[odc] Daily src changes for 2012-03-19

ODC auto at squish.net
Tue Mar 20 07:00:01 GMT 2012


OpenBSD src changes summary for 2012-03-19
==========================================

distrib/miniroot                        regress/lib
sys/arch/amd64/amd64                    sys/arch/amd64/stand/boot
sys/arch/amd64/stand/libsa              sys/arch/i386/i386
sys/arch/mips64/include                 sys/arch/mips64/mips64
sys/arch/sgi/include                    sys/arch/sgi/stand
sys/arch/sgi/stand/boot                 sys/arch/sgi/stand/bootecoff
sys/kern                                sys/sys
usr.bin/kdump                           usr.bin/pkg-config
usr.bin/tmux                            usr.sbin/smtpd
usr.sbin/sysmerge                       

== distrib =========================================================== 01/05 ==

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

miniroot

  ~ list2sh.awk                           

  > Add a comment about harmless warning emitted by tic with -r to the
  > TERMCAP stanza, no longer replace tabset (not needed anymore) and remove
  > blank lines from output termcap file.
  > ok deraadt millert (nicm@)

== regress =========================================================== 02/05 ==

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

lib

  + libpthread/blocked_fifo/Makefile      
  + libpthread/blocked_fifo/blocked_fifo.c

  > Add a (failing) test for a deadlock with fifos and threads.
  > The ``sleep until we have a writer'' behaviour of an open() on a fifo
  > does so with the file descriptor table locked, so if we are waiting for
  > another thread to be our writer we will hang forever.
  > Found this using zotero and firefox.
  > ok fgsch@ (oga@)

  ~ libpthread/Makefile                   

  > Add blocked_fifo. (oga@)

== sys =============================================================== 03/05 ==

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

arch/amd64/amd64

  ~ est.c                                 

  > more intel cpu models, though the low/high est method is largely
  > irrelevant with acpi machines (jsg@)

arch/amd64/stand/boot

  ~ conf.c                                

  > Only change the default boot device to a bootable softraid volume, if we
  > are able to handle the given discipline. At this stage amd64 boot(8) is
  > limited to RAID 1. This makes life easier for users who are booting from
  > a softraid crypto volume, but are loading a kernel from hd0a:/bsd.
  > Tested by jrick at devio.us - thanks! (jsing@)

arch/amd64/stand/libsa

  ~ dev_i386.c                            

  > Only change the default boot device to a bootable softraid volume, if we
  > are able to handle the given discipline. At this stage amd64 boot(8) is
  > limited to RAID 1. This makes life easier for users who are booting from
  > a softraid crypto volume, but are loading a kernel from hd0a:/bsd.
  > Tested by jrick at devio.us - thanks! (jsing@)

arch/i386/i386

  ~ machdep.c                             

  > more intel cpu models, though the low/high est method is largely
  > irrelevant with acpi machines (jsg@)

arch/mips64/include

  ~ cpu.h                                 

  > Use uncached addresses for all exception vectors, when copying our code (or
  > trampolines) to them; this makes sure there is no risk of pending writes
  > being lost when we clear the caches. Of course, this would be a bug in the
  > cache handling routines, but having our vectors correctly set will help
  > debugging the issue.
  > Tested on sgi and loongson. (miod@)

arch/mips64/mips64

  ~ cache_loongson2.S                     ~ cache_r10k.S
  ~ cache_r5k.S                           ~ pmap.c

  > Recent uvm code (and maybe not-so-recent, but it did not explode^WKASSERT
  > at
  > my face then...) depends upon PMAP_PREFER_ALIGN to be a power of two, minus
  > one.
  > On mips64 with 4KB pages, the runtime variable used to compute
  > PMAP_PREFER_ALIGN
  > had the low PAGE_SHIFT bits zeroed (for no good reason I'd say). Don't
  > bother
  > zeroing them anymore. (miod@)

  ~ pmap.c                                ~ vm_machdep.c

  > On sgi, use CKSEG0 addresses whenever possible for pmap_map_direct and u
  > area
  > pages. This will allow R5000-based systems with physical memory fitting in
  > CKSEG0 to use 16KB pages and direct maps (since only XKPHYS accesses
  > trigger
  > the XKPHYS coherency errata on these processors).
  > Tested on IP32, IP30 and IP27 (and loongson too as well). (miod@)

arch/sgi/include

  ~ param.h                               

  > Change the page size computation logic to default to 16KB pages, and only
  > restrict to 4KB if the kernel is configured with R5000 or RM7000 family
  > support.
  > Allow the kernel configuration to override this if it knows better (e.g. an
  > upcoming kernel configuration with support for R5000, but where physical
  > memory on R5000 systems fits within the first 512MB). (miod@)

arch/sgi/stand

  ~ Makefile                              

  > Teach the bootblocks how to load kernel from tftp (i.e. when
  > OSLoadPartition
  > is bootp() instead of a disk). Kind of ugly because of the usual `can't
  > seek'
  > problem causing kernels with symbols to be read from the network twice.
  > While there, add a 32 bit ECOFF boot block, not hooked to the build yet,
  > to be used shortly. (miod@)

arch/sgi/stand/boot

  ~ Makefile                              ~ arcbios.c
  ~ boot.c                                ~ conf.c
  ~ diskio.c                              ~ filesystem.c
  + netfs.c                               + netfs.h
  + netio.c                               

  > Teach the bootblocks how to load kernel from tftp (i.e. when
  > OSLoadPartition
  > is bootp() instead of a disk). Kind of ugly because of the usual `can't
  > seek'
  > problem causing kernels with symbols to be read from the network twice.
  > While there, add a 32 bit ECOFF boot block, not hooked to the build yet,
  > to be used shortly. (miod@)

  ~ arcbios.c                             

  > Prevent the Ethernet driver to be opened twice when using a bootp() path;
  > some ARCBios don't like this. (miod@)

arch/sgi/stand/bootecoff

  + Makefile                              + ld.script

  > Teach the bootblocks how to load kernel from tftp (i.e. when
  > OSLoadPartition
  > is bootp() instead of a disk). Kind of ugly because of the usual `can't
  > seek'
  > problem causing kernels with symbols to be read from the network twice.
  > While there, add a 32 bit ECOFF boot block, not hooked to the build yet,
  > to be used shortly. (miod@)

kern

  ~ kern_event.c                          ~ kern_ktrace.c
  ~ kern_resource.c                       ~ kern_sig.c
  ~ kern_synch.c                          ~ kern_time.c
  ~ sys_generic.c                         ~ vfs_syscalls.c

  > Add tracing and dumping of "pointer to struct" syscall arguments for
  > structs timespec, timeval, sigaction, and rlimit.
  > ok otto@ jsing@ (guenther@)

sys

  ~ ktrace.h                              

  > Add tracing and dumping of "pointer to struct" syscall arguments for
  > structs timespec, timeval, sigaction, and rlimit.
  > ok otto@ jsing@ (guenther@)

== usr.bin =========================================================== 04/05 ==

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

kdump

  ~ kdump.c                               ~ kdump_subr.h
  ~ mksubr                                

  > Add tracing and dumping of "pointer to struct" syscall arguments for
  > structs timespec, timeval, sigaction, and rlimit.
  > ok otto@ jsing@ (guenther@)

pkg-config

  ~ pkg-config                            ~ pkg-config.1

  > GNU's pkg-config added PREFIX/share/pkgconfig/ as a search path 4 years
  > ago. It is mostly used to store arch-independant dot.pc files (although
  > not always); so add this directory to our search path as well (will ease
  > porters' work).
  > While here, also add X11BASE/share/pkgconfig since X.org seems to do the
  > same (req. by matthieu@)
  > ok dcoppa@ fgsch@ phessler@ (ajacoutot@)

tmux

  ~ client.c                              

  > Move MSG_IDENTIFY to the last sent by the client, this will be needed by
  > control clients and irrelevant for others. (nicm@)

== usr.sbin ========================================================== 05/05 ==

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

smtpd

  ~ runner.c                              

  > when a mailer daemon loop is detected, do not insert new bounce in ramqueue
  > fixes an issue experienced by myself and Roman Kravchuck
  > ok eric@ (gilles@)

sysmerge

  ~ sysmerge.sh                           

  > Use $() instead of backticks. (ajacoutot@)

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


More information about the odc mailing list