[odc] Daily src changes for 2007-05-29
ODC
auto at squish.net
Wed May 30 07:00:01 BST 2007
OpenBSD src changes summary for 2007-05-29
==========================================
bin/cat bin/chmod
bin/cp bin/date
bin/dd bin/df
bin/echo bin/ed
bin/expr bin/kill
bin/ln bin/ls
bin/mkdir bin/mv
bin/pax bin/ps
bin/pwd bin/rm
bin/rmdir bin/sleep
bin/stty bin/systrace
bin/test distrib/notes
distrib/sets distrib/zaurus
etc/sensorsd.conf etc/sysctl.conf
include/dlfcn.h lib/libc
libexec/ld.so regress/sys
sbin/disklabel sbin/fsck_ext2fs
sbin/fsck_ffs sbin/growfs
sbin/newfs sbin/sysctl
share/man sys/adosfs
sys/arch/alpha/alpha sys/arch/amd64/amd64
sys/arch/amd64/conf sys/arch/amd64/include
sys/arch/amd64/pci sys/arch/amd64/stand/boot
sys/arch/amd64/stand/cdboot sys/arch/arm/arm
sys/arch/arm/include sys/arch/aviion/aviion
sys/arch/hp300/hp300 sys/arch/hp300/stand/cdboot
sys/arch/hp300/stand/uboot sys/arch/hppa/hppa
sys/arch/hppa/stand/boot sys/arch/hppa64/hppa64
sys/arch/hppa64/stand/boot sys/arch/i386/conf
sys/arch/i386/i386 sys/arch/i386/include
sys/arch/i386/pci sys/arch/i386/stand/boot
sys/arch/i386/stand/cdboot sys/arch/landisk/landisk
sys/arch/luna88k/luna88k sys/arch/m88k/m88k
sys/arch/mac68k/mac68k sys/arch/macppc/dev
sys/arch/macppc/macppc sys/arch/mips64/mips64
sys/arch/mvme68k/dev sys/arch/mvme68k/mvme68k
sys/arch/mvme88k/mvme88k sys/arch/mvmeppc/dev
sys/arch/mvmeppc/mvmeppc sys/arch/powerpc/powerpc
sys/arch/sgi/localbus sys/arch/sgi/sgi
sys/arch/sh/sh sys/arch/solbourne/solbourne
sys/arch/sparc/conf sys/arch/sparc/dev
sys/arch/sparc/include sys/arch/sparc/sparc
sys/arch/sparc64/conf sys/arch/sparc64/dev
sys/arch/sparc64/include sys/arch/sparc64/sparc64
sys/arch/sparc64/stand/ofwboot sys/arch/vax/vax
sys/arch/zaurus/dev sys/arch/zaurus/stand/zboot
sys/compat/linux sys/conf
sys/dev sys/dev/ic
sys/dev/isa sys/dev/pci
sys/dev/usb sys/kern
sys/miscfs/specfs sys/net
sys/netinet sys/nfs
sys/stand sys/sys
sys/ufs/ffs sys/uvm
sys/xfs usr.bin/cdio
usr.bin/cvs usr.bin/diff
usr.bin/join usr.bin/kdump
usr.bin/rcs usr.bin/systat
usr.bin/top usr.sbin/bgpd
usr.sbin/eeprom usr.sbin/hoststatectl
usr.sbin/hoststated usr.sbin/ospfd
usr.sbin/pkg_add usr.sbin/sensorsd
usr.sbin/tcpdump usr.sbin/ypserv
usr.sbin/ztsscale
== bin =============================================================== 01/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/bin
cat
~ cat.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
chmod
~ chgrp.1 ~ chmod.1
~ chown.8
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
cp
~ cp.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
date
~ date.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
dd
~ dd.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
df
~ df.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
~ df.1
> note some flags which differ in behaviour from posix;
> suggested by millert (jmc@)
echo
~ echo.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
ed
~ ed.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
expr
~ expr.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
kill
~ kill.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
ln
~ ln.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
ls
~ ls.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
~ ls.1
> note some flags which differ in behaviour from posix;
> suggested by millert (jmc@)
mkdir
~ mkdir.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
mv
~ mv.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
pax
~ pax.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
ps
~ ps.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
pwd
~ pwd.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
rm
~ rm.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
rmdir
~ rmdir.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
sleep
~ sleep.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
stty
~ stty.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
systrace
~ openbsd-syscalls.c
> remove compat netbsd stuff (otto@)
test
~ test.1
> - consistent STANDARDS blurb
> - note which options are extensions to POSIX
> discussed w/ deraadt otto millert (jmc@)
== distrib =========================================================== 02/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/distrib
notes
~ packages
> Don't suggest running NetBSD/alpha binaries via compat_netbsd anymore.
> (miod@)
sets
~ lists/base/md.alpha ~ lists/base/md.amd64
~ lists/base/md.armish ~ lists/base/md.aviion
~ lists/base/md.cats ~ lists/base/md.hp300
~ lists/base/md.hppa ~ lists/base/md.hppa64
~ lists/base/md.i386 ~ lists/base/md.landisk
~ lists/base/md.luna88k ~ lists/base/md.mac68k
~ lists/base/md.macppc ~ lists/base/md.mvme68k
~ lists/base/md.mvme88k ~ lists/base/md.mvmeppc
~ lists/base/md.sgi ~ lists/base/md.sparc
~ lists/base/md.sparc64 ~ lists/base/md.vax
~ lists/base/md.zaurus ~ lists/comp/mi
~ lists/man/mi
> sync (deraadt@)
zaurus
~ ramdisk/install.md
> adapt ztsscale and the installer to use wscons ioctls for calibration;
> ok deraadt@ (robert@)
== etc =============================================================== 03/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/etc
sensorsd.conf
~ sensorsd.conf
> for sensorsd.conf entries, let's do matching by sensor type if maching by
> complete sysctl name does not yield anything; ok henning, otto (cnst@)
~ sensorsd.conf
> tweaks; help/ok cnst (jmc@)
sysctl.conf
~ sysctl.conf
> put commented out entry for kern.nosuidcoredump=2 in, yes theo & tedu
> (henning@)
== include =========================================================== 04/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/include
dlfcn.h
~ dlfcn.h
> based on a diff from Matt Provost: allow printing of ldd information
> on libraries themselves. Works by setting up the debugging flags then
> calling dlopen() to do the heavy lifting. ok drahn (jason@)
== lib =============================================================== 05/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/lib
libc
~ gen/sysctl.3
> mention new option 2 for nosuidcoredump (tedu@)
~ gen/disklabel.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ gen/disklabel.c
> when producing a v1 label, set the version!!!! (otto@)
== libexec =========================================================== 06/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/libexec
ld.so
~ dlfcn.c ~ ldd/ldd.c
> based on a diff from Matt Provost: allow printing of ldd information
> on libraries themselves. Works by setting up the debugging flags then
> calling dlopen() to do the heavy lifting. ok drahn (jason@)
== regress =========================================================== 07/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/regress
sys
~ kern/mbuf/mbuftest.c
> the MSIZE mbuf constant moved from MD param.h to MI param.h
> ok deraadt@ millert@ (david@)
== sbin ============================================================== 08/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sbin
disklabel
~ disklabel.c ~ editor.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
fsck_ext2fs
~ setup.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
fsck_ffs
~ setup.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
growfs
~ growfs.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
newfs
~ mkfs.c ~ newfs.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
sysctl
~ sysctl.c
> remove the ztsscale sysctl; ok deraadt@ (robert@)
== share ============================================================= 09/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/share
man
~ man4/re.4
> correct frame size limitation in BUGS section (7440 bytes, not 7.5K)
> ok jmc (pvalchev@)
~ man8/man8.hp300/boot_hp300.8
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
~ man9/style.9
> delete some overly fascist rules.. prodded by tom/jmc (deraadt@)
~ man4/tsl.4 ~ man4/iic.4
> normalise the title of tsl(4); ok jmc; 'don't care' kettenis (cnst@)
~ man4/man4.sparc64/Makefile
> led(4). (kettenis@)
~ man4/fxp.4
> +.Cd "iophy* at mii?"
> from brad; ok claudio (jmc@)
+ man4/man4.sparc64/led.4
> led(4). (kettenis@)
~ man4/man4.sparc64/ebus.4
> led.4 here too; plus a tweak (jmc@)
~ man8/afterboot.8
> a `number' is a certain combination of `digits';
> `/etc/mail/localhost.cf' does not has RFC 2476 support disabled by default
> ok jmc@ and millert@ (sobrado@)
~ man4/man4.sparc/le.4 ~ man4/man4.sparc64/le.4
~ man4/man4.sparc64/uperf.4
> use the right capitalization for `EBus' and `SBus'
> ok jmc@ (sobrado@)
~ man4/man4.i386/cpu.4
> Remove option I386_CPU; prompted by jmc@, thanks. (tom@)
~ man4/atalk.4 ~ man4/ifmedia.4
~ man4/man4.hppa/io.4
> zap token ring bits; ok claudio (jmc@)
~ man4/man4.hppa/io.4
> oops; (jmc@)
~ man4/tsl.4
> add HISTORY
> ok jmc@ (jasper@)
== sys =============================================================== 10/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sys
adosfs
~ advfsops.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
arch/alpha/alpha
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ interrupt.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/amd64/amd64
~ gdt.c
> Add a name argument to the RWLOCK_INITIALIZER macro.
> Pick reasonble names for the locks involved..
> ok tedu@, art@ (thib@)
~ pmap.c
> PMAP_STEAL_MEMORY for amd64
> drahn@ ok (art@)
~ pmap.c
> Remove a debugging printf left behind by accident. (art@)
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ identcpu.c + est.c
> Some improvements for better intel cpu support.
> Add EST support from i386, minus the tables
> Also add in support for CPU temperature sensors, based on diff to tech
> by Pierre Riteau.
> ok deraadt gwk (tedu@)
~ identcpu.c
> theo says degrees is spelled degrees (tedu@)
~ conf.c
> Die Die Die wt. (krw@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
~ bus_dma.c
> s/entrace/entrance (not obvious that the code was cut/paste =) (jason@)
~ copy.S
> Improve copyinstr and copyoutstr by erroring out when the userland
> pointer is VM_MAXUSER_ADDRESS right up front, rather than when
> startig the loop. This is how it is done on i386.
> ok weingart@ art@ (tom@)
arch/amd64/conf
~ files.amd64
> Some improvements for better intel cpu support.
> Add EST support from i386, minus the tables
> Also add in support for CPU temperature sensors, based on diff to tech
> by Pierre Riteau.
> ok deraadt gwk (tedu@)
~ files.amd64
> only the i810 is known to be needed for agp, reduce driver set to just that
> (tedu@)
arch/amd64/include
~ pmap.h
> PMAP_STEAL_MEMORY for amd64
> drahn@ ok (art@)
~ cpu.h ~ specialreg.h
> Some improvements for better intel cpu support.
> Add EST support from i386, minus the tables
> Also add in support for CPU temperature sensors, based on diff to tech
> by Pierre Riteau.
> ok deraadt gwk (tedu@)
~ specialreg.h
> put the temperature msr in line with the others (tedu@)
~ cpu.h
> take a lot more stuff out of userland namespace (sysctl cared about ksensor
> use); ok tedu (deraadt@)
arch/amd64/pci
+ agp_machdep.c
> only the i810 is known to be needed for agp, reduce driver set to just that
> (tedu@)
arch/amd64/stand/boot
~ boot.8
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
arch/amd64/stand/cdboot
- cdboot.c ~ Makefile
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
~ Makefile
> forgot to move to using shared boot.c (deraadt@)
arch/arm/arm
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ softintr.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ stubs.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
~ bus_dma.c
> s/entrace/entrance (not obvious that the code was cut/paste =) (jason@)
~ arm32_machdep.c
> Make use of struct wsmouse_calibcoords and add ioctl support
> for getting and passing calibration values instead of using
> sysctl.
> Move the scale struct to uts_softc so that each device can have
> it's own values instead of using the global one.
> Remove the two MD sysctl (ztsscale and ztsrawmode).
> ok deraadt@ (robert@)
arch/arm/include
~ cpu.h
> Make use of struct wsmouse_calibcoords and add ioctl support
> for getting and passing calibration values instead of using
> sysctl.
> Move the scale struct to uts_softc so that each device can have
> it's own values instead of using the global one.
> Remove the two MD sysctl (ztsscale and ztsrawmode).
> ok deraadt@ (robert@)
arch/aviion/aviion
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/hp300/hp300
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ intr.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/hp300/stand/cdboot
~ cdboot.c
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
arch/hp300/stand/uboot
~ uboot.c
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
arch/hppa/hppa
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ autoconf.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
~ mainbus.c
> s/entrace/entrance (not obvious that the code was cut/paste =) (jason@)
arch/hppa/stand/boot
~ boot.8
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
arch/hppa64/hppa64
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ autoconf.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
~ mainbus.c
> s/entrace/entrance (not obvious that the code was cut/paste =) (jason@)
arch/hppa64/stand/boot
~ boot.8
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
arch/i386/conf
~ files.i386
> Fix compilation of kernels without I686_CPU: piixpcib is always
> needed (not just for I686_CPU), but it and ichpcib only call the
> update_cpuspeed functions for I686_CPU.
> ok tedu@ deraadt@ (tom@)
~ GENERIC ~ Makefile.i386
~ RAMDISK ~ RAMDISKB
~ RAMDISKC ~ RAMDISK_CD
> Remove support for 80386 processors. Apologies if you have one of
> the rare 80386-bases system with enough memory, a 387 FPU, a useable
> disk subsystem, and the patience to wait for it to unpack the
> distribution .tgz files.
> approval from art@ and many others (esp. nick@); ok deraadt@ (tom@)
arch/i386/i386
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ apm.c ~ acpi_machdep.c
> Add global variable apm_attached, machine dependant probe routine for ACPI
> will check this flag durring probe, meaning that if the machine has APM
> ACPI will not attach. This should remove one obstacle on the road to
> enabling ACPI by default.
> ok marco, dreaadt, art, krw, art (gwk@)
~ cpu.c ~ lock_machdep.c
~ locore.s ~ machdep.c
~ pmap.c
> Remove support for 80386 processors. Apologies if you have one of
> the rare 80386-bases system with enough memory, a 387 FPU, a useable
> disk subsystem, and the patience to wait for it to unpack the
> distribution .tgz files.
> approval from art@ and many others (esp. nick@); ok deraadt@ (tom@)
~ trap.c
> Remove trapwrite(), which was only used by 80386.
> Prompted by miod@, thanks. (tom@)
~ conf.c
> Die Die Die wt. (krw@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
~ machdep.c ~ mainbus.c
> add support for core cpu temperature sensors.
> from a diff by pierre riteau, but updated to look like the amd64 version
> ok deraadt (tedu@)
~ machdep.c
> s/entrace/entrance (not obvious that the code was cut/paste =) (jason@)
~ locore.s
> Improve copyinstr and copyoutstr by performing the check against
> VM_MAXUSER_ADDRESS before running the loop rather than for every
> byte we copy. Saves 4 bytes in locore :)
> Similar to how this is done on amd64.
> ok weingart@ art@ (tom@)
arch/i386/include
~ endian.h ~ lock.h
~ pmap.h
> Remove support for 80386 processors. Apologies if you have one of
> the rare 80386-bases system with enough memory, a 387 FPU, a useable
> disk subsystem, and the patience to wait for it to unpack the
> distribution .tgz files.
> approval from art@ and many others (esp. nick@); ok deraadt@ (tom@)
~ cpu.h ~ specialreg.h
> add support for core cpu temperature sensors.
> from a diff by pierre riteau, but updated to look like the amd64 version
> ok deraadt (tedu@)
arch/i386/pci
~ ichpcib.c ~ piixpcib.c
> Fix compilation of kernels without I686_CPU: piixpcib is always
> needed (not just for I686_CPU), but it and ichpcib only call the
> update_cpuspeed functions for I686_CPU.
> ok tedu@ deraadt@ (tom@)
arch/i386/stand/boot
~ boot.8
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
arch/i386/stand/cdboot
- cdboot.c ~ Makefile
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
arch/landisk/landisk
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
arch/luna88k/luna88k
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/m88k/m88k
~ m88k_machdep.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
arch/mac68k/mac68k
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ intr.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/macppc/dev
~ macintr.c ~ openpic.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
arch/macppc/macppc
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/mips64/mips64
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ interrupt.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
arch/mvme68k/dev
~ ssh.c ~ sbic.c
> TAILQ_LAST takes two arguments. (pyr@)
arch/mvme68k/mvme68k
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ machdep.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/mvme88k/mvme88k
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/mvmeppc/dev
~ openpic.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
arch/mvmeppc/mvmeppc
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ autoconf.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/powerpc/powerpc
~ mutex.S
> Panic if the locker is self. (drahn@)
arch/sgi/localbus
~ macebus.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
arch/sgi/sgi
~ conf.c
> code typo (deraadt@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/sh/sh
~ interrupt.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ sh_machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/solbourne/solbourne
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/sparc/conf
~ GENERIC ~ RAMDISK
~ SUN4C ~ SUN4M
~ files.sparc
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
arch/sparc/dev
~ obio.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ dma.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ btreg.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ if_le.c
> use the right capitalization for `MBus' and `SBus'
> ok jmc@ (sobrado@)
~ sbus.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ sbusreg.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ sbusvar.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ stp_sbus.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
arch/sparc/include
~ pte.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
arch/sparc/sparc
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ locore.s
> use the right capitalization for `MBus'
> ok jmc@ (sobrado@)
~ autoconf.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ intr.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ iommu.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ intr.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/sparc64/conf
~ GENERIC ~ files.sparc64
> Add led(4). (kettenis@)
arch/sparc64/dev
+ led.c
> New driver for front panel LEDs on the V215/245. (kettenis@)
~ iommu.c ~ iommureg.h
~ iommuvar.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ sbus.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ sbusreg.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ sbusvar.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ stp_sbus.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
arch/sparc64/include
~ bus.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ autoconf.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
arch/sparc64/sparc64
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ disksubr.c
> typo (deraadt@)
~ intr.c
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ timerreg.h
> use the right capitalization for `SBus'
> ok jmc@ (sobrado@)
~ intr.c
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/sparc64/stand/ofwboot
~ boot.c
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
~ ofdev.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
arch/vax/vax
~ disksubr.c
> Refactor readdisklabel() to ensure there is a single point of return, in
> preparation for translating all disk labels visible to the kernel to
> the soon to arrive V1 format.
> ok otto@ deraadt@ (krw@)
~ disksubr.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ machdep.c
> start of dumpconf() unification, which one day we can pray will be MI
> (and part of setroot); ok miod (deraadt@)
arch/zaurus/dev
~ zts.c
> Make use of struct wsmouse_calibcoords and add ioctl support
> for getting and passing calibration values instead of using
> sysctl.
> Move the scale struct to uts_softc so that each device can have
> it's own values instead of using the global one.
> Remove the two MD sysctl (ztsscale and ztsrawmode).
> ok deraadt@ (robert@)
arch/zaurus/stand/zboot
~ boot.8
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
compat/linux
~ files.linux ~ syscalls.master
> linux_getcwd is the same as the system one.
> discovered/ok beck (tedu@)
~ linux_syscall.h ~ linux_syscallargs.h
~ linux_syscalls.c ~ linux_sysent.c
> regen (tedu@)
- linux_getcwd.c
> byebye (tedu@)
~ linux_ipc.c
> adapt from netbsd:
> fold sys_shmat1() back into sys_shmat(), instead add flag for shmget(2)
> to specify that later shmat(2) for the shared memory segment should succeed
> even if the segment would be marked removed; use this to implement the
> Linux-compatible semantics of shmat(2)
> this fixes current opera with shm
> ok millert (sturm@)
conf
~ files
> Move tokenring support to the attic where it can join the cards that where
> decomissioned aeon ago. We will not miss it at all.
> OK dlg@ henning@ and a lot of cheers by other in the room (claudio@)
dev
~ softraid.c
> Sparc64 does not like 32 bit pointers. Bad me.
> help tedu
> ok todd (marco@)
~ ipmi.c
> work around really bad softc abuse by marco; proper fixes for this will
> be part of a later rototilling (deraadt@)
~ ipmi.c
> back out last commit. It helps to compile before commiting. OK beck@ art@
> (claudio@)
~ ipmi.c
> Correct fix for the last backout use strlcpy() to fill dv_xname. OK beck@
> (claudio@)
~ softraid.c ~ softraidvar.h
> Re-add sync.
> ok tedu (marco@)
~ softraid.c
> Fix race in sync code. (marco@)
~ softraid.c
> Kill stale comment (marco@)
~ softraid.c ~ softraidvar.h
> Add shutdownhook for all disciplines (marco@)
~ softraid.c
> sprinkle some SMALL_KERNEL so that it can run on boot media.
> ok todd (marco@)
~ softraid.c ~ softraidvar.h
> Add dirty bit for volumes (marco@)
dev/ic
~ malo.c
> Beef up malo_hexdump() to emulate hexdump -C output. Helps to see hidden
> agendas in buffers that stupidly encode numbers as strings so that we need
> to parse them in the kernel as seen on che(4). OK mglocker@ (claudio@)
~ tropic.c ~ tropicreg.h
~ tropicvar.h
> Move tokenring support to the attic where it can join the cards that where
> decomissioned aeon ago. We will not miss it at all.
> OK dlg@ henning@ and a lot of cheers by other in the room (claudio@)
~ ami.c
> Make stimeout a debug print. It has no value for users to see it.
> prompted by deraadt
> ok dlg (marco@)
- tropic.c - tropicreg.h
- tropicvar.h
> It helps to commit removals from the tree that has the files cvs removed
> instead of one where they where just empty. Figured out by art@ (claudio@)
~ mfi.c ~ ami.c
~ ciss.c
> sprinkle some SMALL_KERNEL to permit bio/softraid to run on boot media
> ok marco@ then suggested deraadt@ (todd@)
dev/isa
~ files.isa ~ if_tr_isa.c
~ if_tribm_isa.c ~ if_trtcm_isa.c
> Move tokenring support to the attic where it can join the cards that where
> decomissioned aeon ago. We will not miss it at all.
> OK dlg@ henning@ and a lot of cheers by other in the room (claudio@)
~ files.isa
> Stray reference in a comment to the dear departed wt. (krw@)
- if_tr_isa.c - if_tribm_isa.c
- if_trtcm_isa.c
> It helps to commit removals from the tree that has the files cvs removed
> instead of one where they where just empty. Figured out by art@ (claudio@)
dev/pci
~ ips.c
> Fix bus_dmamap_sync size argument. (grange@)
~ arc.c
> sprinkle some SMALL_KERNEL to permit bio/softraid to run on boot media
> ok marco@ then suggested deraadt@ (todd@)
dev/usb
~ usbdi.h
> Kill some #if define(__OtherBSD__). OK dlg@ (claudio@)
~ usb_port.h
> set UDCF_DEBUG to 1 when USB_DEBUG is defined.
> ok mjc (mbalmer@)
kern
~ kern_sig.c
> put suid coredumps into /var/crash if kern.nosuidcoredump is set to 2.
> makes debugging easier, docs in a bit
> ok beck deraadt (tedu@)
~ kern_malloc.c ~ kern_sysctl.c
~ tty_pty.c
> Add a name argument to the RWLOCK_INITIALIZER macro.
> Pick reasonble names for the locks involved..
> ok tedu@, art@ (thib@)
~ kern_sig.c
> update comment re nosuidcoredump sysctl
> ok tedu (jcs@)
~ vfs_cache.c ~ vfs_getcwd.c
~ vfs_subr.c
> Step one of some vnode improvements - change getnewvnode to
> actually allocate "desiredvnodes" - add a vdrop to un-hold a vnode held
> with vhold, and change the name cache to make use of vhold/vdrop, while
> keeping track of which vnodes are referred to by which cache entries to
> correctly hold/drop vnodes when the cache uses them.
> ok thib@, tedu@, art@ (beck@)
~ subr_disk.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ sysv_shm.c
> adapt from netbsd:
> fold sys_shmat1() back into sys_shmat(), instead add flag for shmget(2)
> to specify that later shmat(2) for the shared memory segment should succeed
> even if the segment would be marked removed; use this to implement the
> Linux-compatible semantics of shmat(2)
> this fixes current opera with shm
> ok millert (sturm@)
~ vfs_init.c
> Nuke the predefined vattr, it was only used for !DIAGNOSTIC,
> this shrinks the i386 RAMDISK a bit.
> (Using the predefined vattr for GENERIC meant a growth of
> 416bytes without any measurable perfomance gain).
> ok krw@, art@ (thib@)
~ vfs_cache.c
> missing couple of #ifdef DIAGNOSTIC and style nits from art@ (beck@)
~ vfs_bio.c
> Change getnewbuf from taking lots of sleep arguments and then doing magic
> with them and trying to pass an error back after sleep, to just fail with
> NULL and let the caller wait for buffers.
> Introduce buf_wait() that does all the sleep magic and use buf_wait in
> the getnewbuf callers where it matters.
> pedro@ beck@ ok (art@)
~ vfs_bio.c
> I suck. Forgot splx() in the early return path. (art@)
miscfs/specfs
~ spec_vnops.c
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
net
~ pf_ioctl.c
> Add a name argument to the RWLOCK_INITIALIZER macro.
> Pick reasonble names for the locks involved..
> ok tedu@, art@ (thib@)
~ pf.c
> gain us another 10+% of performance.
> boring details:
> long time ago (in r1.313) code was added to handle protocol checksums:
> > Check protocol (TCP/UDP/ICMP/ICMP6) checksums of all incoming packets,
> > and drop packets with invalid checksums. Without such a check, pf would
> > return RST/ICMP errors even for packets with invalid checksums, which
> > could be used to detect the presence of the firewall, reported by
> > "Ed White" in http://www.phrack.org/phrack/60/p60-0x0c.txt.
> that meant we did the checksumming for each and every packet traversing pf.
> now only do the checksumming right before we send an RST back, so in all
> other cases we save that work.
> ok bob theo (henning@)
~ if.c
> Spaces, no binary changes. (claudio@)
~ netisr.h ~ netisr_dispatch.h
> IMP is dead and nothing uses NETISR_IMP so remove it from the netisr list.
> OK henning@ mcbride@ (claudio@)
~ netisr.h
> Use atomic operations to operate on netisr, instead of clearing it at
> splhigh.
> This changes nothing on legacy architectures, but is a bit faster (and
> simpler)
> on the interesting ones. (miod@)
~ if.h
> Define IF_ENQUEUE() and friends as proper C statements using do ... while
> ok henning (uwe@)
~ if_token.h ~ if_tokensubr.c
> Move tokenring support to the attic where it can join the cards that where
> decomissioned aeon ago. We will not miss it at all.
> OK dlg@ henning@ and a lot of cheers by other in the room (claudio@)
~ if_ethersubr.c
> there was code inside
> #if NPF > 0
> to prevent feeding back the mbuf do looutput if we are on simplex
> interfaces and the packet has been routed by pf, whch can lead to a loop
> in weird corner cases.
> apparently nobody triggered these cases in ages, since pf.h was not
> included
> and thus NPF not defined and thus this code not compiled.
> ok theo (henning@)
~ if_fddisubr.c
> now i get my hands dirty in here... from if_ethersubr.c:
> there was code inside
> #if NPF > 0
> to prevent feeding back the mbuf do looutput if we are on simplex
> interfaces and the packet has been routed by pf, which can lead to a loop
> in weird corner cases.
> apparently nobody triggered these cases in ages, since pf.h was not
> included
> and thus NPF not defined and thus this code not compiled. (henning@)
- if_tokensubr.c
> It helps to commit removals from the tree that has the files cvs removed
> instead of one where they where just empty. Figured out by art@ (claudio@)
- if_token.h ~ if_media.h
> remove token ring leftovers, ok mcbride pval (henning@)
netinet
~ ip_input.c ~ ip_output.c
> gain another 5+% in ip forwarding performance.
> boring details:
> skip looking for ipsec tags and descending into ip_spd_lookup if there
> are no ipsec flows, except in one case in ip_output (spotted by markus)
> where we have to if we have a pcb. ip_spd_lookup has the shortcut already,
> but there is enough work done before so that skipping that gains us about
> 5%. ok theo, markus (henning@)
~ ip_carp.c
> Move tokenring support to the attic where it can join the cards that where
> decomissioned aeon ago. We will not miss it at all.
> OK dlg@ henning@ and a lot of cheers by other in the room (claudio@)
nfs
~ nfs_kq.c ~ nfs_node.c
> Add a name argument to the RWLOCK_INITIALIZER macro.
> Pick reasonble names for the locks involved..
> ok tedu@, art@ (thib@)
stand
~ boot/boot.c
> remove support for many alternative kernel choices -- only try /bsd and
> not the other stupid names. that feature was retarded. tested on a few
> architectures by a few people (deraadt@)
- cat.c - copy.c
- ls.c
> these standalone commands are never used anymore; ok tom (deraadt@)
sys
~ rwlock.h
> Add a name argument to the RWLOCK_INITIALIZER macro.
> Pick reasonble names for the locks involved..
> ok tedu@, art@ (thib@)
~ namei.h ~ vnode.h
> Step one of some vnode improvements - change getnewvnode to
> actually allocate "desiredvnodes" - add a vdrop to un-hold a vnode held
> with vhold, and change the name cache to make use of vhold/vdrop, while
> keeping track of which vnodes are referred to by which cache entries to
> correctly hold/drop vnodes when the cache uses them.
> ok thib@, tedu@, art@ (beck@)
~ disklabel.h
> Updated disklabel format to support larger disks and partitions. We
> free room in struct partition by packing fragment size and
> fragments/block more tighlty and use the resulting space to make
> offset and size 48 bits. For the disk part we use spare fields in
> struct disklabel. Kernel converts in-mem copy of the on-disk
> label if needed, disklabel(8) writes new version.
> We are careful to only change fields not used by bootloaders.
> Conception of basic scheme by deraadt.
> ok deraadt@ krw@ (otto@)
~ shm.h
> adapt from netbsd:
> fold sys_shmat1() back into sys_shmat(), instead add flag for shmget(2)
> to specify that later shmat(2) for the shared memory segment should succeed
> even if the segment would be marked removed; use this to implement the
> Linux-compatible semantics of shmat(2)
> this fixes current opera with shm
> ok millert (sturm@)
~ vnode.h
> Nuke the predefined vattr, it was only used for !DIAGNOSTIC,
> this shrinks the i386 RAMDISK a bit.
> (Using the predefined vattr for GENERIC meant a growth of
> 416bytes without any measurable perfomance gain).
> ok krw@, art@ (thib@)
~ disklabel.h
> Collapse the __partition_u1 union in struct disklabel,
> keep p_cpg around since its need. remove p_sgs.
> ok otto@,krw@ (thib@)
ufs/ffs
~ ffs_vfsops.c
> FFS2-aware code for ffs_sbupdate() and ffs_reload().
> Okay otto@ millert@ krw@ beck@ thib@ (pedro@)
~ ffs_vfsops.c
> when mount -f'ing a dirty filesystem, do not clear the unclean flag,
> the fs remains dirty. ok pedro@ (otto@)
uvm
~ uvm_swap.c
> Add a name argument to the RWLOCK_INITIALIZER macro.
> Pick reasonble names for the locks involved..
> ok tedu@, art@ (thib@)
~ uvm_vnode.c
> Kill some simplelock goo that annoyed me while reading.
> ok art@ (thib@)
xfs
~ xfs_node-bsd.c
> Step one of some vnode improvements - change getnewvnode to
> actually allocate "desiredvnodes" - add a vdrop to un-hold a vnode held
> with vhold, and change the name cache to make use of vhold/vdrop, while
> keeping track of which vnodes are referred to by which cache entries to
> correctly hold/drop vnodes when the cache uses them.
> ok thib@, tedu@, art@ (beck@)
~ xfs_node-bsd.c ~ xfs_syscalls-common.c
> make xfs use VT_XFS instead of VT_AFS when getting vnodes
> via getnewvnode();
> ok art@,tedu@
> tested by todd@ (thib@)
== usr.bin =========================================================== 11/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin
cdio
~ rip.c
> reorder code so that the wave header gets the size of the current track
> instead of the previous track. Alexey Vatchenko agrees. (mjc@)
~ cdio.1
> fix some list items; from andres (jmc@)
cvs
~ buf.c ~ diff3.c
~ diff_internals.c ~ rcs.c
~ rcsnum.c ~ util.c
> Since xrealloc dies on failure it is safe to directly assign to the
> original pointer. Theo agrees, and so does the rest of the tree
> (ssh, etc. all do this already).
> Saves a bunch of variables and assignments.
> OK niallo@ (ray@)
~ fatal.c
> I've been running with an abort upon fatal loop, and I think I've
> mostly gotten rid of them. It would be nice to know if there is a
> fatal loop instead of dying quietly, so call errx instead of exit(1).
> errx does not call fatal so it should be fine.
> OK niallo@ (ray@)
~ diff_internals.c
> Get rid of another useless variable, another difference between our
> diff code and diff(1).
> OK niallo@ (ray@)
diff
~ Makefile ~ diff.c
~ diffdir.c ~ diffreg.c
+ xmalloc.c + xmalloc.h
> Bring in some changes from rcsdiff:
> 1. Replace all the e*alloc functions with the x*alloc versions.
> 2. Whitespace syncs according to style.
> 3. Remove the __inline stuff.
> 4. Remove the min/max functions, using the MAX/MIN macros instead.
> OK millert@ (ray@)
join
~ join.1 ~ join.c
> remove -j from SYNOPSIS and usage(); it's deprecated and already
> documented in STANDARDS;
> ok otto (jmc@)
kdump
~ kdump.c
> more parts of netbsd compat that go away (deraadt@)
rcs
~ buf.c ~ diff.c
~ diff3.c ~ rcs.c
~ rcsnum.c ~ rcsutil.c
> Since xrealloc dies on failure it is safe to directly assign to the
> original pointer. Theo agrees, and so does the rest of the tree
> (ssh, etc. all do this already).
> Saves a bunch of variables and assignments.
> OK niallo@ (ray@)
~ diff.c
> Get rid of another useless variable, another difference between our
> diff code and diff(1).
> OK niallo@ (ray@)
systat
~ cmds.c
> otto would prefer only displaying one digit after decimal for refresh rate
> (tedu@)
top
~ Makefile ~ commands.c
~ display.c ~ display.h
~ machine.c ~ machine.h
~ screen.c ~ screen.h
~ top.1 ~ top.c
~ top.h
> Instead of using hand-crafted redraws minimizing screen updates, use
> curses.
> Enables nice things like process highlighting without hurting the brain.
> ok deraadt@ (otto@)
== usr.sbin ========================================================== 12/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin
bgpd
~ session.c
> Correctly encode the End-of-RIB marker capability. It is just 2 bytes not
> 4. OK henning@ (claudio@)
~ rde_update.c
> Correctly fill the MP reach and unreach update messages. Until now
> short updateds included a dummy ::/0 route and the withdraws were totaly
> wrong because the AFI/SAFI header was missing.
> put it in henning@ (claudio@)
eeprom
~ eeprom.8
> use the right capitalization for `EBus' and `SBus'
> ok jmc@ (sobrado@)
hoststatectl
~ hoststatectl.c
> move struct relay to the runtime + config scheme.
> this time around, include hoststatectl changes too. (pyr@)
~ parser.c
> teach hoststatectl's parser about reload. the code to handle it is
> already there. of course this is a dummy for now. (pyr@)
~ hoststatectl.c
> allow the control handling code to send messages back to the parent.
> forward IMSG_CTL_RELOAD which ends up not doing anything for now. (pyr@)
hoststated
~ hoststated.c ~ hoststated.h
~ parse.y ~ relay.c
> move struct relay to the runtime + config scheme.
> this time around, include hoststatectl changes too. (pyr@)
~ hoststated.h ~ parse.y
~ relay.c
> move the ssl cipher suite string to a (small) static charbuf,
> this will make it easier to send the struct over the socket. (pyr@)
~ hoststated.c
> when purging relays, purge sessions as well.
> not needed for the initial purge since no sessions exist but will
> be necessary when reloading the relay process. (pyr@)
~ Makefile ~ hce.c
~ hoststated.c ~ hoststated.conf.5
~ hoststated.h ~ log.c
~ parse.y + check_script.c
> add a new check method which allows to run external scripts/programs
> for custom evaluations.
> pyr agrees to put it in now but to do some improvements of the timeout
> handling later. (reyk@)
~ hoststated.c ~ parse.y
> do not start relay processes when no L7 load balancing is needed.
> ok reyk@ (pyr@)
~ hoststated.h
> put the reload imsg types in right now. it makes my life easier. (pyr@)
~ control.c ~ hoststated.c
~ hoststated.h ~ pfe.c
> allow the control handling code to send messages back to the parent.
> forward IMSG_CTL_RELOAD which ends up not doing anything for now. (pyr@)
~ pfe.c
> when the time comes, let pfe_dispatch_parent be able to service more
> than one message. (pyr@)
~ hoststated.c
> make sure every dispatch function is fed the environment structure. (pyr@)
ospfd
~ Makefile ~ area.c
~ interface.c ~ ospfd.c
~ ospfd.h ~ ospfe.c
~ ospfe.h ~ parse.y
~ printconf.c + carp.c
> Demote support for ospfd. It is possible to specify a demote group on
> interfaces and areas. With this carp setups using ospfd are more reliable
> because we can fail over if the OSPF connectivity is (partially) lost.
> OK norby@ (claudio@)
pkg_add
~ OpenBSD/PkgCfl.pm ~ OpenBSD/Vstat.pm
> don't use 0 as an error number. (espie@)
~ OpenBSD/PackingElement.pm
> downgrade `can't find +DISPLAY' to a warning until I figure things out.
> (espie@)
~ OpenBSD/PackingElement.pm ~ OpenBSD/PackingList.pm
> clean-up SpecialFile a bit: make sure the name passed during creation is
> totally irrelevant.
> Add a new property to packing-lists: infodir, and use it to compute
> fullnames
> for special files.
> Once this is properly synchronized, we can tie infodir to packing-lists,
> and
> remove it from the state... (espie@)
~ OpenBSD/PackingElement.pm ~ OpenBSD/PackingList.pm
> a few more details so that infodirs become useful. (espie@)
~ pkg_create
> use the infodir property of packing-lists and specialfiles instead of
> passing extra parameters around. (espie@)
~ pkg_add
> set_infodir during pkg_add. (espie@)
~ OpenBSD/PackingList.pm
> set_infodir only if plist got defined. (espie@)
~ pkg_add ~ OpenBSD/Add.pm
~ OpenBSD/PackingElement.pm
> don't bypass add_object to create special files, otherwise infodir won't
> get defined.
> Simplify and clean-up register_installation: everything is in the
> packing-list
> so don't pass other arguments. Do not copy arbitrary info_names objects,
> but
> use a copy_info visitor which copies exactly what files are around.
> Use the stored infodir to locate DISPLAY and INSTALL files. (espie@)
~ pkg_add ~ OpenBSD/Add.pm
~ OpenBSD/Delete.pm
> rely on packing-lists knowing where they live, much saner. (espie@)
~ pkg_add
> move code a bit, in preparation for multiple handles in an UpdateSet.
> (espie@)
~ pkg_create
> in checksum creation, allow $base to be undef.
> This allows us to always compute checksums based on $self->fullname,
> so as a result, we no longer need to pass as many parameters around.
> Small negative side-effect: the method names no longer reflect reality
> well, they will need renaming at some point... (espie@)
~ pkg_create
> slightly more descriptive names. (espie@)
~ TODO ~ OpenBSD/Dependencies.pm
> In update mode, restore previous behavior: some updated packages may have
> names that match installed packages, but still mark them to be installed,
> so that they're returned as dependencies, and get updated first.
> Write in TODO that the current situation is brittle, solved dependencies
> should get adjusted in real time to be 100% sure this won't cause
> problems... (espie@)
sensorsd
~ sensorsd.c ~ sensorsd.conf.5
> for sensorsd.conf entries, let's do matching by sensor type if maching by
> complete sysctl name does not yield anything; ok henning, otto (cnst@)
~ sensorsd.conf.5
> tweaks; help/ok cnst (jmc@)
~ sensorsd.c
> Remove watch_cnt in preparation for my bigger patch that makes sensorsd
> usable by default, where we will monitor all sensors that automatically
> provide status, and this watch_cnt won't make much sense. Besides, upon
> startup, sensorsd already shows all sensors that it is going to monitor,
> making this watch_cnt rather unimportant.
> ok henning@ (cnst@)
tcpdump
~ print-cnfp.c
> Do not change snaplen when parsing a netflow packet (canacar@)
ypserv
~ ypserv/ypserv_db.c
> Remove clause 3 and 4. Ok with Chuck. -moj (maja@)
ztsscale
~ ztsscale.c ~ ztsscale.8
> adapt ztsscale and the installer to use wscons ioctls for calibration;
> ok deraadt@ (robert@)
===============================================================================
More information about the odc
mailing list