[owc] Weekly src changes ending 2007-05-20
OWC
auto at squish.net
Mon May 21 07:00:01 BST 2007
OpenBSD src changes summary for 2007-05-13 to 2007-05-20 inclusive
==================================================================
bin bin/rm
bin/systrace distrib/amd64
distrib/sets etc/MAKEDEV.common
etc/etc.vax/MAKEDEV etc/man.conf
etc/wsconsctl.conf gnu/usr.bin/perl
kerberosV lib/libc
lib/libcurses lib/libkvm
lib/libpthread lib/librthread
lib/libskey lib/libutil
libexec libexec/spamd
regress/lib regress/sbin
regress/share regress/usr.sbin
sbin sbin/disklabel
sbin/fsck_ffs sbin/isakmpd
sbin/mount_portal sbin/newfs
sbin/restore sbin/wsconsctl
share/man sys/arch/alpha/include
sys/arch/amd64/amd64 sys/arch/amd64/conf
sys/arch/amd64/include sys/arch/amd64/pci
sys/arch/arm/arm sys/arch/arm/conf
sys/arch/arm/footbridge sys/arch/arm/include
sys/arch/arm/xscale sys/arch/armish/armish
sys/arch/armish/include sys/arch/aviion/include
sys/arch/cats/cats sys/arch/cats/include
sys/arch/hp300/hp300 sys/arch/hp300/include
sys/arch/hppa/conf sys/arch/hppa/dev
sys/arch/hppa/hppa sys/arch/hppa/include
sys/arch/hppa64/include sys/arch/i386/i386
sys/arch/i386/include sys/arch/landisk/include
sys/arch/landisk/landisk sys/arch/luna88k/include
sys/arch/m68k/include sys/arch/m68k/m68k
sys/arch/m88k/include sys/arch/m88k/m88k
sys/arch/mac68k/include sys/arch/mac68k/mac68k
sys/arch/macppc/dev sys/arch/macppc/macppc
sys/arch/mips64/include sys/arch/mips64/mips64
sys/arch/mvme68k/include sys/arch/mvme68k/mvme68k
sys/arch/mvme88k/compile sys/arch/mvme88k/conf
sys/arch/mvme88k/dev sys/arch/mvme88k/include
sys/arch/mvme88k/mvme88k sys/arch/powerpc/include
sys/arch/powerpc/powerpc sys/arch/sgi/conf
sys/arch/sgi/dev sys/arch/sgi/include
sys/arch/sgi/pci sys/arch/sgi/sgi
sys/arch/sh/include sys/arch/sh/sh
sys/arch/sparc/include sys/arch/sparc64/include
sys/arch/sparc64/sparc64 sys/arch/vax/include
sys/arch/vax/vax sys/arch/zaurus/include
sys/arch/zaurus/zaurus sys/conf
sys/dev/cardbus sys/dev/ic
sys/dev/isa sys/dev/pci
sys/dev/pcmcia sys/dev/usb
sys/dev/wscons sys/kern
sys/lib/libkern sys/net
sys/netinet sys/nfs
sys/sys sys/ufs/ufs
sys/uvm usr.bin/asa
usr.bin/compress usr.bin/cvs
usr.bin/env usr.bin/make
usr.bin/man usr.bin/msgs
usr.bin/rcs usr.bin/ssh
usr.bin/sup usr.bin/tip
usr.bin/vi usr.bin/xinstall
usr.sbin usr.sbin/bgpd
usr.sbin/catman usr.sbin/dhcpd
usr.sbin/pkg_add usr.sbin/sasyncd
== bin =============================================================== 01/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/bin
bin
~ Makefile
> prettier Makefiles (deraadt@)
rm
~ rm.1
> simplify the description of -R; from Jeffrey 'jf' Lim (jmc@)
systrace
~ intercept.c
> don't assume 4-byte aligned memory when intercepting a string
> this should get rid of
> systrace: intercept_get_string: ioctl: Invalid argument
> no objections from provos, ok otto (sturm@)
== distrib =========================================================== 02/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/distrib
amd64
~ common/Makefile.inc ~ ramdisk_cd/Makefile
> Make FLOPPYSIZE be in terms of 512 byte blocks like all the other ports.
> This was something I missed in the initial "newfs -c" conversion.
> (millert@)
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/base/mi
~ lists/comp/mi ~ lists/man/mi
> sync (deraadt@)
~ lists/base/md.armish ~ lists/base/md.cats
~ lists/base/md.zaurus ~ lists/comp/md.armish
~ lists/comp/md.cats ~ lists/comp/md.zaurus
> sync (deraadt@)
~ lists/base/md.armish ~ lists/base/md.cats
~ lists/base/md.zaurus
> sync (deraadt@)
~ lists/base/mi
> sync (deraadt@)
~ lists/base/md.landisk ~ lists/comp/md.landisk
> sync (deraadt@)
~ lists/base/md.armish ~ lists/base/md.cats
~ lists/base/md.zaurus ~ lists/comp/md.armish
~ lists/comp/md.cats ~ lists/comp/md.zaurus
> sync (deraadt@)
~ lists/base/mi ~ lists/man/mi
> sync (deraadt@)
== etc =============================================================== 03/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/etc
MAKEDEV.common
~ MAKEDEV.common
> MASSBUSS -> MASSBUS; ok miod (jmc@)
etc.vax/MAKEDEV
~ etc.vax/MAKEDEV
> regen; (jmc@)
man.conf
~ man.conf
> the `doc' entry listed only out of date paths to sendmail docs, so
> remove it (updating it makes little sense);
> ok millert (jmc@)
~ man.conf
> add an entry for `.ms'; ok millert (jmc@)
~ man.conf
> - remove /usr/{X11,contrib,gnu} and /usr/*/man/old from the _default path
> man(1) searches, as well as from the numerical sections list, as they do
> not exist
> - remove the {contrib,new,old} sections, as they point to directories
> that do not exist
> - make the X11 section an alias for the X11R6 section (instead of
> pointing to a non-existent directory)
> ok millert (jmc@)
~ man.conf
> use "gzcat" instead of "gunzip -c", which is clearer, shorter,
> and matches "zcat" (as opposed to uncompress -c");
> grammar tweak whilst here too;
> ok millert (jmc@)
~ man.conf
> - move 3f down the search order list for _subdir
> - update paths for 3{F,f,P,p}
> ok espie millert steven (jmc@)
wsconsctl.conf
~ wsconsctl.conf
> commented out entry to show how to mute the keyboard bell
> plenty ok (tedu@)
== gnu =============================================================== 04/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/gnu
usr.bin/perl
~ lib/ExtUtils/MM_Unix.pm
> pass COPTS thru, okay millert@ (espie@)
~ lib/ExtUtils/MM_Any.pm
> hack generated Makefile so that section is properly 3p.
> Not really clean, but this code is bound to change in the next
> perl version anyways.
> okay millert@ (espie@)
== kerberosV ========================================================= 05/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/kerberosV
kerberosV
~ src/appl/afsutil/pagsh.c
> Check getpwuid() return value for NULL before dereferencing it.
> ok ray@ millert@ (moritz@)
== lib =============================================================== 06/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/lib
libc
~ arch/alpha/string/Makefile.inc ~ arch/amd64/string/Makefile.inc
~ arch/arm/string/Makefile.inc ~ arch/sh/string/Makefile.inc
~ arch/sparc64/string/Makefile.inc ~ arch/vax/string/Makefile.inc
> fix some lint 'xxx used, but not defined' warnings by properly adding
> stuff to LSRCS (otto@)
~ net/gethostnamadr.c ~ net/res_query.c
~ gen/getgrent.c
> Wrap some functions, labels, and variables in #ifdefs corresponding
> to #ifdefs where they are used. Found by lint by bret dot lambert
> at gmail.
> OK deraadt at . (ray@)
~ arch/arm/string/Makefile.inc
> Unbreak libc on arm. -moj (maja@)
~ net/getaddrinfo.c
> Improve reentrancy by not using global variable. Removes two
> wrappers around fopen/fclose. From tbert.
> OK millert@, tedu@, and itojun at . (ray@)
~ arch/arm/string/Makefile.inc ~ arch/vax/string/Makefile.inc
> LSRCS should refer to the C source version of asm files 'put it in'
> deraadt@ (drahn@)
~ gen/psignal.c
> Don't save strlen results in an int. Remove a variable that's used
> only once while here. From tbert.
> OK jaredy at . (ray@)
~ net/getaddrinfo.c
> Remove no-op code. From tbert.
> OK moritz@ and jaredy@ (ray@)
~ net/getaddrinfo.c
> Typos.
> OK jmc at . (ray@)
libcurses
~ tty/lib_mvcur.c
> Check fgets return value, from Charles Longeau.
> OK millert at . (ray@)
libkvm
~ kvm_arm.c
> Kernel crash dump support for arm-based platform, with minimal support in
> libkvm, but all the necessary information for a complete _kvm_kvatop() is
> available in the crash dump. (miod@)
libpthread
~ uthread/uthread_fd.c ~ uthread/uthread_file.c
~ uthread/uthread_info_openbsd.c ~ uthread/uthread_init.c
~ uthread/uthread_kern.c ~ uthread/uthread_priority_queue.c
~ uthread/uthread_rwlock.c ~ uthread/uthread_select.c
~ uthread/uthread_sem.c ~ uthread/uthread_sig.c
~ uthread/uthread_stack.c ~ uthread/uthread_stackseg_np.c
~ uthread/uthread_write.c ~ uthread/uthread_writev.c
> Eliminate many lint warnings by either: using the appropriate type,
> casting when safe or adding ARGSUSED where needed. Reviewed and
> improvements from millert@ and marc at . okay marc@ (kurt@)
librthread
~ rthread.c
> Register the locking hooks with ld.so.
> drahn@ ok (long time ago) (art@)
libskey
~ skeysubr.c
> Check fgets return value, from Charles Longeau.
> OK millert at . (ray@)
libutil
~ fparseln.3
> Make it clear, that the line returned does not have a
> trailing newline in contrast to fgetln(3).
> ok ray@ jmc@ millert@ (moritz@)
== libexec =========================================================== 07/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/libexec
libexec
~ Makefile
> prettier Makefiles (deraadt@)
spamd
~ spamd.8
> spamd-setup needs -b for blacklist-only mode; from Nick Templeton (jmc@)
== regress =========================================================== 08/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/regress
lib
~ libc/Makefile + libc/basename/Makefile
+ libc/basename/basename_test.c + libc/dirname/Makefile
+ libc/dirname/dirname_test.c
> Add basename(3) and dirname(3) regression tests, from tbert.
> OK otto at . (ray@)
sbin
~ newfs/checknewfs
> detect if newfs fails and add an extra test (amd64 floppy) (otto@)
share
~ man/cfman
> remove catman reference from SEE ALSO; (jmc@)
usr.sbin
~ pkg_add/check-name
> fix tests to run now (espie@)
== sbin ============================================================== 09/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sbin
sbin
~ Makefile
> unify the way we prefer other things (deraadt@)
~ Makefile
> prettier Makefiles (deraadt@)
disklabel
~ disklabel.c
> Sync editit() with sendbug.
> OK krw at . (ray@)
fsck_ffs
~ pass5.c
> Fix mismerge; remove statement without effect. OK otto@ pedro@ (millert@)
isakmpd
~ message.c ~ message.h
~ udp_encap.c ~ virtual.c
TAGGED OPENBSD_4_0
> Original Commit by hshoexer@, original commit message:
> ===========================================================================
> Fix interop-issue with vpn peers that start reyking on port 4500 when
> NAT-T is used. Solves problems with cisco and openswan.
> Tested by todd@ (cisco interop), ok ho@
> Original fix with Stefan Roth (stefan dot roth at siemens dot com),
> thanks!
> ===========================================================================
> (ckuethe@)
~ message.c ~ message.h
~ udp_encap.c ~ virtual.c
TAGGED OPENBSD_4_1
> Original Commit by hshoexer@, original commit message:
> ===========================================================================
> Fix interop-issue with vpn peers that start reyking on port 4500 when
> NAT-T is used. Solves problems with cisco and openswan.
> Tested by todd@ (cisco interop), ok ho@
> Original fix with Stefan Roth (stefan dot roth at siemens dot com),
> thanks!
> ===========================================================================
> (ckuethe@)
mount_portal
~ mount_portal.8
> Mention the mfs and portal papers. (jsg@)
newfs
~ newfs.c
> Make usage declaration match prototype; Make it __dead also.
> ok otto@ (thib@)
~ mkfs.c
> Do a bounds check to see if the inode table fits on the disk.
> Prevents a core dump found by djm@ when doing mount_mfs -s 80
> (yeah, that's a tiny file system); ok millert@ pedro@ (otto@)
~ mkfs.c ~ newfs.8
~ newfs.c
> The unit of the -c options is fragments, so adjust man page and
> variable name; correct the loop packing more inodes into the cg:
> in some cases it could put more fragments into the cg than requested;
> give an error if the -c option cannot be honoured. ok millert@ pedro@
> (otto@)
~ mkfs.c ~ newfs.8
~ newfs.c
> backout -c changes; they were not tested on all the media. all newfs
> changes obviously *MUST* be checked against whatever weird things the
> floppies do to fit (deraadt@)
~ newfs.8
> Mention the mfs and portal papers. (jsg@)
~ newfs.8
> kill trailing whitespace; (jmc@)
~ mkfs.c ~ newfs.8
~ newfs.c
> Now that the amd64 floppy and cd image -c setting have been fixed,
> recommit the -c changes. For now, we warn only if we cannot honour -c.
> ok millert@ deraadt@ (otto@)
~ newfs.c
> Update usage wrt -c option when called as mount_mfs. (millert@)
restore
~ restore.c
> g/c unused vars (millert@)
wsconsctl
~ wsconsctl.8 ~ wsconsctl.c
> add a -f option to allow specifing an alternative
> control device; "put it in" miod@ (robert@)
== share ============================================================= 10/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/share
man
~ man9/mutex.9
> reference rwlock(9) (jsg@)
~ man4/umsm.4
> correct tyop. it is called EVDO, not EDVO. (fkr@)
~ man9/atomic.9
> these functions first appeared in 4.1, not 4.0;
> from jung; ok art (jmc@)
~ man9/mutex.9 ~ man9/rwlock.9
> Add notes about sleeping with held locks and interrupt handler usage.
> Based on comments by art, and tweaked by jmc.
> ok art@ (jsg@)
~ man4/pci.4 ~ man4/Makefile
+ man4/sli.4
> manpage for the emulex driver. very basic, it'll get fleshed out as we go
> (dlg@)
~ man4/sli.4 ~ man4/pci.4
> make sli's .Nd match other scsi interfaces; ok dlg (jmc@)
~ man4/isp.4 ~ man4/pci.4
~ man4/sbus.4
> FibreChannel -> Fibre Channel, to be consistent with the rest
> of our pages;
> ok dlg (jmc@)
~ man8/release.8
> Starting with xenocara, /etc/X11/xdm will be overwritten during the
> install phase.
> discussed with, suggesstions by and ok from jmc@ matthieu@ (kili@)
~ man8/man8.vax/MAKEDEV.8
> regen; (jmc@)
~ man4/uts.4
> add the panels made by Hantouch (http://www.hantouch.com/) to the list
> of working devices (robert@)
~ man4/uts.4
> sort the hardware list; (jmc@)
== sys =============================================================== 11/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sys
arch/alpha/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ _types.h
> oops; martin go see what you did here.. (deraadt@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ intr.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/amd64/amd64
~ autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ pmap.c
> Switch amd64 to VM_PAGE_MD. Mostly just imitating i386. flags in pg_flags.
> deraadt@ ok (art@)
~ pmap.c
> Instead of checking whichqs directly, add a "sched_is_idle()" macro to
> sys/sched.h and use that to check if there's something to do.
> kettenis@ thib@ ok (art@)
arch/amd64/conf
~ GENERIC
> remove the COMPAT entries. they are not useful here (tedu@)
~ GENERIC
> aps works on amd64, ok jsg (tedu@)
~ GENERIC ~ files.amd64
> (disabled) option for PCIAGP. at least lets i945 chipsets work
> ok dim (tedu@)
arch/amd64/include
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ pmap.h ~ vmparam.h
> Switch amd64 to VM_PAGE_MD. Mostly just imitating i386. flags in pg_flags.
> deraadt@ ok (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ intr.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/amd64/pci
~ pchb.c
> (disabled) option for PCIAGP. at least lets i945 chipsets work
> ok dim (tedu@)
arch/arm/arm
~ ast.c ~ cpuswitch.S
~ genassym.cf
> Switch arm to __HAVE_CPUINFO. Least effort.
> deraadt@ ok (art@)
~ fault.c
> External bus errors should not occur, so panic. ok deraadt@ (drahn@)
~ pmap.c
> Instead of checking whichqs directly, add a "sched_is_idle()" macro to
> sys/sched.h and use that to check if there's something to do.
> kettenis@ thib@ ok (art@)
~ stubs.c
> Kernel crash dump support for arm-based platform, with minimal support in
> libkvm, but all the necessary information for a complete _kvm_kvatop() is
> available in the crash dump. (miod@)
arch/arm/conf
~ files.footbridge
> Dummy mutex code for arm platforms. ok drahn@ deraadt@ (miod@)
arch/arm/footbridge
~ footbridge_intr.h + footbridge_mutex.c
+ footbridge_mutex.h
> Dummy mutex code for arm platforms. ok drahn@ deraadt@ (miod@)
arch/arm/include
~ _types.h ~ cpu.h
> Switch arm to __HAVE_CPUINFO. Least effort.
> deraadt@ ok (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
+ kcore.h
> Kernel crash dump support for arm-based platform, with minimal support in
> libkvm, but all the necessary information for a complete _kvm_kvatop() is
> available in the crash dump. (miod@)
arch/arm/xscale
~ files.i80321 ~ files.pxa2x0
~ pxa2x0_intr.h + i80321_mutex.c
+ i80321_mutex.h + pxa2x0_mutex.c
+ pxa2x0_mutex.h
> Dummy mutex code for arm platforms. ok drahn@ deraadt@ (miod@)
~ i80321_intr.c ~ pxa2x0_intr.c
~ pxa2x0_intr.h
> Implement splassert for armish and zaurus. (miod@)
arch/armish/armish
~ autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ armish_machdep.c ~ autoconf.c
> Kernel crash dump support for arm-based platform, with minimal support in
> libkvm, but all the necessary information for a complete _kvm_kvatop() is
> available in the crash dump. (miod@)
arch/armish/include
~ _types.h ~ armish_intr.h
~ intr.h + mutex.h
> Dummy mutex code for arm platforms. ok drahn@ deraadt@ (miod@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ armish_intr.h
> Implement splassert for armish and zaurus. (miod@)
arch/aviion/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
arch/cats/cats
~ autoconf.c ~ cats_machdep.c
> Kernel crash dump support for arm-based platform, with minimal support in
> libkvm, but all the necessary information for a complete _kvm_kvatop() is
> available in the crash dump. (miod@)
arch/cats/include
~ _types.h ~ intr.h
+ mutex.h
> Dummy mutex code for arm platforms. ok drahn@ deraadt@ (miod@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
arch/hp300/hp300
~ locore.s ~ trap.c
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
arch/hp300/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ intr.h
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
arch/hppa/conf
~ GENERIC
> Uncomment xl(4). (kettenis@)
~ RAMDISK
> Uncomment entry for xl(4) here too. (kettenis@)
~ RAMDISK
> Uncomment em(4) entry. From brad at comstyle dot com. (kettenis@)
~ GENERIC ~ RAMDISK
> Add some phy's found on certain xl(4) models. From brad at comstyle dot
> com. (kettenis@)
arch/hppa/dev
~ cpudevs.h ~ cpudevs_data.h
> regen (kettenis@)
~ cpudevs
> Add IDs for 8700/pcxw2 cpu and fpu. (kettenis@)
~ dino.c
> Fix io space access.
> tested by krw@, miod@, ok deraadt@ (kettenis@)
arch/hppa/hppa
~ genassym.cf ~ locore.S
~ machdep.c ~ trap.c
> move hppa to __HAVE_CPUINFO
> input from miod@, ok kettenis@ (martin@)
~ machdep.c
> Limit hppa buffer cache to 30%. It used to be 70% because stack garbage
> that was used for minaddr instead of the proper vm_map_min(kernel_map)
> (fixed in this diff).
> From mickey. art@ ok (art@)
arch/hppa/include
~ _types.h ~ cpu.h
~ intr.h
> move hppa to __HAVE_CPUINFO
> input from miod@, ok kettenis@ (martin@)
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ cpu.h
> Userland has the right to compile, too; not sure this will be enough yet
> but deraadt@ insists this goes in now. (miod@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ intr.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/hppa64/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ cpu.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/i386/i386
~ machdep.c
> Enable powernow on amd64 processors in the GENERIC.MP.
> ok tedu (gwk@)
~ machdep.c
> SMALL_KERNEL unused bits (deraadt@)
~ machdep.c
> _bus_dmamap_sync() is unused; ok miod (deraadt@)
~ autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ machdep.c
> make splassert(IPL_NONE) work on i386/MP (the single-processor case is
> a bit more hairy).
> deraadt@ ok (art@)
~ gdt.c
> There's no reason to keep track of gdt_count, we only use it for
> copying the intial gdt when spinning up secondary cpus and then
> we can simply use gdt_size for that.
> From mickey, art@ ok (art@)
~ pmap.c
> Since we're initializing everything in pmap_pinit, it's probably a good
> idea to initialize pm_cpus as well. Otherwise we'll get stray tlb
> shootdowns and pm_cpus is not clear on exit sometimes. (art@)
~ pmap.c
> addess -> address (miod@)
arch/i386/include
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ intr.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/landisk/include
~ intr.h
> forgot this one in my previous commit.
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
~ vmparam.h
> Fix KERNBASE, so that profiling kernels initialize (but profiling is broken
> at the moment). (miod@)
arch/landisk/landisk
~ autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
arch/luna88k/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
arch/m68k/include
~ _types.h ~ asm.h
~ cpu.h ~ param.h
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ psl.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/m68k/m68k
~ genassym.cf ~ m68k_machdep.c
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
arch/m88k/include
~ cpu.h
> Work in progress IPI mechanism, currently only implemented on MVME188, to
> send clock ticks to secondary processors. (miod@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ intr.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
~ lock.h
> Simpler asm constraints for simplelock operations. (miod@)
~ cpu.h
> Send an IPI in signotify() if the process runs on a different processor,
> similar to the fix which went into i386 and amd64 a few weeks ago. (miod@)
arch/m88k/m88k
~ mutex.S
> Oops, correctly handle spl-less mutexes. (miod@)
~ mutex.S
> Revert previous revision, and do it again correctly. (miod@)
~ eh_common.S
> Move proc_do_uret() around so that it can fall through no_ast instead of
> jumping to it. No functional change. (miod@)
~ m88k_machdep.c
> In spl0(), really process soft interrupts at IPL_SOFT instead of whatever
> level we were at. (miod@)
~ db_interface.c
> Force other processors to spin when one is in ddb. (miod@)
~ m88k_machdep.c
> Send an IPI in signotify() if the process runs on a different processor,
> similar to the fix which went into i386 and amd64 a few weeks ago. (miod@)
~ m88100_machdep.c ~ m8820x_machdep.c
> Since we no longer use 3 bits but the whole 7 to get the processor revision
> number, we should test for 10, not 2, as the revision for which the xxx.usr
> errata applies; also, going through the errata, revision 2/10 (1010x) _is_
> affected. (miod@)
arch/mac68k/include
~ intr.h
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
arch/mac68k/mac68k
~ locore.s ~ trap.c
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
~ autoconf.c
> provide dummy device_register() (martin@)
~ intr.c
> On AV-class machines, SCSI interrupts still come from VIA2, not PSC, so
> IPL_BIO should not be modified.
> Found the hard way by martin@ (miod@)
arch/macppc/dev
~ kiicvar.h ~ kiic.c
~ piic.c ~ smu.c
> lockmgr -> rwlock for piic(4), kiic(4) and smu(4);
> smu(4) tested by kettenis@, ok krw@ (thib@)
arch/macppc/macppc
~ cpu.c
> Use value passed as in ca_reg as cpuid instead of unit number.
> ok drahn@ (kettenis@)
~ autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
arch/mips64/include
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ trap.h
> Remove unused badaddr(). (miod@)
arch/mips64/mips64
~ lcore_access.S
> Remove unused badaddr(). (miod@)
arch/mvme68k/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ intr.h
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
arch/mvme68k/mvme68k
~ locore.s ~ trap.c
> switch m68k to __HAVE_CPUINFO
> help miod@, art@
> ok deraadt@ (martin@)
arch/mvme88k/compile
~ .cvsignore
> Add a GENERIC.MP configuration, not for the faint of the heart, if you want
> to play. (miod@)
arch/mvme88k/conf
+ GENERIC.MP
> Add a GENERIC.MP configuration, not for the faint of the heart, if you want
> to play. (miod@)
arch/mvme88k/dev
~ dart.c
> Silence a diagnostic message in multiprocessor kernels, it makes things
> worse. (miod@)
arch/mvme88k/include
~ cpu.h ~ mvme188.h
> Work in progress IPI mechanism, currently only implemented on MVME188, to
> send clock ticks to secondary processors. (miod@)
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ mvme188.h
> Try reducing the number of IPIs, by only reinflicting them to us if we are
> lowering the spl.
> Also, warn and halt in tracks if the interrupt pin of a secondary cpu never
> clears (found the hard way in one of the 2P256 modules here), since there
> is
> nothing better we can do.
> Last, do not attempt to handle VME interrupts on secondary processors yes
> (this confuses the bus, since both processors will perform a VME interrupt
> acknowledge sequence, but only one will succeed).
> This lets processes correctly run on all the available processors in
> single-user mode, but there are still issues to sort out. (miod@)
arch/mvme88k/mvme88k
~ m188_machdep.c ~ m1x7_machdep.c
> Register statclock() at IPL_STATCLOCK, not IPL_CLOCK (no-op since they are
> the same) (miod@)
~ m187_machdep.c ~ m197_machdep.c
> Hold kernel_lock when processing interrupts at a level under IPL_SCHED, as
> done on 188, for MULTIPROCESSOR kernels. (miod@)
~ m188_machdep.c ~ machdep.c
> Work in progress IPI mechanism, currently only implemented on MVME188, to
> send clock ticks to secondary processors. (miod@)
~ m188_machdep.c
> More work in IPI processing: actual processing is now done with interrupts
> enabled, and is delayed if the current spl is not high enough; setipl() and
> raiseipl() will retrigger an IPI interrupt if there are any pending IPIs.
> (miod@)
~ m188_machdep.c
> Try reducing the number of IPIs, by only reinflicting them to us if we are
> lowering the spl.
> Also, warn and halt in tracks if the interrupt pin of a secondary cpu never
> clears (found the hard way in one of the 2P256 modules here), since there
> is
> nothing better we can do.
> Last, do not attempt to handle VME interrupts on secondary processors yes
> (this confuses the bus, since both processors will perform a VME interrupt
> acknowledge sequence, but only one will succeed).
> This lets processes correctly run on all the available processors in
> single-user mode, but there are still issues to sort out. (miod@)
~ m188_machdep.c
> Force other processors to spin when one is in ddb. (miod@)
~ m188_machdep.c
> Oops, get the ``disabling cpu'' message correct. (miod@)
arch/powerpc/include
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ intr.h
> Use atomic operations in set_sint() instead of disabling interrupts in the
> function. ok kettenis@ (miod@)
arch/powerpc/powerpc
~ pmap.c
> Ansi prototypes, not K&R. no binary difference. (drahn@)
arch/sgi/conf
~ GENERIC ~ RAMDISK
> option AHC_CTRL_ID is not used. (miod@)
arch/sgi/dev
~ mavb.c
> Check for codec presence in match(), and don't attach if none is found.
> ok kettenis@ (miod@)
arch/sgi/include
~ mutex.h
> Oops, IPL_NONE is not zero in HEAD. (miod@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ autoconf.h
> Remove unused badaddr(). (miod@)
arch/sgi/pci
- pciide_machdep.c ~ pci_machdep.h
> Remove irrelevant code and defines. (miod@)
arch/sgi/sgi
~ mutex.c
> Oops, IPL_NONE is not zero in HEAD. (miod@)
~ autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
arch/sh/include
~ _types.h ~ cpu.h
~ userret.h
> Switch sh to __HAVE_CPUINFO. The least possible effort for now.
> deraadt@ ok (art@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ vmparam.h
> Fix KERNBASE, so that profiling kernels initialize (but profiling is broken
> at the moment). (miod@)
arch/sh/sh
~ cpu.c ~ genassym.cf
~ vectors.S
> Switch sh to __HAVE_CPUINFO. The least possible effort for now.
> deraadt@ ok (art@)
~ locore_c.c
> Instead of checking whichqs directly, add a "sched_is_idle()" macro to
> sys/sched.h and use that to check if there's something to do.
> kettenis@ thib@ ok (art@)
arch/sparc/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ psl.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/sparc64/include
~ cpu.h
> Move want_resched into struct cpu_info. (kettenis@)
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ psl.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/sparc64/sparc64
~ autoconf.c
> Put names we get from the prom between double quotes when we print them.
> ok deraadt@ (kettenis@)
~ locore.s
> Nuke code to support the Solaris bootloader and old 32-bit bootloaders.
> Remove bits of code that repeat what the bootloader already did for us,
> like setting up the stack and clearing .bss.
> ok deraadt@ (kettenis@)
~ genassym.cf ~ locore.s
~ trap.c
> Move want_resched into struct cpu_info. (kettenis@)
~ locore.s
> We will never call OpenFirmware from 32-bit code.
> ok art@, deraadt@ (kettenis@)
~ locore.s
> Remove unused variable. (kettenis@)
~ pmap.c
> Replace ASSERT with KDASSERT and get rid of its definition; while there,
> remove some #if 0 dead code which won't get ressurrected anytime soon.
> (miod@)
arch/vax/include
~ _types.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ cpu.h ~ intr.h
> Switch vax to __HAVE_CPUINFO, based on an incomplete diff from art. (miod@)
~ _types.h
> Oops, forgot this file in the cpuinfo changes. (miod@)
~ _types.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ intr.h
> splassert_ctl defaults to 1 now, so dont wrap the checks for
> splassert_ctl > 0 in __predict_false().
> ok deraadt@ (thib@)
arch/vax/vax
~ genassym.cf ~ machdep.c
~ subr.s ~ trap.c
> Switch vax to __HAVE_CPUINFO, based on an incomplete diff from art. (miod@)
~ machdep.c
> we need a cpu_info_store; how did miod miss this? (deraadt@)
~ pmap.c
> addess -> address (miod@)
arch/zaurus/include
~ _types.h ~ intr.h
+ mutex.h
> Dummy mutex code for arm platforms. ok drahn@ deraadt@ (miod@)
~ _types.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
arch/zaurus/zaurus
~ autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ autoconf.c ~ zaurus_machdep.c
> Kernel crash dump support for arm-based platform, with minimal support in
> libkvm, but all the necessary information for a complete _kvm_kvatop() is
> available in the crash dump. (miod@)
conf
~ files
> start work on a driver for emulex fibre channel controllers.
> it is split up into a bus independant chunk (dev/ic/sli*) and the attach
> glue (dev/pci/sli_pci.c) cos there are sbus varaints of this chip which i
> hope to get my hands on one day.
> this does nothing except attach to the LP8000 controllers. (dlg@)
~ files
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
dev/cardbus
~ ehci_cardbus.c
> Convert ehci and ucom to rwlock.
> Kill the usb specific lockmgr wrapper as nothing uses it now.
> ok thib@ (jsg@)
dev/ic
~ aic7xxx.c ~ aic7xxx_inline.h
~ aic7xxxvar.h
> delete unused code; ok krw (deraadt@)
+ sli.c + slireg.h
+ slivar.h
> start work on a driver for emulex fibre channel controllers.
> it is split up into a bus independant chunk (dev/ic/sli*) and the attach
> glue (dev/pci/sli_pci.c) cos there are sbus varaints of this chip which i
> hope to get my hands on one day.
> this does nothing except attach to the LP8000 controllers. (dlg@)
~ slireg.h ~ slivar.h
> map the pci BARs that we're interested in. this gives us access to the
> chips registers and io interface, but whats actually in them is still a
> mystery to me. (dlg@)
~ slireg.h
> add definitions for the bits in the "register" bus_space region. (dlg@)
~ gemreg.h
> spelling in comment fix (dlg@)
~ sli.c ~ slivar.h
> hook up interrupts, but a little different to what i usually do. this maps
> the interrupt before the bus independant code is attached, and establishes
> the interrupt after that. (dlg@)
~ slivar.h
> stash the dma tag in sli_softc (dlg@)
~ slireg.h
> command ids for mailbox commands (dlg@)
~ xl.c
> My 3c905C needs a brief pause after reset for PIO too, at least on hppa.
> tested by thib@, ok deraadt@ (kettenis@)
dev/isa
~ aps.c
> the thinkpad 60 models are in "state" 5.
> ok deraadt jasper jsg mk (tedu@)
dev/pci
~ viaenv.c
> Remove crufty unused code.
> ok kettenis@ krw@ henning@ (jsg@)
~ ahc_pci.c
> delete unused code; ok krw (deraadt@)
~ if_msk.c
> Properly wrap mii_tick() up in splnet().
> This seems to fix a bizzare problem I was having for some time with msk on
> MP amd64 where the machine would reboot itself under load.
> At least sk(4) and rl(4) code seems to have the potential for a similiar
> problem.
> ok kettenis@ (jsg@)
~ files.pci + sli_pci.c
> start work on a driver for emulex fibre channel controllers.
> it is split up into a bus independant chunk (dev/ic/sli*) and the attach
> glue (dev/pci/sli_pci.c) cos there are sbus varaints of this chip which i
> hope to get my hands on one day.
> this does nothing except attach to the LP8000 controllers. (dlg@)
~ sli_pci.c
> map the pci BARs that we're interested in. this gives us access to the
> chips registers and io interface, but whats actually in them is still a
> mystery to me. (dlg@)
~ if_tht.c
> add support for multicast addresses. this adds the add/del multi handlers
> to the ioctl, and programs the multicast filter in tht_iff.
> we are not using the perfect multicast filters, only the imperfect one. i
> think there is a requirement that addresses in the perfect filter be in
> order, but im not sure how to do that nicely (yet).
> this has taken a long time because i dont know how to write c. (dlg@)
~ if_tht.c
> support changing the mtu up to the maximum frame size it currently
> supports (MCLBYTES). (dlg@)
~ if_tht.c
> copy the lladdr from the arpcom struct onto the hardware when the interface
> is brought up.
> this driver now supports all the features our network layer expects of it.
> (dlg@)
~ sli_pci.c
> hook up interrupts, but a little different to what i usually do. this maps
> the interrupt before the bus independant code is attached, and establishes
> the interrupt after that. (dlg@)
~ sli_pci.c
> stash the dma tag in sli_softc (dlg@)
~ pcidevs
> Add Radeon X700 PCIE, Radeon X700 PCIE Secondary and Marvell 88SE6145 SATA.
> (moritz@)
~ pcidevs.h ~ pcidevs_data.h
> regen (moritz@)
~ ehci_pci.c
> Convert ehci and ucom to rwlock.
> Kill the usb specific lockmgr wrapper as nothing uses it now.
> ok thib@ (jsg@)
~ arc.c
> Add missing arguments to printf.
> OK tedu@, dlg at . (ray@)
dev/pcmcia
~ pcmciadevs.h
> regen (mglocker@)
~ pcmciadevs
> Add vendor and product; AmbiCom, Inc. WL54CF. (mglocker@)
dev/usb
~ umsm.c
> sync comment with reality (fkr@)
~ if_axe.c ~ if_axereg.h
> Convert from lockmgr to rwlock.
> Initial diff from krw@, uneeded lock_mii()/unlock_mii() on all commands
> in existing code removed by me so it doesn't try to recurse.
> ok krw@ (jsg@)
~ uts.c
> fix the check of wsmouse_calibcoords values and
> invert the X or Y coordinates to make the top left
> corner (0,0). (robert@)
~ ehci.c ~ ehcivar.h
~ ucom.c ~ usb_port.h
> Convert ehci and ucom to rwlock.
> Kill the usb specific lockmgr wrapper as nothing uses it now.
> ok thib@ (jsg@)
~ usbdevs.h ~ usbdevs_data.h
> regen (robert@)
~ usbdevs
> add Brother MFC 210C and Logitech QuickCam Notebook Deluxe IDs;
> from brad. (robert@)
~ usbdivar.h
> addess -> address (miod@)
dev/wscons
~ wsmux.c
> wsmouse.h is needed for the NWSMOUSE test to actually work
> found by and ok art (tedu@)
kern
~ kern_rwlock.c
> print lock name in some panic messages so we know what's up
> ok krw marco pedro (tedu@)
~ subr_autoconf.c
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ subr_disk.c
> now that setroot() prints nice things like:
> root on sd0a swap on sd0b dump on sd0b
> we can skip printing:
> rootdev=0x400 rrootdev=0xd00 rawdev=0xd02
> ok miod, no objections from other lazy slackers (deraadt@)
- subr_mutex.c
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ init_main.c ~ kern_clock.c
~ kern_fork.c ~ kern_ktrace.c
~ kern_subr.c ~ kern_synch.c
~ kern_sysctl.c ~ sched_bsd.c
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ vfs_subr.c
> Collapse struct v_selectinfo in struct vnode, remove the
> simplelock and reuse the name for the selinfo member.
> Clean-up accordingly.
> ok tedu@,art@ (thib@)
~ subr_disk.c
> in case of pre-configured nfs, we must assume that the bootdv is the rootdv
> (deraadt@)
~ sched_bsd.c
> Widen the SCHED_LOCK in two cases to protect p_estcpu and p_priority.
> kettenis@ ok (art@)
~ subr_disk.c
> Dont write to the string passed to parsedisk(), ok deraadt@ (drahn@)
lib/libkern
~ arch/hppa/bcopy.m4
> move hppa to __HAVE_CPUINFO
> input from miod@, ok kettenis@ (martin@)
net
~ if_bridge.c
> Update the ifp of bridge cache entries if the entry is not static. This
> makes carp(4) fail-over work over bridge(4). OK michele@ jason@ (claudio@)
~ if_ethersubr.c
> count the number of multicast ranges that are kept in the list of multicast
> addresses in struct arpcom. this lets a nic driver easily see if it wants
> allmulti behaviour, which in turn means we can clean some code up.
> ok jason@ claudio@ norby@ (dlg@)
~ if_bridge.c ~ if_ethersubr.c
> fix return values in bridge_output.
> OK claudio@ jason@ (michele@)
~ if_spppsubr.c
> Handle IP address stuff in tls/tlf state messages
> instead of open/close events. Fixes PR 5457 where
> an extra open event arrived due to an authentication
> timeout/retransmission and confused things. Tested
> by RD Thrush and mpf@, discussions and ok mpf@ (canacar@)
netinet
~ if_ether.h
> count the number of multicast ranges that are kept in the list of multicast
> addresses in struct arpcom. this lets a nic driver easily see if it wants
> allmulti behaviour, which in turn means we can clean some code up.
> ok jason@ claudio@ norby@ (dlg@)
nfs
~ nfs_kq.c
> Collapse struct v_selectinfo in struct vnode, remove the
> simplelock and reuse the name for the selinfo member.
> Clean-up accordingly.
> ok tedu@,art@ (thib@)
sys
~ exec_elf.h
> uint32_t is not spelled u_long. makes the 32-bit NO_ADDR potentially
> useful on
> 64-bit machines, even though it's not used. ok art deraadt millert tom
> (tedu@)
~ device.h
> kill __HAVE_DEVICE_REGISTER by requiring all architectures to have a
> device_register() function -- even if it does nothing. reduces the
> cpp-based blather different between architectures
> idea ok'd by miod; tested on all architectures (except a few miod will
> need to cleanup because he has them) (deraadt@)
~ mutex.h
> Remove the MI implementation of mutexes and remove the __HAVE_MUTEX
> option. Every architecture implements mutexes now. (art@)
~ kernel.h ~ proc.h
~ sched.h
> The world of __HAVEs and __HAVE_NOTs is reducing. All architectures
> have cpu_info now, so kill the option.
> eyeballed by jsg@ and grange@ (art@)
~ vnode.h
> Collapse struct v_selectinfo in struct vnode, remove the
> simplelock and reuse the name for the selinfo member.
> Clean-up accordingly.
> ok tedu@,art@ (thib@)
~ sched.h
> Instead of checking whichqs directly, add a "sched_is_idle()" macro to
> sys/sched.h and use that to check if there's something to do.
> kettenis@ thib@ ok (art@)
ufs/ufs
~ ufs_vnops.c
> Collapse struct v_selectinfo in struct vnode, remove the
> simplelock and reuse the name for the selinfo member.
> Clean-up accordingly.
> ok tedu@,art@ (thib@)
uvm
~ uvm_glue.c
> Instead of a silly loop with goto, just use while(1). (art@)
~ uvm_page.c
> Instead of checking whichqs directly, add a "sched_is_idle()" macro to
> sys/sched.h and use that to check if there's something to do.
> kettenis@ thib@ ok (art@)
== usr.bin =========================================================== 12/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin
asa
~ asa.c
> Fix fgetln(3) handling of lines without trailing newline.
> From Tobias Stoeckmann.
> ok ray@ millert@ (moritz@)
~ asa.c
> - Rewrite argument handling to not pass closed fd's to asa()
> - Reorder fgetln() handling to prevent duplicate code
> - Some more cleanup
> Initial diff from Tobias Stoeckmann.
> ok jaredy@ (moritz@)
compress
~ Makefile ~ zmore
~ zmore.1
> Add support for zmore being called as zless,
> same behaviour but pager defaults to less instead of more.
> Incorporating feedback from millert@ and jmc@
> ok millert@ simon@ (jsg@)
~ zmore.1
> since less(1) and more(1) are the same point, don;t list them both
> in SEE ALSO; (jmc@)
cvs
~ client.c
> open() returns -1 on error, not NULL. From Tobias Stoeckmann. (xsa@)
~ remote.c
> In remote setup, write sent files to inlog (if specified) instead of
> outlog.
> Matches GNU CVS' behaviour. From Tobias Stoeckmann. (xsa@)
~ entries.c
> Figure out DST by setting tm_isdst element to -1.
> From Tobias Stoeckmann. OK otto at . (xsa@)
~ client.c
> Remove nonsensical dereference. From tbert.
> OK niallo && xsa (ray@)
env
~ env.1
> rewrite DIAGNOSTICS to avoid confusion about which application returns
> what, and when;
> started by Jeffrey 'jf' Lim, with some correction from Ingo Schwarze;
> ok otto (jmc@)
make
~ lowparse.c
> Remove extra return statement, from tbert from lint.
> OK jaredy@ and moritz@ (ray@)
man
~ man.1
> improve the descriptions of "section" and "subsection"; ok millert (jmc@)
msgs
~ msgs.c
> Check getpwuid() return value for NULL before dereferencing it.
> ok ray@ millert@ (moritz@)
rcs
~ rcs.c
> Don't free unused variable, from tbert.
> OK niallo && xsa (ray@)
ssh
~ log.c
> save and restore errno when logging; ok deraadt@ (djm@)
~ sftp-server.c
> bz#1286 stop reading and processing commands when input or output buffer
> is nearly full, otherwise sftp-server would happily try to grow the
> input/output buffers past the maximum supported by the buffer API and
> promptly fatal()
> based on patch from Thue Janus Kristensen; feedback & ok dtucker@ (djm@)
~ sshconnect2.c
> fall back to gethostname() when the outgoing connection is not
> on a socket, such as is the case when ProxyCommand is used.
> Gives hostbased auth an opportunity to work; bz#616, report
> and feedback stuart AT kaloram.com; ok markus@ (djm@)
~ monitor.c
> pass received SIGINT from monitor to postauth child so it can clean
> up properly. bz#1196, patch from senthilkumar_sen AT hotpop.com;
> ok markus@ (djm@)
~ sshconnect2.c
> djm owes me a vb and a tism cd for breaking ssh compilation (jolan@)
sup
~ src/supcmisc.c
> Check getpwnam() return value for NULL before dereferencing it.
> ok ray@ millert@ (moritz@)
tip
~ acu.c ~ cmds.c
~ cu.c ~ log.c
~ remote.c ~ tip.c
~ tip.h ~ value.c
~ vars.c
> Remove useless NOxxx defines and use NULL instead.
> No binary change.
> OK ray@ (moritz@)
~ aculib/hayes.c
> Add missing ssize_t rlen, used when DEBUG is defined, and only
> define c when DEBUG is undefined. From Pav Lucistnik <pav at
> FreeBSD>.
> While here, correct format strings to match ssize_t.
> OK moritz at . (ray@)
vi
~ common/api.c ~ common/line.c
~ common/main.c ~ common/put.c
~ ex/ex_global.c
> use sys/queue macros instead of accessing fields directly.
> no binary change.
> ok krw@ (pyr@)
xinstall
~ xinstall.c
> Allow stripping of files, even if the destination filename
> starts with a dash.
> ok millert@ jaredy@ ray@ (moritz@)
== usr.sbin ========================================================== 13/13 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin
usr.sbin
~ Makefile
> prettier Makefiles (deraadt@)
bgpd
~ printconf.c
> printconf didn't handle the 'down' flag for peers
> From: Stuart Henderson <stu at spacehopper.org> (henning@)
~ bgpd.h
> Graceful restart is no longer a draft so point it to the right RFC 4724.
> Additionally our as4byte support is based on draft-ietf-idr-as4bytes-13.
> (claudio@)
~ session.c
> draft-ietf-idr-restart -> RFC 4724 and bump draft-ietf-idr-as4bytes to
> version 13. (claudio@)
catman
- Makefile - TODO
- catman.8 - catman.c
- pathnames.h
> nothing uses catman, so kill it; ok millert deraadt (jmc@)
dhcpd
~ pfutils.c
> use the correct fd for poll. we're waiting on the pipe, not pf
> ok ckuethe deraadt (tedu@)
pkg_add
~ OpenBSD/PackageRepository.pm ~ OpenBSD/PackageRepositoryList.pm
> provide match_spec, findstem, find_partialstem at the repository list
> level,
> with ld -L like semantics: we stop at the first repository which has the
> corresponding packages.. (espie@)
~ OpenBSD/PackingElement.pm
> normalize quoting in extra info. (espie@)
~ OpenBSD/PackageRepositoryList.pm
> tweak _first_of so it can take an extra $filter routine, so that we
> get the first set of results that are non empty after filtering
> properly. (espie@)
~ OpenBSD/PackageLocator.pm
> allow filtering while calling the Locator. (espie@)
~ pkg_add ~ OpenBSD/Update.pm
> turn the updater into an object (espie@)
~ OpenBSD/Update.pm
> move most of the code into one process_package method (espie@)
~ OpenBSD/Update.pm
> reorg the update code some more, so that it goes through a filtering
> of candidates. (espie@)
~ OpenBSD/PackageLocator.pm
> have PackageLocator defer to PackageRepositoryList, so that package
> availability uses normal linker rules. (espie@)
~ OpenBSD/PackageName.pm ~ OpenBSD/PkgSpec.pm
> move towards unifying stem searches and pkgspec searches. (espie@)
~ OpenBSD/PackageRepository.pm
> make caching of stemlists implicit (espie@)
~ OpenBSD/PackageLocator.pm ~ OpenBSD/PackageName.pm
~ OpenBSD/PackageRepositoryList.pm
> also create PartialStem objects, and create a generic
> PackageLocator->match($o), where $o can be a spec, a stem, a partialstem.
> (espie@)
~ pkg_add ~ pkg_info
~ OpenBSD/Dependencies.pm ~ OpenBSD/PackageInfo.pm
~ OpenBSD/PackageName.pm ~ OpenBSD/PackageRepository.pm
~ OpenBSD/PackageRepositoryList.pm ~ OpenBSD/Update.pm
> unify object lookup as locator->match($o) or $repo->match($o) (espie@)
~ OpenBSD/PackageLocator.pm ~ OpenBSD/PackageRepository.pm
~ OpenBSD/PackageRepositoryList.pm
> kill unused stuff (espie@)
~ OpenBSD/PackageName.pm
> kill more stuff (espie@)
~ OpenBSD/Update.pm
> next -> return (espie@)
~ OpenBSD/PackageLocator.pm ~ OpenBSD/PackageRepository.pm
~ OpenBSD/PackageRepositoryList.pm
> remove some specialness of the Filtering stuff. (espie@)
~ OpenBSD/Update.pm
> simplify updater slightly (espie@)
~ pkg_add ~ pkg_delete
~ OpenBSD/Delete.pm ~ OpenBSD/Dependencies.pm
~ OpenBSD/PackageInfo.pm ~ OpenBSD/PackageRepositoryList.pm
~ OpenBSD/RequiredBy.pm
> tidy (espie@)
~ OpenBSD/PackageInfo.pm ~ OpenBSD/PackageName.pm
~ OpenBSD/PackageRepository.pm
> maintain stemlist for installed packages dynamically (espie@)
~ Makefile ~ OpenBSD/PackageRepository.pm
+ OpenBSD/PackageRepository/Installed.pm
> move code around to allow OO access to installed packages always. (espie@)
~ OpenBSD/PackageRepository/Installed.pm
> load PackageLocation just in time. Control what we import from
> PackageInfo. (espie@)
~ OpenBSD/PackageRepository/Installed.pm
> two repos: new(1) gives all packages. new() zaps .libs and friends.
> (espie@)
~ pkg ~ pkg_info
> use Repository::Installed a bit more (espie@)
~ pkg ~ pkg_add
~ pkg_info ~ OpenBSD/Dependencies.pm
~ OpenBSD/PackageName.pm ~ OpenBSD/PackingElement.pm
~ OpenBSD/PkgCfl.pm ~ OpenBSD/PkgSpec.pm
~ OpenBSD/Replace.pm
~ OpenBSD/PackageRepository/Installed.pm
> name tweaks: match -> match_list, match_repo -> match (espie@)
~ pkg ~ pkg_add
~ pkg_info ~ OpenBSD/Dependencies.pm
~ OpenBSD/PackageInfo.pm ~ OpenBSD/PackageName.pm
~ OpenBSD/PackingElement.pm ~ OpenBSD/PkgCfl.pm
~ OpenBSD/PkgSpec.pm ~ OpenBSD/Update.pm
> put search objects into a search class. (espie@)
~ Makefile ~ pkg
~ pkg_add ~ pkg_info
~ OpenBSD/Dependencies.pm ~ OpenBSD/PackageName.pm
~ OpenBSD/PackingElement.pm ~ OpenBSD/PkgCfl.pm
~ OpenBSD/PkgSpec.pm + OpenBSD/Search.pm
> put Search objects into their own file, finally (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
~ OpenBSD/PkgCfl.pm ~ OpenBSD/Replace.pm
~ OpenBSD/Search.pm ~ OpenBSD/Update.pm
~ OpenBSD/PackageRepository/Installed.pm
> redo search filters as real objects. Accordingly, rename match_list to
> filter. (espie@)
~ OpenBSD/Search.pm
> for completion, allow Stem object to filter as well. (espie@)
~ OpenBSD/PackageRepository.pm
> move parsing of http problems where it belongs. (espie@)
~ OpenBSD/PackageRepository.pm ~ OpenBSD/PackageRepositoryList.pm
> move the retry code down at the http/ftp repository level, where it
> belongs. (espie@)
~ OpenBSD/PackageLocator.pm ~ OpenBSD/PackageRepositoryList.pm
> kill "available" routine, which is no longer used. (espie@)
~ Makefile ~ pod/OpenBSD::PkgSpec.pod
+ pod/OpenBSD::Search.pod
> document search objects. Adjust PkgSpec documentation (espie@)
~ OpenBSD/PackageLocation.pm
> cosmetic tweak (espie@)
~ OpenBSD/Dependencies.pm
> use the solver object to split the code into smaller chunks (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
> move verbose printing into a separate method. (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
> put the solver into its own class. (espie@)
~ OpenBSD/SharedLibs.pm
> remove unneeded quotes (espie@)
~ OpenBSD/PackageInfo.pm
> fix dependencies (espie@)
~ OpenBSD/Delete.pm
> do not copy pkgname, since we're going to erase it, and packing-lists
> now detect duplicate stuff. (espie@)
~ pkg_create
> make sure to give all special files to root/wheel, and strip them of write
> and suid bits. (espie@)
~ pkg_info
> require the right code at the right location.
> problem noticed by deanna@ (espie@)
~ OpenBSD/PackageLocation.pm
> allows for incomplete locations (without arch) and for repositories to
> print PackageLocations (to be cleaned up later). (espie@)
~ OpenBSD/PackageLocation.pm
> oops, committed debug code, sorry. (espie@)
~ OpenBSD/PackageName.pm
> handle normalization of flavors in PackageName objects. (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
> tweak the solver object a bit, now that I have a clearer idea of what
> it's going to be used for. (espie@)
~ OpenBSD/Dependencies.pm
> make the solved dependencies more explicit, match them back to the spec
> they match. (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
> access solver internals in a clean way through dependencies, uses it
> internally.
> store solver objects along with solved_dependencies.
> Eventually, all solved_dependencies requests will use solver objects
> instead. (espie@)
~ pkg_add
> use solver object in all direct instances of solved_dependencies. (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
> library lookup becomes a real method (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
> move the library lookup code into Solver, so that I can turn it into
> methods as well... (espie@)
~ OpenBSD/PackingElement.pm
> make sure that mkfontdir and friends get their arguments (espie@)
~ OpenBSD/PackageRepository.pm
~ OpenBSD/PackageRepository/Installed.pm
~ OpenBSD/PackageRepository/SCP.pm ~ OpenBSD/PackageRepository/Source.pm
> name each urlscheme (espie@)
~ OpenBSD/PackageLocation.pm ~ OpenBSD/PackageRepository.pm
~ OpenBSD/PackageRepository/Installed.pm
> new methods: repositories can create url. Rename PackageLocation->stringize
> to url, since that's what it does. Use the new methods to shorten code.
> (espie@)
~ OpenBSD/PackageRepository.pm ~ OpenBSD/PackageRepository/SCP.pm
> do not pass the urlscheme to the constructors, since they each have
> a known urlscheme.
> Recognize inst: and file: as specific schemes that refer to given
> repositories, for future extension. (espie@)
~ OpenBSD/Dependencies.pm ~ pkg_add
> we don't need to walk the full dependency tree each time we don't find
> a library. Instead, put the state of the closure computation in the solver
> object, and only proceed until we find the library. We just need an extra
> marker (known) to be aware of packages that may contain libraries, but have
> not been used for extra dependencies yet. (espie@)
~ OpenBSD/Dependencies.pm
> cut dependency code in smaller chunks (espie@)
~ pkg_info
~ OpenBSD/PackageRepository/Installed.pm
> normalize installed names (espie@)
~ OpenBSD/PackageInfo.pm
> also strip +CONTENTS (espie@)
~ pkg_info
> show the url for the package we're dealing with. Unconfuse installed vs.
> uninstalled stuff. (espie@)
~ Makefile ~ OpenBSD/PackageRepository.pm
~ OpenBSD/PackingElement.pm ~ OpenBSD/Search.pm
~ OpenBSD/PackageRepository/Source.pm
> Add pkgpath hints to pkgspec. Make PackageRepository/Source functional:
> if a pkgpath is hinted at, then we build that package, and we delegate
> to the corresponding File: repository.
> Some XXX: we need to unlock the database temporarily, as the build
> *requires* shared access to stuff that is going on (and possibly write
> access for installing subpackages, so this cannot work 100% yet).
> Still, this allows the transparent build of one single package, at least.
> (espie@)
~ OpenBSD/PackageLocator.pm ~ OpenBSD/PackageRepository.pm
~ OpenBSD/PackageRepositoryList.pm
~ OpenBSD/PackageRepository/Installed.pm
> don't pass srcpath around, since we don't use it. (espie@)
~ OpenBSD/PackageRepository/SCP.pm
> fix scp url parsing and printing (espie@)
~ OpenBSD/PackageRepositoryList.pm
> and I forgot this one (espie@)
~ OpenBSD/PackageLocator.pm ~ OpenBSD/Search.pm
~ OpenBSD/PackageRepository/Installed.pm
> add match_locations/filter_locations methods to search stuff that returns
> full pkg locations instead of package names. (espie@)
~ pkg_add ~ OpenBSD/Dependencies.pm
> move the code that handles dependencies registration into Dependencies
> (espie@)
~ OpenBSD/Replace.pm
> zap old depend code I missed (espie@)
~ OpenBSD/Replace.pm
> remove extra forwarding method that is no longer needed (espie@)
~ OpenBSD/Replace.pm
> libs_to_check ? actually, never use it... (espie@)
~ pkg_add ~ OpenBSD/Replace.pm
> compute list of dependencies to adjust just-in-time.
> This lets the pkg_add regress pass, finally. (espie@)
~ pkg_delete
> both pkg_add and pkg_delete use Vstat, make it explicit. (espie@)
~ pkg_add ~ pkg_delete
> placeholders for UpdateSet, once we have it. (espie@)
~ pkg_add ~ pkg_delete
> create specific `state' objects for pkg_add and pkg_delete, so that I can
> write shortcuts for common behavior, and start hiding some details.
> (espie@)
~ pkg_add
> move deptree and installed into state object (espie@)
~ OpenBSD/PackingElement.pm
> make the package tools more aware of manpages: handle names correctly,
> and figure out how to process manpages. (espie@)
sasyncd
~ conf.y
> Check getpwnam() return value for NULL before dereferencing it.
> ok ray@ millert@ (moritz@)
===============================================================================
More information about the owc
mailing list