt to restart the transmission when OACTIVE is set > There's not much point in adding an additional condition to the > ifq_restart call so just get rid of those bits. Also while here, > remove a debugging message and make sure to cleanup the whole RNDIS > packet before sending it out to the hypervisor. (mikeb@) dev/usb ~ uaudio.c > Use memcpy() to copy structures into freshly malloc'ated buffers, > as source and destination never overlap in this case. From > Michael W. Bombardieri <mb at ii.net>. (ratchov@) ~ usbdevs > Add Sierra Wireless EM7455 LTE module. > from Bryan Vyhmeister, ok sthen@ (gerhard@) ~ usbdevs_data.h ~ usbdevs.h > regen (gerhard@) net ~ pf_if.c ~ pf_ioctl.c > Kill recursive splsoftnet()s. > ok bluhm@ (mpi@) ~ bpf.c > Allow bpf_allocbufs() to fail when allocating memory. > This will help trading the KERNEL_LOCK for a mutex. > ok bluhm@ (mpi@) ~ bpf.h ~ if_switch.c ~ if_switch.h ~ switchofp.c > Add new DLT_OPENFLOW link-type to allow using tcpdump to debug switch(4), > eg. tcpdump -y openflow -i switch0 > Includes a minor bump for libpcap. > Feedback and OK rzalamena@ (reyk@) ~ bpf.c > Use goto in bpf{read,write}() to ease review of locked sections. > While here properly account for used reference in bpfwrite(). > ok bluhm@ (mpi@) ~ if_ppp.c > Kill outdated comment talking about sofnet. (mpi@) ~ if_spppsubr.c > if_link_state_change() doesn't need to be called under splsoftnet(), > it just schedule a task. (mpi@) ~ if_gre.c > Convert gre_send_keepalive() to timeout_set_proc(9) as it calls > ip_output(). (mpi@) netinet ~ tcp_input.c ~ tcp_var.h > Kill recursive splsoftnet()s. > While here keep local definitions local. > ok bluhm@ (mpi@) ~ ip_icmp.c > Inherit route label when creating dynamic routes for path MTU. > From Rivo Nurges; OK claudio@ mpi@ phessler@ (bluhm@) ~ ip_icmp.c > Bring icmp6_mtudisc_clone() in line with icmp_mtudisc_clone(). The > IPv4 dynamic route inherits the priority. Only clone from a valid > IPv6 route. Do not use splsoftnet() in IPv6. Some stylistic changes > to make the functions similar. > OK mpi@ (bluhm@) ~ tcp_input.c > Be consistent and do not use braces for single line statements. > Prodded by and ok bluhm@ (mpi@) netinet6 ~ icmp6.c > Inherit route label when creating dynamic routes for path MTU. > From Rivo Nurges; OK claudio@ mpi@ phessler@ (bluhm@) ~ icmp6.c > Bring icmp6_mtudisc_clone() in line with icmp_mtudisc_clone(). The > IPv4 dynamic route inherits the priority. Only clone from a valid > IPv6 route. Do not use splsoftnet() in IPv6. Some stylistic changes > to make the functions similar. > OK mpi@ (bluhm@) == usr.bin =========================================================== 06/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin tmux ~ cmd-find.c ~ cmd-select-pane.c ~ tmux.h ~ window.c > The target validity check used window_pane_visible but that may be false > if the pane is zoomed, so instead add a new function to just check if > the pane is actually on screen (most commands still want to accept panes > invisible by zoom). Also reject panes outside the window for various > special targets. Problem reported by Sean Haugh. (nicm@) ~ window.c > Initialize modeprefix to 1 properly in window_pane_create, from Artem > Fokin. (nicm@) ~ cmd-new-session.c ~ server-client.c > Do not stop dragging when the wheel is pressed, from Artem Fokin. (nicm@) ~ cmd-rename-session.c > Don't error if renaming a session to its current name, from shik dot > chen at gmail dot com. (nicm@) ~ tty.c > Fix calculation of whether we need a region for drawing a cell (only if > full width and at the edge of the region), otherwise clear the region > entirely. (nicm@) == usr.sbin ========================================================== 07/07 == http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.sbin smtpd ~ bounce.c ~ filter.c ~ ioev.c ~ ioev.h ~ mda.c ~ mta_session.c ~ smtp_session.c > pass the user pointer as parameter to the io callback instead of having > the user dereference the io structure. > ok millert@ gilles@ (eric@) switchd ~ proc.c ~ proc.h > Sync switchd(8) and vmd(8) with the latest proc.c changes. This sync > basically imports the imsg_flush() fix from reyk@ in httpd(8). > ok reyk@ (rzalamena@) syspatch ~ syspatch.8 ~ syspatch.sh > /bsd.rollbackXX -> bsd.syspatchXX to make it more obvious where it comes > from > and less prone to conflict. > Simplify ls_installed while here. (ajacoutot@) ~ syspatch.sh > If we install a new kernel, add /bsd twice (for size checking) when: > - we are on an MP system (/bsd.mp does not exist there) > - /bsd.syspatchXX is not present (create_rollback will add it) (ajacoutot@) ~ syspatch.sh > Merge comments altogether, it's easier to understand this way. > And drop some duplicate while here. (ajacoutot@) tcpdump ~ interface.h ~ print-ofp.c ~ print-tcp.c ~ tcpdump.c > Add new DLT_OPENFLOW link-type to allow using tcpdump to debug switch(4), > eg. tcpdump -y openflow -i switch0 > Includes a minor bump for libpcap. > Feedback and OK rzalamena@ (reyk@) vmd ~ proc.c ~ proc.h > Sync switchd(8) and vmd(8) with the latest proc.c changes. This sync > basically imports the imsg_flush() fix from reyk@ in httpd(8). > ok reyk@ (rzalamena@) =============================================================================== _______________________________________________ odc mailing list [email protected] http://www.squish.net/mailman/listinfo/odc
