[odc] Daily src changes for 2005-10-08

ODC auto at squish.net
Sun Oct 9 08:00:36 BST 2005


OpenBSD src changes summary for 2005-10-08
==========================================

bin/ksh                                 sys/arch/i386/conf
sys/arch/macppc/pci                     sys/arch/macppc/stand
sys/arch/powerpc/powerpc                sys/dev/ic
sys/dev/isa                             sys/dev/microcode
sys/dev/pci                             sys/kern
sys/msdosfs                             sys/net
usr.bin/cvs                             usr.bin/rcs
usr.sbin/tcpdump                        

== bin =============================================================== 01/04 ==

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

ksh

  ~ var.c                                 

  > Tell user which value of index is out of bounds. ok deraadt@ (otto@)

  ~ c_sh.c                                

  > Silence bogus gcc warning. (otto@)

== sys =============================================================== 02/04 ==

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

arch/i386/conf

  ~ GENERIC                               ~ RAMDISK_CD

  > enable lge (brad@)

arch/macppc/pci

  ~ vgafb.c                               

  > Disable caching on vga buffer on 64bit machines. currently it causes
  > hangs. (drahn@)

  ~ mpcpcibus.c                           

  > if find_node_intr() doesn't find a match, try to use the contents of
  > 'interrupts'. Also fix a bug where a signed value with -1 in it is compared
  > against 'sizeof(....)'. (drahn@)

arch/macppc/stand

  ~ boot.c                                

  > Preallocate the space where the kernel would live, needs to be at least
  > as large as the ramdisk kernel, however currently on GENERIC this causes
  > about 4MB loss of avail mem, once the kernel is loaded, the remainder
  > needs to be returned. (drahn@)

arch/powerpc/powerpc

  ~ pmap.c                                

  > 64 bit compat. clean up the 32/64 code paths so less decision points exist
  > allow more than 256MB ram on G5, (still 2G limit) by creating PTE entries
  > dynamically for all physical memory. (drahn@)

dev/ic

  ~ siopreg.h                             

  > Remove some trailing whitespace in license. From perry via NetBSD.
  > Note we are now sync'd up to latest (r1.16) NetBSD source. (krw@)

  ~ siopvar_common.h                      

  > Remove some trailing whitespace in license. From perry via NetBSD. Add
  > SF_CHIP_AAIP const which will be used later. From bouyer via NetBSD.
  > Note we are now sync'd up to latest (r1.32) NetBSD source, except
  > r1.28 and r1.29 will be sync'd separately since they actually do
  > something. (krw@)

  ~ siopvar.h                             

  > Remove some trailing whitespace in license. From perry via NetBSD.
  > Note we are now sync'd up to latest (r1.21) NetBSD source. (krw@)

  ~ siop_common.c                         

  > If SF_CHIP_AAIP is set then disable AIP to avoid confusing
  > pre-Ultra160 drives (bouyer via NetBSD from FreeBSD). Affects
  > 1010-66 only. (krw@)

  ~ siop.c                                

  > 'dma' -> 'DMA' (wiz via NetBSD), save a test after TAILQ_FOREACH() by
  > initializing siop_cmd to NULL first (matt via NetBSD), remove trailing
  > whitespace (perry vis NetBSD). (krw@)

  ~ siop_common.c                         

  > More trailing whitespace removal (perry via NetBSD). (krw@)

dev/isa

  ~ viasio.c                              

  > dmesg output nit. (grange@)

dev/microcode

  ~ siop/oosiop.out                       ~ siop/osiop.out
  ~ siop/siop.out                         

  > Sync with r1.7 of ncr53cxxx.c. Changes to comments only. (krw@)

  ~ siop/ncr53cxxx.c                      

  > Some whitespace/syntax tweaks from bouyer & simonb via NetBSD. Also
  > 'Tweak siop scripts assembler for new oosiop.' from tsutsui via
  > NetBSD. Syncs up to latest (r1.14) in NetBSD.
  > No changes to generated .out files. (krw@)

  ~ siop/siop.ss                          

  > More trailing whitespace removal (perry via NetBSD). (krw@)

dev/pci

  ~ if_em.c                               

  > - return from em_start() if not IFF_RUNNING.
  > - remove unnecessary em_start_locked()/em_init_locked().
  > - remove unnecessary spl usage in em_start().
  > - fix spl usage in em_ioctl().
  > ok krw@ (brad@)

  ~ if_em.c                               

  > fix spl usage in em_init(). (brad@)

  ~ if_vge.c                              

  > Some fixes for vge(4)'s multicast handling, including..
  > - Fix IFF_ALLMULTI handling
  > - Stop endless loop in multicast hash table filter mode
  > From christos NetBSD
  > ok pvalchev@ (brad@)

  ~ if_lge.c                              

  > remove unused VLAN headers. (brad@)

  ~ if_lge.c                              

  > - remove return at end of void function
  > - function return value on separate line
  > - ANSI protos (brad@)

  ~ siop_pci_common.h                     

  > Remove some trailing whitespace in license. From perry via NetBSD.
  > Some whitespace tweaks to minimize delta. Note we are now sync'd up
  > to latest (r1.6) NetBSD source. (krw@)

  ~ siop_pci_common.c                     

  > ANSI function decs (thorpej via NetBSD), remove trailing whitespace
  > (perry via NetBSD), add SF_CHIP_AAIP flag (still unused) to 1010-66
  > (bouyer via NetBSD), add error handling if PCI memory type is invalid
  > (christos via NetBSD).  Note we are now sync'd up to latest (r1.25)
  > NetBSD source. (krw@)

  ~ siop_pci.c                            

  > ANSI function decs (thorpej via NetBSD), remove trailing whitespace
  > (perry via NetBSD). Note we are now sync'd up to latest (r1.18) NetBSD
  > source. (krw@)

  ~ if_lge.c                              

  > a little better like this. (brad@)

  ~ if_lge.c                              

  > stop bzero'ing the softc in lge_attach() and tada lge(4) now works
  > to some extent. (brad@)

kern

  ~ vfs_bio.c                             

  > Revert revision 1.39. Under heavy load, it could cause severe corruption
  > in the buffer lists by removing a buffer from the hash twice. Problem
  > identified in discussion with Alexander Bluhm <Alexander_Bluhm at genua.de  > (pedro@)

msdosfs

  ~ msdosfs_vnops.c                       

  > When moving a directory to under a new parent, write the correct
  > cluster number into our . pointer.	This fixes filesystem corruption
  > seen under these circumstances.
  > Testing nick@, krw@, ian@ and others - thanks.
  > "i agree" pedro@; "get it in" deraadt@ (tom@)

net

  ~ if_pppoe.c                            

  > Prevent information leak by using a random hunique value instead of
  > the address of the softc. Reported by Peter Phillip via PR-4501. (canacar@)

== usr.bin =========================================================== 03/04 ==

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

cvs

  ~ diff.c                                

  > KNF;
  > from sir reyk@ (joris@)

  ~ diff.c                                

  > add missing newline, fixes D_NORMAL diff output;
  > ok niallo@ (joris@)

rcs

  ~ ci.c                                  

  > spacing. (niallo@)

  ~ ci.c                                  

  > KNF; (niallo@)

  ~ ci.c                                  

  > remember to NUL-terminate logbuf in checkin_getlogmsg() (niallo@)

  ~ ci.c                                  

  > - add support for -u and -l options
  > - KNF (niallo@)

  ~ ci.c                                  

  > - add support for getting the log message from the user interactively,
  > if it is not supplied on the command line.
  > - correct a couple of typos in comments. (niallo@)

  ~ ci.c                                  

  > sync checkin_usage() with manual page (niallo@)

  + ci.1                                  

  > add a basic manual page for `ci' command. (niallo@)

  ~ ci.c                                  

  > remove useless comment about command line options. not needed since this
  > is all in the manual page now. (niallo@)

  ~ ci.1                                  

  > add ident(1) to SEE ALSO
  > pointed out by joris@ (niallo@)

  ~ co.1                                  ~ rcs.1

  > add ident(1) to SEE ALSO here as well; (joris@)

  + rcsclean.1                            + rcsdiff.1

  > initial manpages for rcsclean and rcsdiff; (joris@)

  ~ rcsdiff.1                             ~ rcsdiff.c

  > support unified, context and rcsdiff output
  > (-u -c -n flags) (joris@)

  ~ rcsdiff.c                             

  > - add support for diffing working file against arbitrary revision, eg
  > rcsdiff -r1.1 <file>
  > - add support for diffing arbitrary revision against arbitrary revision, eg
  > rcsdiff -r1.1 -r1.3 <file>
  > ok joris@ (niallo@)

  ~ rcsdiff.c                             

  > kill extra ']' (joris@)

  ~ rcsdiff.c                             

  > sync rcsdiff_usage() (niallo@)

  ~ rcsdiff.1                             

  > - document -r flag
  > - add ident(1) to SEE ALSO (niallo@)

  ~ ci.1                                  

  > new sentence, new line; (jmc@)

  ~ rcsdiff.1                             

  > tweaks; (jmc@)

  ~ co.1                                  

  > fix SEE ALSO; (jmc@)

  ~ rcs.1                                 

  > fix SEE ALSO; (jmc@)

  ~ rcsclean.1                            

  > new sentence, new line;
  > grammar; (jmc@)

== usr.sbin ========================================================== 04/04 ==

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

tcpdump

  ~ print-ip.c                            ~ print-ip6.c

  > Check IP version when decoding v4 and v6 packets.
  > From Jason Wright via PR-4531. (canacar@)

  ~ Makefile                              ~ interface.h
  ~ print-ether.c                         ~ print-ppp.c
  + print-mpls.c                          

  > Add a best effort mpls decoder. From Jason L. Wright.
  > Since the encapsulated protocol information is not always
  > available in the MPLS tag stack. The decoder attempts
  > to guess the protocol.
  > ok brad@ (canacar@)

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



More information about the odc mailing list