[odc] Daily src changes for 2007-05-08

ODC auto at squish.net
Wed May 9 07:00:01 BST 2007


OpenBSD src changes summary for 2007-05-08
==========================================

bin/ls                                  distrib/sets
sbin/wsconsctl                          share/man
sys/arch/i386/i386                      sys/arch/i386/isa
sys/arch/luna88k/include                sys/arch/macppc/macppc
sys/arch/macppc/pci                     sys/arch/mvme68k/include
sys/arch/mvme88k/include                sys/arch/sparc/conf
sys/arch/sparc/include                  sys/arch/sparc/sparc
sys/dev                                 sys/dev/cardbus
sys/dev/eisa                            sys/dev/ic
sys/dev/isa                             sys/dev/pci
sys/dev/pcmcia                          sys/dev/usb
sys/dev/wscons                          sys/kern
sys/net                                 sys/netinet
sys/netinet6                            sys/scsi
sys/sys                                 usr.sbin/pkg_add

== bin =============================================================== 01/06 ==

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

ls

  ~ ls.1                                  

  > small update, now that -g does something; ok millert (jmc@)

== distrib =========================================================== 02/06 ==

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

sets

  ~ lists/base/md.hppa                    ~ lists/comp/md.hppa

  > sync (deraadt@)

  ~ lists/base/md.hp300                   ~ lists/base/md.mac68k
  ~ lists/base/md.mvme68k                 ~ lists/comp/md.hp300
  ~ lists/comp/md.mac68k                  ~ lists/comp/md.mvme68k

  > sync (deraadt@)

  ~ lists/base/md.hppa64                  ~ lists/base/md.mvmeppc
  ~ lists/comp/md.hppa64                  ~ lists/comp/md.mvmeppc

  > sync (deraadt@)

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

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

wsconsctl

  ~ mouse.c                               ~ util.c
  ~ wsconsctl.h                           

  > Introduce the mouse.rawmode and mouse.scale fields for
  > touchscreen calibration via ioctls.
  > ok miod@ (robert@)

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

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

man

  ~ man5/pf.conf.5                        

  > Document the fact that 'allow-opts' applies to IPv6 now as well.
  > ok jmc@ dhartmei@ henning@ deraadt@ claudio@ (mcbride@)

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

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

arch/i386/i386

  ~ cpu.c                                 

  > cpu_init() is only used for the MP case (for now) (deraadt@)

  ~ process_machdep.c                     

  > process_s87_to_xmm() is only needed if PTRACE (deraadt@)

arch/i386/isa

  ~ pccom.c                               

  > put more foo_detach() routines where they are used (deraadt@)

  ~ pccom.c                               

  > move com_activate() to where it belongs (deraadt@)

  ~ pccom.c                               

  > backout while the amd64 puc situation is rethought (deraadt@)

arch/luna88k/include

  ~ autoconf.h                            

  > remove more junk in the setroot() code path (deraadt@)

arch/macppc/macppc

  ~ autoconf.c                            

  > uninitialized variable passed to setroot() (deraadt@)

  ~ autoconf.c                            

  > no need to print boot device name twice (deraadt@)

arch/macppc/pci

  ~ ht.c                                  

  > correct dmesg output (deraadt@)

arch/mvme68k/include

  ~ autoconf.h                            

  > remove more junk in the setroot() code path (deraadt@)

arch/mvme88k/include

  ~ autoconf.h                            

  > remove more junk in the setroot() code path (deraadt@)

arch/sparc/conf

  ~ Makefile.sparc                        

  > Switch sparc to __HAVE_CPUINFO.
  > miod@ tested (since I hacked it up blindly) and ok. (art@)

  ~ Makefile.sparc                        

  > Remove cross-compiling stuff that is neither correct nor should have been
  > committed. (art@)

arch/sparc/include

  ~ _types.h                              ~ cpu.h
  ~ psl.h                                 

  > Switch sparc to __HAVE_CPUINFO.
  > miod@ tested (since I hacked it up blindly) and ok. (art@)

arch/sparc/sparc

  ~ cpu.c                                 ~ cpuvar.h
  ~ genassym.cf                           ~ locore.s
  ~ trap.c                                

  > Switch sparc to __HAVE_CPUINFO.
  > miod@ tested (since I hacked it up blindly) and ok. (art@)

dev

  ~ softraid.c                            ~ softraidvar.h

  > Assemble disk from user supplied metadata.
  > Verify all metada to make sure it is correct; this includes sorting the
  > chunks
  > Add 2 new functions for printing uuid and metadata. (marco@)

dev/cardbus

  ~ if_dc_cardbus.c                       

  > dc_detach() is only used by cardbus code, so move it there; ok jsg
  > (deraadt@)

  ~ if_rl_cardbus.c                       

  > rl_detach() is only used for cardbus case (deraadt@)

  ~ com_cardbus.c                         ~ if_fxp_cardbus.c

  > put more foo_detach() routines where they are used (deraadt@)

  ~ com_cardbus.c                         ~ if_fxp_cardbus.c

  > repair com/fxp foo_detach() stuff, the hardest part being the pccom mess
  > (deraadt@)

  ~ com_cardbus.c                         

  > move com_activate() to where it belongs (deraadt@)

  ~ com_cardbus.c                         

  > backout while the amd64 puc situation is rethought (deraadt@)

dev/eisa

  ~ aha1742.c                             

  > ahb_debug should be inside #ifdef AHBDEBUG (deraadt@)

dev/ic

  ~ dc.c                                  

  > dc_detach() is only used by cardbus code, so move it there; ok jsg
  > (deraadt@)

  ~ siop.c                                

  > turn off SIOP_STATS; ok krw (deraadt@)

  ~ ami.c                                 

  > ami chips have no mechanism to take commands off them once we've submitted
  > them. this means that we cant reliably complete an io before the chip says
  > we can because it can possibly complete later and overwrite memory it no
  > longer owns, or write garbage to disk.
  > so this diff forces the timeout on a scsi io to be as long as the chip
  > should ever take, which is sixty seconds. after much discussion with marco
  > we also decided to add a bit more so the time spent in the software runq
  > wouldnt affect the ios lifetime too much.
  > instead of completing ios out of the timeout, this simply warns the
  > operator to check the state of the volumes if things are starting to slow
  > down.
  > ok marco@ (dlg@)

  ~ wdc.c                                 ~ wdcvar.h

  > atapi_print() is never used (deraadt@)

  ~ wdc.c                                 ~ wdcvar.h

  > wdcrestart() is never used (deraadt@)

  ~ rtl81x9.c                             ~ rtl81x9reg.h

  > rl_detach() is only used for cardbus case (deraadt@)

  ~ fxp.c                                 ~ fxpvar.h
  ~ com.c                                 ~ comvar.h

  > put more foo_detach() routines where they are used (deraadt@)

  ~ com.c                                 ~ comvar.h

  > move com_activate() to where it belongs (deraadt@)

  ~ re.c                                  

  > non-static inline functions waste space (deraadt@)

  ~ com.c                                 ~ comvar.h

  > backout while the amd64 puc situation is rethought (deraadt@)

dev/isa

  ~ if_we.c                               

  > old school media handling is no longer used; ok jsg (deraadt@)

  ~ fd.c                                  

  > 2nd one found:  If you really wish to make a function __inline (which will
  > be called once), please also make it static so that the compiler does not
  > generate a 2nd one to export to outside callers!  that is a stupid waste
  > of space. (deraadt@)

dev/pci

  ~ pcidevs                               

  > add Radeon RX1650 XT, Radeon RX1650 XT Secondary and SoundBlaster X-Fi
  > (robert@)

  ~ pcidevs.h                             ~ pcidevs_data.h

  > regen (robert@)

  ~ if_tht.c                              

  > write the mac back to the chip when we bring it up. (dlg@)

  ~ if_tht.c                              

  > add the multicast filter registers (dlg@)

  ~ if_tht.c                              

  > make a start at dealing with interface flags. this toggles the hardware
  > promiscuity depending on the if_flags IFF_PROMISC bit. (dlg@)

  ~ if_tl.c                               

  > #if 0 a function that is only ever called in #if 0 (deraadt@)

dev/pcmcia

  ~ if_ne_pcmcia.c                        

  > places they get called are #ifdef notdef, so the code itself can be too
  > (deraadt@)

  ~ if_ep_pcmcia.c                        

  > #ifdef notdef functions which are only called in #ifdef notdef contexts
  > (deraadt@)

dev/usb

  ~ uts.c                                 

  > Make use of struct wsmouse_calibcoords and add ioctl support
  > for getting and passing calibration values. Move the scale
  > struct to uts_softc so that each device can have it's own values
  > instead of using the global one.
  > ok miod@ (robert@)

dev/wscons

  ~ wsconsio.h                            

  > add swapxy, resx and resy to struct wsmouse_calibcoords,
  > so we can store these values here; ok miod@ (robert@)

kern

  ~ kern_sig.c                            

  > First step in making the SCHED_LOCK not recursive.
  > - Defer the sending of SIGCHLD and waking the parent when a process goes
  > to sleep. We set a flag in the process "P_STOPPED" and schedule a
  > timeout that checks all processes for this flag and sends SIGCHLD
  > and wakes the parent on those processes that have this flag set.
  > - Delay the waking of the parent in psignal on SIGCONT until we've
  > released the SCHED_LOCK at the end.
  > - In issignal, check for P_SINTR as a way to see we were called from
  > the tsleep path with SCHED_LOCK locked and don't skip locking
  > in that case. (This is slightly ugly, but it works until we have
  > some better way).
  > miod@ ok (after finding and fixing an MD bug on sgi) (art@)

  ~ kern_sig.c                            

  > Needs an extern decl for generic soft interrupts platforms. (miod@)

  ~ vfs_conf.c                            

  > struct mount *rootfs is never used (deraadt@)

  ~ kern_lock.c                           

  > lockmgr_printinfo() is only called from #ifdef DIAGNOSTIC positions, so
  > #ifdef DIAGNOSTIC it too (deraadt@)

net

  ~ route.c                               

  > rtalloc2() and  rtalloc_noclone() inside NPF > 0 (deraadt@)

  ~ route.c                               

  > wildcard is nowhere used anymore. Figured out by deraadt@ (claudio@)

  ~ pf.c                                  

  > Routing headers are dangerous. Deal with them the same way as IPv4 options:
  > drop all packets with routing headers unless the matching rule explicitly
  > specifies 'allow-opts'.
  > ok dhartmei@ henning@ deraadt@ claudio@ (mcbride@)

  ~ pf.c                                  

  > block ALL packets with rthdr0 in pf_test6(). We already do this
  > in ip6_input(), but pf_test6() is called from bridge code as well.
  > ok dhartmei@ henning@ deraadt@ claudio@ (mcbride@)

netinet

  ~ tcp_subr.c                            

  > variables used by #ifdef code should be inside #ifdef too (deraadt@)

netinet6

  ~ ip6_input.c                           

  > KILL all IPv6 packets with the rthdr0 extension header or multiple routing
  > headers, regardless of forwarding path. It's the sane thing to do.
  > ip6_check_rthdr0() function from claudio@
  > ok deraadt@ claudio@ henning@ (mcbride@)

scsi

  ~ scsiconf.c                            

  > all scsidebug_*-using code is under #ifdef, so the variables themselves
  > should be too (deraadt@)

sys

  ~ proc.h                                

  > First step in making the SCHED_LOCK not recursive.
  > - Defer the sending of SIGCHLD and waking the parent when a process goes
  > to sleep. We set a flag in the process "P_STOPPED" and schedule a
  > timeout that checks all processes for this flag and sends SIGCHLD
  > and wakes the parent on those processes that have this flag set.
  > - Delay the waking of the parent in psignal on SIGCONT until we've
  > released the SCHED_LOCK at the end.
  > - In issignal, check for P_SINTR as a way to see we were called from
  > the tsleep path with SCHED_LOCK locked and don't skip locking
  > in that case. (This is slightly ugly, but it works until we have
  > some better way).
  > miod@ ok (after finding and fixing an MD bug on sgi) (art@)

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

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

pkg_add

  ~ OpenBSD/PackageLocator.pm             

  > spec match works better with a list (espie@)

  ~ OpenBSD/PackageRepository.pm          

  > allow match_spec on repository (espie@)

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


More information about the odc mailing list