[odc] Daily src changes for 2004-09-17

[ODC]

OpenBSD src changes summary for 2004-09-17
==========================================

etc/rc                                  etc/rc.conf
games/hack                              sbin/isakmpd
sbin/pflogd                             share/man
sys/arch/i386/conf                      sys/arch/mips64/include
sys/arch/mips64/mips64                  sys/arch/sgi/conf
sys/dev/cardbus                         sys/dev/pci
sys/kern                                sys/net
sys/netinet                             sys/sys
usr.sbin/bgpd                           usr.sbin/pkg_add

== etc =============================================================== 01/06 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/etc

rc

  ~ rc                                    

  > Set up alternate keyboard encoding earlier in rc.
  > ok miod@ henning@ deraadt@ (mcbride@)

rc.conf

  ~ rc.conf                               

  > correct comment: spamd flags in spamd(8), not spamd-setup(8);
  > from frank bax on misc@;
  > ok deraadt@ (jmc@)

== games ============================================================= 02/06 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/games

hack

  ~ hack.zap.c                            

  > dothrow() calls bhit() w/ both null hit functions thus have to check before
  > calling those; from Juha Erkkila <erkkila at cc.jyu.fipr3923 (mickey@)

== sbin ============================================================== 03/06 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/sbin

isakmpd

  ~ message.c                             

  > Permit next payload type NAT-OA. Noted by Kamel Messaoudi. (ho@)

  ~ init.c                                

  > #include <stdlib.h> for srandom(). (ho@)

  ~ exchange.c                            ~ ike_quick_mode.c
  ~ ipsec.c                               ~ key.c
  ~ pf_key_v2.c                           

  > Missing #ifdefs. (ho@)

  ~ isakmpd.c                             

  > avoid signal race.
  > ok ho@ otto@ (hshoexer@)

pflogd

  ~ pflogd.c                              

  > ugly spacing (deraadt@)

== share ============================================================= 04/06 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/share

man

  ~ man4/options.4                        

  > Remove option EON from kernel and options(4)
  > ok henning@ (msf@)

  ~ man4/man4.hppa/uturn.4                

  > tweaks; ok mickey@ (jmc@)

== sys =============================================================== 05/06 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/sys

arch/i386/conf

  ~ DISKLESS                              

  > Remove option EON from kernel and options(4)
  > ok henning@ (msf@)

arch/mips64/include

  ~ param.h                               

  > Crank MSIZE and NMBCLUSTERS, per other 64bit arches. (miod@)

  ~ pmap.h                                ~ vmparam.h

  > Long-due mips pmap potpourri:
  > - un-__P()ee.
  > - remove splimp() protection.
  > - allow pmap_extract() to return FALSE, dammit!
  > XXX some code under arch/mips64 considers pa == 0 as failure, instead
  > of checking the return value of pmap_extract(). Free commits.
  > - do not peek at uvm vm_page flags for ref/mod accounting, but use real
  > MD flags in pv lists, and implement pmap_is_* and pmap_clear_*, so uvm
  > (which knows better) does the right thing.
  > - debug code #ifdef PMAPDEBUG, not #ifdef DIAGNOSTIC.
  > - merge pmap_init() and pmap_pinit(), pmap_destroy() and pmap_release(),
  > and update comments.
  > - better pmap_enter() logic, from NetBSD.
  > - do not use IS_VM_PHYSADDR() to know if we can use PHYS_TO_VM_PAGE()
  > unchecked. Instead, compute PHYS_TO_VM_PAGE(), and check it against
  > NULL. Saves a vm_physseg_find() call each time.
  > - pass vm_page_t to internal function whenever possible, rather than
  > paddr_t, so that PHYS_TO_VM_PAGE() needs not be computed again.
  > - in the same way, do not compute PHYS_TO_KSEG0(foo) several times in a
  > row (though PHYS_TO_KSEG0 is trivial).
  > - allow userspace address to cross 0x80000000 (may be useful in the
  > future).
  > - implement PMAP_CANFAIL.
  > As a bonus:
  > - switch from HAVE_PMAP_PHYSSEG to HAVE_VM_PAGE_MD.
  > - remove dead (pre-pmap_pv_pool) code.
  > - KNF fixes. (miod@)

arch/mips64/mips64

  ~ lcore_ddb.S                           

  > Kill an assembler warning. (miod@)

  ~ sendsig.c                             

  > Properly setup register ra in sendsig() in 64bit world. (miod@)

  ~ pmap.c                                ~ trap.c

  > Long-due mips pmap potpourri:
  > - un-__P()ee.
  > - remove splimp() protection.
  > - allow pmap_extract() to return FALSE, dammit!
  > XXX some code under arch/mips64 considers pa == 0 as failure, instead
  > of checking the return value of pmap_extract(). Free commits.
  > - do not peek at uvm vm_page flags for ref/mod accounting, but use real
  > MD flags in pv lists, and implement pmap_is_* and pmap_clear_*, so uvm
  > (which knows better) does the right thing.
  > - debug code #ifdef PMAPDEBUG, not #ifdef DIAGNOSTIC.
  > - merge pmap_init() and pmap_pinit(), pmap_destroy() and pmap_release(),
  > and update comments.
  > - better pmap_enter() logic, from NetBSD.
  > - do not use IS_VM_PHYSADDR() to know if we can use PHYS_TO_VM_PAGE()
  > unchecked. Instead, compute PHYS_TO_VM_PAGE(), and check it against
  > NULL. Saves a vm_physseg_find() call each time.
  > - pass vm_page_t to internal function whenever possible, rather than
  > paddr_t, so that PHYS_TO_VM_PAGE() needs not be computed again.
  > - in the same way, do not compute PHYS_TO_KSEG0(foo) several times in a
  > row (though PHYS_TO_KSEG0 is trivial).
  > - allow userspace address to cross 0x80000000 (may be useful in the
  > future).
  > - implement PMAP_CANFAIL.
  > As a bonus:
  > - switch from HAVE_PMAP_PHYSSEG to HAVE_VM_PAGE_MD.
  > - remove dead (pre-pmap_pv_pool) code.
  > - KNF fixes. (miod@)

  ~ busdma.c                              ~ vm_machdep.c

  > Now that we can check for pmap_extract() failure properly, do so. (miod@)

  ~ vm_machdep.c                          

  > In vmapbuf(), pass VM_PROT_READ | VM_PROT_WRITE as flags hints too. (miod@)

arch/sgi/conf

  ~ RAMDISK                               

  > Remove option EON from kernel and options(4)
  > ok henning@ (msf@)

dev/cardbus

  ~ cardbusdevs                           

  > add Linksys vendor ID and Linksys PCM200 ID
  > From: Chuck Gagnon <gagnocg at mac dot com> (brad@)

  ~ cardbusdevs.h                         ~ cardbusdevs_data.h

  > regen (brad@)

dev/pci

  ~ amdpm.c                               ~ piixpm.c

  > Some cleanup:
  > - don't mix unsigned and u_int across the code
  > - un'static some funcs
  > ok art@ (grange@)

kern

  ~ kern_tc.c                             

  > Some cleanup:
  > - don't mix unsigned and u_int across the code
  > - un'static some funcs
  > ok art@ (grange@)

net

  ~ if_pfsync.c                           ~ pf.c

  > Clean up reference counting wrt state creation and destruction. Fixes
  > problems with adaptive timeouts, max-states limits, and rules not being
  > freed from memory.
  > Diff from Chris Pascoe.
  > ok henning@ dhartmei@ (mcbride@)

netinet

  ~ in_proto.c                            

  > Remove option EON from kernel and options(4)
  > ok henning@ (msf@)

sys

  ~ mbuf.h                                

  > move MINCLSIZE back to preferring clusters; there is some fallout because
  > the other parts of this were not done.  This is a conservative decision
  > for this release. (deraadt@)

== usr.sbin ========================================================== 06/06 ==

  http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin

bgpd

  ~ imsg.c                                

  > Fix log message (claudio@)

pkg_add

  ~ pkg_create                            

  > warn for dirrm. (espie@)

===============================================================================