[odc] Daily src changes for 2007-10-24

ODC auto at squish.net
Thu Oct 25 07:00:01 BST 2007


OpenBSD src changes summary for 2007-10-24
==========================================

lib/libc                                regress/usr.bin
sbin/pfctl                              share/man
sys/arch/amd64/amd64                    sys/arch/amd64/include
sys/arch/i386/i386                      sys/arch/i386/include
sys/arch/m88k/include                   sys/arch/m88k/m88k
sys/arch/mips64/mips64                  sys/arch/sgi/sgi
sys/arch/sparc64/dev                    sys/dev/i2c
sys/dev/ic                              sys/dev/pci
usr.bin/make                            usr.bin/pctr
usr.bin/ssh                             usr.bin/vmstat
usr.sbin/ripd                           

== lib =============================================================== 01/07 ==

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

libc

  ~ arch/m88k/SYS.h                       ~ arch/m88k/sys/syscall.S

  > Change the syscall invocation sequences to no longer fetch the last few
  > arguments from the stack, since the kernel will now copyin() them when
  > necessary.
  > This makes all system calls (but mmap()) slightly faster.
  > WARNING! After this commit, your binaries must run against a kernel
  > featuring m88k/m88k/trap.c r1.34 or better - i.e. a 4.1 or later kernel.
  > (miod@)

== regress =========================================================== 02/07 ==

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

usr.bin

  ~ ssh/sftp-cmds.sh                      ~ ssh/sftp-glob.sh
  ~ ssh/test-exec.sh                      

  > comprehensive tests for sftp escaping its interaction with globbing;
  > ok dtucker@ (djm@)

== sbin ============================================================== 03/07 ==

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

pfctl

  ~ pfctl.c                               

  > HW_PHYSMEM is unsigned
  > yuck & ok henning@ (wilfried@)

== share ============================================================= 04/07 ==

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

man

  ~ man4/man4.amd64/pctr.4                ~ man4/man4.i386/pctr.4

  > Remove idle loop counter.
  > ok art deraadt (mikeb@)

== sys =============================================================== 05/07 ==

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

arch/amd64/amd64

  ~ pctr.c                                

  > Don't spam the dmesg.
  > ok deraadt (mikeb@)

  ~ pctr.c                                

  > Remove idle loop counter.
  > ok art deraadt (mikeb@)

arch/amd64/include

  ~ pctr.h                                

  > Remove idle loop counter.
  > ok art deraadt (mikeb@)

arch/i386/i386

  ~ pctr.c                                

  > Don't spam the dmesg.
  > ok deraadt (mikeb@)

  ~ locore.s                              ~ pctr.c

  > Remove idle loop counter.
  > ok art deraadt (mikeb@)

arch/i386/include

  ~ pctr.h                                

  > Remove idle loop counter.
  > ok art deraadt (mikeb@)

arch/m88k/include

  ~ cpu.h                                 

  > Turn curcpu() into an inline function instead of a macro relying on a GCC
  > extension. (miod@)

arch/m88k/m88k

  ~ m88k_machdep.c                        

  > Remove sir_lock, superseded by the atomic bit operations. (miod@)

  ~ process.S                             

  > Rely on 16 byte pcb alignment, and use double loads and stores during
  > context switches. Should have been commited ages ago (when pcb alignment
  > was fixed). No functional change. (miod@)

arch/mips64/mips64

  ~ context.S                             

  > In cpu_switchto(), load pmap->pm_tlbpid manually, instead of expecting
  > pmap_update() to return with this value in v0. A void function. Bwahahaha.
  > (miod@)

  ~ cache_r5k.S                           

  > pipleine -> pipeline (miod@)

arch/sgi/sgi

  ~ genassym.cf                           

  > In cpu_switchto(), load pmap->pm_tlbpid manually, instead of expecting
  > pmap_update() to return with this value in v0. A void function. Bwahahaha.
  > (miod@)

arch/sparc64/dev

  ~ pcf8591_envctrl.c                     

  > Add support for power supply temperature sensors on the E450. (kettenis@)

  ~ pcfiic_ebus.c                         

  > Enable power supply temperature sensors on the E450. (kettenis@)

dev/i2c

  ~ i2c_scan.c                            

  > skip the ignored addresses completely (now covers the eeprom range, too); 
  > ok deraadt (cnst@)

  ~ asc7621.c                             

  > Uncomment PECI mode diagnosting bits.
  > ok deraadt (mikeb@)

  ~ i2c_scan.c                            

  > fix r1.101 (2007-10-07) regression of not calling the iic_dump() for a
  > successfully identified sensor chip that was not claimed by any driver;
  > whilst at it, get rid of the 'if defined' madness
  > ok deraadt (cnst@)

  ~ i2c_scan.c                            

  > make older gcc happier (deraadt@)

  ~ i2c_scan.c                            

  > remove the unused i2cbus_attach_args param from the probe functions;  ok
  > deraadt (cnst@)

dev/ic

  ~ rtw.c                                 

  > Unconditionally initialize all channels, regardless of locale.
  > Shaves 258 bytes from i386 kernel.
  > Feedback, guidance, testing, and ok jsg@ (blambert@)

dev/pci

  ~ mbg.c                                 

  > Define a bitmaks of status bits we're interested in, output/log all
  > relevant status bits, not only the first one. (mbalmer@)

== usr.bin =========================================================== 06/07 ==

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

make

  ~ main.c                                

  > fix logic of command line/ MAKEFLAGS parsing.
  > Forces make to see the -j4 in MAKEFLAGS and to actually pass it to submakes
  > (espie@)

pctr

  ~ pctr.c                                

  > Remove idle loop counter.
  > ok art deraadt (mikeb@)

ssh

  ~ sftp.c                                

  > rework argument splitting and parsing to cope correctly with common shell
  > escapes and make handling of escaped characters consistent with sh(1) and
  > between sftp commands (especially between ones that glob their arguments
  > and ones that don't).
  > parse command flags using getopt(3) rather than hand-rolled parsers.
  > ok dtucker@ (djm@)

  ~ scp.c                                 

  > factor out network read/write into an atomicio()-like function, and use it
  > to handle short reads, apply bandwidth limits and update counters.
  > make network IO non-blocking, so a small trickle of reads/writes has a
  > chance of updating the progress meter; bz #799
  > ok dtucker@ (djm@)

vmstat

  ~ dkstats.c                             

  > add missing checks against NULL after realloc calls
  > Ok gilles@ moritz@ (chl@)

== usr.sbin ========================================================== 07/07 ==

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

ripd

  ~ rde_rib.c                             

  > Don't use inet_ntoa() twice in a printf like function. So instead of
  > using inet_ntoa() on the netmask use mask2prefixlen(). (claudio@)

  ~ rde.c                                 ~ ripd.c
  ~ ripe.c                                

  > Sync a bit more with ospfd. Do not fatal if one of the pipes is closed,
  > instead do a "nice" shutdown. Ignore SIGHUP in the childs even though we
  > have no reload support (yet). Remove some redundant code and clean up some
  > types. (claudio@)

  ~ parse.y                               ~ neighbor.c
  ~ interface.c                           

  > Garbage collect unused stuff found by lint. (claudio@)

  ~ ripd.h                                

  > ctl_rt flags should be u_int16_t -- noticed by lint (claudio@)

  ~ buffer.c                              

  > The index should be unsigned int -- another lint finding (claudio@)

  ~ control.c                             

  > Use an ssize_t instead of int like in all other imsg handlers. Hint by lint
  > (claudio@)

  ~ ripd.c                                

  > Add the 5 needed lines to support -D makro=value option. Another surprising
  > finding by lint. (claudio@)

  ~ ripd.8                                

  > Document -D. Stolen from bgpd.8 (claudio@)

  ~ kroute.c                              

  > An u_int8_t is enough as prefixlen. Another lint warning gone. (claudio@)

  ~ rde.c                                 ~ rde.h
  ~ rde_rib.c                             

  > The metric is a 8bit value ranging from 1 to INFINITY (defined as a
  > whopping
  > 16). So use u_int8_t everywhere. (claudio@)

  ~ auth.c                                ~ message.c
  ~ packet.c                              ~ ripe.h

  > Make the packet buffer u_int8_t * everywhere. This is binary data and not
  > a string. (claudio@)

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


More information about the odc mailing list