[owc] Weekly src changes ending 2007-04-08
OWC
auto at squish.net
Mon Apr 9 08:03:29 BST 2007
OpenBSD src changes summary for 2007-04-01 to 2007-04-08 inclusive
==================================================================
bin/cp bin/pax
distrib/notes distrib/sets
etc/etc.vax/MAKEDEV etc/etc.vax/MAKEDEV.md
games/adventure games/arithmetic
games/atc games/gomoku
games/hack games/hangman
games/hunt games/quiz
games/sail games/snake
games/wump gnu/usr.bin/gcc
gnu/usr.sbin/sendmail lib/libc
lib/libdes lib/libpthread
lib/libssl libexec/ftpd
libexec/ld.so libexec/tftpd
sbin/ccdconfig sbin/fdisk
sbin/isakmpd sbin/newfs
sbin/pflogd share/man
sys/arch/amd64/amd64 sys/arch/amd64/stand/libsa
sys/arch/hp300/hp300 sys/arch/hppa/hppa
sys/arch/hppa64/hppa64 sys/arch/i386/i386
sys/arch/i386/include sys/arch/i386/stand/libsa
sys/arch/m68k/m68k sys/arch/mac68k/mac68k
sys/arch/macppc/dev sys/arch/macppc/macppc
sys/arch/mips64/include sys/arch/mvme68k/mvme68k
sys/arch/powerpc/powerpc sys/arch/sh/include
sys/arch/sparc64/conf sys/arch/sparc64/dev
sys/arch/sparc64/include sys/arch/sparc64/sparc64
sys/arch/vax/include sys/arch/vax/vax
sys/compat/freebsd sys/conf
sys/dev sys/dev/acpi
sys/dev/ata sys/dev/cardbus
sys/dev/ic sys/dev/microcode
sys/dev/pci sys/dev/usb
sys/kern sys/miscfs/deadfs
sys/msdosfs sys/scsi
sys/sys sys/ufs/ffs
sys/ufs/ufs sys/uvm
usr.bin/awk usr.bin/calendar
usr.bin/compress usr.bin/file
usr.bin/fold usr.bin/fstat
usr.bin/ftp usr.bin/ktrace
usr.bin/login usr.bin/mail
usr.bin/sendbug usr.bin/sort
usr.bin/top usr.bin/unifdef
usr.bin/ypwhich usr.sbin/acpidump
usr.sbin/afs usr.sbin/bgpctl
usr.sbin/bgpd usr.sbin/httpd
usr.sbin/lpr usr.sbin/ospfd
usr.sbin/pkg_add usr.sbin/spamdb
usr.sbin/user
== bin =============================================================== 01/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/bin
cp
~ cp.1 ~ utils.c
> when copying regular files, look for blocks of zeroes. when we see them,
> just lseek ahead to create a sparse file. saves disk and time.
> note this behavior in the man page.
> ok? deraadt millert (tedu@)
pax
~ gen_subs.c
> For verbose archive listings, use -> for symlinks, not =>,
> as per POSIX. From Igor Sobrado via jmc at . otto@ (millert@)
== distrib =========================================================== 02/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/distrib
notes
~ macppc/hardware
> Add acx to the supported HW list and remove the [*] from the malo entries,
> they have no meaning in this file. (claudio@)
~ i386/hardware
> Cleanup some flags. uath, pgt and malo are included in GENERIC. (claudio@)
sets
~ lists/base/md.alpha ~ lists/base/md.amd64
~ lists/base/md.armish ~ lists/base/md.aviion
~ lists/base/md.cats ~ lists/base/md.hp300
~ lists/base/md.hppa ~ lists/base/md.hppa64
~ lists/base/md.i386 ~ lists/base/md.landisk
~ lists/base/md.luna88k ~ lists/base/md.mac68k
~ lists/base/md.macppc ~ lists/base/md.mvme68k
~ lists/base/md.mvme88k ~ lists/base/md.mvmeppc
~ lists/base/md.sgi ~ lists/base/md.sparc
~ lists/base/md.sparc64 ~ lists/base/md.vax
~ lists/base/md.zaurus ~ lists/comp/mi
~ lists/man/mi
> sync (deraadt@)
== etc =============================================================== 03/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/etc
etc.vax/MAKEDEV
~ etc.vax/MAKEDEV
> sync (todd@)
etc.vax/MAKEDEV.md
~ etc.vax/MAKEDEV.md
> fix wrong major # for xfs (todd@)
== games ============================================================= 04/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/games
adventure
~ adventure.6
> .Ar for arguments; from Igor Sobrado (jmc@)
arithmetic
~ arithmetic.6
> -o and -r take arguments; from Igor Sobrado (jmc@)
atc
~ atc.6 ~ main.c
> tidy up synopsis and usage(); from Igor Sobrado (jmc@)
gomoku
~ gomoku.6 ~ main.c
> sort options + kill usage() verbosity; from Igor Sobrado (jmc@)
hack
~ hack.6
> tidy up synopsis; from Igor Sobrado
> tweak AUTHORS whilst here (jmc@)
hangman
~ hangman.6
> -d takes an argument; from Igor Sobrado (jmc@)
hunt
~ hunt/hunt.c ~ hunt/hunt.6
> sort options; from Igor Sobrado (jmc@)
~ huntd/huntd.6 ~ huntd/driver.c
> tidy up synopis and usage(); sort options; from Igor Sobrado (jmc@)
quiz
~ quiz.6
> -i takes an argument; from Igor Sobrado (jmc@)
sail
~ sail.6
> tidy up synopsis and sort options; from Igor Sobrado (jmc@)
snake
~ snake.6 ~ snake.c
> tidy up synopsis and usage(); from Igor Sobrado (jmc@)
wump
~ wump.6 ~ wump.c
> provide some missing arguments and tidy up usage(); from Igor Sobrado
> (jmc@)
== gnu =============================================================== 05/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/gnu
usr.bin/gcc
~ gcc/config/pa/openbsd.h
> Make sure all public declarations are declared .globl; remove code to
> declare functions .export, which is now redundant. This makes symbol
> aliases created with attribute("alias") visible.
> ok deraadt@ (kettenis@)
usr.sbin/sendmail
~ KNOWNBUGS ~ RELEASE_NOTES
~ cf/README ~ cf/m4/proto.m4
~ cf/m4/version.m4 ~ contrib/qtool.pl
~ doc/op/op.me ~ include/libmilter/mfdef.h
~ include/sm/conf.h ~ include/sm/errstring.h
~ libmilter/engine.c
~ libmilter/docs/smfi_addrcpt_par.html
~ libmilter/docs/smfi_getsymval.html ~ libmilter/docs/smfi_version.html
~ libmilter/docs/xxfi_body.html ~ libsm/config.c
~ libsm/memstat.c ~ libsm/t-memstat.c
~ libsm/t-sem.c ~ sendmail/conf.c
~ sendmail/daemon.c ~ sendmail/deliver.c
~ sendmail/headers.c ~ sendmail/helpfile
~ sendmail/mailq.8 ~ sendmail/map.c
~ sendmail/milter.c ~ sendmail/queue.c
~ sendmail/recipient.c ~ sendmail/sendmail.h
~ sendmail/srvrsmtp.c ~ sendmail/version.c
+ include/sm/os/sm_os_qnx.h
> Update to sendmail-8.14.1 (millert@)
== lib =============================================================== 06/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/lib
libc
~ stdio/printf.3
> clarify return values, prompted by moritz; ok moritz jmc (deraadt@)
libdes
~ read_pwd.c
> Add proper checks against fgets failure. From Charles Longeau.
> OK moritz@, millert@, and jaredy at .ay@)
libpthread
~ uthread/uthread_fork.c
> Upon fork() set the child's view of the stored file status flags inherited
> from the parent to all non-blocking. This prevents the child from restoring
> the file to be blocking upon close() and causing the parent to deadlock
> when it later attempts to utilize the blocking fd. Corrects the netbeans
> build failure caused by the jdk forkAndExec() function that replaces the
> std fd's before execvp(). okay marc@ (kurt@)
libssl
~ src/crypto/ui/ui_openssl.c
> Add proper checks against fgets failure. From Charles Longeau.
> OK moritz@, millert@, and jaredy at .ay@)
== libexec =========================================================== 07/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/libexec
ftpd
~ monitor.c
> Remove lint comment, which does not apply anymore. (moritz@)
ld.so
~ loader.c
> Since the kernel allows argv[0] == NULL, dont core in ld.so
> diff from Matthew Dempsky, with reminder from theo. (drahn@)
~ dlfcn.c ~ ld.so.1
~ loader.c ~ resolve.h
> add support for LD_TRACE_LOADED_OBJECTS_{FMT1/FMT2/PROGNAME}.
> Functionality
> matches and expands on the old a.out ld.so formatting. ok drahn (jason@)
~ ld.so.1
> s/library/object for the format string stuff (jason@)
~ ld.so.1
> properly escape "%O"; ok jason (jmc@)
tftpd
~ tftpd.c
> be more forgiving about tftp requests that come from the 255.255.255.255,
> as is of particular concern with the sun v215; tested by kettenis, no
> regressions spotted by pyr (deraadt@)
~ tftpd.c
> only do the SO_REUSEADDR/SO_REUSEPORT if we are going to bind; pointed out
> by millert (deraadt@)
== sbin ============================================================== 08/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sbin
ccdconfig
~ ccdconfig.c
> use rwlocks instead of hand-rolled. make the softc private parts private.
> this means making ccdconfig "SMALL" to turn off the kvm groveling nonsense
> ok deraadt millert pedro todd (tedu@)
fdisk
~ part.c
> Thinkpad -> ThinkPad; from Igor Sobrado
> ok millert jasper mbalmer (jmc@)
isakmpd
~ isakmpd.c
> Don't let -r fall through to the next case block,
> if INSECURE_RAND is defined.
> ok hshoexer@ (moritz@)
~ monitor.c
> When setting all signals to their default handlers, start
> with signal 1, since there's no signal 0.
> ok hshoexer@ (moritz@)
~ monitor.c
> Don't append the errno string in a log_error() call,
> since it will be automatically be appended.
> ok hshoexer@ (moritz@)
~ x509.c
TAGGED OPENBSD_4_0
> Make isakmpd(8) useable with CaCert signed certificates and add a fix from
> -current.
> MFC:
> revision 1.108
> date: 2007/03/03 20:03:03; author: tom; state: Exp; lines: +7 -2
> There may be more than one item in the subjectAltName (cropping up
> with CACert certificates) so don't require the reported length to be
> exactly equal to the length of the data, but accept it if it's <=
> the length of the data (i.e. we just use the first alt name). The
> purpose of the check is to make sure we don't try to read beyond the
> data we actually have.
> ok cloder@ hshoexer@
> ----------------------------
> revision 1.107
> date: 2006/09/19 10:48:41; author: otto; state: Exp; lines: +3 -3
> Use S_IS* macros insted of masking with S_IF* flags. The latter may
> have multiple bits set, which leads to surprising results. Spotted
> by/partly
> from Paul Stoeber, more to come. ok ho@ miod@ hshoexer@ (mbalmer@)
~ x509.c
TAGGED OPENBSD_3_9
> Make isakmpd(8) usable with CaCert signed certificates and add a fix from
> -current.
> MFC:
> revision 1.108
> date: 2007/03/03 20:03:03; author: tom; state: Exp; lines: +7 -2
> There may be more than one item in the subjectAltName (cropping up
> with CACert certificates) so don't require the reported length to be
> exactly equal to the length of the data, but accept it if it's <=
> the length of the data (i.e. we just use the first alt name). The
> purpose of the check is to make sure we don't try to read beyond the
> data we actually have.
> ok cloder@ hshoexer@
> ----------------------------
> revision 1.107
> date: 2006/09/19 10:48:41; author: otto; state: Exp; lines: +3 -3
> Use S_IS* macros insted of masking with S_IF* flags. The latter may
> have multiple bits set, which leads to surprising results. Spotted
> by/partly
> from Paul Stoeber, more to come. ok ho@ miod@ hshoexer@ (mbalmer@)
~ monitor.c
TAGGED OPENBSD_3_9
> o Kill another strerror() from a call to log_error(),
> which already adds the errno string.
> o Avoid closing fd, if it's -1.
> o Don't replace illegal pathes with /dev/null in
> m_priv_local_sanitize_path(). All callers skip it
> anyways, in the failure case.
> ok hshoexer@ (moritz@)
~ pf_key_v2.c
TAGGED OPENBSD_3_9
> Fix lint comments. s/Fall through/FALLTHROUGH/.
> ok hshoexer@ (moritz@)
newfs
~ mkfs.c ~ newfs.8
~ newfs.c
> Modern disks do bad sector replacement themselves. There is no
> need to reserve space for them in the filesystem itself.
> OK deraadt@ pedro@ (millert@)
~ newfs.c ~ mkfs.c
> Minor cleanup: remove rcsids, remove #ifdef'd out code, COMPAT is
> the default, quiet gcc. OK deraadt@ (millert@)
~ mkfs.c ~ newfs.8
~ newfs.c
> Remove obsolete options for superblock fields that are no longer used.
> OK deraadt @, OK in principle tedu@ (millert@)
~ newfs.8
> Remove part of the -x description that I missed. (millert@)
~ newfs.c
> Use strtonum(), not atoi(). OK deraadt, with help from otto@ (millert@)
pflogd
~ pflogd.8 ~ pflogd.c
> Adds a pidfile argument to pflogd so that individual
> instances can be rotated via newsyslog.
> From Berk Demir < bdd at mindcast org >
> ok henning@ (canacar@)
~ pflogd.8 ~ pflogd.c
> - use .Bk/.Ek to avoid SYNOPSIS splitting
> - put -p in the correct place
> - sync the -p argument name and tweak its description
> - add -p to usage() (jmc@)
== share ============================================================= 09/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/share
man
~ man8/release.8
> XENOCARA_TOP and XENOCARA_OBJDIR were renamed XSRCDIR and XOBJDIR
> respectiveley. with tweaks from jmc@, ok krw@ todd@ (matthieu@)
~ man5/mk.conf.5
> Document X related options. ok and tweaks jmc@ (matthieu@)
~ man5/mk.conf.5
> kill trailing whitespace; (jmc@)
~ man4/man4.macppc/daca.4
> DAC3550A is found in PowerBook2,1 and PowerBook2,2. (xsa@)
~ man4/man4.macppc/asms.4
> sudden motion sensor is also found in PowerBook5,8 and PowerBook6,7. (xsa@)
~ man4/man4.macppc/daca.4
> no need for a new section; (jmc@)
~ man4/man4.macppc/daca.4
> revert previous: at least it's consistent (jmc@)
~ man9/file.9 ~ man9/hardclock.9
~ man9/mbuf.9 ~ man9/microtime.9
~ man9/printf.9
> put "CODE REFERENCES" before "SEE ALSO" for consistency
> ok jmc@ (jasper@)
~ man4/sd.4
> i have a bad habit of not documenting extra interfaces i add...
> i added an ioctl to get product information from disks. the sd driver
> supports it, so say so. (dlg@)
~ man5/bsd.port.mk.5
> Document CHECK_LIB_DEPENDS.
> Based on an idea by bernd and nikolai. (espie@)
~ man4/man4.macppc/snapper.4
> add more supported models. (xsa@)
- man4/bcw.4
> After been attacked by Michael Buesch <mb at bu3sch.deecause we initially
> were using some of their routines in the bcw driver, I decided to stop
> working on it. To avoid any further license chit chat I plain drop the
> driver. (mglocker@)
~ man4/Makefile
> Remove bcw(4) man page. (mglocker@)
~ man4/cardbus.4 ~ man4/pci.4
> no more bcw (deraadt@)
~ man4/man4.sparc64/ebus.4
> Mention ebus at mainbustachment. (miod@)
~ man7/ports.7
> document STARTDIR (espie@)
~ man4/pgt.4
> Sync with i386.html; The NETGEAR WG511 made in China is a SoftMAC one
> and therefore not supported yet. (mglocker@)
~ man8/afterboot.8
> since both daily(8) and afterboot(8) contain instructions for
> /altroot daily backup, remove the (obviously) duplicated text
> from this page and point instead to daily(8);
> from jared r r spiegel (jmc@)
~ man4/pgt.4
> revert previous, since it made the hardware list too long (in terms of
> screen width). instead change the descriptive string to a shorter
> term to avoid the problem.
> ok mglocker (jmc@)
== sys =============================================================== 10/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sys
arch/amd64/amd64
~ pmap.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
arch/amd64/stand/libsa
~ biosdev.c
> Thinkpad -> ThinkPad; from Igor Sobrado
> ok millert jasper mbalmer (jmc@)
arch/hp300/hp300
~ trap.c
> Remove some leftovers from the userret cleanups.
> No need to save sticks anymore.
> miod@ ok and test (art@)
arch/hppa/hppa
~ pmap.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
arch/hppa64/hppa64
~ pmap.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
arch/i386/i386
~ genassym.cf ~ locore.s
~ machdep.c
> Make the ast on i386 per-process instead of per-cpu. This makes
> signal delivery more reliable in some cases when a process switches
> cpu.
> kettenis@ ok (art@)
~ pmap.c ~ pmapae.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
arch/i386/include
~ cpu.h ~ intr.h
~ proc.h
> Make the ast on i386 per-process instead of per-cpu. This makes
> signal delivery more reliable in some cases when a process switches
> cpu.
> kettenis@ ok (art@)
arch/i386/stand/libsa
~ biosdev.c
> Thinkpad -> ThinkPad; from Igor Sobrado
> ok millert jasper mbalmer (jmc@)
arch/m68k/m68k
~ pmap_motorola.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
arch/mac68k/mac68k
~ trap.c
> Remove some leftovers from the userret cleanups.
> No need to save sticks anymore.
> miod@ ok and test (art@)
arch/macppc/dev
~ if_mc.c ~ mesh.c
> unmapiodev() takes the va as its first argument, not the pa.
> ok gwk@ (kettenis@)
arch/macppc/macppc
~ mainbus.c
> A fix for smp old world macs such as the 9500MP and 9600MP which lack the
> /cpus openfirmware node of the new world macs, and only have one cpu node.
> We look at a register of the memory controller called hammerhead for the
> arbitration bit being set indicating a two way machine. Quad processor
> old worlds made by e.g. Daystar Digital would require a different
> heurisitic.
> This heuristic comes from linux/netbsd.
> ok kettenis@ and drahn@ (gwk@)
arch/mips64/include
~ param.h
> delete userland DELAY() version; ok miod (deraadt@)
arch/mvme68k/mvme68k
~ trap.c
> Remove some leftovers from the userret cleanups.
> No need to save sticks anymore.
> miod@ ok and test (art@)
arch/powerpc/powerpc
~ trap.c
> Remove some leftovers from the userret cleanups.
> No need to save sticks anymore.
> miod@ ok and test (art@)
arch/sh/include
~ atomic.h
> Disable interrupts around bit operations for propert atomicity against
> them. (miod@)
arch/sparc64/conf
~ files.sparc64
> Add support for ebus at mainbus, as found on machines with pyro(4) like the
> v215. (kettenis@)
~ GENERIC ~ RAMDISK
> Enable pyro(4) and ebus(4) at mainbus(4). Reorder bus drivers such that
> the order in GENERIC and RAMDISK is identical. (kettenis@)
~ GENERIC
> ral(4) reported to work on sparc64 by Maxim Belooussov
> <belooussov at gmail.com > so enable it. (jsg@)
arch/sparc64/dev
~ pci_machdep.c
> Shift tags when converting them into an address offset. Necessary to
> support
> access to PCIe config space, which is larger than normal PCI config space.
> (kettenis@)
~ pyro.c
> Shift pci tags for PCIe. (kettenis@)
~ pyro.c
> Provide (non functional) dma tag. (kettenis@)
~ pyro.c ~ pyrovar.h
> Add code to establish interrupts. (kettenis@)
~ iommu.c
> Return EFBIG in case not enough segments are present to load a DMA request
> instead of ENOMEM and remove/shortcut the additional debug printfs.
> Some network drivers will try to load the mbuf chain and linearize the
> mbufs
> if EFBIG is returned.
> OK miod@ kettenis@ (claudio@)
~ pci_machdep.c
> Remove bogus code. Turning on parity for pci busses may be a good idea,
> but
> this is not the way to do it. And given the way Sun host bridges crap out
> upon access to pci config space for nonexistent devices, this code is
> rightout harmful for machines where device 0 isn't present on some busses.
> Tested by many; ok deraadt@ (kettenis@)
~ pyro.c ~ pyrovar.h
> Add iommu glue. (kettenis@)
~ com_ebus.c
> Match su16552. (kettenis@)
~ ebus.c ~ ebusreg.h
~ ebusvar.h + ebus_mainbus.c
> Add support for ebus at mainbus, as found on machines with pyro(4) like the
> v215. (kettenis@)
~ iommu.c
> Correctly set the TSB size on pyro(4). (kettenis@)
~ ebus_mainbus.c
> Fix interrupts. (kettenis@)
~ ebus_mainbus.c
> Remove unused variable. (kettenis@)
arch/sparc64/include
~ pci_machdep.h
> Shift tags when converting them into an address offset. Necessary to
> support
> access to PCIe config space, which is larger than normal PCI config space.
> (kettenis@)
~ autoconf.h
> Replace code that tries to find the boot device by pretending to walk the
> device tree and matching locators in the bootpath string components with
> code that tries to match the PROM node of the devices to bootpath
> components.
> This way we don't need a list of possible boot devices that needs to be
> tweaked whenever we try to support new Sun hardware (or when dlg plugs a
> new fancy storage controller in a sparc64 machine).
> Tested by many. (kettenis@)
arch/sparc64/sparc64
~ clock.c
> Fix DEBUG built.
> ok art@ (kettenis@)
~ pmap.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
~ pmap.c
> Fix a comment. 64KB is enough for the cpu_info structure. 64MB would be a
> bit lavish. OK jason@ (claudio@)
~ ofw_machdep.c
> Add support for PCIe. Don't map interrupts that are already mapped
> correctly.
> Tested by deraadt@, nick@, sturm@, naddy@ and others. (kettenis@)
~ autoconf.c
> Replace code that tries to find the boot device by pretending to walk the
> device tree and matching locators in the bootpath string components with
> code that tries to match the PROM node of the devices to bootpath
> components.
> This way we don't need a list of possible boot devices that needs to be
> tweaked whenever we try to support new Sun hardware (or when dlg plugs a
> new fancy storage controller in a sparc64 machine).
> Tested by many. (kettenis@)
~ clock.c
> USECPERSEC is no longer used. (kettenis@)
arch/vax/include
~ intr.h
> splx() does not need to return a value. (miod@)
~ cpu.h
> Reuse symbolic values for splsoftXXX() inlines, instead of hardcoding them.
> No functional change. (miod@)
~ atomic.h
> Wrap bit operations between splhigh()/splx() for atomicity wrt interrupts.
> Not enough for multiprocessor, but we're not there yet anyway. (miod@)
arch/vax/vax
~ trap.c
> Remove some leftovers from the userret cleanups.
> No need to save sticks anymore.
> miod@ ok and test (art@)
~ db_machdep.c
> splx() does not need to return a value. (miod@)
compat/freebsd
~ freebsd_syscall.h ~ freebsd_syscallargs.h
~ freebsd_syscalls.c ~ freebsd_sysent.c
> regen (tedu@)
~ syscalls.master ~ freebsd_misc.c
> freebsd added some MAP_NOCORE flag. we don't like it, so mask it out.
> started by claudio. ok deraadt millert (tedu@)
conf
~ files
> After been attacked by Michael Buesch <mb at bu3sch.deecause we initially
> were using some of their routines in the bcw driver, I decided to stop
> working on it. To avoid any further license chit chat I plain drop the
> driver. (mglocker@)
dev
~ ccd.c ~ ccdvar.h
> use rwlocks instead of hand-rolled. make the softc private parts private.
> this means making ccdconfig "SMALL" to turn off the kvm groveling nonsense
> ok deraadt millert pedro todd (tedu@)
~ ipmi.c
> x4100 machines need a hold off in between reads and write to the bmc
> (marco@)
dev/acpi
~ acpidock.c
> Fix a strange typo and don't set sc_docked in acpidock_dockctl(). It's
> being handled by acpidock_status().
> From Matthew R. Dempsky who had another chunk in his diff that I didn't
> include. It's correct, but it seems to cause some strange regression on
> my X60 that I need to look into first. (mk@)
dev/ata
~ atascsi.h
> Read log page 10h to determine the NCQ error, instead of aborting all
> active commands when an error is encountered.
> ok dlg@ (pascoe@)
~ atascsi.h
> some registers are shared amongst various sata controllers, so stick
> definitions of them here. this starts with the SStatus register. (dlg@)
~ atascsi.h
> device signatures appear to be common to sata controllers too. (dlg@)
~ ata.c
> Use CMD_OK instead of 0, no binary change, from mickey (pedro@)
dev/cardbus
- if_bcw_cardbus.c ~ files.cardbus
> After been attacked by Michael Buesch <mb at bu3sch.deecause we initially
> were using some of their routines in the bcw driver, I decided to stop
> working on it. To avoid any further license chit chat I plain drop the
> driver. (mglocker@)
dev/ic
~ bcw.c ~ bcwvar.h
> Remove dead code. Comments. (mglocker@)
~ silireg.h
> Correct some sili port offsets.
> ok dlg@ (jsg@)
~ bcwvar.h
> There is no need to keep the intrstr in the drivers softc struct.
> (mglocker@)
~ bcwvar.h
> Cleanup sc_softc. (mglocker@)
~ bcw.c
> Some sorting. More comments. (mglocker@)
~ bcw.c
> Write MAC and BSSID into devices template RAM at init time. (mglocker@)
~ rtw.c
> Similar to other drivers call if_start() at the end of the rx interrupt
> because
> ieee80211_input() may enqueue packets.
> OK jsg@ (claudio@)
~ mpi.c
> modernise scsi_inquiry. the length field has grown and now theres pages to
> query.
> ok krw@ (dlg@)
~ acx.c
> Really drop data packets if the station is not associated. It does not
> make sense to keep these packets around and send them out somewhen later
> when they are no longer valid. OK mglocker@ (claudio@)
~ sili.c
> do a global reset properly. (dlg@)
~ silireg.h
> port control clear bits (dlg@)
~ silireg.h
> interrupt enable set/clear defns (dlg@)
~ sili.c
> implement sili_pwait_{ne,eq} for busy waiting on register values.
> fill in sili_ata_probe a bit so it actually moves the port out of the reset
> state and waits for the SStatus register to show that a device has been
> detected.
> the next step is to issue a soft reset and get the device signature, but i
> need a code path for commands (like a soft reset) first. (dlg@)
~ bcw.c ~ bcwreg.h
~ bcwvar.h
> Debugging. Fix a lot wrong registers. Fix a couple of /* XXX */ and
> /* TODO */ for the radio / PHY init. (mglocker@)
~ bcw.c
> Activate interference mitigation. (mglocker@)
~ silireg.h
> start defining the scatter/gather bits and the port request blocks (PRBs).
> (dlg@)
~ sili.c ~ silireg.h
> finish resetting the device by sending a soft reset via the post_direct
> mechanism, and then read the device signature. (dlg@)
~ silireg.h
> control bits in the prb. shuffle things around a bit... (dlg@)
~ sili.c
> DPRINTF will be useful soon, so stick it in (dlg@)
~ sili.c
> rearrange the busywait funcs slightly (dlg@)
~ sili.c
> sili_post_direct lets you submit a prb without all the goop of a ccb and
> xfer. (dlg@)
~ sili.c
> this is a relatively big commit that implements a basic io path.
> it does this:
> - limits the ncmds advertised to atascsi to 1 (cos that lets us write dumb
> but working code)
> - the port probe translates device signatures to device types now
> - defines a basic ccb structure
> - the port probe allocates a ccb if it decides that the port is going to
> be used
> - takes an io from atascsi, fills in the prb, does a direct command
> submission and then polls for completion (all without generating
> interrupts, its kinda cool)
> - limits the sgl to only two elements, which sucks, but it lets us get the
> inquiry and disk attach done. note, the sgl structures are like a smaller
> version of mpis sgl chains, but with a better layout. sgl chains suck
> btw.
> - frees the ccb when the port is freed. (dlg@)
- bcw.c - bcwreg.h
- bcwvar.h
> After been attacked by Michael Buesch <mb at bu3sch.deecause we initially
> were using some of their routines in the bcw driver, I decided to stop
> working on it. To avoid any further license chit chat I plain drop the
> driver. (mglocker@)
~ sili.c
> copy the dmamem wrapper into sili for use for allocating hba memory. this
> has journeyed from ami into mpi, mfi, arc, vic, ahci, and now sili so far.
> (dlg@)
~ sili.c
> white space fix (dlg@)
~ rtw.c
> Correct order of bus_space_barrier() parameters.
> From brad (jsg@)
~ silivar.h
> sizeofa will give us the number of elements in an array. use it when
> iterating over the pci device ids. (dlg@)
~ sili.c ~ silireg.h
> this diff covers two changes, but theyre very tightly tied together.
> implement long sgls by allocating 512 bytes per command, which gives us 7
> scatter gather tables to fill in per command. we can now do proper sized
> io. commands are still polled though, so its very very slow ;)
> since you cant submit more than 128 bytes using direct command submission,
> we needed to implement indirect submission too.
> this almost worked, but i gave the controller to pascoe before i got it
> working. he did the fixes to my dumb dumb mistakes. thanks pascoe. thascoe.
> (dlg@)
~ silireg.h
> Correct offset for global interrupt status register. (pascoe@)
~ sili.c
> Add infrastructure to handle more than one ccb. (pascoe@)
~ sili.c
> Sync the command corresponding to our ccb, not slot 0. (pascoe@)
~ sili.c
> Helper macros to provide the port name in debug messages. (pascoe@)
~ silireg.h
> Add some more device status related register definitions. (pascoe@)
~ sili.c
> Use the rx_count saved in the LRAM command slot entry to determine the
> xfer residual. (pascoe@)
~ sili.c
> Permit PACKET command issue. (pascoe@)
~ sili.c
> Fill out control information for ATAPI commands. (pascoe@)
~ sili.c
> Split command completion out into a separate function, working towards
> IRQ based completion. (pascoe@)
~ sili.c
> Rework command issue/completion flow to be more like AHCI's and switch to
> using interrupts to detect command completion.
> Report on command timeouts (but no recovery yet). (pascoe@)
~ sili.c
> Add transfer failure detection and recovery from non-fatal errors.
> (pascoe@)
~ sili.c
> Whitespace cleanup. No code change. (pascoe@)
~ sili.c
> Always assume write commands had zero residual; the chip only counts
> received bytes. (pascoe@)
~ sili.c
> sili uses one queue for NCQ and legacy commands. No need for the
> AHCI-style
> indirection when determining which commands are complete. (pascoe@)
~ sili.c
> Track the order that CCBs are submitted in, so that after an error we are
> sure to reissue commands in the correct order. Defer completion of
> commands
> when we an error has occurred so that new commands don't go active before
> queeud ones. (pascoe@)
~ sili.c
> Add support for NCQ error recovery and enable NCQ command submission.
> (pascoe@)
~ sili.c
> Change the API to sili_port_intr so it can timeout a command and reuse the
> existing reset/requeue infrastructure, and implement command timeouts.
> sili_port_intr now always processes all pending CCBs, so switch to use the
> auto-clearing interrupt status reads. (pascoe@)
dev/microcode
- bcw/Makefile - bcw/build/Makefile
- bcw/build/build.c - bcw/extract/Makefile
- bcw/extract/extract.c
> We also don't need the microcode builder anymore. (mglocker@)
~ aic7xxx/Makefile ~ atmel/Makefile
~ bnx/Makefile ~ cirruslogic/Makefile
~ fxp/Makefile ~ kue/Makefile
~ neomagic/Makefile ~ ral/Makefile
~ rum/Makefile ~ siop/Makefile
~ symbol/Makefile ~ tigon/Makefile
~ typhoon/Makefile ~ uyap/Makefile
~ yds/Makefile ~ zydas/Makefile
> Allow usb microcode to be built in a cross environment. System endian has
> been
> factored out of the result binaries already. Still need to address non usb.
> ok deraadt. (drahn@)
dev/pci
~ if_bcw_pci.c
> There is no need to keep the intrstr in the drivers softc struct.
> (mglocker@)
~ ahci.c
> Read log page 10h to determine the NCQ error, instead of aborting all
> active commands when an error is encountered.
> ok dlg@ (pascoe@)
~ if_bge.c
> On sparc64, prefer the local-mac-address provided by the PROM. Makes sure
> the
> MAC address on the builtin ports of the Sun Fire V215 match what is used by
> the PROM.
> ok deraadt@ (kettenis@)
~ sili_pci.c
> start defining the scatter/gather bits and the port request blocks (PRBs).
> (dlg@)
- if_bcw_pci.c ~ files.pci
> After been attacked by Michael Buesch <mb at bu3sch.deecause we initially
> were using some of their routines in the bcw driver, I decided to stop
> working on it. To avoid any further license chit chat I plain drop the
> driver. (mglocker@)
~ pcidevs
> Add some more Intel LPC parts. (jsg@)
~ pcidevs.h ~ pcidevs_data.h
> regen (jsg@)
~ pcidevs
> More sili additions. (jsg@)
~ pcidevs.h ~ pcidevs_data.h
> regen (jsg@)
~ sili_pci.c
> Match even more ids: 3131/3531 and some rebadged Intel/Adaptec
> parts that show up in Linux driver.
> ok dlg@ (jsg@)
~ ahci.c
> Wait until we have idled the port before obtaining the error CCB, so as
> to avoid tripping a KASSERT in the case when there is more than one
> outstanding command.
> ok dlg@ (pascoe@)
~ sili_pci.c
> sizeofa will give us the number of elements in an array. use it when
> iterating over the pci device ids. (dlg@)
~ if_bce.c
> Use pci_matchbyid() rather than handrolled equivalent. From Brad.
> tested & ok todd@ (krw@)
~ if_bce.c
> Nuke NetBSD defines APRINTF_NORMAL/_ERROR and just use the printf they were
> defined to. No functional change.
> From Brad. (krw@)
~ if_bce.c
> ANSI-fy. No functional change.
> From Brad. (krw@)
~ ahci.c
> Add compile-time support for coalescing command interrupts to reduce the
> overall interrupt rate.
> #define AHCI_COALESCE to enable. (pascoe@)
~ ahci.c
> Whitespace/long line fixups. No code change. (pascoe@)
dev/usb
~ if_wi_usb.c
> s/multpile/multiple/
> From Dunceor. (ray@)
~ uhci.c
> Correct URL to UHCI spec.
> From dunceor at gmail.comsg@)
~ if_axereg.h
> Data direction was not properly encoded in (unused)
> AX88178 write nodeid command.
> From Hans Petter Selasky <hselasky at c2i.netjsg@)
~ if_cdce.c ~ usbdevs
> laurence tratt says his cable mostly works with this
> ok millert (tedu@)
~ usbdevs.h ~ usbdevs_data.h
> regen (tedu@)
kern
~ init_main.c ~ kern_exit.c
~ kern_fork.c ~ kern_proc.c
~ kern_prot.c ~ kern_sig.c
~ kern_synch.c
> Start moving state that is shared among threads in a process into
> a new struct. Instead of doing a huge rename and deal with the fallout
> for weeks, like other projects that need no mention, we will slowly and
> carefully move things out of struct proc into a new struct process.
> - Create struct process and the infrastructure to create and remove them.
> - Move threads in a process into struct process.
> deraadt@, tedu@ ok (art@)
~ kern_exit.c
> oursleves -> ourselves (pedro@)
~ kern_exit.c
> Don't remove the process from the threads queue in proc_zap() as that
> currently already happens in exit1(), okay art@ (pedro@)
~ kern_time.c
> Implement clock_gettime(CLOCK_PROF) since it's so simple.
> deraadt@ ok (art@)
~ kern_malloc_debug.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
~ kern_rwlock.c
> Implement RW_DOWNGRADE that downgrades an exclusive lock to a shared lock
> without letting any other exclusive locks in between. As opposed to
> upgrading
> locks, this is easy and solves real problems.
> deraadt@ ok (art@)
~ kern_exit.c
> jason crawford noticed that the rthreads diff didn't compile with rthreads!
> (tedu@)
miscfs/deadfs
~ dead_vnops.c
> ansify and deregister, no binary change, okay mickey@ bluhm@ (pedro@)
msdosfs
~ msdosfs_vnops.c
> When running out of space to place a directory entry in msdosfs_readdir(),
> save the correct offset in case the directory has support for long file
> names, and return it to the caller so she can proceed from a valid point.
> From Alexey Vatchenko, okay tedu at .edro@)
scsi
~ safte.c ~ scsi_all.h
~ scsi_base.c
> modernise scsi_inquiry. the length field has grown and now theres pages to
> query.
> ok krw@ (dlg@)
~ scsiconf.h ~ scsi_base.c
> add a wrapper for querying vpd inquiry pages (dlg@)
~ scsi_all.h
> define the vpd inquiry page code. (dlg@)
~ sd.c
> implement the DK INFO ioctl so userland can query the disks product,
> vendor, etc details. (dlg@)
~ sd.c
> sprinkle some ansification around.
> no binary change, and ok krw@ marco@ (dlg@)
sys
~ dkio.h
> provide an ioctl interface for userland to ask disk drivers about various
> product details. (dlg@)
~ proc.h
> Start moving state that is shared among threads in a process into
> a new struct. Instead of doing a huge rename and deal with the fallout
> for weeks, like other projects that need no mention, we will slowly and
> carefully move things out of struct proc into a new struct process.
> - Create struct process and the infrastructure to create and remove them.
> - Move threads in a process into struct process.
> deraadt@, tedu@ ok (art@)
~ proc.h
> abouse -> abuse (pedro@)
~ rwlock.h
> Implement RW_DOWNGRADE that downgrades an exclusive lock to a shared lock
> without letting any other exclusive locks in between. As opposed to
> upgrading
> locks, this is easy and solves real problems.
> deraadt@ ok (art@)
ufs/ffs
~ ffs_softdep.c
> Back out revision 1.70. By asynchronously writing the bitmaps to disk
> upon unmount, we were generating rollbacks that wouldn't be taken care
> of, as well as leaving dangling items in softdep's worklist. (pedro@)
ufs/ufs
~ ufs_vnops.c
> vrele() the destination directory vnode in ufs_rename() if the call to
> ufs_checkpath() fails, okay krw@ tom@ thib@ deraadt@ (pedro@)
uvm
~ uvm_amap.c ~ uvm_anon.c
~ uvm_aobj.c ~ uvm_fault.c
~ uvm_km.c ~ uvm_loan.c
~ uvm_map.c ~ uvm_page.c
~ uvm_page.h ~ uvm_page_i.h
~ uvm_pager.c ~ uvm_pdaemon.c
~ uvm_pglist.c ~ uvm_vnode.c
> Mechanically rename the "flags" and "version" fields in struct vm_page
> to "pg_flags" and "pg_version", so that they are a bit easier to work with.
> Whoever uses generic names like this for a popular struct obviously doesn't
> read much code.
> Most architectures compile and there are no functionality changes.
> deraadt@ ok ("if something fails to compile, we fix that by hand") (art@)
~ uvm_map.h ~ uvm_map_i.h
> Switch vm_map lock to rwlock.
> Use a simple "rw_exit_read(); rw_enter_write();" for lock upgrade, since
> that's what lockmgr did anyway.
> deraadt@ ok (art@)
== usr.bin =========================================================== 11/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin
awk
~ maketab.c
> Do not touch uninitialized variable, if sscanf(3) fails.
> ok millert@ (moritz@)
calendar
~ calendars/calendar.history
> world water day (jasper@)
~ calendars/calendar.usholiday
> reflect new stupid DST US/Canada rules; #5433 John L. Scarfone (deraadt@)
compress
~ main.c
> Fix PR 5417; gunzip should set the timestamp based on the
> compressed file by default. OK otto@ (millert@)
~ compress.1 ~ gzip.1
> Clarify the behavior of the -N and -n flags. (millert@)
~ main.c
> Rename "nosave" variable to "storename" in hopes of clarifying
> the code a bit. OK otto@ (millert@)
~ compress.1
> tweak previous; (jmc@)
~ gzip.1
> oops, tweak previous here too; (jmc@)
file
~ magdir/audio
> detect MP3 file with ID3 v2 tag
> ok deraadt (sturm@)
fold
~ fold.c
> use strtonum, from jason dixon (tedu@)
fstat
~ fstat.c
> use strtonum, from jason dixon (tedu@)
ftp
~ main.c
> use strtonum, from jason dixon (tedu@)
ktrace
~ ktrace.1
> fix the description of -f; from Igor Sobrado (jmc@)
login
~ login.1 ~ login.c
> tweak the -LR options, after similar changes to tcpdrop(8); ok deraadt
> (jmc@)
mail
~ cmd1.c
> make it work with height 4 (division by zero)
> ok millert@ (martynas@)
sendbug
~ sendbug.c
> Only include newest dmesg.
> ``Come on, take some risks'' deraadt at .ay@)
~ sendbug.1 ~ sendbug.c
> Attach dmesg to sendbug reports by default. Use -D flag to disable.
> OK deraadt@ and millert at .ay@)
~ sendbug.c
> argc does not include __progname after subtracting optind. (ray@)
~ Makefile ~ sendbug.c
> Don't shadow variables. Enable -Wshadow. (ray@)
~ sendbug.c
> Warn if one of the required fields is not filled out. Requested
> by millert at .ay@)
~ sendbug.c
> Use __progname and warn instead of fprintf where appropriate. (ray@)
~ sendbug.c
> De-lint. (ray@)
~ sendbug.c
> Sprinkle const poison, remove unneeded variable. (ray@)
sort
~ init.c
> The scanf(3) family of functions can return EOF (-1) if an input
> failure occurs. So check it, before blindly adding it to some
> pointer.
> ok millert@ (moritz@)
top
~ top.c
> tidy up usage(); partly from Mark Lumsden (jmc@)
~ screen.c
> garbage collect unused var ospeed; from Mark Lumsden (otto@)
~ machine.c
> missing prototypes; from Mark Lumsden with a twist by me. (otto@)
~ utils.c
> unneeded include; -Wall -Wmissing-prototypes is happy; from Mark Lumsden
> (otto@)
unifdef
~ unifdef.1 ~ unifdef.c
> tidy up synopsis and usage(); sort options;
> from Igor Sobrado, tweaked by myself (jmc@)
ypwhich
~ ypwhich.1 ~ ypwhich.c
> tidy up synopsis and usage();
> includes bits of a diff from Igor Sobrado and some help from otto
> ok otto (jmc@)
== usr.sbin ========================================================== 12/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin
acpidump
~ acpi_user.c
> Don't get upset about short sized input files.
> ok moritz@ (jsg@)
afs
~ src/lib/ko/kocell.c
> remove extra test, but do it correctly this time (tedu@)
bgpctl
~ bgpctl.c
> Include update statistics in the bgpctl show neighbor output. (claudio@)
~ bgpctl.c ~ parser.c
~ parser.h
> Implement "bgpctl show neighbor <peer> terse" this will print all
> statistics
> in an easily parseable form. This output can be used to implement SNMP MIBs
> or rrdtool/mrtg update scripts. OK henning@ (claudio@)
~ bgpctl.8
> Document the terse format outputs. Additional input by jmc@ (claudio@)
~ bgpctl.8
> tweak previous; ok claudio (jmc@)
bgpd
~ rde_rib.c
> Typo. (claudio@)
~ rde.c ~ rde.h
~ rde_update.c ~ session.h
> Count the updates and withdraws for additional statisic gathering.
> Discussed with henning@ (claudio@)
httpd
~ src/modules/standard/mod_cern_meta.c
> another extra test i missed before, from charles longeau (tedu@)
lpr
~ common_source/displayq.c ~ common_source/rmjob.c
> trailing "\n" not needed in fatal(); ok millert@ (stevesk@)
~ lpc/cmds.c ~ lpd/printjob.c
~ lpr/lpr.c
> KNF, spaces; ok millert@ (stevesk@)
~ common_source/displayq.c
> I have an HP LaserJet (P2015dn) whose LPR implementation may not end
> the send queue state command stream with '\n'; check for this case
> and print '\n' if needed. Without this you may see something like:
> $ lpq
> queue empty$
> ok millert@ (stevesk@)
ospfd
~ rde_spf.c
> Use fatal() if calloc fails instead of fatalx(). (claudio@)
~ rde.c
> There is no need to originate inter-area routes to the backbone.
> Inter-area routes are already comming from the backbone so readding them
> is wrong.
> OK norby@ (claudio@)
~ rde.c
> Type 4 (router) summery LSA need to be sent out with an LS ID equal to the
> ASBR router ID. Using the avertising router for this only works for intra-
> area routes but fails horribly for inter-area ones. Luckily RIB router
> entries
> use the router ID as prefix so just use that instead -- this simplifies the
> code nicly. Figured out the hardway by pyr at .sted and OK pyr@ (claudio@)
~ rde.c ~ rde.h
> remove an unused function.
> ok claudio@ (pyr@)
pkg_add
~ OpenBSD/PackingList.pm
> remove warning (exiting thru next) (espie@)
spamdb
~ spamdb.8
> document correctly WHITE format; ok beck
> fixes system/5432 from Ullrich 'Jason' Haag (jmc@)
user
~ user.c
> remove some extra tests, from charles longeau (tedu@)
===============================================================================
More information about the owc
mailing list