[odc] Daily src changes for 2009-03-23

ODC auto at squish.net
Tue Mar 24 07:00:02 GMT 2009


OpenBSD src changes summary for 2009-03-23
==========================================

etc/security                            lib/libc
sbin/disklabel                          share/man
share/zoneinfo                          sys/arch/amd64/amd64
sys/arch/amd64/conf                     sys/arch/arm/arm
sys/arch/zaurus/stand/zbsdmod           sys/dev/mii
sys/dev/pci                             sys/dev/usb
sys/kern                                sys/net
sys/scsi                                sys/sys
sys/uvm                                 usr.bin/cvs
usr.bin/ssh                             usr.sbin/config
usr.sbin/smtpd                          

== etc =============================================================== 01/07 ==

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

security

  ~ security                              

  > Small rewording to remove ambiguity.
  > A package can actually "change" after an update even if its version does
  > not and yet it will not appear in the security(8) output.
  > wording by eric@
  > ok mbalmer@ (ajacoutot@)

== lib =============================================================== 02/07 ==

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

libc

  ~ gen/fts.3                             

  > add a HISTORY section to fts(3).
  > diff proposed by an anonymous user on bugs@; tweaked by jmc@;
  > verified using the superb 4BSD collection at pdp-11.org.ru (thanks form@).
  > ok jmc@ (sobrado@)

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

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

disklabel

  ~ disklabel.8                           

  > tweak the description of -A a little more;
  > feedback/ok krw (jmc@)

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

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

man

  ~ man7/hier.7                           

  > some smtpd bits; ok jacekm gilles (jmc@)

  ~ man4/Makefile                         + man4/otus.4

  > otus(4), a driver for Atheros AR9001U USB IEEE 802.11 devices.
  > more work is required but basic operations work.
  > requires a non-free firmware to operate.
  > partly based on source code released under the ISC by Atheros
  > Communications for Linux, although I had to rewrite almost everything
  > (actually I only used some .h files from the Atheros driver.)
  > there also exists a rewrite of the Atheros driver for Linux (ar9170)
  > but the guy decided to make the code less free by wrapping the GPL
  > around the ISC.
  > committed over a NETGEAR WNDA3100.
  > ok deraadt@ (damien@)

  ~ man4/usb.4                            

  > Xr otus(4) (deraadt@)

zoneinfo

  ~ datfiles/africa                       ~ datfiles/asia
  ~ datfiles/southamerica                 

  > Update to tzdata2009d from elsie.nci.nih.gov (millert@)

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

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

arch/amd64/amd64

  ~ pmap.c                                

  > Processor affinity for processes.
  > - Split up run queues so that every cpu has one.
  > - Make setrunqueue choose the cpu where we want to make this process
  > runnable (this should be refined and less brutal in the future).
  > - When choosing the cpu where we want to run, make some kind of educated
  > guess where it will be best to run (very naive right now).
  > Other:
  > - Set operations for sets of cpus.
  > - load average calculations per cpu.
  > - sched_is_idle() -> curcpu_is_idle()
  > tested, debugged and prodded by many@ (art@)

arch/amd64/conf

  ~ Makefile.amd64                        

  > Add missing dependency generation for assym.h (...the lack of which
  > made testers of art's affinity diff go insane)
  > ok krw@ miod@ (guenther@)

arch/arm/arm

  ~ pmap.c                                

  > Processor affinity for processes.
  > - Split up run queues so that every cpu has one.
  > - Make setrunqueue choose the cpu where we want to make this process
  > runnable (this should be refined and less brutal in the future).
  > - When choosing the cpu where we want to run, make some kind of educated
  > guess where it will be best to run (very naive right now).
  > Other:
  > - Set operations for sets of cpus.
  > - load average calculations per cpu.
  > - sched_is_idle() -> curcpu_is_idle()
  > tested, debugged and prodded by many@ (art@)

arch/zaurus/stand/zbsdmod

  ~ compat_linux.h                        

  > sys/types.h before other things (deraadt@)

dev/mii

  ~ atphy.c                               

  > use auto-negotiation for manual media type selection. This fixes
  > establishment of 10/100Mbps link on ale(4).
  > From FreeBSD (kevlo@)

dev/pci

  ~ ips.c                                 

  > Reorganize NBIO and SMALL_KERNEL handling again but this time without
  > ramdisk breakage. (grange@)

  ~ if_msk.c                              

  > Prevent null pointer dereference that could happen when we process an RX
  > interrupt that was queued while stopping the interface. (kettenis@)

  ~ if_msk.c                              ~ if_mskvar.h

  > Switch msk(4) over to using MCLGETI when allocating mbufs for the RX ring.
  > This gets rid of one of the evil jumbo allocators! (kettenis@)

dev/usb

  ~ files.usb                             + if_otus.c
  + if_otusreg.h                          

  > otus(4), a driver for Atheros AR9001U USB IEEE 802.11 devices.
  > more work is required but basic operations work.
  > requires a non-free firmware to operate.
  > partly based on source code released under the ISC by Atheros
  > Communications for Linux, although I had to rewrite almost everything
  > (actually I only used some .h files from the Atheros driver.)
  > there also exists a rewrite of the Atheros driver for Linux (ar9170)
  > but the guy decided to make the code less free by wrapping the GPL
  > around the ISC.
  > committed over a NETGEAR WNDA3100.
  > ok deraadt@ (damien@)

kern

  ~ vfs_lockf.c                           

  > advisory locks should return ENOLCK if we're out of locks.
  > ok deraadt@ fgs@ (otto@)

  ~ kern_fork.c                           ~ kern_sched.c
  ~ kern_synch.c                          ~ sched_bsd.c

  > Processor affinity for processes.
  > - Split up run queues so that every cpu has one.
  > - Make setrunqueue choose the cpu where we want to make this process
  > runnable (this should be refined and less brutal in the future).
  > - When choosing the cpu where we want to run, make some kind of educated
  > guess where it will be best to run (very naive right now).
  > Other:
  > - Set operations for sets of cpus.
  > - load average calculations per cpu.
  > - sched_is_idle() -> curcpu_is_idle()
  > tested, debugged and prodded by many@ (art@)

  ~ vfs_bio.c                             

  > fix buffer cache pending writs statistic so it does not go negative.
  > this ensures we ignore counting any buffers returning through biodone()
  > for which B_PHYS has been set - which should be set on all transfers
  > that manually do raw io bypassing the buffer cache by setting up their
  > own buffer and calling strategy..
  > ok thib@, todd@, and now that he is a buffer cache and nfs hacker oga@
  > (beck@)

net

  ~ if_pfsync.c                           

  > wait an appropriate amount of time before giving up on a bulk update,
  > rather than giving up after a hardcoded 5 seconds (which is generally much
  > too short an interval for a bulk update).
  > pointed out by david@, eyeballed by mcbride@ (dlg@)

scsi

  ~ iscsi.h                               

  > actually follow rfc 3720, from Remco. (dlg@)

sys

  ~ proc.h                                ~ sched.h

  > Processor affinity for processes.
  > - Split up run queues so that every cpu has one.
  > - Make setrunqueue choose the cpu where we want to make this process
  > runnable (this should be refined and less brutal in the future).
  > - When choosing the cpu where we want to run, make some kind of educated
  > guess where it will be best to run (very naive right now).
  > Other:
  > - Set operations for sets of cpus.
  > - load average calculations per cpu.
  > - sched_is_idle() -> curcpu_is_idle()
  > tested, debugged and prodded by many@ (art@)

uvm

  ~ uvm_meter.c                           ~ uvm_page.c

  > Processor affinity for processes.
  > - Split up run queues so that every cpu has one.
  > - Make setrunqueue choose the cpu where we want to make this process
  > runnable (this should be refined and less brutal in the future).
  > - When choosing the cpu where we want to run, make some kind of educated
  > guess where it will be best to run (very naive right now).
  > Other:
  > - Set operations for sets of cpus.
  > - load average calculations per cpu.
  > - sched_is_idle() -> curcpu_is_idle()
  > tested, debugged and prodded by many@ (art@)

  ~ uvm_device.c                          

  > Remove space added in last commit. Pointed out by miod at . (oga@)

  ~ uvm_swap.c                            ~ uvm_swap_encrypt.c
  ~ uvm_swap_encrypt.h                    

  > turn a for (i = 0; i < size; i++) arc4random(); loop into arc4random_buf().
  > Since that function is now so small (2 lines), inline it into it's only
  > user.
  > Shaves some bytes (104 on amd64).
  > ok deraadt@, blambert at . djm@ liked an earlier diff. (oga@)

  ~ uvm_swap_encrypt.h                    

  > Whitespace nit in previous. (oga@)

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

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

cvs

  ~ diff.c                                

  > any directory messages are, believe it or not, suppose to go to LP_ERR
  > instead of LP_NOTICE, while here modify output for non-existing
  > files to match "? foobar" instead of "cannot find foobar".
  > makes our output a little less alien and a little bit more gnuish. (joris@)

  ~ update.c                              

  > LP_NOTICE should be LP_ERR when showing directory messages. (joris@)

  ~ getlog.c                              

  > add rlog -d support, ported from openrcs by
  > Josh Elsasser <josh at elsasser.org> (joris@)

  ~ getlog.c                              

  > remove debug, bleh (joris@)

  ~ init.h                                

  > simplify (joris@)

  ~ buf.c                                 

  > kill SIZE_LEFT() macro, no longer used (joris@)

ssh

  ~ ssh-agent.c                           

  > Fixed a possible out-of-bounds memory access if the environment variable
  > SHELL is shorter than 3 characters.
  > with input by and ok dtucker (tobias@)

  ~ ssh-agent.c                           

  > My previous commit didn't fix the problem at all, so stick at my first
  > version of the fix presented to dtucker.
  > Issue notified by Matthias Barkhoff (matthias dot barkhoff at gmx dot de).
  > ok dtucker (tobias@)

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

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

config

  ~ config.8                              

  > correct example; from Sebastian Rother (jmc@)

smtpd

  ~ mda.c                                 

  > Fixed a leaking of a fd each time a file is delivered to a mbox, which
  > was introduced in revision 1.7.
  > ok and log message by gilles (tobias@)

  ~ makemap/Makefile                      

  > good god no you do not use LDFLAGS to pull in libraries (deraadt@)

  ~ smtpctl/Makefile                      ~ smtpd/Makefile

  > repair library use.  this was detected on the vax, a nice static linking
  > architecture.  it was silently creating broken code on other architectures.
  > (deraadt@)

  ~ makemap.8                             ~ newaliases.8
  ~ smtpctl.8                             ~ smtpctl.c
  ~ smtpd.8                               

  > various minor improvements; ok jacekm gilles (jmc@)

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


More information about the odc mailing list