[owc] Weekly src changes ending 2007-10-14
OWC
auto at squish.net
Mon Oct 15 07:00:01 BST 2007
OpenBSD src changes summary for 2007-10-07 to 2007-10-14 inclusive
==================================================================
bin/chio distrib/notes
distrib/sets distrib/sgi
etc/Makefile etc/changelist
etc/etc.sgi/disktab etc/ftpusers
etc/group etc/mail/aliases
etc/master.passwd etc/mtree/special
etc/ospf6d.conf etc/rc
etc/rc.conf kerberosV
lib/libc lib/libkvm
lib/libossaudio lib/libssl
regress/sbin regress/sys
regress/usr.bin sbin/ifconfig
sbin/iopctl sbin/ipsecctl
sbin/nmeaattach sbin/pfctl
share/man sys/arch/alpha/alpha
sys/arch/alpha/conf sys/arch/alpha/pci
sys/arch/amd64/amd64 sys/arch/amd64/conf
sys/arch/arm/arm sys/arch/arm/include
sys/arch/arm/xscale sys/arch/armish/conf
sys/arch/aviion/aviion sys/arch/hp300/hp300
sys/arch/hppa/hppa sys/arch/hppa/include
sys/arch/i386/conf sys/arch/i386/i386
sys/arch/i386/include sys/arch/i386/pci
sys/arch/luna88k/luna88k sys/arch/m68k/include
sys/arch/m68k/m68k sys/arch/m88k/include
sys/arch/m88k/m88k sys/arch/mac68k/dev
sys/arch/mac68k/mac68k sys/arch/macppc/compile
sys/arch/macppc/conf sys/arch/macppc/macppc
sys/arch/mips64/mips64 sys/arch/mvme68k/mvme68k
sys/arch/mvme88k/mvme88k sys/arch/powerpc/conf
sys/arch/powerpc/include sys/arch/powerpc/powerpc
sys/arch/sgi/include sys/arch/sgi/localbus
sys/arch/sgi/sgi sys/arch/sh/include
sys/arch/sh/sh sys/arch/solbourne/solbourne
sys/arch/sparc/conf sys/arch/sparc/dev
sys/arch/sparc/sparc sys/arch/sparc64/compile
sys/arch/sparc64/conf sys/arch/sparc64/dev
sys/arch/sparc64/sparc64 sys/arch/vax/include
sys/arch/vax/mscp sys/arch/vax/vax
sys/conf sys/dev
sys/dev/acpi sys/dev/i2c
sys/dev/ic sys/dev/isa
sys/dev/mii sys/dev/ofw
sys/dev/onewire sys/dev/pci
sys/dev/pcmcia sys/dev/usb
sys/kern sys/lib/libkern
sys/net sys/net80211
sys/netinet sys/nfs
sys/sys sys/ufs/ext2fs
sys/ufs/ufs sys/uvm
sys/xfs usr.bin/bgplg
usr.bin/calendar usr.bin/compile_et
usr.bin/cvs usr.bin/ftp
usr.bin/gzsig usr.bin/make
usr.bin/pcc usr.bin/sudo
usr.bin/xlint usr.sbin/acpidump
usr.sbin/bgpctl usr.sbin/bgpd
usr.sbin/dhcpd usr.sbin/dvmrpd
usr.sbin/hostapd usr.sbin/hoststated
usr.sbin/httpd usr.sbin/ifstated
usr.sbin/ndp usr.sbin/ntpd
usr.sbin/ospf6ctl usr.sbin/ospf6d
usr.sbin/ospfctl usr.sbin/ospfd
usr.sbin/ripd usr.sbin/tcpdump
usr.sbin/ypbind
== bin =============================================================== 01/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/bin
chio
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
== distrib =========================================================== 02/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/distrib
notes
~ i386/hardware ~ macppc/hardware
~ sparc64/hardware
> Correct name GPS170PCI instead of GPS170. (mbalmer@)
sets
~ lists/man/mi
> sync (deraadt@)
~ lists/comp/md.amd64 ~ lists/etc/mi
> sync (deraadt@)
~ lists/comp/mi
> sync (deraadt@)
~ lists/comp/mi
> sync (deraadt@)
~ lists/base/md.amd64 ~ lists/base/md.sgi
> sync (deraadt@)
~ lists/man/mi
> sync (deraadt@)
sgi
+ iso/Makefile
> a start at full sgi install cd (deraadt@)
~ Makefile
> make obj in iso (deraadt@)
~ iso/Makefile
> install to right place (deraadt@)
~ iso/Makefile
> the iso can be smaller (deraadt@)
== etc =============================================================== 03/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/etc
Makefile
~ Makefile
> Add a sample configuration file for the newly imported ospf6d(8) daemon.
> ok dlg@ (norby@)
changelist
~ changelist
> Add a sample configuration file for the newly imported ospf6d(8) daemon.
> ok dlg@ (norby@)
etc.sgi/disktab
~ etc.sgi/disktab
> entry for the install##.iso file we now create (deraadt@)
~ etc.sgi/disktab
> the iso can be smaller (deraadt@)
ftpusers
~ ftpusers
> Add the user _ospf6d otherwise the newly imported ospf6d daemon will not
> start.
> ok dlg@ (norby@)
group
~ group
> Add the user _ospf6d otherwise the newly imported ospf6d daemon will not
> start.
> ok dlg@ (norby@)
mail/aliases
~ mail/aliases
> Add the user _ospf6d otherwise the newly imported ospf6d daemon will not
> start.
> ok dlg@ (norby@)
master.passwd
~ master.passwd
> Add the user _ospf6d otherwise the newly imported ospf6d daemon will not
> start.
> ok dlg@ (norby@)
mtree/special
~ mtree/special
> Add a sample configuration file for the newly imported ospf6d(8) daemon.
> ok dlg@ (norby@)
ospf6d.conf
+ ospf6d.conf
> Add a sample configuration file for the newly imported ospf6d(8) daemon.
> ok dlg@ (norby@)
rc
~ rc
> Add the user _ospf6d otherwise the newly imported ospf6d daemon will not
> start.
> ok dlg@ (norby@)
rc.conf
~ rc.conf
> Add the user _ospf6d otherwise the newly imported ospf6d daemon will not
> start.
> ok dlg@ (norby@)
== kerberosV ========================================================= 04/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/kerberosV
kerberosV
~ src/doc/setup.texi
> Typo, from Mark Peoples (biorn@)
== lib =============================================================== 05/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/lib
libc
~ locale/runetype.h
> attribute packed -> __packed
> OK millert (gilles@)
~ crypt/crypt.3
> add a RETURN VALUES section.
> ok jmc@ (pyr@)
~ regex/regcomp.c
> add a proper test around allocset
> use a simpler way to allocated memory in allocset, mostly done by
> replacing malloc/realloc dance, by only one realloc
> add comments about variables that are not used uninitialized, even
> if gcc told the contrary
> another malloc/realloc -> realloc change
> Work initially started by otto@, and then I joined him
> ok otto@ ray@ (chl@)
~ net/gethostnamadr.c ~ net/getrrsetbyname.c
~ net/res_debug.c
> use RRSIG instead of SIG for DNSSEC. ok djm@ (jakob@)
~ gen/rewinddir.c ~ gen/telldir.h
> fix typo in _telldir_unlocked function prototype, to avoid warning.
> while there, put all function prototypes in header file.
> ok kurt@ (chl@)
libkvm
~ kvm_proc.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
libossaudio
~ ossaudio.c
> - make SNDCTL_DSP_GETOSPACE much more efficient by using AUDIO_GETPRINO
> instead of AUDIO_GETINFO
> - make SNDCTL_DSP_GETISPACE return correct values by using AUDIO_GETRRINFO
> instead of AUDIO_GETINFO
> - remove calls to setblocksize() from SNDCTL_DSP_GET[IO]SPACE. changing
> the blocksize inside play or record loops, where these ioctls are most
> often used, can have serious negative consequences. the blocksize
> should be "fixed" to a power of 2 before these ioctls are called by
> using SNDCTL_DSP_SETFRAGMENT or SNDCTL_DSP_GETBLKSIZE.
> heavily tested by myself. no negative feedback from anyone else.
> ok ratchov (jakemsr@)
libssl
~ src/ssl/ssl_lib.c
> Replace use of strcpy(3) and other pointer goo in
> SSL_get_shared_ciphers() with strlcat(3).
> ok deraadt@ markus@ (moritz@)
~ src/ssl/ssl_lib.c
TAGGED OPENBSD_4_0
> MFC, fix by moritz:
> The SSL_get_shared_ciphers() function contains an off-by-one overflow.
> (henning@)
~ src/ssl/ssl_lib.c
TAGGED OPENBSD_4_1
> MFC, fix by moritz:
> The SSL_get_shared_ciphers() function contains an off-by-one overflow.
> (henning@)
~ src/ssl/ssl_lib.c
TAGGED OPENBSD_4_2
> MFC, fix by moritz:
> The SSL_get_shared_ciphers() function contains an off-by-one overflow.
> (henning@)
== regress =========================================================== 06/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/regress
sbin
~ pfctl/Makefile + pfctl/pfail53.in
+ pfctl/pfail53.ok
> we decided numbers used as strings is wrong (deraadt@)
~ pfctl/Makefile ~ pfctl/pf95.in
~ pfctl/pf95.ok + pfctl/pf95.include
> regression test for include directive (if anyone has a better way to do
> this messy include file copy, let me know) (deraadt@)
sys
~ uvm/misc/misc.c
> Do not attempt to work on more than SHMMAXPGS pages, makes this run
> unmodified
> on vax. (miod@)
usr.bin
- pcc/cc/Makefile - pcc/cc/ccom/Makefile
- pcc/cc/ccom/const001.c - pcc/cc/ccom/darray001.c
- pcc/cc/ccom/enum001.c - pcc/cc/ccom/init001.c
- pcc/cc/ccom/init002.c - pcc/cc/ccom/init003.c
- pcc/cc/ccom/init004.c - pcc/cc/ccom/tmpalloc001.c
- pcc/cc/cpp/Makefile - pcc/cc/cpp/skip001.c
- pcc/cc/cpp/skip001.out - pcc/cc/cpp/subdir/include.h
~ pcc/Makefile + pcc/ccom/Makefile
+ pcc/ccom/const001.c + pcc/ccom/darray001.c
+ pcc/ccom/enum001.c + pcc/ccom/init001.c
+ pcc/ccom/init002.c + pcc/ccom/init003.c
+ pcc/ccom/init004.c + pcc/ccom/tmpalloc001.c
+ pcc/cpp/Makefile + pcc/cpp/res1
+ pcc/cpp/res2 + pcc/cpp/res3
+ pcc/cpp/res4 + pcc/cpp/res5
+ pcc/cpp/res6 + pcc/cpp/res7
+ pcc/cpp/res8 + pcc/cpp/res9
+ pcc/cpp/skip001.c + pcc/cpp/skip001.out
+ pcc/cpp/test1 + pcc/cpp/test2
+ pcc/cpp/test3 + pcc/cpp/test4
+ pcc/cpp/test5 + pcc/cpp/test6
+ pcc/cpp/test7 + pcc/cpp/test8
+ pcc/cpp/test9 + pcc/cpp/subdir/include.h
> and reorg this tree as well (otto@)
~ xlint/test-11.c
> attribute packed -> __packed
> OK millert (gilles@)
~ cvs/Makefile
> New OpenCVS regression tests.
> OK niallo@, xsa@ (tobias@)
== sbin ============================================================== 07/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sbin
ifconfig
~ ifconfig.c
> Fix range check for carp vhid: vhid 0 isn't valid (joel@)
iopctl
~ iopctl.c
> attribute packed -> __packed
> OK millert (gilles@)
ipsecctl
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ ipsecctl.c ~ ipsecctl.h
~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
nmeaattach
~ nmeaattach.c
> Unconditionally call the TIOCSTSTAMP ioctl, this way calling nmeaattach(8)
> without the '-t' option can be used to turn of tty timestamping.
> problem noticed by sthen@, ok sthen, deraadt (mbalmer@)
pfctl
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ parse.y
> support an include directive; file of course must also be "secure" like
> the main configuration file; ok henning (deraadt@)
~ parse.y ~ pfctl.c
~ pfctl_parser.h
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
== share ============================================================= 08/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/share
man
~ man4/man4.i386/Makefile + man4/man4.i386/glxpcib.4
> Add support for the PCI-ISA bridge, timecounter, and watchdog timer
> found in the CS5536 companion chip to the AMD Geode LX-800 processor.
> glxpcib(4) driver written by mickey, with some tweaks by me. Manpage
> nits by jmc.
> Tested on the Soekris net5501 (by me) and the Alix 2a2
> (by earxtacy at dixinet.com)
> ok kettenis (mbalmer@)
~ man4/Makefile ~ man4/iic.4
+ man4/spdmem.4
> Port of Nicolas Joly, Paul Goyette & Tobias Nygren's
> spdmem(4) driver for Serial Presence Detect memory information. (jsg@)
~ man4/man4.i386/Makefile + man4/man4.i386/amdpcib.4
> manual page for an upcoming driver.
> "committ it" deraadt. (mbalmer@)
~ man4/pci.4
> mention amdpcib(4) and glxpcib(4). (mbalmer@)
~ man4/isa.4
> mention amdpcib(4) and glxpcib(4) (mbalmer@)
~ man4/man4.i386/amdpcib.4
> mention the LPC bridge part. (mbalmer@)
~ man4/pci.4
> for miscellaneous devices section:
> - sync Nd for amdpcib(4)
> - add pcib(4)
> - adjust the width size of the list accordingly (jmc@)
~ man4/pci.4 ~ man4/man4.i386/amdpcib.4
> AMD8111 -> AMD-8111 (jmc@)
~ man4/man4.amd64/Makefile + man4/man4.amd64/amdpcib.4
> amd64 needs the page too (deraadt@)
~ man4/audio.4
> explain that AUDIO_GET[PR]RINFO were added for porting 3rd party
> software and should be avoided in new code
> suggested by ratchov@, help from jmc@ (jakemsr@)
~ man4/uslcom.4
> attach uslcom to the track systems traqmate
> ok jsg (jcs@)
~ man4/mbg.4
> The PCI based GPS card is GPS170PCI, not GPS170, which is a different
> receiver in "Europa"-card size. (mbalmer@)
~ man4/man4.amd64/amdpcib.4 ~ man4/man4.i386/amdpcib.4
~ man4/man4.i386/glxpcib.4 ~ man4/man4.i386/gscpcib.4
~ man4/man4.i386/ichpcib.4 ~ man4/man4.i386/piixpcib.4
> - do not talk about "pcib functionality": these pages already document what
> the devices are
> - no need to Xr pcib either
> ok mbalmer deraadt (jmc@)
~ man4/man4.alpha/intro.4 ~ man4/man4.amd64/intro.4
~ man4/man4.i386/intro.4 ~ man4/man4.macppc/intro.4
~ man4/man4.sparc64/intro.4
> add iic(4) to the hardware list; ok deraadt (jmc@)
~ man4/man4.armish/intro.4
> (missed in previous commit)
> add iic(4) to the hardware list; ok deraadt (jmc@)
~ man4/man4.alpha/intro.4 ~ man4/man4.amd64/intro.4
~ man4/man4.armish/intro.4 ~ man4/man4.hppa/intro.4
~ man4/man4.i386/intro.4 ~ man4/man4.landisk/intro.4
~ man4/man4.macppc/intro.4 ~ man4/man4.sparc64/intro.4
~ man4/man4.zaurus/intro.4
> add onewire(4) to the hardware list; (jmc@)
~ man4/rgephy.4
> Mention RTL8211B support. (jsg@)
~ man9/Makefile ~ man9/ctxsw.9
> Update to match cpu_switchto reality. (art@)
- man9/ctxsw.9 ~ man9/Makefile
+ man9/mi_switch.9
> rename man page to a real function name (deraadt@)
~ man5/pf.conf.5
> support an include directive; file of course must also be "secure" like
> the main configuration file; ok henning (deraadt@)
~ man4/acx.4 ~ man4/bwi.4
~ man4/malo.4 ~ man4/pgt.4
> Bump firmware version; Adapt to new package mechanism by rebuilding
> those firmware packages. Solves ``Warning: obsolete construct: @ignore''.
> (mglocker@)
~ man4/zyd.4
> add GW-US54GD to the list of supported devices (fkr@)
~ man4/Makefile ~ man4/iic.4
+ man4/wbng.4
> wbng(4) manual page for Winbond W83793G; 'commit it' deraadt (cnst@)
~ man5/pf.conf.5
> include in bnf, wanted by jmc (deraadt@)
~ man4/wbng.4
> keep the Nd lines consistent; (jmc@)
== sys =============================================================== 09/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sys
arch/alpha/alpha
~ genassym.cf ~ locore.s
~ machdep.c ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Fix cpu_exit() comments to be more closer to reality. (miod@)
arch/alpha/conf
~ GENERIC
> enable spdmem where it matters (deraadt@)
arch/alpha/pci
~ mcpcia.c
> More simple memset(,0,) -> M_ZERO changes. In this batch move to
> size(*p) as the first malloc() parameter where p is declared locally
> and thus easy to check. Add M_ZERO to gpe_table allocation in acpi.c
> even though there is no obvious bzero or memset nearby. (krw@)
arch/amd64/amd64
~ cpu.c ~ genassym.cf
~ locore.S ~ mptramp.S
~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Fix cpu_exit() comments to be more closer to reality. (miod@)
- Locore.c
> Remove leftovers art forgot to prune... (miod@)
arch/amd64/conf
~ GENERIC ~ files.amd64
> Add a driver, amdpcib(4), for the AMD 8111 series LPC bridge and HPET
> written
> by mickey, man page by me. Help, suggestions by Theo and jmc.
> Enable this driver and glxpcib(4) in all RAMDISK kernel, but glxpcib(4)
> does not provide the watchdog timer in the RAMDISK kernels.
> Prodded by deraadt.
> E (mbalmer@)
~ GENERIC
> enable spdmem where it matters (deraadt@)
~ files.amd64
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ GENERIC
> enable wbng (deraadt@)
arch/arm/arm
~ cpuswitch.S ~ genassym.cf
~ pmap.c ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Fix cpu_exit() comments to be more closer to reality. (miod@)
~ cpuswitch.S
> There is no need to fiddle with spl in cpu_idle_{enter,leave}, actually.
> (miod@)
arch/arm/include
~ pmap.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/arm/xscale
~ pxa2x0_apm_asm.S
> Align the suspend saved data to a cache line boundary, apparently prevents
> issues upon resume. (miod@)
arch/armish/conf
~ GENERIC
> enable spdmem where it matters (deraadt@)
~ GENERIC ~ files.armish
> enable puc(4) for GENERIC on armish
> ok deraadt@ (sthen@)
arch/aviion/aviion
~ locore.S ~ machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ machdep.c
> Enable interrupts in secondary processors before invoking cpu_switchto(),
> rather the expecting it to do this for us. (miod@)
arch/hp300/hp300
~ locore.s ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Do not splhigh() before invoking sched_exit(), sched_exit() will do it
> better. (miod@)
arch/hppa/hppa
~ genassym.cf ~ locore.S
~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/hppa/include
~ cpu.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/i386/conf
~ GENERIC ~ files.i386
> Add support for the PCI-ISA bridge, timecounter, and watchdog timer
> found in the CS5536 companion chip to the AMD Geode LX-800 processor.
> glxpcib(4) driver written by mickey, with some tweaks by me. Manpage
> nits by jmc.
> Tested on the Soekris net5501 (by me) and the Alix 2a2
> (by earxtacy at dixinet.com)
> ok kettenis (mbalmer@)
~ GENERIC ~ RAMDISK
~ RAMDISKB ~ RAMDISKC
~ RAMDISK_CD ~ files.i386
> Add a driver, amdpcib(4), for the AMD 8111 series LPC bridge and HPET
> written
> by mickey, man page by me. Help, suggestions by Theo and jmc.
> Enable this driver and glxpcib(4) in all RAMDISK kernel, but glxpcib(4)
> does not provide the watchdog timer in the RAMDISK kernels.
> Prodded by deraadt.
> E (mbalmer@)
~ GENERIC
> enable spdmem where it matters (deraadt@)
~ GENERIC
> isa0 at amdpcib? got skipped, sigh (deraadt@)
~ GENERIC
> enable wbng (deraadt@)
arch/i386/i386
~ trap.c
> Change argsize from size_t to short, since it only stores a short
> value. Fixes a format string error.
> Pointed out by deraadt, OK miod. (ray@)
~ cpu.c ~ genassym.cf
~ locore.s ~ mptramp.s
~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Fix cpu_exit() comments to be more closer to reality. (miod@)
arch/i386/include
~ frame.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/i386/pci
+ glxpcib.c
> Add support for the PCI-ISA bridge, timecounter, and watchdog timer
> found in the CS5536 companion chip to the AMD Geode LX-800 processor.
> glxpcib(4) driver written by mickey, with some tweaks by me. Manpage
> nits by jmc.
> Tested on the Soekris net5501 (by me) and the Alix 2a2
> (by earxtacy at dixinet.com)
> ok kettenis (mbalmer@)
~ glxpcib.c
> Add a driver, amdpcib(4), for the AMD 8111 series LPC bridge and HPET
> written
> by mickey, man page by me. Help, suggestions by Theo and jmc.
> Enable this driver and glxpcib(4) in all RAMDISK kernel, but glxpcib(4)
> does not provide the watchdog timer in the RAMDISK kernels.
> Prodded by deraadt.
> E (mbalmer@)
arch/luna88k/luna88k
~ locore.S ~ machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ autoconf.c ~ machdep.c
> Enable interrupts in secondary processors before invoking cpu_switchto(),
> rather the expecting it to do this for us. (miod@)
arch/m68k/include
~ cpu.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/m68k/m68k
~ genassym.cf ~ proc_subr.s
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
- proc_subr.s
> One more cpu_switchto() leftover. (miod@)
arch/m88k/include
~ cpu.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ cpu.h
> It is no longer necessary to fiddle with spl in cpu_idle_{enter,leave} now
> that proc_trampoline has been fixed. (miod@)
arch/m88k/m88k
~ genassym.cf ~ m88k_machdep.c
~ process.S ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Do not splhigh() before invoking sched_exit(), sched_exit() will do it
> better. (miod@)
~ eh_common.S
> Be sure to spl0() in proc_trampoline, so that kernel threads start at
> IPL_NONE. (miod@)
~ process.S
> It is no longer necessary to fiddle with spl in cpu_idle_{enter,leave} now
> that proc_trampoline has been fixed. (miod@)
arch/mac68k/dev
~ if_mc.c
> 'expresion' -> 'expression'. Reported by Jung on tech at . (krw@)
arch/mac68k/mac68k
~ locore.s ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Do not splhigh() before invoking sched_exit(), sched_exit() will do it
> better. (miod@)
arch/macppc/compile
~ .cvsignore
> GENERIC.MP dirs will show up here some time (deraadt@)
arch/macppc/conf
~ GENERIC
> enable spdmem where it matters (deraadt@)
~ GENERIC ~ files.macppc
> Provide a way to attach spdmem(4) by faking an I2C bus with EEPROMs with
> contents from OpenFirmware.
> ok deraadt@ (kettenis@)
arch/macppc/macppc
~ cpu.c ~ genassym.cf
~ locore.S
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ locore.S
> Disable PSL_POW code for now in cpu_idle_cycle, because it does not yet
> do the cache flush / instruction cache disable / tlb invalidate dance.
> For some reason in a switchto world this suddenly matters or machines
> become really slow running NFS... huh? figured out with kettenis (deraadt@)
~ openfirm.c
> Implement OF_getproplen().
> ok deraadt@ (kettenis@)
~ mainbus.c ~ mem.c
> Provide a way to attach spdmem(4) by faking an I2C bus with EEPROMs with
> contents from OpenFirmware.
> ok deraadt@ (kettenis@)
arch/mips64/mips64
~ context.S ~ db_machdep.c
~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Do not splhigh() before invoking sched_exit(), sched_exit() will do it
> better. (miod@)
~ context.S
> There is no need to fiddle with spl in cpu_idle_{enter,leave}, actually.
> (miod@)
~ interrupt.c
> registred -> registered (miod@)
arch/mvme68k/mvme68k
~ locore.s ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Do not splhigh() before invoking sched_exit(), sched_exit() will do it
> better. (miod@)
arch/mvme88k/mvme88k
~ locore.S ~ machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ autoconf.c ~ machdep.c
> Enable interrupts in secondary processors before invoking cpu_switchto(),
> rather the expecting it to do this for us. (miod@)
arch/powerpc/conf
~ files.powerpc
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/powerpc/include
~ pcb.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/powerpc/powerpc
~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
- Locore.c
> Remove leftovers art forgot to prune... (miod@)
arch/sgi/include
~ intr.h
> registred -> registered (miod@)
arch/sgi/localbus
~ macebus.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ macebus.c
> Various typos in comments; Joel Sing (miod@)
~ macebus.c
> Disable timer/compare interrupts on the macebus. This prevents interrupt
> storms from occurring on IRQ 6. ok miod@ (jsing@)
arch/sgi/sgi
~ genassym.cf
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/sh/include
~ cpu.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/sh/sh
~ genassym.cf ~ locore_c.c
~ locore_subr.S ~ sh_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/solbourne/solbourne
~ locore.s
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/sparc/conf
~ files.sparc
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/sparc/dev
~ obio.c
> A few trailing bzero/memset -> M_ZERO occurrences, cast removal and
> size(*p) usage. (krw@)
arch/sparc/sparc
~ locore.s ~ locore2.c
~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
- locore2.c
> Remove leftovers art forgot to prune... (miod@)
arch/sparc64/compile
~ .cvsignore
> prepare for the future (deraadt@)
arch/sparc64/conf
~ GENERIC
> enable spdmem where it matters (deraadt@)
~ files.sparc64
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/sparc64/dev
~ ofwi2c.c
> Don't skip EEPROMs in ofw i2c scan,
> match on sparc64 SPD name in spdmem.
> Tested on a V120.
> ok deraadt@ (jsg@)
~ sbus.c
> A few trailing bzero/memset -> M_ZERO occurrences, cast removal and
> size(*p) usage. (krw@)
~ psycho.c ~ pyro.c
~ schizo.c
> A few trailing bzero/memset -> M_ZERO occurrences, cast removal and
> size(*p) usage. This should be the last of the simple malloc()
> changes. (krw@)
~ schizo.c ~ pyro.c
> malloc+bzero -> malloc+M_ZERO. Use 'malloc(sizeof(*p) ...' idiom.
> This completes my bzero/memset(,0,) -> M_ZERO sweep of the tree.
> ok kettenis@ (krw@)
arch/sparc64/sparc64
~ machdep.c
> A few trailing bzero/memset -> M_ZERO occurrences, cast removal and
> size(*p) usage. (krw@)
~ locore.s ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Fix cpu_exit() comments to be more closer to reality. (miod@)
- locore2.c
> Remove leftovers art forgot to prune... (miod@)
~ autoconf.c
> Don't bail out after finding the first cpu. (kettenis@)
arch/vax/include
~ cpu.h ~ macros.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
arch/vax/mscp
~ mscp.c
> A few trailing bzero/memset -> M_ZERO occurrences, cast removal and
> size(*p) usage. (krw@)
arch/vax/vax
~ subr.s ~ vm_machdep.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ vm_machdep.c
> Do not splhigh() before invoking sched_exit(), sched_exit() will do it
> better. (miod@)
conf
~ files
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ newvers.sh
TAGGED OPENBSD_4_2
> enter -stable (henning@)
dev
~ ipmi.c
> More simple memset(,0,) -> M_ZERO changes. In this batch move to
> size(*p) as the first malloc() parameter where p is declared locally
> and thus easy to check. Add M_ZERO to gpe_table allocation in acpi.c
> even though there is no obvious bzero or memset nearby. (krw@)
~ bio.c ~ rnd.c
> MALLOC -> malloc
> ok krw@ (gilles@)
dev/acpi
~ acpi.c ~ acpicpu.c
~ acpimadt.c ~ acpiprt.c
~ dsdt.c
> More simple memset(,0,) -> M_ZERO changes. In this batch move to
> size(*p) as the first malloc() parameter where p is declared locally
> and thus easy to check. Add M_ZERO to gpe_table allocation in acpi.c
> even though there is no obvious bzero or memset nearby. (krw@)
~ acpi.c
> When walking the tree yo find _PRT methods, check whether the device is
> actually enabled and functioning. Similar to what we do for _INI.
> ok weingart@, beck@ (kettenis@)
dev/i2c
~ i2c_scan.c
> add support for scanning the spdmem range of addresses, using a different
> probe function; based on earlier work by jsg, and retested by us (deraadt@)
~ files.i2c + spdmem.c
> Port of Nicolas Joly, Paul Goyette & Tobias Nygren's
> spdmem(4) driver for Serial Presence Detect memory information. (jsg@)
~ spdmem.c
> Print CAS latency for DDR/DDR2.
> Shuffle strings around to look pretty as suggested by deraadt. (jsg@)
~ spdmem.c
> Indicate if the module is registered, DDR/DDR2 only at the moment.
> Tested by ian@ wilfried@ and deraadt@ (jsg@)
~ spdmem.c
> Don't skip EEPROMs in ofw i2c scan,
> match on sparc64 SPD name in spdmem.
> Tested on a V120.
> ok deraadt@ (jsg@)
~ spdmem.c
> Correct frequency for SDRAM. (jsg@)
~ spdmem.c
> To identity PC133 we need to check cycle time as modules
> often identify as PC100. (jsg@)
~ spdmem.c
> Print registered/CAS latency information for SDRAM. (jsg@)
~ i2c.c ~ i2c_scan.c
> we have been knocking out i2c when ipmi exists on a machine (because
> there are ipmi bmc's that apparently violate i2c bus master locking).
> but those problems bite us only for sensors which we re-scan every few
> seconds, and then we hit the ipmi bmc's races. for non-sensor devices
> (like spdmem) it is probably ok to look once. let's try that method
> for a while and see if anyone sees breakage. (deraadt@)
~ spdmem.c
> correct : printing; ok jsg (deraadt@)
~ spdmem.c
> hypenate memory type for readability. (jsg@)
~ i2c_scan.c
> probe for Winbond W83793G; tested by jon.steel at esentire.com; suggestions/ok
> deraadt@ (cnst@)
~ spdmem.c
> really correct : printing; ok deraadt@ (cnst@)
~ thmc50.c
> print the name of the chip on attach; ok deraadt (cnst@)
~ files.i2c + w83793g.c
> wbng(4) driver for Winbond W83793G Hardware Monitor; 'commit it' deraadt
> (cnst@)
~ w83793g.c
> fix newlines (deraadt@)
~ spdmem.c
> correctly print PC and PC2- prefixes; ok jsg (deraadt@)
dev/ic
~ sili.c
> Use %zu for printing size_t values.
> OK dlg. (ray@)
~ if_wi_hostap.c
> MALLOC -> malloc
> ok krw@ (gilles@)
~ ar5xxx.c ~ ar5xxx.h
> The newer single chip Atheros wireless chipsets like the AR5424,
> AR2423 etc. are mostly compatible to the AR5212 but use a different
> algorithm to set the 2GHz RF channel, that's why they didn't work in
> OpenBSD. I figured out that the channels were set with an offset,
> setting channel 11 in the driver caused the hardware to set channel 5
> etc. Because I didn't figure out the pattern to fix the algoritm yet,
> I fixed it in a workaroundish way by defining a small "table" with
> offsets for the 11b channels to get the right results. For example, if
> we want to set channel 11 (2462MHz), we add an offset of -30MHz, and
> feed the result (2432MHz ^= channel 5) into the unmodified
> AR5212/AR5112 RF setup function.
> Long description for a commit message, but it needed some time to
> figure it out. It is still not perfect, needs some more work, and it
> doesn't work in all cases; but it allows to use newer chipsets in 11b
> mode restricted to 1 or to 2Mbit/s. 11a mode seems to work without
> problems so far. (reyk@)
~ ar5xxx.c
> add the AR5416 and AR5418 device IDs (needs some more testing) (reyk@)
~ fxp.c ~ elink3.c
~ ath.c
> remove unneeded declarations that shadows existing vars; ok by many.
> (fgsch@)
~ acx.c
> Fix a panic which happens with ACX100 chipsets in ieee80211_add_xrates()
> because the drivers tries to set extended rates for a 11b chip.
> Tested and OK winiger@ (mglocker@)
dev/isa
~ aha.c
> MALLOC -> malloc
> ok krw@ (gilles@)
dev/mii
~ rgephy.c ~ rgephyreg.h
> Add support for the RTL8211B PHY based on a patch by
> Michael Eisele for FreeBSD.
> This fixes nfe/rgephy for ckuethe & couderc.
> Tested on re/rgephy by pyr. (jsg@)
~ brgphy.c
> Add support for BCM5906.
> ok deraadt@ (kettenis@)
dev/ofw
~ openfirm.h
> Implement OF_getproplen().
> ok deraadt@ (kettenis@)
dev/onewire
~ onewire.c
> MALLOC -> malloc
> ok krw@ (gilles@)
dev/pci
+ amdpcib.c
> Add a driver, amdpcib(4), for the AMD 8111 series LPC bridge and HPET
> written
> by mickey, man page by me. Help, suggestions by Theo and jmc.
> Enable this driver and glxpcib(4) in all RAMDISK kernel, but glxpcib(4)
> does not provide the watchdog timer in the RAMDISK kernels.
> Prodded by deraadt.
> E (mbalmer@)
~ if_san_xilinx.c ~ qli_pci.c
> More simple memset(,0,) -> M_ZERO changes. In this batch move to
> size(*p) as the first malloc() parameter where p is declared locally
> and thus easy to check. Add M_ZERO to gpe_table allocation in acpi.c
> even though there is no obvious bzero or memset nearby. (krw@)
~ pcidevs
> GPS170 -> GPS170PCI
> ok claudio (mbalmer@)
~ pcidevs.h ~ pcidevs_data.h
> regen (mbalmer@)
~ mbg.c
> GPS170 -> GPS170PCI
> ok claudio (mbalmer@)
~ if_bge.c
> Use %zu for printing size_t values.
> OK dlg. (ray@)
~ maestro.c
> MALLOC -> malloc
> ok krw@ (gilles@)
~ pciide.c
> malloc+bzero -> malloc+M_ZERO. Use 'malloc(sizeof(*p) ...' idiom. (krw@)
~ pcidevs.h ~ pcidevs_data.h
> regen (dlg@)
~ pcidevs
> add the mellanox vendor id (dlg@)
~ pcidevs.h ~ pcidevs_data.h
> regen (dlg@)
~ pcidevs
> add the pci product id for the Mellanox ConnectX EN adapters. (dlg@)
~ azalia.c ~ azalia.h
~ azalia_codec.c
> Rework the debug dmesg to perform a more complete codec dump.
> Tested by jmc, ckuethe, kettenis and jakemsr. (deanna@)
~ if_vge.c
> Fix some obviously bogus code in vge_newbuf(). Should fix PR 5582.
> ok claudio@ (kettenis@)
~ pcidevs ~ pcidevs.h
~ pcidevs_data.h ~ pucdata.c
> Add device ID used by Commell MP-954GPS
> ok deraadt@ (sthen@)
~ pcidevs.h ~ pcidevs_data.h
> sync (deraadt@)
~ if_bgereg.h
> Make BGE_JUMBO_FRAMELEN big enough to include vlan tag. Fixes problems
> with
> receiving jumbo frames on bge(4).
> ok krw@ (kettenis@)
~ pcidevs ~ if_em_hw.h
> Add some more Intel em(4) PCI ids.
> From FreeBSD
> ok dlg@ (brad@)
~ pcidevs.h ~ pcidevs_data.h
> regen (brad@)
~ if_bge.c ~ if_bgereg.h
> Add support for BCM5906.
> ok deraadt@ (kettenis@)
~ if_skreg.h
> Add Yukon-2 PHY powerdown bits. (kettenis@)
~ azalia_codec.c
> audio(4)-compatible mixer items for ALC883.
> Tested jmc, jakemsr (deanna@)
~ azalia.c
> The spec says to wait 40 microseconds for the run bit to clear before
> performing a stream reset. Fixes stream reset errors on ALC883,
> reported by jmc.
> Tested jmc, jakemsr, ok ratchov (deanna@)
dev/pcmcia
~ if_malo.c ~ if_maloreg.h
~ if_malovar.h
> Make pcmcia malo(4) work with trunk(4) by:
> Setting interface ethernet address from link level address.
> Adding firmware command to set ethernet address. (mglocker@)
~ if_malo.c
> Give the device a bit time to finish the reset when stopped. Avoids
> helper firmware load timeouts when device is brought up very fast again,
> like seen with trunk(4). (mglocker@)
~ if_malo.c ~ if_malovar.h
> Load the firmware files just one time per device lifecycle (attach time)
> instead on every `ifconfig up'. (mglocker@)
~ if_malo.c ~ if_malovar.h
> Setup channels with OFDM flags. Permit user to set fix rates. (mglocker@)
~ if_malo.c
> Ops, forgot debug printf. (mglocker@)
dev/usb
~ usbdevs
> track systems traqmate gps data acquisition system (jcs@)
~ usbdevs.h ~ usbdevs_data.h
> regen (jcs@)
~ uslcom.c
> support 230400 speed
> ok jsg (jcs@)
~ uslcom.c
> attach uslcom to the track systems traqmate
> ok jsg (jcs@)
~ umass_scsi.c
> malloc+memset(,0,) -> malloc+M_ZERO. 'Fixes' not zeroing of all
> allocated memory, as I can't see an obvious initialization of the
> memory elsewhere. NetBSD uses M_ZERO here. (krw@)
~ usbdevs.h ~ usbdevs_data.h
> sync (deraadt@)
~ usbdevs
> apple products; from dunceor (deraadt@)
~ usbdevs.h ~ usbdevs_data.h
> sync (deraadt@)
~ usbdevs
> oops (deraadt@)
~ usbdevs
> shortening (deraadt@)
~ usbdevs.h ~ usbdevs_data.h
> sync (deraadt@)
~ usb_subr.c ~ usbdi.c
> serialize usbd_probe_and_attach(); only let one usb device be match'd and
> attached at a time; ok kettenis (deraadt@)
~ if_atu.c ~ if_aue.c
~ if_axe.c ~ if_cdce.c
~ if_cdcef.c ~ if_cue.c
~ if_kue.c ~ if_ral.c
~ if_rum.c ~ if_uath.c
~ if_udav.c ~ if_upl.c
~ if_url.c ~ if_wi_usb.c
~ if_zyd.c ~ moscom.c
~ uark.c ~ uaudio.c
~ uberry.c ~ ubsa.c
~ ubt.c ~ uchcom.c
~ udcf.c ~ udsbr.c
~ ueagle.c ~ uftdi.c
~ ugen.c ~ uhidev.c
~ uhub.c ~ uipaq.c
~ ulpt.c ~ umass.c
~ umct.c ~ umidi.c
~ umodem.c ~ umsm.c
~ uow.c ~ uplcom.c
~ urio.c ~ usb_subr.c
~ uscanner.c ~ uslcom.c
~ usscanner.c ~ uts.c
~ uvisor.c ~ uvscom.c
~ uyap.c
> treat usb vendor/product names as a locator, and have usbd_print handle it,
> so that it shows up before the :. as a result, all the usb devices do not
> need to have name printing code anymore. all this now works and prints
> nicely because usbd_probe_and_attach() is serialized. ok kettenis
> (deraadt@)
~ usbdevs
> Add two Gude time signal station receivers. (mbalmer@)
~ usbdevs.h ~ usbdevs_data.h
> sync. (mbalmer@)
~ usbdevs
> add device id for Planex GWUS54GD, from freebsd
> ok damien@ (fkr@)
~ usbdevs.h ~ usbdevs_data.h
> regen (fkr@)
~ if_zyd.c
> attach Planex GWUS54GD, from freebsd
> ok damien@ (fkr@)
kern
~ subr_prf.c
> Introduce %z for printf(9), like %z for printf(3). Used to print
> size_t and ssize_t variables.
> OK miod and cthulhu. (ray@)
~ init_main.c ~ kern_clock.c
~ kern_exit.c ~ kern_fork.c
~ kern_proc.c ~ kern_synch.c
~ kern_sysctl.c ~ sched_bsd.c
+ kern_sched.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ sched_bsd.c
> sched_lock_idle and sched_unlock_idle are obsolete now. (art@)
~ sysv_msg.c ~ sysv_sem.c
> remove tests for EIDRM and ENOMSG existence now that they're in the tree.
> miod@ ok. (fgsch@)
lib/libkern
~ arch/hppa64/Makefile.inc
> Uncomment rule to build bcopy.S, and use that as our bcopy(9)
> implementation. (kettenis@)
~ arch/hppa64/bcopy.m4
> Make this actually work by using the right register numbers. In the
> conversion
> from hppa the fact that t1-t4 actually number down from r22-r19 got somehow
> lost. (kettenis@)
net
~ if_sl.c ~ if_fddisubr.c
> remove unneeded declarations that shadows existing vars; ok by many.
> (fgsch@)
~ bridgestp.c
> 'syncronize' -> 'synchronize'. From Diego Casati. (krw@)
net80211
~ ieee80211_proto.c
> Get trunk(4) working again with 802.11 devices. Spotted by ajacoutot at .
> Diff has been done in co-work with damien@, jsg@, reyk@, and claudio at .
> OK damien@ (mglocker@)
netinet
~ ip_ah.c
> MALLOC+bzero -> malloc+M_ZERO. Don't forget FREE->free this time.
> 'Fixes' two cases where the zeroing did not cover all the memory
> allocated.
> As with many of the last commits, looked at by chl@ in an earlier
> version. (krw@)
~ ip_ipsp.c
> MALLOC+bzero -> malloc+M_ZERO. Don't forget FREE->free this time.
> As with many of the last commits, looked at by chl@ in an earlier
> version. (krw@)
nfs
~ nfs_subs.c
> Make sure nfsrv_v2errmap[] really contains ELAST initialized values.
> (miod@)
~ nfs_subs.c
> shrink nfsrv_v2errmap[] and check table size more carefully to avoid
> dependency on ELAST; ok miod thib (deraadt@)
~ nfs_socket.c ~ nfs_subs.c
~ nfs_var.h
> Remove alot of dead kerberos code (add sane comments too).
> Cleanup and partly redo the way we create the RPC header, by having
> nfsm_rpchead() do a bit more work. Right now this is pretty RPCAUTH_UNIX
> centric, but since it is the only auth method we support right now thats
> fine.
> Make sure we can never generate a zero xid, thats forbidden by the RFC.
> Misc cleanup.
> tested by a few. (thib@)
~ nfs_syscalls.c ~ nfs_var.h
> Remove now unused functions dealing with kerberos
> authentication.
> ok beck@ (thib@)
sys
~ proc.h ~ sched.h
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
~ sched.h
> sched_lock_idle and sched_unlock_idle are obsolete now. (art@)
ufs/ext2fs
~ ext2fs_lookup.c
> MALLOC+memset(,0,) -> malloc+M_ZERO. Don't forget FREE->free this time.
> (krw@)
ufs/ufs
~ ufs_dirhash.c
> Initialize dh_blkfree with zero's.
> ok beck@ (krw@)
uvm
~ uvm_meter.c
> Make context switching much more MI:
> - Move the functionality of choosing a process from cpu_switch into
> a much simpler function: cpu_switchto. Instead of having the locore
> code walk the run queues, let the MI code choose the process we
> want to run and only implement the context switching itself in MD
> code.
> - Let MD context switching run without worrying about spls or locks.
> - Instead of having the idle loop implemented with special contexts
> in MD code, implement one idle proc for each cpu. make the idle
> loop MI with MD hooks.
> - Change the proc lists from the old style vax queues to TAILQs.
> - Change the sleep queue from vax queues to TAILQs. This makes
> wakeup() go from O(n^2) to O(n)
> there will be some MD fallout, but it will be fixed shortly.
> There's also a few cleanups to be done after this.
> deraadt@, kettenis@ ok (art@)
xfs
~ xfs_vnodeops-common.c
> More simple memset(,0,) -> M_ZERO changes. In this batch move to
> size(*p) as the first malloc() parameter where p is declared locally
> and thus easy to check. Add M_ZERO to gpe_table allocation in acpi.c
> even though there is no obvious bzero or memset nearby. (krw@)
~ xfs_syscalls-common.c
> A memset(,0,) -> M_ZERO change which also fixes a possible bug with
> malloc'ing one struct size but zero'ing the memory with another struct
> size.
> Feedback from todd@ tedu@ miod@ (krw@)
== usr.bin =========================================================== 10/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin
bgplg
~ bgplg.c
> Limit the allowed characters in a request to [a-zA-Z0-9-_.:/= ] everything
> else will cause an "invalid character in input" error.
> Fixes xss issue noticed by Anton Karpov.
> OK henning@, sthen@ (claudio@)
calendar
~ calendars/calendar.computer
> date UCB rescinded term 3; Mikolaj Kucharski (deraadt@)
compile_et
~ compile_et.1
> delete ugly $Header stuff; pr 5595 (deraadt@)
cvs
~ cvs.c
> break after exit() is useless
> ok ray@ (chl@)
~ client.c
> nb is an int, so it should be compared to an int.
> ok ray@ (chl@)
~ commit.c
> do not try to reset the default branch if the RCS file
> does not exist, fixes a segfault when commiting a new file.
> found by tobias@ who is going to make a nice little regression test for
> this. (joris@)
~ util.c
> cvs import must not parse files in local CVS directory.
> OK niallo@ (tobias@)
~ rcs.c
> Properly resolve branches, i.e. supplying branch base revision (e.g. 1.1.1)
> should resolve to latest branch revision (e.g. 1.1.1.1). (tobias@)
~ rcs.c
> Return NULL if a revision has been supplied that is higher than HEAD.
> This one and previous OK niallo@, "makes sense to me" xsa@ (tobias@)
~ file.c
> Do not warn user about files that are "no longer in the repository"
> during cvs import.
> OK niallo@ (tobias@)
~ annotate.c
> various style cleanups:
> - Copyright order
> - rev -> cvs_specified_tag
> - crev -> rev
> - line instead of alines[i] in for-loop
> OK niallo@ (tobias@)
~ annotate.c
> Do not use working HEAD (file_rcsrev), instead use file_rcs->rf_head to
> print revision info.
> OK niallo@ (tobias@)
~ annotate.c
> Properly parse supplied symbols with -r argument.
> OK niallo@ (tobias@)
~ annotate.c ~ rcs.c
~ rcs.h
> Added proper support for branch revisions in annotate.
> OK niallo@, twice :) (tobias@)
ftp
~ cookie.c
> Simplify the code that strips out \r by using strcspn.
> ok millert@ (pyr@)
gzsig
~ extern.h ~ sign.c
~ verify.c
> use __progname where needed
> ok otto@ (jasper@)
make
~ job.c
> rename functions for clarity.
> Add a global `got_signals' to quick-path thru handle_all_signals.
> remove most of the #ifdef USE_PGRP cruft: always be ready to handle those
> signals, just do not define a handler for them yet. (espie@)
~ job.c
> just-in-time signal handling: do not setup the handlers until we need them.
> (espie@)
~ job.c
> rename jobs to more descriptive runningJobs (and easier to grep) (espie@)
~ job.c
> minor refactor: introduce function start_queued_job, to make it more
> apparent
> JobStart and JobRestart often do the same thing.
> Move JobMakeArgv call around for similar reasons.
> Introduce new list (queuedJobs) we don't use yet. (espie@)
~ job.c
> don't bother returning anything from JobStart, we don't look at it anyways
> (espie@)
~ job.c
> factor out identical code (espie@)
~ job.c
> cut up JobStart, prepare the job to be run in a separate function (espie@)
~ compat.c ~ job.c
> allows node to be run twice, by not destroying the cmdnode contents.
> This way, you can run `make .BEGIN'. Okay, this will run .BEGIN twice.
> (espie@)
pcc
~ cc/cpp/scanner.l
> From master repo:
> > Don't let pragmas pass through if #ifdef'd away, from Stefan Kempf.
> (ragge@)
- arch/nova/README
> and another stray file (otto@)
- arch/m16c/TODO - arch/mips/TODO
> garbage collect a few files (otto@)
- arch/m16c/code.c - arch/m16c/local.c
- arch/m16c/local2.c - arch/m16c/macdefs.h
- arch/m16c/order.c - arch/m16c/table.c
- arch/mips/code.c - arch/mips/local.c
- arch/mips/local2.c - arch/mips/macdefs.h
- arch/mips/order.c - arch/mips/table.c
- arch/nova/code.c - arch/nova/local.c
- arch/nova/local2.c - arch/nova/macdefs.h
- arch/nova/order.c - arch/nova/table.c
- arch/pdp10/code.c - arch/pdp10/local.c
- arch/pdp10/local2.c - arch/pdp10/macdefs.h
- arch/pdp10/order.c - arch/pdp10/table.c
- arch/vax/code.c - arch/vax/local.c
- arch/vax/local2.c - arch/vax/macdefs.h
- arch/vax/order.c - arch/vax/table.c
- arch/x86/code.c - arch/x86/local.c
- arch/x86/local2.c - arch/x86/macdefs.h
- arch/x86/order.c - arch/x86/table.c
- cc/cc/Makefile - cc/cc/cc.1
- cc/cc/cc.c - cc/ccom/Makefile
- cc/ccom/Makefile.inc - cc/ccom/ccom.1
- cc/ccom/cgram.y - cc/ccom/gcc_compat.c
- cc/ccom/init.c - cc/ccom/inline.c
- cc/ccom/main.c - cc/ccom/optim.c
- cc/ccom/pass1.h - cc/ccom/pftn.c
- cc/ccom/scan.l - cc/ccom/stabs.c
- cc/ccom/symtabs.c - cc/ccom/trees.c
- cc/ccom/vax/Makefile - cc/ccom/x86/Makefile
- cc/cpp/Makefile - cc/cpp/cpp.1
- cc/cpp/cpp.c - cc/cpp/cpp.h
- cc/cpp/cpy.y - cc/cpp/scanner.l
- cc/cpp/tests/res1 - cc/cpp/tests/res2
- cc/cpp/tests/res3 - cc/cpp/tests/res4
- cc/cpp/tests/res5 - cc/cpp/tests/res6
- cc/cpp/tests/res7 - cc/cpp/tests/res8
- cc/cpp/tests/res9 - cc/cpp/tests/test1
- cc/cpp/tests/test2 - cc/cpp/tests/test3
- cc/cpp/tests/test4 - cc/cpp/tests/test5
- cc/cpp/tests/test6 - cc/cpp/tests/test7
- cc/cpp/tests/test8 - cc/cpp/tests/test9
- os/dragonfly/ccconfig.h - os/linux/ccconfig.h
- os/netbsd/ccconfig.h - os/netbsd/f77config.h
- os/none/ccconfig.h - os/openbsd/ccconfig.h
- os/openbsd/f77config.h ~ Makefile
~ cc/Makefile + ccconfig.h
+ f77config.h + cc/cc.1
+ cc/cc.c + ccom/Makefile
+ ccom/Makefile.inc + ccom/ccom.1
+ ccom/cgram.y + ccom/gcc_compat.c
+ ccom/init.c + ccom/inline.c
+ ccom/main.c + ccom/optim.c
+ ccom/pass1.h + ccom/pftn.c
+ ccom/scan.l + ccom/stabs.c
+ ccom/symtabs.c + ccom/trees.c
+ ccom/vax/Makefile + ccom/x86/Makefile
+ cpp/Makefile + cpp/cpp.1
+ cpp/cpp.c + cpp/cpp.h
+ cpp/cpy.y + cpp/scanner.l
+ m16c/TODO + m16c/code.c
+ m16c/local.c + m16c/local2.c
+ m16c/macdefs.h + m16c/order.c
+ m16c/table.c + mips/TODO
+ mips/code.c + mips/local.c
+ mips/local2.c + mips/macdefs.h
+ mips/order.c + mips/table.c
+ nova/README + nova/code.c
+ nova/local.c + nova/local2.c
+ nova/macdefs.h + nova/order.c
+ nova/table.c + pdp10/code.c
+ pdp10/local.c + pdp10/local2.c
+ pdp10/macdefs.h + pdp10/order.c
+ pdp10/table.c + vax/code.c
+ vax/local.c + vax/local2.c
+ vax/macdefs.h + vax/order.c
+ vax/table.c + x86/code.c
+ x86/local.c + x86/local2.c
+ x86/macdefs.h + x86/order.c
+ x86/table.c
> reorg pcc tree; requested by and ok deraadt@; ok ragge@ (otto@)
~ ccom/Makefile.inc ~ ccom/cgram.y
~ ccom/pass1.h ~ ccom/pftn.c
~ ccom/scan.l ~ ccom/trees.c
> Merge from ragge's repo:
> Add initial support for packed/aligned/rename pragmas.
> Still missing: Support for architectures with strict alignment.
> TODO: Cleanup the rename stuff so that it's not depending on gcc
> compat. (otto@)
~ ccom/gcc_compat.c
> merge from ragge's repo:
> Add __inline__ as gcc-compatible keyword. (otto@)
~ cc/cc.c
> merge from ragge's repo:
> Bugfix for .i files, by Jeremy Reed. (otto@)
~ config.h ~ cpp/cpp.c
~ mip/regs.c
> get rid of the configure remains (otto@)
~ cc/cc.c
> Remove two unneeded macro definitions, replace strdup() with copy() which
> does error checking.
> otto@, ragge@ ok (stefan@)
~ cc/cc.c ~ ccom/main.c
> signal handles must not call exit() or use stdio; ok otto (deraadt@)
~ ccom/trees.c
> initilize sflags and only print n_rval when initialized; tracking by
> Laurent Desnogues. ok ragge@ (otto@)
~ ccom/trees.c
> oops, fix merge error (otto@)
~ cpp/cpp.h ~ cpp/cpy.y
~ cpp/scanner.l
> rewrite of constant expression evaluation; taking into account
> signed vs unsigned and shortcutting of && and ||
> ok ragge@ (otto@)
~ ccom/pftn.c
> fix va_arg in conditional expressions; from ragge@ (otto@)
~ cc/cc.1
> From master repo:
> > Document -X flag. ok otto@ (stefan@)
~ cpp/cpp.c ~ cpp/cpp.h
~ cpp/cpy.y
> From master repo:
> > Add zero checks when dividing or taking modulo. ok otto@ (stefan@)
~ cpp/cpp.c ~ cpp/cpp.h
> Fix broken build. Reported by Gregory McGarry on pcc-list.
> ok otto@ (stefan@)
~ x86/local2.c ~ x86/macdefs.h
~ x86/table.c
> From master repo:
> Use xor to clear 32/16-bit registers instead of mov. ok ragge@, otto@
> (stefan@)
~ cc/Makefile
> don't hardcode TARGMACH (otto@)
~ ccom/Makefile
> by default just build ccom for the current host (otto@)
sudo
~ sudoers
> add SSH_AUTH_SOCK to env_keep.
> ok millert@ (steven@)
xlint
~ lint1/param.h
> pcc also knows about quads; ok cloder@ (otto@)
== usr.sbin ========================================================== 11/11 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin
acpidump
~ acpidump.h ~ aml/aml_memman.h
> attribute packed -> __packed
> OK millert (gilles@)
bgpctl
~ parser.c
> do not fallthrough (deraadt@)
bgpd
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ bgpd.h ~ config.c
~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
dhcpd
~ options.c
> "Minimum IP MTU" means what it says. Ensure that packets returned by
> dhcpd are the minimum size or larger no matter what the client thinks
> the minimum allowable size is. Found by Nahuel Riva and Gera Richarte.
> Fix by millert at . (krw@)
~ options.c
TAGGED OPENBSD_4_0
> merge fix from current, prompted by deraadt. original commit message:
> "Minimum IP MTU" means what it says. Ensure that packets returned by
> dhcpd are the minimum size or larger no matter what the client thinks
> the minimum allowable size is. Found by Nahuel Riva and Gera Richarte.
> Fix by millert at . (ckuethe@)
~ options.c
TAGGED OPENBSD_4_1
> merge fix from current, prompted by deraadt. original commit message:
> "Minimum IP MTU" means what it says. Ensure that packets returned by
> dhcpd are the minimum size or larger no matter what the client thinks
> the minimum allowable size is. Found by Nahuel Riva and Gera Richarte.
> Fix by millert at . (ckuethe@)
~ options.c
TAGGED OPENBSD_4_2
> merge fix from current, prompted by deraadt. original commit message:
> "Minimum IP MTU" means what it says. Ensure that packets returned by
> dhcpd are the minimum size or larger no matter what the client thinks
> the minimum allowable size is. Found by Nahuel Riva and Gera Richarte.
> Fix by millert at . (ckuethe@)
dvmrpd
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ dvmrpd.c ~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
hostapd
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
hoststated
~ check_icmp.c
> Do not notify twice about the same timeout in check_icmp, this previously
> corrupted statistics.
> Found out by Piotr Sikora < piotr at sikora.nu >. (pyr@)
~ parse.y
> use macro argument (deraadt@)
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ hoststated.h ~ hoststated.c
~ hce.c
> Silence some lint(1) warnings
> ok pyr@ (blambert@)
~ check_script.c
> avoid errno trashing in signal handler (deraadt@)
~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
httpd
~ src/modules/ssl/mod_ssl.c
> 'expresion' -> 'expression'. Reported by Jung on tech at . (krw@)
ifstated
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
ndp
~ gmt2local.c ~ gmt2local.h
> delete ugly $Header stuff; pr 5595 (deraadt@)
ntpd
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
ospf6ctl
+ Makefile + ospf6ctl.8
+ ospf6ctl.c + ospf6ctl.cat8
+ parser.c + parser.h
> Controller for the recently imported ospf6d(8) daemon.
> Currently simple stuff like show interface works.
> Not yet connected to the builds.
> ok claudio@ (norby@)
~ ospf6ctl.c
> Properly format the output of "show interface".
> In order to make room for the longer IPv6 addresses in the output,
> the two collumns nc and ac must go. Neighbor and Adjacent Neighbor Count
> is still avalable in "show interface detail".
> ok claudio@ (norby@)
~ ospf6ctl.c
> Sanitize the output of "show interface detail". (norby@)
~ ospf6ctl.c ~ parser.c
> please lint (deraadt@)
- ospf6ctl.cat8
> should not be in the tree (deraadt@)
ospf6d
+ Makefile + area.c
+ buffer.c + carp.c
+ control.c + control.h
+ database.c + hello.c
+ imsg.c + in_cksum.c
+ interface.c + iso_cksum.c
+ kroute.c + log.c
+ log.h + lsack.c
+ lsreq.c + lsupdate.c
+ name2id.c + neighbor.c
+ ospf6.h + ospf6d.8
+ ospf6d.c + ospf6d.conf.5
+ ospf6d.h + ospfe.c
+ ospfe.h + packet.c
+ parse.y + printconf.c
+ rde.c + rde.h
+ rde_lsdb.c + rde_spf.c
> Welcome ospf6d
> The new ospf6d daemon will support OSPFv3, basically OSPF for IPv6
> networks.
> It is heavily based on ospfd(8), it is more or less a copy and paste of it.
> Currently some unneeded stuff has been removed and the trasition from
> IPv4 to IPv6 has begun.
> ospf6d is not very usefull at the moment, it is being imported to allow
> more
> people to work on it concurrently.
> Not yet connected to the builds.
> ok claudio@ dlg@ (norby@)
~ ospf6d.8
> bad place for a stray space; (jmc@)
~ log.c
> Use a small cyclic buffer ring in log_in6addr() so that the function can be
> used more then once per printf. The current limit is 4. (claudio@)
~ ospf6.h ~ parse.y
~ ospf6d.h
> The router dead_interval switched from 32bit to a 16bit value in OSPFv3.
> OK norby@ (claudio@)
~ interface.c ~ ospfe.c
~ ospfe.h
> Instead of IP_RECVIF we use IPV6_RECVPKTINFO to get the ifindex and the
> destination address of incomming packets. This also removes the need for
> IP_HDRINCL. Additionally use IPV6_CHECKSUM to let the kernel do the
> necessary packet checksumming, way easier than doing it in userland.
> OK norby@ (claudio@)
~ hello.c ~ ospfe.h
~ packet.c
> Remaining bit of last night work. Make packet reception use all the goodies
> introduced (checksumming is no longer needed, no IP header checking is
> needed,
> get a in6_pktinfo with the ifindex and dest addr, cleanup and remove a lot
> of other code). With this ospf6d is able to send and receive first hello
> packets. OK norby@ (claudio@)
- in_cksum.c ~ Makefile
> Bye bye in_cksum.c, you're not needed here anymore. (claudio@)
~ interface.c
> Unneeded and to noisy debug output (claudio@)
~ hello.c
> Turn on the V6 and E bits in hello messages. This makes other OSPFv3
> routers
> a bit more happy. The E bit should be turned of in stub areas but there is
> no
> support for those at the moment. OK norby@ (claudio@)
~ kroute.c
> Address scope was probably the most stupid idea comming out of IPv6.
> Abusing the 3rd and 4th byte of a ff02::/32 address to store the scope is
> wrong wrong wrong. Depending on the calls it is not possible to compare
> addresses with IN6_ARE_ADDR_EQUAL(). Remove the scope hack when fetching
> interface addresses so that we never rely on that inside ospf6d.
> OK norby@ (claudio@)
~ database.c ~ hello.c
~ lsack.c ~ lsreq.c
~ lsupdate.c ~ ospfe.h
~ packet.c
> send_packet() now takes a struct in6_addr as destination instead of a
> struct sockaddr_in6 and builds the struct sockaddr_in6 internaly adding
> scope
> if necessary. While there switch to sendto() we don't need any of the
> sendmsg()
> features here.
> OK norby@ (claudio@)
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ database.c ~ neighbor.c
~ ospfe.h
> From ospfd:
> Do not overload nbr->options with the dd exchange bits. nbr->options is
> used
> by the hello protocol. Instead add a dd_more flag that is used together
> with
> (the renamed) dd_master flag. (claudio@)
~ area.c ~ database.c
~ hello.c ~ ospf6d.c
~ ospf6d.h ~ ospfe.c
~ parse.y ~ rde.c
> From ospfd:
> Bye bye global ospf options. OSPF options are per area (at least the one
> flag that we set). So introduce a area_ospf_options() function that will
> return the correct flags for each area. This makes stub area support a lot
> easier.
> Don't check for OSPF_OPTION_E in the parent. OSPF_OPTION_E is per area and
> so the parent process has no way to know if it should redistribute or not.
> (claudio@)
~ area.c ~ database.c
~ ospf6d.h ~ ospfe.h
> area_ospf_options() should return the options in network byte order.
> (claudio@)
~ ospf6d.h
> Last commit added way to much stuff. Revert the readd of global options.
> (claudio@)
~ hello.c
> The person that came up with the glorious idea to define a 24bit field
> should
> be hanged and shot. Add a ntohl() around the area_ospf_options() call.
> (claudio@)
~ ospfe.c ~ rde.c
> Disable some code that is currently far from working and results in
> frequent
> session resets because of bad packets. With this it seems we survive the
> hello and database exchange phases. A closer look at the DR and BDR
> calculation
> is still needed. (claudio@)
~ interface.c ~ neighbor.c
~ ospf6d.h ~ ospfe.h
> In OSPFv3 auth crypt is no longer (actually all the auth code is gone) so
> kill crypt_seq_num as well. (claudio@)
~ hello.c
> OSPFv3 runs on links not on networks so there is no need for a mask in the
> hello protocol. Kill no longer needed code. (claudio@)
~ interface.c ~ ospf6d.c
~ ospf6d.h
> No need to store the interface mask. An interface represents a link and is
> only addressed via its link local address. All networks on top of it are
> handled separately via type 8 Link-LSA. (claudio@)
~ kroute.c ~ ospf6d.h
~ parse.y
> There is no need to support interface em0:<IP> anymore. Remove most of the
> code the rest will be reused somehow. (claudio@)
~ ospf6d.8
> superceed -> supersede; from Tamas TEVESZ (jmc@)
~ ospf6d.c ~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
~ control.c ~ ospfe.c
~ ospfe.h
> From ospfd: Funny typo, it is fib not fip so adjust function name.
> (claudio@)
~ ospf6d.8
> 'syncronize' -> 'synchronize'. From Diego Casati. (krw@)
~ log.h ~ log.c
~ packet.c
> Switch log_sockaddr() to take a void * so it is not necessary to cast the
> sockaddr_in6 or sockaddr_in to a sockaddr. (claudio@)
ospfctl
~ ospfctl.c
> Cleanup the way we display timers.
> It does not make sense to display timer = 0 as "Stopped". When a timer is
> 0 it usually means that it is getting reset very soon. Display the string
> "00:00:00" instead.
> ok claudio@ (norby@)
~ ospfctl.c ~ parser.c
> please lint (deraadt@)
ospfd
~ database.c ~ neighbor.c
~ ospfe.h
> Do not overload nbr->options with the dd exchange bits. nbr->options is
> used
> by the hello protocol. Instead add a dd_more flag that is used together
> with
> (the renamed) dd_master flag.
> Tested and OK norby@ (claudio@)
~ ospfd.c
> Don't check for OSPF_OPTION_E in the parent. OSPF_OPTION_E is per area and
> so the parent process has no way to know if it should redistribute or not.
> Let the RDE decide. OK norby@ (claudio@)
~ area.c ~ database.c
~ hello.c ~ ospfd.h
~ ospfe.c ~ parse.y
~ rde.c
> Bye bye global ospf options. OSPF options are per area (at least the one
> flag that we set). So introduce a area_ospf_options() function that will
> return the correct flags for each area. This makes stub area support a lot
> easier.
> OK norby@ (claudio@)
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ ospfd.8
> superceed -> supersede; from Tamas TEVESZ (jmc@)
~ control.c ~ ospfe.c
~ ospfe.h
> Funny typo, it is fib not fip so adjust function name. (claudio@)
~ ospfd.c ~ parse.y
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
~ ospfd.8
> 'syncronize' -> 'synchronize'. From Diego Casati. (krw@)
~ ospfd.c
> Missing space (claudio@)
ripd
~ parse.y
> next step in the yylex unification: handle quoted strings in a nicer
> fashion
> as found in hoststated, and make all the code diff as clean as possible. a
> few issues remain mostly surrounding include support, which will likely be
> added to more of the grammers soon.
> ok norby pyr, others (deraadt@)
~ parse.y ~ ripd.c
> in all these programs using the same pfctl-derived parse.y, re-unify the
> yylex implementation and the code which interacts with yylex. this also
> brings the future potential for include support to all of the parsers.
> in the future please do not silly modifications to one of these files
> without checking if you are de-unifying the code.
> checked by developers in all these areas. (deraadt@)
tcpdump
~ INSTALL ~ README
~ addrtoname.c ~ addrtoname.h
~ appletalk.h ~ bootp.h
~ bpf_dump.c ~ decnet.h
~ ethertype.h ~ extract.h
~ fddi.h ~ gmt2local.c
~ gmt2local.h ~ gnuc.h
~ interface.h ~ ipx.h
~ llc.h ~ machdep.c
~ machdep.h ~ nameser.h
~ netbios.h ~ ntp.h
~ parsenfsfh.c ~ print-arp.c
~ print-atalk.c ~ print-atm.c
~ print-bgp.c ~ print-bootp.c
~ print-carp.c ~ print-cdp.c
~ print-decnet.c ~ print-dhcp6.c
~ print-domain.c ~ print-dvmrp.c
~ print-enc.c ~ print-ether.c
~ print-etherip.c ~ print-fddi.c
~ print-hsrp.c ~ print-icmp.c
~ print-igrp.c ~ print-ike.c
~ print-ip.c ~ print-ipsec.c
~ print-ipx.c ~ print-isoclns.c
~ print-krb.c ~ print-l2tp.c
~ print-llc.c ~ print-lwres.c
~ print-netbios.c ~ print-ntp.c
~ print-null.c ~ print-ospf.c
~ print-ospf6.c ~ print-pflog.c
~ print-pfsync.c ~ print-pim.c
~ print-ppp.c ~ print-raw.c
~ print-rip.c ~ print-sl.c
~ print-smb.c ~ print-snmp.c
~ print-sunrpc.c ~ print-tcp.c
~ print-tftp.c ~ print-timed.c
~ print-udp.c ~ print-udpencap.c
~ print-vrrp.c ~ print-wb.c
~ savestr.c ~ savestr.h
~ setsignal.c ~ setsignal.h
~ smbutil.c ~ tcpdump.c
~ util.c
> trash $Header goo which is just annoying; 5595 (deraadt@)
ypbind
~ ypbind.c
> missing newlines in printfs (deraadt@)
===============================================================================