[owc] Weekly src changes ending 2007-11-11
OWC
auto at squish.net
Mon Nov 12 07:00:01 GMT 2007
OpenBSD src changes summary for 2007-11-04 to 2007-11-11 inclusive
==================================================================
bin/cp bin/dd
bin/ed bin/md5
bin/ps distrib/notes
distrib/sets etc/Makefile
etc/etc.mvme88k/Makefile.inc etc/etc.sparc64/Makefile.inc
etc/mtree/special etc/rc
games/atc games/hunt
gnu/lib/libstdc++ gnu/usr.bin/perl
lib/libc lib/libossaudio
lib/libpcap regress/usr.bin
sbin/ifconfig sbin/mount
sbin/newfs sbin/sysctl
share/man share/mk
share/zoneinfo sys/arch/alpha/alpha
sys/arch/alpha/tc sys/arch/amd64/amd64
sys/arch/amd64/conf sys/arch/aviion/aviion
sys/arch/hp300/dev sys/arch/hp300/hp300
sys/arch/hp300/include sys/arch/i386/conf
sys/arch/i386/i386 sys/arch/luna88k/luna88k
sys/arch/m68k/include sys/arch/m68k/m68k
sys/arch/m88k/include sys/arch/m88k/m88k
sys/arch/mac68k/mac68k sys/arch/macppc/conf
sys/arch/macppc/dev sys/arch/macppc/macppc
sys/arch/macppc/pci sys/arch/mvme68k/mvme68k
sys/arch/mvme88k/include sys/arch/mvme88k/mvme88k
sys/arch/mvmeppc/mvmeppc sys/arch/powerpc/include
sys/arch/powerpc/powerpc sys/arch/sgi/localbus
sys/arch/sparc64/dev sys/arch/sparc64/include
sys/arch/sparc64/sparc64 sys/arch/zaurus/dev
sys/ddb sys/dev/acpi
sys/dev/atapiscsi sys/dev/eisa
sys/dev/ic sys/dev/isa
sys/dev/mii sys/dev/pci
sys/dev/raidframe sys/dev/tc
sys/dev/usb sys/kern
sys/net sys/netinet
sys/nfs sys/scsi
sys/sys sys/ufs/ffs
usr.bin/cvs usr.bin/ftp
usr.bin/make usr.bin/pcc
usr.bin/stat usr.bin/top
usr.bin/uniq usr.sbin/hoststated
usr.sbin/mailwrapper usr.sbin/pkg_add
== bin =============================================================== 01/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/bin
cp
~ cp.c
> some small nonfunctional tweaks. ok beck millert (tedu@)
dd
~ dd.c
> remove "unused variable" warning appearing when building release
> ok millert@ (chl@)
ed
~ io.c
> fix format string
> ok ray@ gilles@ (chl@)
md5
~ md5.c
> properly exit >0 when encountering bad md5 in cheklist
> ok gilles@ millert@
> From Tim van der Molen <tbvdm.lists at xs4all.nl> (chl@)
ps
~ print.c
> fix format string
> ok ray@ gilles@ (chl@)
== distrib =========================================================== 02/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/distrib
notes
~ m4.common
> Grammer nit ("all ... has been extracted" -> "all ... have been
> extracted". Pointed out by smallm in PR 5622. (krw@)
~ i386/hardware ~ macppc/hardware
~ sparc64/hardware
> Mention the Meinberg PCI509 card and sort list to make it look like in the
> manual page. (mbalmer@)
~ i386/hardware
> Mention the PEX511 card here. (mbalmer@)
~ sparc64/hardware
> SMP systems should no longer considered as unsupported. (miod@)
~ mvme88k/contents ~ mvme88k/hardware
> Build and advertize bsd.mp on mvme88k. (miod@)
~ i386/hardware ~ macppc/hardware
> Mention the Meinberg USB5131. (mbalmer@)
sets
~ lists/comp/mi
> sync (deraadt@)
~ lists/base/mi ~ lists/man/mi
> sync (pvalchev@)
== etc =============================================================== 03/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/etc
Makefile
~ Makefile
> Change chio.conf's group ownership to operator and mode to 644.
> At this time, there is no sensitive information in that file.
> ok beck@, millert@, jdixon@, deraadt@ (merdely@)
etc.mvme88k/Makefile.inc
~ etc.mvme88k/Makefile.inc
> Build and advertize bsd.mp on mvme88k. (miod@)
etc.sparc64/Makefile.inc
~ etc.sparc64/Makefile.inc
> Missing PHONY target. (miod@)
mtree/special
~ mtree/special
> Change chio.conf's group ownership to operator and mode to 644.
> At this time, there is no sensitive information in that file.
> ok beck@, millert@, jdixon@, deraadt@ (merdely@)
rc
~ rc
> -w flag for wsconsctl is deprecated; from Tim van der Molen (jmc@)
== games ============================================================= 04/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/games
atc
~ log.c
> fix format string
> ok ray@ gilles@ (chl@)
hunt
~ huntd/answer.c
> fix format string
> ok ray@ gilles@ (chl@)
== gnu =============================================================== 05/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/gnu
lib/libstdc++
~ Makefile.bsd-wrapper
> make sure configure does not hardcode MAKE in its generated makefile.
> (GNU people high on LSD... ? ??? ) (espie@)
usr.bin/perl
~ regcomp.c
> Fix potential heap overflow given a specially crafted regexp.
> From Tavis Ormandy; adapted from fixes in the perl 5.9.x codebase.
> (millert@)
~ MANIFEST ~ lib/File/Temp.pm
+ lib/File/Temp/t/cmp.t + lib/File/Temp/t/seekable.t
> update File::Temp to CPAN version 0.18 as other modules start to use it
> ok millert@ (simon@)
== lib =============================================================== 06/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/lib
libc
~ sys/intro.2
> - sync the error descriptions w/ those in errlist
> - escape "No" incase groff thinks it's a macro call (jmc@)
~ sys/stat.2
> sync the comments for struct stat w/ <sys/stat.h>; (jmc@)
~ gen/fts.c
> malloc/memset -> calloc
> ok ray@ (chl@)
~ gen/syslog.c
> Stash value of log_tag before calling closelog_r() and restore it after.
> When closelog() was modified to clear log_tag it inadvertantly made
> syslog_r() (which calls closelog_r at the end) clear log_tag as well.
> OK miod@ dhartmei@ (millert@)
~ stdlib/strtod.3
> strtod may return HUGE_VAL on overflow, which is defined in math.h.
> Therefore added math.h to SYNPOSIS.
> OK millert@ (tobias@)
libossaudio
~ ossaudio.c
> - when we need to do both AUDIO_SETINFO and AUDIO_GETINFO to emulate
> a single OSS ioctl, return the return value and errno from from
> AUDIO_SETINFO if there was an error, since it's more informative
> than the return value and errno from AUDIO_GETINFO. also, don't
> bail if AUDIO_SETINFO fails, because OSS ioctls should return the
> current value of the parameter they are setting, even if there was
> an error.
> - avoid a possible (though highly unlikely) divide by zero in
> SNDCTL_DSP_GET[IO]SPACE.
> ok ratchov@ (jakemsr@)
libpcap
~ gencode.c
> fix format string
> ok ray@ gilles@ (chl@)
== regress =========================================================== 07/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/regress
usr.bin
~ pcc/ccom/Makefile + pcc/ccom/switch001.c
+ pcc/ccom/switch002.c + pcc/ccom/switch003.c
> Add testcases for switch. ok otto@ (stefan@)
== sbin ============================================================== 08/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sbin
ifconfig
~ ifconfig.8
> talk about "vlan-tag" consistently; from Tero Koskinen (jmc@)
mount
~ mount.8
> - clarify the descriptions of "sync" and "async"
> - "async" and "softdep" are mutually exclusive (from simon)
> - change the "no" example to sth potentially less confusing (and safer)
> some feedback from millert
> ok simon (jmc@)
~ mount.8
> FFS_SOFTUPDATES:
> - no need to document needing a config option when it is already enabled
> by default in GENERIC
> - move to format similar to "async"
> - grammar tweak whilst here
> agreed by miod (i think)
> ok millert (jmc@)
newfs
~ newfs.c
> forgotten arg to printf, Tobias Ulmer (otto@)
sysctl
~ sysctl.8 ~ sysctl.c
> do not document -w: it is for compat only. just like we did for audioctl,
> mixerctl, and wsconsctl.
> also move to "name=value", as for the other *ctl apps.
> problem spotted by Tim van der Molen, though this might not be the fix
> he wanted; (jmc@)
== share ============================================================= 09/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/share
man
~ man4/arc.4
> we support the 1202 and first revision of the 1200. (dlg@)
~ man4/mbg.4
> Mention the PCI509 card. (mbalmer@)
~ man4/mbg.4
> Mention the PEX511 card. (mbalmer@)
~ man4/mbg.4
> Tweaks. (mbalmer@)
~ man4/uscanner.4
> - from FreeBSD:
> add support for Epson multifunction USB devices (i.e. scanner/printer/card
> readers) by adding their IDs to the table of recognised devices (because
> we don't have a "scanner" class) and preventing uscanner attach routine to
> reconfigure the whole USB device while we are dealing only with one of its
> interfaces.
> - add the Epson Stylus 3850/4050 ID in the process
> feedback from mbalmer@ and jmc@
> ok jsg@ (ajacoutot@)
~ man4/uscanner.4
> a little tidy up of text; (jmc@)
~ man5/pf.conf.5
> when "max <number>" is exceeded, packets are not dropped - rather they
> fail to match;
> from Doichin Dokov
> diff from henning and myself (jmc@)
~ man4/Makefile + man4/umbg.4
> Add a manual page for umbg(4). (mbalmer@)
~ man4/usb.4
> Mention the umbg(4) driver. (mbalmer@)
~ man4/umbg.4
> Document the timedelta sensor status and the "Signal" sensor. (mbalmer@)
~ man4/usb.4
> List time receiver devices in their own section, like we do in pci(4).
> ok jmc. (mbalmer@)
~ man4/uscanner.4
> - add support for more multifunction Epson devices (from FreeBSD and SANE).
> I'm not adding any new !MF USB scanner IDs as those can be handled using
> libusb and sane-backends (from ports).
> - bring the uscanner man page up-to-date with actually supported scanners
> ok winiger@ mbalmer@ jmc@, input and ok jsg@ (ajacoutot@)
~ man4/aue.4
> 2 port usb hub with ethernet in it.. (deraadt@)
~ man4/uftdi.4
> support little I-O DATA UPS GPS thingy (deraadt@)
~ man4/uscanner.4
> tidy up the hardware list a little; ok ajacoutot (jmc@)
~ man4/mbg.4
> Update after recently committed changes to the mbg(4) driver code.
> (mbalmer@)
~ man4/mbg.4 ~ man4/umbg.4
> tweak; (jmc@)
mk
~ bsd.lkm.mk
> Always compile modules with -mcmodel=kernel on amd64.
> Fix parentheses while here.
> Reminded by deraadt@ (mikeb@)
zoneinfo
~ datfiles/asia ~ datfiles/northamerica
> update to tzdata2007i (millert@)
== sys =============================================================== 10/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sys
arch/alpha/alpha
~ db_disasm.c
> temporary ddb buffers need to be a bit larger on 64 bit systems to cope
> with
> some radix configurations. (miod@)
arch/alpha/tc
~ cfb.c ~ ioasic.c
~ scc.c ~ tc_3000_300.c
~ tc_3000_500.c ~ tc_conf.h
> Get rid of TC_IPL_xxx values and tc_intrlevel_t, and use IPL_xxx and int.
> No functional change. (miod@)
arch/amd64/amd64
~ db_disasm.c
> temporary ddb buffers need to be a bit larger on 64 bit systems to cope
> with
> some radix configurations. (miod@)
arch/amd64/conf
~ RAMDISK ~ RAMDISK_CD
> ramdisks should have the same behaviour as GENERIC (deraadt@)
arch/aviion/aviion
~ locore.S ~ machdep.c
> Remove the now unused idle_u, and call the secondary processors startup
> stack a startup stack. (miod@)
arch/hp300/dev
~ apci.c ~ dca.c
> Get rid of IPL_TTYNOBUF, it was handled exactly like IPL_TTY for years.
> (miod@)
~ dma.c
> IPL_xxx values on hp300 are logical values (although some of them happen to
> match the real hardware level used for them).
> So instead of keeping a few loose variables to store the runtime computed
> psr bits to use for the logical IPL levels, just keep a complete array of
> all psr values. Thus we can have a correct splassert_check(). (miod@)
arch/hp300/hp300
~ intr.c
> Get rid of IPL_TTYNOBUF, it was handled exactly like IPL_TTY for years.
> (miod@)
~ intr.c
> Move splassert_check() from common m68k code to each m68k-based
> architecture,
> as its implementation may vary accross platforms (and will very soon).
> (miod@)
~ intr.c
> IPL_xxx values on hp300 are logical values (although some of them happen to
> match the real hardware level used for them).
> So instead of keeping a few loose variables to store the runtime computed
> psr bits to use for the logical IPL levels, just keep a complete array of
> all psr values. Thus we can have a correct splassert_check(). (miod@)
arch/hp300/include
~ intr.h
> Get rid of IPL_TTYNOBUF, it was handled exactly like IPL_TTY for years.
> (miod@)
~ intr.h
> IPL_xxx values on hp300 are logical values (although some of them happen to
> match the real hardware level used for them).
> So instead of keeping a few loose variables to store the runtime computed
> psr bits to use for the logical IPL levels, just keep a complete array of
> all psr values. Thus we can have a correct splassert_check(). (miod@)
arch/i386/conf
~ RAMDISK ~ RAMDISKB
~ RAMDISKC ~ RAMDISK_CD
> ramdisks should have the same behaviour as GENERIC (deraadt@)
~ GENERIC
> Activate umbg(4) in i386.
> ok dlg. (mbalmer@)
arch/i386/i386
~ locore.s
> I screwed the pooch. Fix the stack offsets.
> "That's much better." kettenis@ (weingart@)
arch/luna88k/luna88k
~ locore.S ~ machdep.c
> Remove the now unused idle_u, and call the secondary processors startup
> stack a startup stack. (miod@)
arch/m68k/include
~ psl.h
> Fix PSLTOIPL and IPLTOPSL to only use three bits of interrupt mask. (miod@)
arch/m68k/m68k
~ m68k_machdep.c
> Move splassert_check() from common m68k code to each m68k-based
> architecture,
> as its implementation may vary accross platforms (and will very soon).
> (miod@)
arch/m88k/include
~ cpu.h
> Remove the now unused idle_u, and call the secondary processors startup
> stack a startup stack. (miod@)
~ cpu.h
> Keep a pending software interrupts mask per processor, instead of having it
> global; and only schedule software interrupts on the currently running cpu.
> (miod@)
arch/m88k/m88k
~ pmap.c
> Comment out the pmap fine grained locking stuff, it is not necessary for
> now
> because of the global lock. It will get enabled again when locking work
> progresses. (miod@)
~ genassym.cf ~ m88k_machdep.c
> Remove the now unused idle_u, and call the secondary processors startup
> stack a startup stack. (miod@)
~ process.S
> Be sure to pmap_deactivate() a process during context switches, so that
> the cpu which runs it is accounted correctly in MP kernels. (miod@)
~ m88k_machdep.c
> Keep a pending software interrupts mask per processor, instead of having it
> global; and only schedule software interrupts on the currently running cpu.
> (miod@)
~ m88k_machdep.c
> On MULTIPROCESSOR kernels, don't forget to grab the kernel lock when
> processing soft interrupts; and there was much rejoicing. (miod@)
~ eh_common.S
> Do not bother checking for curproc != NULL if we know a trap comes from
> usermode, since curproc can not be NULL outside the kernel. (miod@)
~ m8820x_machdep.c
> In dma_cachectl*(), try and perform fewer remote processor operations
> whenever
> possible. (miod@)
~ m8820x_machdep.c
> In dma_cachectl(), flush unconditionnaly on all processors, regardless of
> the
> cpu bitmask of the pmap. (miod@)
~ db_interface.c
> Give more information in ``machine cpu'' under ddb. (miod@)
arch/mac68k/mac68k
~ intr.c
> Move splassert_check() from common m68k code to each m68k-based
> architecture,
> as its implementation may vary accross platforms (and will very soon).
> (miod@)
arch/macppc/conf
~ GENERIC
> Activate umbg(4) on macppc.
> ok jsg, dlg. (mbalmer@)
arch/macppc/dev
~ awacs.c ~ i2s.c
> audio encoding conversion cleanup
> - the endianness of the conversions don't depend on the endianness
> of machine the conversions are built on, but the endianness of the
> audio data itself. choose encoding conversions explicitly, instead
> of relying on #defines based on the endianness of the machine.
> - replace home-grown conversions with comparable conversions in
> auconv.c and mulaw.c
> - use the proper conversion for ulinear_be:16 -> slinear_le:16 in
> auixp(4)
> thanks ajacoutot@ and sthen@ for !x86 testing (jakemsr@)
arch/macppc/macppc
~ machdep.c
> replace even more ctob/btoc with ptoa/atop (martin@)
arch/macppc/pci
~ vgafb_pci.c ~ vgafb_pcivar.h
> fix and use DEVICE_IS_VGA_PCI, makes vgafb_pci_match() much simpler
> ok oga@ brad@ (martin@)
arch/mvme68k/mvme68k
~ machdep.c
> Move splassert_check() from common m68k code to each m68k-based
> architecture,
> as its implementation may vary accross platforms (and will very soon).
> (miod@)
arch/mvme88k/include
~ mvme188.h
> Use two software interrupt sources per processor for IPIs, instead of only
> one, so that we can have maskable and unmaskable IPIs. Make the clock ipis
> maskable, and masked at IPL_CLOCK and above. This allows us to get rid
> of the retrig hack in setipl(). (miod@)
arch/mvme88k/mvme88k
~ locore.S ~ machdep.c
> Remove the now unused idle_u, and call the secondary processors startup
> stack a startup stack. (miod@)
~ m188_machdep.c
> Be more careful when handling ipis:
> - only process the pending ipis once per external interrupt, at the
> beginning.
> - use the ipl we were at when the interrupt occured, not the ipl at which
> we enabled interrupts again, in order to decide whether we can run
> hardclock
> or statclock. (miod@)
~ m188_machdep.c
> Use two software interrupt sources per processor for IPIs, instead of only
> one, so that we can have maskable and unmaskable IPIs. Make the clock ipis
> maskable, and masked at IPL_CLOCK and above. This allows us to get rid
> of the retrig hack in setipl(). (miod@)
~ m188_machdep.c
> Rework ipi handling to delay acking the maskable ipi interrupt, so that it
> gets handled like a real hardware interrupt (which it is supposed to mimic
> anyway). (miod@)
arch/mvmeppc/mvmeppc
~ autoconf.c ~ machdep.c
~ ppc1_machdep.c
> replace even more ctob/btoc with ptoa/atop (martin@)
arch/powerpc/include
~ param.h
> replace even more ctob/btoc with ptoa/atop (martin@)
arch/powerpc/powerpc
~ pmap.c
> replace even more ctob/btoc with ptoa/atop (martin@)
arch/sgi/localbus
~ macebus.c
> Add spurious interrupt detection to macebus. This allows for the detection
> of interrupt storms and enables the user to "boot poweroff" from ddb,
> instead of pulling the powercord.
> Initial code heavily reworked by miod@
> ok miod@ (jsing@)
arch/sparc64/dev
~ fb.c
> weel -> well (jsing@)
arch/sparc64/include
~ lock.h
> Sprinkle a few holy membars around memory writes.
> ok kettenis@ (miod@)
arch/sparc64/sparc64
~ locore.s ~ vm_machdep.c
> Close a race where we might save/drop the fpu state of the wrong process in
> the ipi handlers. (kettenis@)
~ locore.s
> Call prom_set_trap_table for secondary CPUs too. (kettenis@)
~ cpu.c
> Use the interrupt stack as the initial stack for spinning up secondary
> CPUs.
> Inspired by a change in NetBSD pointed out by miod at . (kettenis@)
~ cpu.c
> For single-processor kernels, only match the CPU we're running on.
> (kettenis@)
~ locore.s
> Remove a fair amount of duplicated code by making cpu_mp_startup call
> cpu_initialize. (kettenis@)
~ clock.c ~ locore.s
> Replace next_tick() with simpler C code that I can actually understand.
> (kettenis@)
arch/zaurus/dev
~ zaurus_audio.c
> audio encoding conversion cleanup
> - the endianness of the conversions don't depend on the endianness
> of machine the conversions are built on, but the endianness of the
> audio data itself. choose encoding conversions explicitly, instead
> of relying on #defines based on the endianness of the machine.
> - replace home-grown conversions with comparable conversions in
> auconv.c and mulaw.c
> - use the proper conversion for ulinear_be:16 -> slinear_le:16 in
> auixp(4)
> thanks ajacoutot@ and sthen@ for !x86 testing (jakemsr@)
ddb
~ db_command.c ~ db_examine.c
~ db_print.c ~ db_write_cmd.c
> temporary ddb buffers need to be a bit larger on 64 bit systems to cope
> with
> some radix configurations. (miod@)
dev/acpi
~ acpitimer.c
> Check for non-zero address if mapping to the extended addresses.
> Patch from Oleg Safiullin (a.k.a form@)
> ok kettenis canacar weingart (mikeb@)
~ dsdt.c
> Use the acpi_{acquire,release}_global_lock() to implement the global
> functionality for ACPI. Has not shown any regressions, and may solve
> some EC race conditions.
> Ok gwk@, canacar@, ckuethe@ (weingart@)
~ acpi.c
> Establish the acpi interrupt earlier. Fixes problems with some
> GENERIC.MP kernels not getting acpi interrupts. This is a
> workaround until the real issue with late establishment is fixed.
> ok weingart@, beck@, gwk@ (canacar@)
~ acpi.c
> move acpi_read_pmreg() and acpi_write_pmreg() out of SMALL_KERNEL so
> that acpi ramdisks can compile (obvious solution) (deraadt@)
~ acpi.c
> all ACPI_ENABLE kernels need acpi_interrupt() (deraadt@)
~ acpibat.c ~ dsdt.c
~ dsdt.h
> convert the value to a string instead of using aml_val2str()
> which returns invalid pointers if the value is not a string, causing
> acpibat to panic, as reported by mbalmer@ and Alexey Suslikov.
> tested by Alexey Suslikov, ok marco@ (canacar@)
~ dsdt.c
> Pass correct pointers to memcpy when concatanating strings/buffers
> ok weingart@, marco@ (canacar@)
~ dsdt.c
> do not check malloc return value against NULL, as M_WAITOK is used
> ok marco@ (chl@)
dev/atapiscsi
~ atapiscsi.c
> More scsi_done() at SPLBIO. (krw@)
dev/eisa
~ aha1742.c
> More scsi_done() at SPLBIO. If you poll by manually calling the
> interrupt function, wrap the call in splbio/splx! (krw@)
~ uha_eisa.c
> More scsi_done() at SPLBIO. Wrap another interrupt function call while
> polling in splbio/splx. (krw@)
dev/ic
~ iha.c ~ oosiop.c
~ osiop.c
> Take a bit more care to ensure that SCSI_POLL commands always return
> COMPLETE and never SUCCESSFULLY_QUEUED. (krw@)
~ bha.c
> More scsi_done() at SPLBIO. If you poll by manually calling the
> interrupt function, wrap the call in splbio/splx! (krw@)
~ aic6360.c
> More scsi_done() at SPLBIO. Wrap another interrupt function call while
> polling in splbio/splx. And wrap scsi_done() itself since it appears
> reachable from multiple places. (krw@)
~ dpt.c
> More scsi_done() at SPLBIO. Wrap another interrupt function call while
> polling in splbio/splx. Wrap one call to dpt_done_ccb() not already at
> SPLBIO. (krw@)
~ adv.c ~ adw.c
> More scsi_done() at SPLBIO. Wrap more interrupt function calls while
> polling in splbio/splx. (krw@)
~ ncr53c9x.c
> More scsi_done() at SPLBIO. Wrap more interrupt function calls while
> polling in splbio/splx. Put belts and suspenders splbio/splx around
> the scsi_done() call since it may be possible to get there other ways.
> Partly from thib@ (krw@)
~ mfi.c
> More scsi_done() at SPLBIO. (krw@)
~ cac.c
> More scsi_done() at SPLBIO. (krw@)
~ oosiop.c
> More SPLBIO around scsi_done. (krw@)
~ malo.c
> - Fix command timeouts in certain scenarios (like under X11)
> - Properly issue the RTS-Threshold firmware command
> ok claudio@ (mglocker@)
~ gdt_common.c ~ gdtvar.h
> The interrupt routine doesn't need to explicity invoke splbio/splx.
> But the polling routine needs to wraps its call of the interrupt
> routine with splbio/splx. This ensures scsi_done() is properly
> protected.
> While in the polling routine, fix the timing code so milliseconds are
> not treated as microseconds. (krw@)
dev/isa
~ ad1848.c ~ ess.c
~ sbdsp.c
> audio encoding conversion cleanup
> - the endianness of the conversions don't depend on the endianness
> of machine the conversions are built on, but the endianness of the
> audio data itself. choose encoding conversions explicitly, instead
> of relying on #defines based on the endianness of the machine.
> - replace home-grown conversions with comparable conversions in
> auconv.c and mulaw.c
> - use the proper conversion for ulinear_be:16 -> slinear_le:16 in
> auixp(4)
> thanks ajacoutot@ and sthen@ for !x86 testing (jakemsr@)
~ wds.c
> More scsi_done() at SPLBIO. If you poll by manually calling the
> interrupt function, wrap the call in splbio/splx! (krw@)
~ seagate.c
> More scsi_done() at SPLBIO. (krw@)
dev/mii
~ miidevs.h
> regen (kettenis@)
~ miidevs
> Add Enable/Agere LU3X31T2. (kettenis@)
dev/pci
~ ahci.c
> when the ati ahci stuff is in the ide compat mode, it can sometimes cause
> a wdc(4) controller to appear. this code disables the compat mode using
> some magic niklas gleaned from the linux driver. (dlg@)
~ arc.c
> the 1200 and 1202 shoudl work fine too. (dlg@)
~ arc.c
> very initial foundation for splitting things up to support both the
> traditional intel iop and the new marvell one used on the ARC-1200 rev B.
> (dlg@)
~ arc.c
> start providing function pointers to differentiate between the intel and
> marvell boards. (dlg@)
~ pcidevs.h ~ pcidevs_data.h
> regen. (mbalmer@)
~ pcidevs
> Add the Meinberg PCI509 card, a 5V DCF77 radio clock. (mbalmer@)
~ mbg.c
> Add support for the Meinberg PCI509 5V DCF77 time signal station receiver.
> With feedback from kettenis (mbalmer@)
~ auixp.c ~ cs4280.c
~ neo.c ~ sv.c
> audio encoding conversion cleanup
> - the endianness of the conversions don't depend on the endianness
> of machine the conversions are built on, but the endianness of the
> audio data itself. choose encoding conversions explicitly, instead
> of relying on #defines based on the endianness of the machine.
> - replace home-grown conversions with comparable conversions in
> auconv.c and mulaw.c
> - use the proper conversion for ulinear_be:16 -> slinear_le:16 in
> auixp(4)
> thanks ajacoutot@ and sthen@ for !x86 testing (jakemsr@)
~ arc.c
> wait for the marvell firmware to become ok. (dlg@)
~ qli_pci.c
> scsi_done() must be called at SPLBIO. (krw@)
~ ahci.c
> recognise (and use) ahci 1.2 controllers.
> diff from Henrik Gustafsson (dlg@)
~ pcidevs
> add "82452KX/GX Orion Extended Express Processor to PCI Bridge", from
> NetBSD.
> ok mbalmer@ (oga@)
~ pcidevs.h ~ pcidevs_data.h
> regen. (oga@)
~ pcidevs.h ~ pcidevs_data.h
> regen. (mbalmer@)
~ pcidevs
> Add the Meinberg PEX511 PCI-express DCF77 radio clock. (mbalmer@)
~ mbg.c
> Add support for the Meinberg PEX511 PCI Express DCF77 radio clock.
> Only print firmware ID when compiled with -DMBG_DEBUG, as suggested by
> miod. (mbalmer@)
~ azalia_codec.c
> Generalize most of the apple-specific code, since it works fine on
> other machines with STAC9221.
> Tested by Vladas Urbonas on XPS M1210.
> no objections jakemsr@ (deanna@)
~ pcidevs.h ~ pcidevs_data.h
> regen (brad@)
~ pcidevs
> rename the Intel 82452 hostbridge chipset entry.
> ok oga@ dlg@ (brad@)
~ mbg.c
> Fix a comment. (mbalmer@)
~ pcidevs
> Add Exar UARTs XR17C152, XR17C154 and XR17C158.
> From Daniel Neri (daniel dot neri at sigicom dot se) via PR/5618. (fgsch@)
~ pucdata.c
> Add Exar UARTs XR17C152, XR17C154 and XR17C158.
> From Daniel Neri (daniel dot neri at sigicom dot se) via PR/5618. (fgsch@)
~ pcidevs.h ~ pcidevs_data.h
> regen. (fgsch@)
~ eso.c ~ esoreg.h
~ esovar.h
> bring in ~7 years worth of improvements and bug fixes from NetBSD
> user noticible highlights:
> - recording works
> - setting the blockize works correctly
> - added master volume mute
> ok ratchov (jakemsr@)
~ mbg.c
> Remove all code that prints to the console or writes to the log after
> initialization. The behaviour of the Signal sensor is now the same as
> for umbg(4), the sensor status indicates if the clock is synchronized,
> free running or if there is a hardware problem. (mbalmer@)
dev/raidframe
~ rf_openbsdkintf.c
> Fix some cpu_switchto fallout. Reading disklabels to find raid
> components was causing some systems to hang. The hangs were due to
> tsleep() being called before it was prudent to do so.
> Create a startup hook to find the raid components when it is safe to
> encounter tsleep() but before root is mounted.
> Noticed and fix tested by Josh Grosse (who proved the cpu_switchto
> link), merdely@, fkr at . (krw@)
~ rf_openbsdkintf.c
> Oops. Stray debug printf snuck in. (krw@)
dev/tc
~ if_fta.c ~ if_le_ioasic.c
~ if_le_tc.c ~ tcds.c
~ tcvar.h
> Get rid of TC_IPL_xxx values and tc_intrlevel_t, and use IPL_xxx and int.
> No functional change. (miod@)
dev/usb
~ usb_subr.c
> improve/repair locator printing; ok jsg (deraadt@)
~ uaudio.c
> audio encoding conversion cleanup
> - the endianness of the conversions don't depend on the endianness
> of machine the conversions are built on, but the endianness of the
> audio data itself. choose encoding conversions explicitly, instead
> of relying on #defines based on the endianness of the machine.
> - replace home-grown conversions with comparable conversions in
> auconv.c and mulaw.c
> - use the proper conversion for ulinear_be:16 -> slinear_le:16 in
> auixp(4)
> thanks ajacoutot@ and sthen@ for !x86 testing (jakemsr@)
~ uscanner.c ~ usbdevs
> - from FreeBSD:
> add support for Epson multifunction USB devices (i.e. scanner/printer/card
> readers) by adding their IDs to the table of recognised devices (because
> we don't have a "scanner" class) and preventing uscanner attach routine to
> reconfigure the whole USB device while we are dealing only with one of its
> interfaces.
> - add the Epson Stylus 3850/4050 ID in the process
> feedback from mbalmer@ and jmc@
> ok jsg@ (ajacoutot@)
~ usbdevs.h ~ usbdevs_data.h
> sync (ajacoutot@)
~ ubt.c
> Wrong logic in ubt_xmit_acl_start(); reported by Dmitry Komissaroff
> (dxi, mail dot ru), thanks! (miod@)
~ files.usb + umbg.c
> umbg(4) is a driver to support the USB attached Meinberg USB5131 DCF77
> radio clock. umbg(4) implements a timedelta sensor and reports the
> signal quality in the Signal sensor as percentage. The signal sensor
> status
> further indicates if the clock is free running (WARN).
> ok dlg, jsg (mbalmer@)
~ umbg.c
> knf. (mbalmer@)
~ umbg.c
> Clean up some debugging left-overs. (mbalmer@)
~ usbdevs ~ uscanner.c
> - add support for more multifunction Epson devices (from FreeBSD and SANE).
> I'm not adding any new !MF USB scanner IDs as those can be handled using
> libusb and sane-backends (from ports).
> - bring the uscanner man page up-to-date with actually supported scanners
> ok winiger@ mbalmer@ jmc@, input and ok jsg@ (ajacoutot@)
~ usbdevs.h ~ usbdevs_data.h
> - sync (ajacoutot@)
~ usbdevs.h ~ usbdevs_data.h
> sync (deraadt@)
~ if_aue.c ~ usbdevs
> 2 port usb hub with ethernet in it.. (deraadt@)
~ usbdevs ~ uftdi.c
> support little I-O DATA UPS GPS thingy (deraadt@)
~ usbdevs.h ~ usbdevs_data.h
> sync (martin@)
kern
~ subr_disk.c
> Make disklabel aware of NTFS.
> ok miod@ krw@ (jsing@)
net
~ pf_if.c
> Check for pfi_kif's that are neither groups nor have a real interface.
> This prevents a null-deref when empty groups are used in set loginterface.
> Fixes PR 5628 as reported by Andreas Bihlmaier.
> Bad mpf :(
> OK henning@ (mpf@)
~ pf.c
> Don't leak pfstatekey upon insert conflict (most often caused via pfsync).
> tested david@, ok dhartmei@ (pascoe@)
netinet
~ if_ether.c
> Do not panic in arpintr() if IF_DEQUEUE returns NULL, this is perfectly
> legal
> and can happen on slow mp systems under a lot of network load; ok kettenis@
> (miod@)
nfs
~ nfs_subs.c
> deinline nfs_attrtimeo(); and remove some associated
> #ifdef goo.
> ok krw@, blambert@ (thib@)
~ nfsm_subs.h ~ nfs_serv.c
> Expand the nfsm_srvfillattr() macro and
> nuke it's definition.
> OK krw@, blambert@ (thib@)
~ nfsm_subs.h ~ nfs_serv.c
> more low hanging fruit; Expand the nfsm_srvwcc_data() macro
> and nuke it's definition.
> OK krw@, blamers@ (thib@)
~ nfs_serv.c
> Fix unsafe queue macro use
> ok thib@ 'emphatic' ok beck@ (blambert@)
scsi
~ scsiconf.c ~ sd.c
~ sdvar.h
> Fix SDF_DIRTY handling, eliminate useless SDF_FLUSHING. The sd_flush()
> called from the last sdclose() on a device will now reset SDF_DIRTY
> after submitting the SYNCHRONIZE CACHE command. sddone() need not
> worry about SDF_DIRTY since it was never called for the SYNCHRONIZE
> CACHE command anyway.
> This eliminates a spurious SYNCHRONIZE CACHE command being issued for
> every sd device from sd_shutdown().
> ok dlg@ (krw@)
~ sd.c
> Revert useless textual optimization that resulted in 16 bytes being
> sent to the adapter instead of the 10 bytes the scsi_sychronize_cache
> command actually occupies. This freaked out the ahc on jsing@'s SGI
> and probably others.
> Found, isolated, fix tested, and ok jsing@ (krw@)
sys
~ disklabel.h
> Make disklabel aware of NTFS.
> ok miod@ krw@ (jsing@)
ufs/ffs
~ ffs_alloc.c
> Code to read in the cylinder groups is duplicated in a few spots,
> factor it into a new function, ffs_cgread(); and use that.
> From mickey;
> OK art@, toby@ (thib@)
== usr.bin =========================================================== 11/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin
cvs
~ rcs.c
> Return latest revision of default branch when HEAD is requested and no
> newer revisions are available (for example multiple imports).
> Reported by ckuethe@, OK niallo@ (tobias@)
~ rcs.c
> Stick at GNU cvs output format of docdate.
> OK niallo@ (tobias@)
~ config.c
> An umask is octet not decimal. This means we cannot use strtonum() to
> parse it due to base 10. Instead strtol() with base 8 must be used.
> (tobias@)
~ file.c
> Fix for a rather unintuitive behaviour, this way it can be reproduced:
> cvs checkout src/sbin/swapon
> cvs checkout -rHEAD src/sbin/swapon
> If no date or revision has been supplied, swapon.8 won't be checked out,
> but if a revision (even if it is HEAD) or a date (-D) has been supplied,
> swapon.8 will be checked out due to modifications in Attic/ after it has
> been removed from tree. (tobias@)
~ server.c
> Parse CVSROOT/config and set umask on server-side, too. (tobias@)
~ rcs.c
> Apply umask on RCS file modes. The umask can be changed in CVSROOT/config
> and should therefore be handled. (tobias@)
~ client.c
> If a file has modifications in it (i.e. cvs update on modified files),
> reapply umask as done with unmodified files. Sticks at GNU cvs behaviour.
> (tobias@)
~ rcs.c ~ rcs.h
> Added CVS keyword "Locker". Although it is a dummy function, it's already
> enough for OpenBSD's src/. (tobias@)
~ entries.c
> CVS/Entries needs at least one line specifying a 'D'irectory. If the
> current directory has subdirectories, there is no need to apply another
> 'D' line. Sticks at GNU cvs behaviour. (tobias@)
~ rcs.c
> Do not (re)expand CVS keywords if it is requested to keep old keywords.
> (tobias@)
~ rcs.c
> Max. 80 chars per line in rcs_kwexp_line. Saved a tab by redoing
> for (...) {
> if (something) {
> ..
> }
> }
> into
> for (...) {
> if (!something)
> continue;
> ..
> }
> All other lines which were still too long have been splitted up.
> No functional change (beside of if-statement). (tobias@)
ftp
~ ftp.1
> agressively -> aggressively; (jmc@)
make
~ var.c
> prevent gcc from complaining (espie@)
~ job.c ~ make.c
> get make.c in a more readable state, by extracting code into separate
> functions.
> Restore a big more debug: if DEBUG(JOB), print commands (unexpanded).
> (espie@)
~ compat.c ~ engine.c
~ gnode.h ~ make.c
~ suff.c ~ targ.c
> simplify the way we deal with implicit rules and handle $<.
> Having an `iParents' field is actually backwards, it's ways simpler to
> store the pointer in the child, as an impliedsrc, and to set the variable
> just in time along with all the rest in DoAllVar.
> This is simpler, and it should allow us to call SuffFindDeps much later.
> (espie@)
~ suff.c
> bug-fix: try the empty suffix when we cannot find any transformation, not
> when we cannot find any suffix.
> This allows Makefiles like:
> .SUFFIXES: .in .sh
> all: truc.sh
> truc.sh.in:
> echo "bonjour" >$@
> .in:
> sed -e 's/a/b/' <$< >$@
> to work (before that, the empty suffix would not be tried, since truc.sh
> matches the .sh suffix, even though there is NO transformation). (espie@)
~ arch.c ~ compat.c
~ engine.c ~ gnode.h
~ job.c ~ make.c
~ targ.c
> rename make -> must_make, made -> built_status
> to make them easier to find in source files. (espie@)
~ targ.c
> clearer debug code (espie@)
~ compat.c ~ gnode.h
~ make.c ~ targ.c
> UNMADE -> UNKNOWN (espie@)
pcc
~ mip/protos.h ~ mip/pass2.h
> Update from master repo:
> > Remove some unused prototypes. (ragge@)
~ mip/manifest.h
> iUpdate from master repo:
> > kflag exists in both passes. (ragge@)
~ i386/local2.c
> Update from master repo:
> > Print string identifier in case OREG has a relocatable offset. (ragge@)
~ i386/code.c ~ i386/local.c
> Update from master repo:
> > Work in progress: PIC code for i386.
> > Still missing: ebx as argument to functions. (ragge@)
~ powerpc/code.c
> Pull from master repo:
> Fix defalign(). ok otto@, gmcgarry (stefan@)
~ mip/manifest.h
> Pull from master repo:
> Include string.h. ok ragge@, otto@ (stefan@)
~ cc/cc.c
> Pull from master repo:
> Get rid of ifdef'd out code. Replace nodup() call with equivalent code
> at its only invocation.
> ok ragge@, otto@ (stefan@)
~ cpp/scanner.l
> Pull from master repo:
> ++ and -- are not allowed in preprocessor expressions. Based on a
> diff by Jan Kryl.
> ok ragge@, otto@ (stefan@)
stat
~ stat.1
> adjust an awkward piece of grammar;
> problem spotted by, and ok, jsing (jmc@)
~ stat.1
> relate fields to stat(2).
> with help from jmc@ for wording and nroff magic (espie@)
top
~ top.c
> Get rid of the very annoying warning message + delay when a command
> line argument is wrong. Just display the message in the message area.
> ok deraadt@ sobrado@ simon@ henning@ (otto@)
~ top.c
> Also show warning message for -U and -p in message area; noted by Mark
> Lumsden (otto@)
~ top.c
> prevent segv on wrong sort name; Mark Lumsden. (otto@)
uniq
~ uniq.1 ~ uniq.c
> Allow -c in conjunction with -d (and -u).
> If field skipping is active, don't skip beyond the last non-blank
> of a field, and use isblank() instead of isspace() -- POSIX states
> that the longest string matching [[:blank:]]*[^[:blank:]]* makes
> up a field.
> Don't skip over trailing newlines.
> Help, testing and ok millert@ (kili@)
~ uniq.1 ~ uniq.c
> Mention that -cd and -cu are possible, and that it's an extension to POSIX.
> discussed with jmc and otto
> ok jmc (kili@)
== usr.sbin ========================================================== 12/12 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin
hoststated
~ parse.y
> Not using all defined protocols is not a hard error. Just print a warning
> but start anyway. OK reyk@ (claudio@)
~ log.c ~ hoststated.c
> the last change to enable logging to stderr on startup also enabled
> annoying debugging messages on the console by default. since we do
> not want to see debugging messages unless specified by the "-d" flag,
> add an extra debugging level "debug > 1" to be checked in log_debug().
> (reyk@)
mailwrapper
~ mailwrapper.c
> avoid coredump when ther;s only one token on a line; ok millert@ (otto@)
pkg_add
~ pkg_add.1
> document Ev FTP_KEEPALIVE;
> from giovanni, tweaked by espie and myself (jmc@)
~ pkg_create.1
> Slight rewording of @depend description (this -> that)
> ok espie (sthen@)
===============================================================================
More information about the owc
mailing list