[owc] Weekly src changes ending 2008-03-16
OWC
auto at squish.net
Mon Mar 17 07:00:01 GMT 2008
OpenBSD src changes summary for 2008-03-09 to 2008-03-16 inclusive
==================================================================
bin/df distrib/miniroot
distrib/notes gnu/usr.bin/binutils
gnu/usr.bin/perl include/stdlib.h
kerberosV lib/libc
lib/libpthread libexec/ftpd
libexec/login_skey libexec/login_tis
libexec/tftp-proxy libexec/tftpd
regress/sys regress/usr.bin
sbin/dhclient sbin/isakmpd
sbin/mount_portal sbin/nfsd
sbin/pflogd sbin/ping6
sbin/sysctl share/man
sys/arch/alpha/stand/boot sys/arch/alpha/stand/netboot
sys/arch/amd64/conf sys/arch/amd64/pci
sys/arch/hppa/conf sys/arch/i386/conf
sys/arch/i386/pci sys/arch/mac68k/mac68k
sys/arch/mvme68k/stand/libsa sys/arch/sparc/dev
sys/arch/sparc/sparc sys/arch/sparc64/conf
sys/arch/sparc64/dev sys/arch/sparc64/include
sys/arch/sparc64/sparc64 sys/compat/common
sys/compat/osf1 sys/conf
sys/dev sys/dev/ic
sys/dev/mii sys/dev/pci
sys/dev/sbus sys/dev/usb
sys/kern sys/net
sys/netinet sys/netinet6
sys/nfs sys/sys
sys/ufs/ffs usr.bin/audioctl
usr.bin/column usr.bin/cvs
usr.bin/ftp usr.bin/mixerctl
usr.bin/ssh usr.bin/vmstat
usr.sbin/arp usr.sbin/bgpd
usr.sbin/bind usr.sbin/ifstated
usr.sbin/ospf6d usr.sbin/ospfd
usr.sbin/pcidump usr.sbin/pkg_add
usr.sbin/ppp usr.sbin/relayd
usr.sbin/ripd usr.sbin/route6d
usr.sbin/sensorsd usr.sbin/snmpd
usr.sbin/syslogd usr.sbin/tcpdump
usr.sbin/traceroute6
== bin =============================================================== 01/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/bin
df
~ df.c
> actually print the now available large numbers (otto@)
== distrib =========================================================== 02/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/distrib
miniroot
~ install.sub
> Consolidate FTP option handling, fixing one missing set of options
> and adding a new question during install to allow the setting of
> FTP's -k option in situations where intervening routers/firewalls
> are prematurely killing connections.
> Original idea and diff from espie@
> ok espie@ (krw@)
~ install.sub
> Remove question about keep alive option.
> Requested by deraadt@ (krw@)
~ install.sub
> When installing via NFS, only retry the mount five times (instead of the
> default 10,000 times).
> ok deraadt@ krw@ thib@ (jsing@)
notes
~ alpha/hardware ~ hppa/hardware
~ macppc/hardware ~ sparc64/hardware
~ zaurus/hardware
> +Atlantis Land A02-UP1-W54 (jmc@)
~ alpha/install
> Tweak dhcpd configuration to please old DEC 3000 SRM; reported by
> Maurice Janssen, thanks! (miod@)
~ m4.common ~ alpha/contents
~ amd64/contents ~ armish/contents
~ aviion/contents ~ hp300/contents
~ hppa/contents ~ i386/contents
~ landisk/contents ~ luna88k/contents
~ mac68k/contents ~ macppc/contents
~ mvme68k/contents ~ mvme88k/contents
~ sgi/contents ~ sparc/contents
~ sparc64/contents ~ vax/contents
~ zaurus/contents
> Sync sets sizes with release bits. (miod@)
~ armish/prep ~ armish/whatis
> Layout tweaks. (miod@)
~ amd64/install ~ armish/install
~ hp300/install ~ hppa/install
~ i386/install ~ landisk/install
~ luna88k/install ~ mac68k/install
~ macppc/install ~ mvme68k/install
~ mvme88k/install ~ sgi/install
~ sparc/install ~ sparc64/install
~ vax/install ~ zaurus/install
> Suggest vt220 as a TERM value for serial-console-in-an-xterm setup
> everywhere
> instead of a mix of vt100 and vt220. (miod@)
~ hppa/hardware
> Mention 16MB of memory is necessary now. (miod@)
~ i386/contents ~ i386/hardware
~ i386/whatis
> No more 80386 support, and you really don't want to try and install the
> system unless you have over 16MB of RAM. Here's a nickel kid, get yourself
> a
> better computer. (miod@)
~ mvme88k/whatis ~ sgi/upgrade
~ sparc64/hardware ~ sparc64/prep
~ vax/hardware ~ zaurus/hardware
> Minor tweaks, including behind-the-scenes fixes. (miod@)
== gnu =============================================================== 03/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/gnu
usr.bin/binutils
~ ld/emulparams/elf64btsmip_obsd.sh ~ ld/emulparams/elf64ltsmip_obsd.sh
~ ld/emulparams/elf_obsd.sh ~ ld/emulparams/hppaobsd.sh
> Add __data_start symbol to all ELF archs to consistently mark the beginning
> of the writable sections. This is useful for garbage collectors such as
> boehm-gc to locate these sections in a uniform way.
> input and okay drahn@, miod@ (kurt@)
usr.bin/perl
~ perl.c
> This should set msg_control using CMSG_SIZE(), and not assume that
> sizeof(buffer) is the right size (alignments can mess it up). Code is
> apparently not used, but it is better if all examples show the correct
> idiom
> ok hshoexer (deraadt@)
== include =========================================================== 04/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/include
stdlib.h
~ stdlib.h
> diff from djm@ committed at his request:
> introduce two new APIs for requesting strong random numbers:
> arc4random_buf() - fill an arbitrary memory range with random numbers
> arc4random_uniform() - return a uniformly distributed random number
> below
> a specified upper bound, avoiding the bias that comes from a naive
> "arc4random() % upper_bound" construction.
> these mirror similarly-named functions in the kernel;
> lots of discussion deraadt@ mcbride@ (otto@)
== kerberosV ========================================================= 05/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/kerberosV
kerberosV
~ src/kcm/connect.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ src/kcm/connect.c
> seperate out space vs len calculations; ok hshoexer (deraadt@)
== lib =============================================================== 06/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/lib
libc
~ gen/auth_subr.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ gen/auth_subr.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
~ string/memset.c
> Convert c to unsigned char, like it says in the manual. Also add
> cast to make it explicit.
> Found by lint, OK millert. (ray@)
~ string/swab.c
> - len is size_t, but n uses len and is an int. Matching those types
> should be good, plus it prevents weird things from happening if
> len > INT_MAX.
> - Since n is now size_t, compare it against 0 instead of >= 0.
> - temp is used to store individual bytes, so use char instead
> (matches fp and tp).
> - millert noted that the comma operator may not guarantee order of
> execution, so replace with semicolons.
> Found by lint, OK millert. (ray@)
~ sys/Makefile.inc + sys/statvfs.c
> statvfs support (otto@)
~ crypt/arc4random.3 ~ crypt/arc4random.c
> diff from djm@ committed at his request:
> introduce two new APIs for requesting strong random numbers:
> arc4random_buf() - fill an arbitrary memory range with random numbers
> arc4random_uniform() - return a uniformly distributed random number
> below
> a specified upper bound, avoiding the bias that comes from a naive
> "arc4random() % upper_bound" construction.
> these mirror similarly-named functions in the kernel;
> lots of discussion deraadt@ mcbride@ (otto@)
~ sys/statvfs.c
> indent (deraadt@)
~ gen/sysconf.3 ~ gen/sysconf.c
> Add the semi-standard _SC_PHYS_PAGES and _SC_AVPHYS_PAGES, sysconf(3)
> variable.
> ok espie@ (kettenis@)
~ shlib_version
> bump (otto@)
~ sys/statfs.2
> sync with sys/statfs.h; ok jmc@ (otto@)
~ crypt/Makefile.inc
> Add missing MLINKS (deraadt@)
~ crypt/arc4random.3
> - add NAME entries for arc4random_buf and arc4random_uniform
> - simplify a macro call (Do/Dc -> Dq) (jmc@)
libpthread
~ shlib_version
> bump (otto@)
== libexec =========================================================== 07/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/libexec
ftpd
~ monitor_fdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ monitor_fdpass.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
login_skey
~ login_skey.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ login_skey.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
login_tis
~ login_tis.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ login_tis.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
tftp-proxy
~ tftp-proxy.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ tftp-proxy.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
tftpd
~ tftpd.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ tftpd.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
== regress =========================================================== 08/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/regress
sys
~ kern/unfdpass/unfdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ kern/unfdpass/expected ~ kern/unfdpass/unfdpass.c
> Repair usage of CMSG_SPACE and CMSG_LEN. While there, send three fds
> instead of just two as this decreases the propability that things just
> work although the sizes are wrong (ie. 8 aligns correctly on both 32 and 64
> bit
> platforms even with wrong usage of CMSG_{LEN,SPACE} whereas 12 doesn't).
> (hshoexer@)
usr.bin
~ cvs/Makefile + cvs/setdate.pl
+ cvs/test_cvs_import_01_seed1.txt,v
+ cvs/test_opencvs_import_01_seed1.txt,v
> Further improve regress for OpenCVS by including possibility to change and
> check dates in RCS files (needed for various -D cases). As a reference,
> import has been adjusted.
> Also, it is no longer needed to supply -f option if CVS is specified
> with make.
> ok joris (tobias@)
== sbin ============================================================== 09/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sbin
dhclient
~ dhclient.c
> Code was not passing script exit status; fix with macro.
> deraadt@ OK (hugh@)
isakmpd
~ monitor_fdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ monitor_fdpass.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
mount_portal
~ activate.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ activate.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
nfsd
~ nfsd.c
> make sure we start 4 servers on udp if called with
> no arguments, like the man page promises.
> ok deraadt@, blambert@
> From tcuji at pwns.ms (thib@)
pflogd
~ privsep_fdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ privsep_fdpass.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
ping6
~ ping6.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ ping6.c
> Fix several CMSG-related bugs. Chaining CMSG's is tricky since you need to
> use CMSG_SPACE for all except the last one, which should be CMSG_LEN. This
> makes calculate .msg_controllen tricky. Also make sure that we can receive
> multiple CMSG's again.
> ok deraadt@, hshoexer@ (kettenis@)
sysctl
~ sysctl.8 ~ sysctl.c
> "-Aa" are not required flags; sort synopses; improve description of "-a"
> written with invaluable advice from jmc@
> ok jmc@ (sobrado@)
== share ============================================================= 10/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/share
man
~ man9/ieee80211_radiotap.9
> DLT_IEEE_80211_RADIO -> DLT_IEEE802_11_RADIO;
> from Sam Banks, freebsd docs/121477 (jmc@)
~ man4/man4.sparc64/vbus.4 ~ man4/man4.sparc64/intro.4
> standard format for vbus.4, and bring it into intro.4; (jmc@)
~ man4/man4.sparc64/vbus.4 ~ man4/man4.sparc64/vcons.4
~ man4/man4.sparc64/vrtc.4
> - kill some unneeded .Pp
> - tweak SEE ALSO for vcons.4 (jmc@)
~ man4/man4.sparc64/Makefile + man4/man4.sparc64/vpci.4
> vpci(4) (kettenis@)
~ man4/pci.4
> add vpci; ok kettenis (jmc@)
~ man4/ugen.4
> source code should fit on a 80-column display; add some paragraph breaks
> ok jmc@ (sobrado@)
~ man4/umidi.4
> suggest tweaking the "usb descriptor" switch (if any) if the device
> attaches
> as ugen(4) rather than umidi(4).
> from jmc@ and jakemsr@ (ratchov@)
~ man3/CMSG_DATA.3
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ man4/man4.sparc64/Makefile + man4/man4.sparc64/prtc.4
> prtc(4) (kettenis@)
~ man4/ip.4
> IP_MAX_MEMBERSHIPS was bumped to 4095 in -r1.74; from Matthew Dempsky
> (jmc@)
~ man4/cdce.4
> xref cdcef(4). ok mbalmer, jmc (sthen@)
~ man4/man4.sparc/tvtwo.4 ~ man4/man4.sparc64/tvtwo.4
> Document video modes. (miod@)
~ man3/CMSG_DATA.3
> show the correct idiom (deraadt@)
~ man4/man4.sparc/tvtwo.4 ~ man4/man4.sparc64/tvtwo.4
> tweak previous; (jmc@)
~ man4/agp.4
> agp now attaches at vga. (oga@)
~ man4/agp.4 ~ man4/vga.4
> some more agp/vga bits; (jmc@)
== sys =============================================================== 11/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sys
arch/alpha/stand/boot
~ newvers.sh
> use the right capitalization for platform names
> ok jsing@ (sobrado@)
arch/alpha/stand/netboot
~ if_prom.c ~ newvers.sh
> use the right capitalization for platform names
> ok jsing@ (sobrado@)
arch/amd64/conf
~ GENERIC ~ files.amd64
> Make agp(4) attach at vga(4) instead of pchb(4). This is because sometimes
> agp and drm need to use the same memory mapping, the best way to deal
> with that is to allow them to share (that's coming later), for this to
> work cleanly we move the attach point of agp.
> Ideally most agp drivers would attach at pchb, with only agp_i810
> (and any that work similarly) attaching at vga, but this will do for
> now.
> ok kettenis@, miod at . (oga@)
arch/amd64/pci
~ pchb.c
> Make agp(4) attach at vga(4) instead of pchb(4). This is because sometimes
> agp and drm need to use the same memory mapping, the best way to deal
> with that is to allow them to share (that's coming later), for this to
> work cleanly we move the attach point of agp.
> Ideally most agp drivers would attach at pchb, with only agp_i810
> (and any that work similarly) attaching at vga, but this will do for
> now.
> ok kettenis@, miod at . (oga@)
arch/hppa/conf
~ RAMDISK
> Uncomment ehci(4). (kettenis@)
arch/i386/conf
~ GENERIC ~ files.i386
> Make agp(4) attach at vga(4) instead of pchb(4). This is because sometimes
> agp and drm need to use the same memory mapping, the best way to deal
> with that is to allow them to share (that's coming later), for this to
> work cleanly we move the attach point of agp.
> Ideally most agp drivers would attach at pchb, with only agp_i810
> (and any that work similarly) attaching at vga, but this will do for
> now.
> ok kettenis@, miod at . (oga@)
~ files.i386
> for some reason the pchb dependency on agp got left in. kill it. (oga@)
arch/i386/pci
~ pchb.c
> Make agp(4) attach at vga(4) instead of pchb(4). This is because sometimes
> agp and drm need to use the same memory mapping, the best way to deal
> with that is to allow them to share (that's coming later), for this to
> work cleanly we move the attach point of agp.
> Ideally most agp drivers would attach at pchb, with only agp_i810
> (and any that work similarly) attaching at vga, but this will do for
> now.
> ok kettenis@, miod at . (oga@)
arch/mac68k/mac68k
~ dpme.h
> use the right capitalization for platform names
> ok jsing@ (sobrado@)
arch/mvme68k/stand/libsa
~ Makefile
> use muldi3.c to avoid putting instructions 68060 does not support into the
> boot media; from miod, tested by martin, this is just making release.
> (deraadt@)
arch/sparc/dev
~ tvtwo.c
> Support the old-prom XVideo boards by rewriting the sbus range registers;
> tested by Thorsten Glaser <tg at mirbsd.de> on sparc; while there, pretty and
> shrink dmesg output. (miod@)
arch/sparc/sparc
~ locore.s
> After 15 years of fun, fix Torek's ovbcopy() operation when copying shorts
> backwards. (miod@)
arch/sparc64/conf
~ GENERIC
> Allow clock to attach to any fhc?, not just fhc0. There might be backup
> clocks on other boards that we want to use if the one on fhc0 failed.
> tested by miod@ (kettenis@)
~ files.sparc64
> Add hvcall.S if option SUN4V is defined. (kettenis@)
~ files.sparc64
> The IOMMU code is needed for schizo(4) and pyro(4) too. (kettenis@)
~ files.sparc64
> Add files for vpci(4). (kettenis@)
~ files.sparc64
> Add prtc(4). (kettenis@)
~ GENERIC ~ RAMDISK
> Add prtc(4). (kettenis@)
~ files.sparc64
> Add some e10k support code. (kettenis@)
arch/sparc64/dev
~ fhc.c
> Check "status" property. (kettenis@)
~ sbus.c
> Check "status" property. (kettenis@)
+ vbusvar.h
> Virtual bus for virtual devices on sun4v. (kettenis@)
+ viommu.c + viommuvar.h
> sun4v hypervisor IOMMU code. (kettenis@)
+ vpci.c
> Driver for the virtual PCI host bridge on sun4v. (kettenis@)
~ iommuvar.h
> Add some members needed by the sunv IOMMU code. (kettenis@)
+ prtc.c
> First shot at a driver to get the time-of-day on the e10k. (kettenis@)
+ starfire.c + starfire.h
> Add some e10k support code. (kettenis@)
~ sbus.c ~ psycho.c
> Set up interrupt translation for e10k. (kettenis@)
arch/sparc64/include
~ conf.h
> Add vcons(4). (kettenis@)
~ intr.h
> Introduce a per-handler interrupt acknowledgement function. (kettenis@)
~ cpu.h
> Make GENERIC.MP work on the e10k. The e10k is a bit funky since UPA only
> supports 32 ports, and a machine with up to 64 CPUs obviously needs more.
> So the machine has a special ASIC that does port translation, and because
> of that we need to distinguish between port ID's and interrupt target ID's.
> (kettenis@)
arch/sparc64/sparc64
~ conf.c
> Add vcons(4). (kettenis@)
~ intr.c ~ locore.s
~ genassym.cf
> Introduce a per-handler interrupt acknowledgement function. (kettenis@)
~ locore.s
> Remove code to set mmu context to 0. It should already be set to 0 at that
> point; this code is probably a leftover from some code that tried to find
> out the number of available context dynamically.
> tested by ckuethe@, jsg@, sthen@ (kettenis@)
~ autoconf.c
> Attach prtc(4) if no (hardware) real-time clock was found. (kettenis@)
~ db_interface.c
> Make this catch up with some changed structs. Fix indentation while I'm
> there. (kettenis@)
~ autoconf.c
> Don't even try to attach pcons(4) if we have a real console. (kettenis@)
~ cpu.c ~ ipifuncs.c
~ pmap.c ~ vm_machdep.c
> Make GENERIC.MP work on the e10k. The e10k is a bit funky since UPA only
> supports 32 ports, and a machine with up to 64 CPUs obviously needs more.
> So the machine has a special ASIC that does port translation, and because
> of that we need to distinguish between port ID's and interrupt target ID's.
> (kettenis@)
~ cpu.c
> Include <sparc64/dev/starfire.h> instead of having our own definitions.
> (kettenis@)
compat/common
+ vfs_syscalls_o43.c
> Widen some struct statfs fields to support large filesystem stata
> and add some to be able to support statvfs(2). Do the compat dance
> to provide backward compatibility. ok thib@ miod@ (otto@)
compat/osf1
~ README.syscalls
> use the right capitalization for platform names
> ok jsing@ (sobrado@)
conf
~ files
> Make agp(4) attach at vga(4) instead of pchb(4). This is because sometimes
> agp and drm need to use the same memory mapping, the best way to deal
> with that is to allow them to share (that's coming later), for this to
> work cleanly we move the attach point of agp.
> Ideally most agp drivers would attach at pchb, with only agp_i810
> (and any that work similarly) attaching at vga, but this will do for
> now.
> ok kettenis@, miod at . (oga@)
~ GENERIC ~ files
> Widen some struct statfs fields to support large filesystem stata
> and add some to be able to support statvfs(2). Do the compat dance
> to provide backward compatibility. ok thib@ miod@ (otto@)
dev
~ audio.c
> Set the default high water mark to the buffer size, instead of the buffer
> size minus one block. Otherwise, recording cannot work with 32kB blocks.
> ok jakemsr@ (ratchov@)
dev/ic
~ re.c ~ rtl81x9reg.h
> Set the RX FIFO threshold to no RX threshold for re(4) adapters.
> This resolves a number of various bad symptoms experienced by
> numerous users especially with the adapters at Gig speed.
> Tested by quite a few users.
> From FreeBSD/NetBSD
> ok deraadt@ (brad@)
~ re.c
> Set the RX FIFO threshold to no RX threshold for re(4) adapters.
> This resolves a number of various bad symptoms experienced by
> numerous users especially with the adapters at Gig speed.
> Tested by quite a few users.
> From FreeBSD/NetBSD
> ok deraadt@ (brad@)
~ acx.c ~ atw.c
~ pgt.c
> Do not increment the input packet counter in these drivers as
> ieee80211_input() already does this.
> Mentioned by Peter Philipp in PR 5248.
> ok claudio@ mglocker@ (brad@)
~ vga.c
> Make agp(4) attach at vga(4) instead of pchb(4). This is because sometimes
> agp and drm need to use the same memory mapping, the best way to deal
> with that is to allow them to share (that's coming later), for this to
> work cleanly we move the attach point of agp.
> Ideally most agp drivers would attach at pchb, with only agp_i810
> (and any that work similarly) attaching at vga, but this will do for
> now.
> ok kettenis@, miod at . (oga@)
dev/mii
~ amphy.c ~ gentbi.c
~ nsphy.c ~ nsphyter.c
> Fix comment typo, of -> if.
> ok sthen@ (brad@)
dev/pci
~ emuxki.c
> the Dell SoundBlaster Live! 5.1 cards differ enough from "regular"
> SB Live!/5.1/Audigy/2/4 cards that both OSS and ALSA provide
> separate modules from their emu10k1 drivers to support them.
> and in PR 5143, we see that emu(4) does not support these Dell
> SBLive! cards, so don't attach emu(4) to them.
> ok ratchov, brad (jakemsr@)
~ agp.c ~ agpvar.h
~ files.agp ~ vga_pci.c
> Make agp(4) attach at vga(4) instead of pchb(4). This is because sometimes
> agp and drm need to use the same memory mapping, the best way to deal
> with that is to allow them to share (that's coming later), for this to
> work cleanly we move the attach point of agp.
> Ideally most agp drivers would attach at pchb, with only agp_i810
> (and any that work similarly) attaching at vga, but this will do for
> now.
> ok kettenis@, miod at . (oga@)
dev/sbus
~ tvtwo.c
> Support the old-prom XVideo boards by rewriting the sbus range registers;
> tested by Thorsten Glaser <tg at mirbsd.de> on sparc; while there, pretty and
> shrink dmesg output. (miod@)
dev/usb
~ if_cdce.c
> when attaching a cdce(4) device that does not have a proper mac address,
> use the tv_usec part of a getmicrotime(9) call instead of 'int ticks' to
> create a mac address. as ticks is 0 during boot, this resulted in both
> ends
> of a USB link cable to have the same mac address.
> problem found/diff tested winiger, feedback/ok deraadt (mbalmer@)
~ if_cdce.c
> Make sure the most often changing bits of tv_usec are used for lladdr
> generation.
> ok brad (mbalmer@)
kern
~ kern_clock.c
> In statclock(), fix local index type in the profiling code. Avoids
> out-of-bounds array accesses on some platforms. (miod@)
~ syscalls.master ~ vfs_bio.c
~ vfs_subr.c
> Widen some struct statfs fields to support large filesystem stata
> and add some to be able to support statvfs(2). Do the compat dance
> to provide backward compatibility. ok thib@ miod@ (otto@)
~ init_sysent.c ~ syscalls.c
~ syscalls.conf
> regen (otto@)
net
~ if_pppoe.c
> Include relay-sid in the response if one was received.
> Reported and tested by Zach Wilkinson
> ok brad@, claudio@ (canacar@)
~ if_pppoe.c
> Do not allow pppoe(4) interfaces to be attached to anything but Ethernet
> or VLAN interfaces.
> Based on a similar change from NetBSD.
> ok canacar@ (brad@)
netinet
~ ip_id.c
> Because the ip_id code initialisation is a specific case of shuffling
> a set of incrementing integers (and not an arbitrary set of values) it
> is possible to populate the array as we shuffle it in a single forward
> pass. Clever optimisation from didickman AT gmail.com;
> ok deraadt@ mcbride@ (djm@)
~ ip_id.c
> off by one at end of array (djm@)
~ ip_id.c
> revert - I'm a dumbfuck who doesn't know his own API (djm@)
netinet6
~ in6.h
> use the right capitalization for platform names
> ok jsing@ (sobrado@)
nfs
~ nfs_serv.c ~ nfs_vfsops.c
> Widen some struct statfs fields to support large filesystem stata
> and add some to be able to support statvfs(2). Do the compat dance
> to provide backward compatibility. ok thib@ miod@ (otto@)
sys
~ param.h
> forgot to update the param symbols to 4.3 (deraadt@)
~ _types.h ~ mount.h
~ types.h ~ vnode.h
+ statvfs.h
> Widen some struct statfs fields to support large filesystem stata
> and add some to be able to support statvfs(2). Do the compat dance
> to provide backward compatibility. ok thib@ miod@ (otto@)
~ syscall.h ~ syscallargs.h
> regen (otto@)
~ unistd.h
> Add the semi-standard _SC_PHYS_PAGES and _SC_AVPHYS_PAGES, sysconf(3)
> variable.
> ok espie@ (kettenis@)
ufs/ffs
~ ffs_vfsops.c
> Widen some struct statfs fields to support large filesystem stata
> and add some to be able to support statvfs(2). Do the compat dance
> to provide backward compatibility. ok thib@ miod@ (otto@)
== usr.bin =========================================================== 12/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin
audioctl
~ audioctl.1
> synchronize the synopsis and usage; improve description of flag "-a"
> written with invaluable advice from jakemsr@, jmc@, ratchov@
> ok jmc@ (sobrado@)
column
~ column.1
> EXAMPLES:
> - actually say what the example does
> - use a display instead of .Dl, so we don;t have to escape everything
> - use a prompt and indicate line break
> - use "echo" instead of "printf", so we don;t need to specify newline
> additionally, use .Ex whilst in here;
> ok otto (jmc@)
cvs
~ client.c
> Prevent sending of "Directory" messages for arguments in remote setup if
> the specified files (or directories) do not exist.
> ok joris (tobias@)
~ README
> Get README up to date: OpenCVS is linked to build now, but man pages are
> not installed yet.
> ok joris, xsa (tobias@)
~ buf.c ~ file.c
~ logmsg.c ~ update.c
> While allocation memory, make sure that file sizes are smaller than
> SIZE_MAX.
> ok joris (tobias@)
~ file.c ~ file.h
> rename check_dir_tag to user_supplied to reflect what the flag
> actually means and is used for. (joris@)
~ diff3.c
> unused var (joris@)
~ util.c
> be sure to create CVS/Root if we added a new directory to the repository.
> (joris@)
~ admin.c ~ file.c
~ file.h
> pass user_supplied to struct cvs_file so it can be used
> in the callbacks to verify if this file was specified on the command line.
> (joris@)
~ file.c
> Don't force ourself to support CVS_USE_WDIR commands (release, update, ...)
> only if we are in top directory of a repository.
> ok joris (tobias@)
~ file.c
> cf->user_supplied cannot be used in cvs_file_classify for all commands
> (like import, add, commit), so revert this change until a universal
> solution is found. (tobias@)
~ file.c
> do not empty the list in cvs_file_walklist() this is the caller
> his responsibility, allows us to reuse the same file list
> for multiple callbacks. (joris@)
~ admin.c ~ checkout.c
~ commit.c ~ cvs.h
~ file.c ~ repository.c
~ repository.h ~ tag.c
> proper repository locking:
> - all read operations now look for a lock, and wait if present but never
> try to lock the tree themselfs anymore.
> - all write operations lock the tree where needed.
> - commit locks all relevant directories before even attempting to start.
> (joris@)
~ client.c
> revert one of tobias his changes, it broke remote for several commands.
> (joris@)
~ add.c
> do not use CR_REPO in a remote setup, that will obviously fail. (joris@)
~ util.c
> be sure to build CVS/Root when updating and building directories (-d)
> (joris@)
~ update.c
> when in server mode, be sure the client picks up any new directories
> if they are available if we run with -d, however do not send the
> stuff when running a checkout so empty directories do not even appear
> in our working copy, saves us the whole pruning trouble at the end with co.
> (joris@)
~ update.c
> simplify logic in cvs_update_leavedir() (joris@)
~ update.c
> locate the first occurance of ':' in the given -j arguments
> instead of the last one, otherwise we cannot properly separate
> the tag from the given date (if any). (joris@)
~ file.c
> only compare file types if the file/directory exists both on disk
> and in the CVS/Entries file, otherwise ignore it so we correctly
> remove it from CVS/Entries.
> spotted by krw@ (joris@)
~ rcs.c
> Fix segmentation fault when running checkout -p on a file without
> specified head revision.
> ok joris (tobias@)
~ tag.c
> missing rcsnum_free();
> from Igor Zinovik (joris@)
~ rcs.c
> properly initialize variables (joris@)
~ file.c
> use RCS_PARSE_FULLY for CVS_OP_ADMIN as well. (joris@)
~ diff.c
> when a file was locally removed and a diff was done against
> an opencvs server (with either gnu or opencvs as the client),
> no actual diff output was produced. traced to the typo fixed
> in this diff.
> ok joris (sthen@)
ftp
~ ftp.1
> tweak previous; (jmc@)
~ ftp.1 ~ ftp.c
> turn on keepalive by default (one byte every 60 seconds)
> okay beck@, deraadt@, krw@ (espie@)
mixerctl
~ mixerctl.1
> synchronize the synopsis and usage; improve description of flag "-a"
> written with invaluable advice from jakemsr@, jmc@, ratchov@
> ok jmc@ (sobrado@)
ssh
~ monitor_fdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ monitor_fdpass.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
vmstat
~ vmstat.c
> use ellipsis to show that more than one disk is allowed
> ok jmc@ (sobrado@)
== usr.sbin ========================================================== 13/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin
arp
~ arp.c
> "usage:" should precede the first statement only
> ok jmc@ (sobrado@)
bgpd
~ buffer.c ~ imsg.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ buffer.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
~ imsg.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
bind
~ lib/isc/unix/privsep_fdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ lib/isc/shuffle.c
> Because the shuffle code initialisation is a specific case of shuffling
> a set of incrementing integers (and not an arbitrary set of values) it
> is possible to populate the array as we shuffle it in a single forward
> pass. Clever optimisation from didickman AT gmail.com;
> ok deraadt@ mcbride@
> (same change as netinet/ip_id.c) (djm@)
~ lib/isc/unix/privsep_fdpass.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
ifstated
~ parse.y
> Fix a null deref in link_state, which happens on config
> files which refer to undeclared states.
> popfile() needs to be called later.
> Found by and OK cnst@ (mpf@)
ospf6d
~ packet.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ packet.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
ospfd
~ packet.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ packet.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
pcidump
~ pcidump.c
> if the device supports the PCI capabilities list and the verbose flag
> is specified, print all PCI capabilities by name instead of just the
> initial capabilities pointer.
> ok dlg@ kettenis@ (reyk@)
pkg_add
~ pkg_add.1 ~ pkg_delete.1
> tweak previous; (jmc@)
~ OpenBSD/Error.pm
> print better error message if child dies... with extra dark magic to grab
> signal names from POSIX. (espie@)
~ OpenBSD/Delete.pm
> oops (espie@)
ppp
~ ppp/bundle.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ ppp/bundle.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
relayd
~ hce.c
> nuke unused variable. (pyr@)
~ buffer.c ~ imsg.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ buffer.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
~ imsg.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
ripd
~ packet.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ packet.c
> Repair more msg_controllen dealing with structures or arrays of
> descriptors; ok hshoexer, also looked at by kettenis and henning (deraadt@)
route6d
~ route6d.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ route6d.c
> repair msg_controllen and also an errant buf[CMSG_LEN(..)] decl (deraadt@)
sensorsd
~ sensorsd.c ~ sensorsd.conf.5
> Allow a program invoked on state change to receive sensor status. Perhaps
> you might want to toggle an error light when a sensor is not OK. Perhaps
> you might want to schedule a shutdown if a sensor is reporting bad news.
> Now you can do this, and cancel that pending shutdown (or turn off the
> error
> light) if the sensor decides all is well.
> ok mbalmer (who came up with an almost identical diff months ago)
> useful feedback and generally positive responses from deraadt, henning, msf
> (ckuethe@)
~ sensorsd.conf.5
> document the just-committed %s token.
> ok mbalmer (ckuethe@)
~ sensorsd.c ~ sensorsd.conf.5
> Add a "%l" token to signal whether a sensor value is in or out of bounds,
> without regard for the specific value. It's a big heavy binary hammer...
> ok & style feedback from cnst (ckuethe@)
~ sensorsd.conf.5
> - punctuate and order the tokens list, to stop it looking scrappy
> - it doesn;t make sense to list esm(4) in SEE ALSO (jmc@)
~ sensorsd.c ~ sensorsd.conf.5
> explicitly report whether the lower or upper limit is exceeded; ok sthen@
> ckuethe@ (cnst@)
~ sensorsd.8 ~ sensorsd.c
> Reduce the number of unnecessary time(3)/gettimeofday(2) calls, and make
> sure
> that the time always goes forward, so reports are neither duplicated nor
> lost.
> Report state changes stabilised through dampening immediately, instead of
> delay-
> ing them until the next reporting window; previously, it was common for
> check()
> to lag one second behind report(), hence the initial report was delayed one
> extra minute (this then reduces the number of sleep(3)/nanosleep(2) calls,
> too).
> ok ckuethe; some man-page suggestions jmc (cnst@)
~ sensorsd.8
> text simplification; (jmc@)
~ sensorsd.c
> s/273.16/273.15/g, then refactor the formulae to ensure that the
> double to int64_t conversions happen sooner rather than later
> discussed with / ok by kettenis (cnst@)
snmpd
~ mib.c ~ mib.h
> provide a partial implementation of the HOST-RESOURCES-MIB. this implements
> the hrStorage part, which is enough to see how much space is used on your
> filesystems. makes my nms happy.
> ok reyk@ (dlg@)
~ ber.c
> Return an error if more data is requested even though the buffer is empty.
> Fixes an infinite loop seen by cloder at . OK cloder@, reyk@ (claudio@)
~ ber.c
> SNMP has a restricted BER encoding especially all encodings use the
> definite-length from. So bail out with an error if this is not the case.
> OK cloder@, reyk@
> cvs: ----------------------------------------------------------------------
> (claudio@)
~ snmpe.c
> if (class != BER_CLASS_UNIVERSAL || type != BER_TYPE_SEQUENCE)
> not
> if (class != BER_CLASS_UNIVERSAL && type != BER_TYPE_SEQUENCE)
> the class and type need to be of correct.
> OK reyk@ (claudio@)
~ snmpe.c
> mostly stylistic - always print the client host name in snmpe error
> messages and make sure that the host string is initialized. (reyk@)
~ ber.c
> Check that the current ber element is a sequence or a set when '{' or '('
> is used.
> OK reyk@ (claudio@)
~ ber.c
> fix a memleak in the ber_read_elements() error case.
> ok claudio@ (reyk@)
~ ber.c
> remove noisy warning message
> ok claudio@ (reyk@)
~ buffer.c ~ imsg.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ ber.3 ~ ber.c
~ ber.h
> add a new ber function ber_add_astring() that strdups the string and
> sets the be_free flag (which is required in some cases). this will
> make it easier because it is done manually at some places in the code.
> discussed with dlg (reyk@)
~ mib.c
> move zerodotzero up so more stuff can use it
> this is an ok bit of a bigger diff that reyk wanted split up (dlg@)
~ buffer.c ~ imsg.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
~ mib.c
> provide sizeofa for counting the number of elements in an array. use it for
> passing the length of the mib in all these sysctl calls.
> ok reyk@ (dlg@)
~ snmpd.c
> usage statements are preceeded with usage:
> ok reyk@ (dlg@)
~ mib.c ~ mib.h
> implement the hrSWRun portion of the HOST-RESOURCES mib. you can see whats
> running now, which is another thing my nms at work likes.
> ok reyk@ (dlg@)
syslogd
~ privsep_fdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ privsep_fdpass.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
~ privsep.c
> syslogd leaves zombies around if multiple |/pathto/mylogprog
> children died in a row.
> Do waitpid(2) in a loop until there's nothing left.
> OK henning@, millert@ (mpf@)
~ privsep.c
> avoid errno trashing, ok mpf henning (deraadt@)
tcpdump
~ privsep_fdpass.c
> Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
> an extensive discussion with otto, kettenis, millert, and hshoexer
> (deraadt@)
~ privsep_fdpass.c
> Repair the simple cases for msg_controllen where it should just be
> CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
> of alignment; ok kettenis hshoexer (deraadt@)
traceroute6
~ traceroute6.c
> space allocated is sum of CMSG_SPACE()'s. length given to the kernel
> is the sum of all the CMSG_SPACE()'s, except for the last element use
> CMSG_LEN(). Insane, right?
> ok kettenis hshoexer (deraadt@)
===============================================================================
More information about the owc
mailing list