[odc] Daily src changes for 2007-11-28
ODC
auto at squish.net
Thu Nov 29 07:00:01 GMT 2007
OpenBSD src changes summary for 2007-11-28
==========================================
etc/hoststated.conf gnu/usr.bin/binutils
gnu/usr.bin/gcc gnu/usr.bin/ld
gnu/usr.bin/perl gnu/usr.sbin/mkhybrid
lib/libc share/man
sys/arch/i386/conf sys/arch/i386/i386
sys/arch/i386/include sys/arch/i386/isa
sys/arch/i386/pci sys/arch/sparc/dev
sys/arch/sparc/include sys/arch/sparc/sparc
sys/arch/sparc64/include sys/arch/sparc64/sparc64
sys/compat/common sys/compat/freebsd
sys/conf sys/crypto
sys/dev/ata sys/dev/ic
sys/dev/onewire sys/dev/pci
sys/dev/usb sys/kern
sys/lib/libkern sys/nfs
sys/sys sys/ufs/mfs
usr.bin/ftp usr.bin/make
usr.sbin/hoststated usr.sbin/procmap
usr.sbin/pstat usr.sbin/sensorsd
== etc =============================================================== 01/07 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/etc
hoststated.conf
~ hoststated.conf
> extend proxy example (reyk@)
== gnu =============================================================== 02/07 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/gnu
usr.bin/binutils
~ gdb/Makefile.in
> fix race: install-all can only be run *after* all is finished, otherwise
> we may end up installing gdb while it's been rm -f.
> okay kettenis@ (espie@)
usr.bin/gcc
~ gcc/Makefile.in
> add dependencies for protector.o, preventing it from being built too early.
> Okay miod@ (espie@)
usr.bin/ld
~ m88k/md.c ~ m88k/md.h
> Fix md_swap{in,out}_reloc. (miod@)
usr.bin/perl
~ Makefile.bsd-wrapper1
> insert .WAIT so that stuff is built first, and checked afterwards, instead
> of everything happening simultaneously. (espie@)
usr.sbin/mkhybrid
~ src/libfile/apprentice.c
> use sizeof (buf) instead of hardcoded value
> use strcspn to properly overwrite '\n' in fgets returned buffer
> ok ray@ (chl@)
== lib =============================================================== 03/07 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/lib
libc
~ locale/setlocale.c
> remove unused function
> ok otto@ (chl@)
~ stdio/vfprintf.c
> %hhd was being printed as if %hd was specified
> ok millert tedu (deraadt@)
== share ============================================================= 04/07 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/share
man
~ man4/options.4
> COMPAT_{23,25,35} should not be used directly (deraadt@)
~ man4/man4.amd64/Makefile
> add apm (tedu@)
~ man4/acpi.4 ~ man4/man4.i386/apm.4
+ man4/man4.amd64/apm.4
> first attempt to consolidate some of the apm/acpi documentation.
> ok deraadt (tedu@)
~ man4/man4.i386/apm.4
> provide a hint that apm might be acpi on some machines (tedu@)
~ man4/man4.zaurus/apm.4
> history section is totally wrong (tedu@)
- man4/man4.sgi/crime.4 - man4/man4.sgi/dpclock.4
- man4/man4.sgi/dsclock.4 - man4/man4.sgi/gio.4
- man4/man4.sgi/haltwo.4 - man4/man4.sgi/hpc.4
- man4/man4.sgi/imc.4 - man4/man4.sgi/newport.4
- man4/man4.sgi/pic.4 - man4/man4.sgi/wdsc.4
~ man4/man4.sgi/Makefile
> Remove man pages that do not relate to OpenBSD/sgi. Spotted by jmc at .
> ok miod@ (jsing@)
~ man9/workq_add_task.9
> i'm always behind on the doc updates (tedu@)
~ man4/acpi.4
> fix SEE ALSO; (jmc@)
~ man4/man4.amd64/apm.4
> tweak previous; (jmc@)
~ man4/man4.sgi/intro.4
> remove already commented out entries, of devices which we do not support;
> (jmc@)
~ man4/man4.sgi/Makefile + man4/man4.sgi/mainbus.4
> Document the SGI mainbus device. With feedback from jmc at .
> ok miod@ (jsing@)
~ man4/man4.sgi/gbe.4 ~ man4/man4.sgi/macebus.4
> Cleanup and add references to mainbus(4).
> ok jmc@ miod@ (jsing@)
~ man9/workq_add_task.9
> readjust list width after previous; (jmc@)
~ man4/man4.i386/cpu.4
> cpu options are gone now. (tedu@)
~ man9/workq_add_task.9
> i prefer our ISC license. (dlg@)
~ man9/Makefile ~ man9/tsleep.9
> <oga> art write me a manpage
> <art> What? Write it yourself.
> <oga> sudo art write me a manpage.
> <art> ok
> Document msleep(9). (art@)
== sys =============================================================== 05/07 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/sys
arch/i386/conf
~ GENERIC ~ RAMDISK
~ RAMDISKB ~ RAMDISKC
~ RAMDISK_CD ~ files.i386
> quite a bit of simplification by removing cpu classes.
> also assume that 386 cpus are really unknown, and promote them to 486
> instead of panic. (tedu@)
~ Makefile.i386
> basic support for building an i386 kernel with the native toolchain
> on amd64. helpful for test compiling. i have not attempted to boot
> such a kernel. you shouldn't either.
> ok deraadt (tedu@)
~ GENERIC
> Initial import of the DRM (direct rendering manager).
> This is the kernel part necessary for DRI support in X. Disabled for now
> because it still has a few bugs, but now I can work on it in tree. Also
> requires the requisite bits in X, which are currently under discussion
> on how to deal with them with privsep. ported from a combination of the
> free and netbsd implementations.
> Known bugs:
> 1) only the first occurence of X in any session will have dri, after
> that something prevents it working.
> 2) if the machine does not have a dri capable card, the kernel panics.
> Something's up in one of the probe functions. I haven't been able to
> find it though.
> 3) radeon cards need to be forced to use PCI mode otherwise they get
> into an infinite loop.
> This is known to at least kinda work with SiS, radeons in pci mode and
> intel cards.
> ok deraadt, kinda ok art, a few other people had a quick look. (oga@)
arch/i386/i386
~ autoconf.c ~ cpu.c
~ locore.s ~ machdep.c
~ pmap.c ~ procfs_machdep.c
~ via.c
> quite a bit of simplification by removing cpu classes.
> also assume that 386 cpus are really unknown, and promote them to 486
> instead of panic. (tedu@)
~ machdep.c
> doh! (tedu@)
~ ioapic.c
> Don't free the old idt vector if we reuse it.
> ok deraadt@ (kettenis@)
~ conf.c
> Initial import of the DRM (direct rendering manager).
> This is the kernel part necessary for DRI support in X. Disabled for now
> because it still has a few bugs, but now I can work on it in tree. Also
> requires the requisite bits in X, which are currently under discussion
> on how to deal with them with privsep. ported from a combination of the
> free and netbsd implementations.
> Known bugs:
> 1) only the first occurence of X in any session will have dri, after
> that something prevents it working.
> 2) if the machine does not have a dri capable card, the kernel panics.
> Something's up in one of the probe functions. I haven't been able to
> find it though.
> 3) radeon cards need to be forced to use PCI mode otherwise they get
> into an infinite loop.
> This is known to at least kinda work with SiS, radeons in pci mode and
> intel cards.
> ok deraadt, kinda ok art, a few other people had a quick look. (oga@)
arch/i386/include
~ cpu.h ~ cpufunc.h
> quite a bit of simplification by removing cpu classes.
> also assume that 386 cpus are really unknown, and promote them to 486
> instead of panic. (tedu@)
arch/i386/isa
~ clock.c ~ npx.c
> quite a bit of simplification by removing cpu classes.
> also assume that 386 cpus are really unknown, and promote them to 486
> instead of panic. (tedu@)
arch/i386/pci
~ elan520.c ~ ichpcib.c
~ piixpcib.c
> quite a bit of simplification by removing cpu classes.
> also assume that 386 cpus are really unknown, and promote them to 486
> instead of panic. (tedu@)
arch/sparc/dev
~ xd.c ~ xy.c
> ctob/btoc -> ptoa/atop
> from Rodolfo Gouveia (martin@)
arch/sparc/include
~ param.h
> ctob/btoc -> ptoa/atop
> from Rodolfo Gouveia (martin@)
arch/sparc/sparc
~ machdep.c ~ pmap.c
~ vm_machdep.c
> ctob/btoc -> ptoa/atop
> from Rodolfo Gouveia (martin@)
arch/sparc64/include
~ atomic.h
> Use cas/casx instead of casa/casxa. (kettenis@)
~ ctlreg.h
> Garbage collect sparc_cas() and sparc_casx(). We have equivalent functions
> in <machine/atomic.h> now and they're not used anywhere in our tree.
> (kettenis@)
arch/sparc64/sparc64
~ cpu.c
> Give UltraSPARC-IV CPUs a chance again to attach on GENERIC. (kettenis@)
~ pmap.c
> Don't complain about removing wired pages; this is expected to happen with
> ftruncate(2) on mlock(2)'ed pages. (kettenis@)
compat/common
~ compat_vm.c
> compat_25_sys_omsync is unused (deraadt@)
~ kern_info_09.c
> delete the compat_09_sys_uname code here (deraadt@)
compat/freebsd
~ freebsd_misc.c ~ syscalls.master
> COMPAT_09 uname is only used by freebsd compat, so move it to the right
> place (deraadt@)
~ freebsd_syscall.h ~ freebsd_syscallargs.h
~ freebsd_syscalls.c ~ freebsd_sysent.c
> sync (deraadt@)
conf
~ GENERIC
> COMPAT_{23,25,35} should not be used directly (deraadt@)
~ files
> Initial import of the DRM (direct rendering manager).
> This is the kernel part necessary for DRI support in X. Disabled for now
> because it still has a few bugs, but now I can work on it in tree. Also
> requires the requisite bits in X, which are currently under discussion
> on how to deal with them with privsep. ported from a combination of the
> free and netbsd implementations.
> Known bugs:
> 1) only the first occurence of X in any session will have dri, after
> that something prevents it working.
> 2) if the machine does not have a dri capable card, the kernel panics.
> Something's up in one of the probe functions. I haven't been able to
> find it though.
> 3) radeon cards need to be forced to use PCI mode otherwise they get
> into an infinite loop.
> This is known to at least kinda work with SiS, radeons in pci mode and
> intel cards.
> ok deraadt, kinda ok art, a few other people had a quick look. (oga@)
crypto
~ crypto.c ~ cryptodev.h
> finish conversion to workq. remove list remnants, and put spl in the right
> places. handle the no workq case here. ok deraadt (tedu@)
dev/ata
~ atascsi.c
> make ata controllers protect their own command lists so atascsi doesnt have
> to continually go to splbio to ensure its safe to work on them. shrinks
> code a little. (dlg@)
~ atascsi.c ~ atascsi.h
> when we probe a port send an ata inquiry to the device and cache
> the result in the ata_port struct. use this cached inquiry data
> when building the replies to scsi commands rather than fetching a
> new copy of ata inq every time.
> this shrinks the code by 100 lines. it also avoids a malloc in a
> scsi io path, which could be issued at a time when malloc isnt
> guaranteed to succeed, but a real scsi io definitely should. (dlg@)
dev/ic
~ sili.c
> make ata controllers protect their own command lists so atascsi doesnt have
> to continually go to splbio to ensure its safe to work on them. shrinks
> code a little. (dlg@)
dev/onewire
~ owtemp.c
> my brother, David Fries david at fries dot net, noticed that if you put
> your
> sensors in the freezer, they return negative results, making unsigned
> values
> bad to store negative numbers and then do math on. Another popular os had
> this same bug fixed by the same person.
> ok grange@ (todd@)
dev/pci
~ if_vic.c
> use correct data type for the pci address (bus_addr_t)
> From mickey (reyk@)
~ ahci.c
> make ata controllers protect their own command lists so atascsi doesnt have
> to continually go to splbio to ensure its safe to work on them. shrinks
> code a little. (dlg@)
~ ahci.c
> the ccb lists protect themselves, so we dont have to check if we're at
> splbio when getting and putting ata xfers. (dlg@)
~ agp.c ~ agp_i810.c
> do not check malloc return value against NULL, as M_WAITOK is used
> ok oga@ (chl@)
~ agp_i810.c ~ vga_pci.c
+ drm/ati_pcigart.c + drm/drm_agpsupport.c
+ drm/drm_auth.c + drm/drm_bufs.c
+ drm/drm_context.c + drm/drm_dma.c
+ drm/drm_drawable.c + drm/drm_drv.c
+ drm/drm_fops.c + drm/drm_ioctl.c
+ drm/drm_irq.c + drm/drm_lock.c
+ drm/drm_memory.c + drm/drm_pci.c
+ drm/drm_scatter.c + drm/drm_sysctl.c
+ drm/drm_vm.c + drm/i915_dma.c
+ drm/i915_drv.c + drm/i915_irq.c
+ drm/i915_mem.c + drm/mach64_dma.c
+ drm/mach64_drv.c + drm/mach64_irq.c
+ drm/mach64_state.c + drm/mga_dma.c
+ drm/mga_drv.c + drm/mga_irq.c
+ drm/mga_state.c + drm/mga_warp.c
+ drm/r128_cce.c + drm/r128_drv.c
+ drm/r128_irq.c + drm/r128_state.c
+ drm/r300_cmdbuf.c + drm/radeon_cp.c
+ drm/radeon_drv.c + drm/radeon_irq.c
+ drm/radeon_mem.c + drm/radeon_state.c
+ drm/savage_bci.c + drm/savage_drv.c
+ drm/savage_state.c + drm/sis_drv.c
+ drm/sis_ds.c + drm/sis_mm.c
+ drm/tdfx_drv.c + drm/via_dma.c
+ drm/via_drv.c + drm/via_ds.c
+ drm/via_irq.c + drm/via_map.c
+ drm/via_mm.c + drm/via_verifier.c
+ drm/via_video.c + drm/drm.h
+ drm/drmP.h + drm/drm_atomic.h
+ drm/drm_linux_list.h + drm/drm_pciids.h
+ drm/drm_sarea.h + drm/i915_drm.h
+ drm/i915_drv.h + drm/mach64_drm.h
+ drm/mach64_drv.h + drm/mga_drm.h
+ drm/mga_drv.h + drm/mga_ucode.h
+ drm/r128_drm.h + drm/r128_drv.h
+ drm/r300_reg.h + drm/radeon_drm.h
+ drm/radeon_drv.h + drm/savage_drm.h
+ drm/savage_drv.h + drm/sis_drm.h
+ drm/sis_drv.h + drm/sis_ds.h
+ drm/tdfx_drv.h + drm/via_3d_reg.h
+ drm/via_drm.h + drm/via_drv.h
+ drm/via_ds.h + drm/via_mm.h
+ drm/via_verifier.h
> Initial import of the DRM (direct rendering manager).
> This is the kernel part necessary for DRI support in X. Disabled for now
> because it still has a few bugs, but now I can work on it in tree. Also
> requires the requisite bits in X, which are currently under discussion
> on how to deal with them with privsep. ported from a combination of the
> free and netbsd implementations.
> Known bugs:
> 1) only the first occurence of X in any session will have dri, after
> that something prevents it working.
> 2) if the machine does not have a dri capable card, the kernel panics.
> Something's up in one of the probe functions. I haven't been able to
> find it though.
> 3) radeon cards need to be forced to use PCI mode otherwise they get
> into an infinite loop.
> This is known to at least kinda work with SiS, radeons in pci mode and
> intel cards.
> ok deraadt, kinda ok art, a few other people had a quick look. (oga@)
+ drm/files.drm
> oga just showed the worst case of "commit and run for beer" i've ever
> seen. unbreak the tree by adding a missing file (hope it is the only one)
> (deraadt@)
dev/usb
~ usb.h
> add the device and interface class codes for USB Video Class (UVC)
> devices
> ok deraadt@ (robert@)
kern
~ syscalls.master
> COMPAT_09 uname is only used by freebsd compat, so move it to the right
> place (deraadt@)
~ init_sysent.c ~ syscalls.c
> sync (deraadt@)
~ kern_workq.c
> remove the DIRECTOK flag. the caller semantics were too weird, and it's a
> bad idea to have functions that react unpredictably. (tedu@)
~ sys_pipe.c
> When updating the timestamps on pipes, use getnanotime() instead of more
> accurate but more expensive nanotime(), the loss of precision shouldn't
> matter.
> Inspired from the other *BSD which did a similar change.
> ok tedu@ millert@ henning@ deraadt@ (miod@)
~ uipc_usrreq.c ~ vfs_syscalls.c
> a few more places where getmicrotime is good enough. ok miod (tedu@)
~ vfs_conf.c
> g/c mfs_mountroot() and mfs_initminiroot(). We don't support root on
> mfs. OK deraadt@ (millert@)
~ kern_synch.c
> Add msleep. This is identical to tsleep but it takes a mutex as a
> parameter. The mutex is unlocked just before sleep and relocked after
> unless P_NORELOCK is in flags, in which case it is left unlocked.
> ok art at . (oga@)
lib/libkern
~ arch/alpha/bcopy.S ~ arch/alpha/htonl.S
~ arch/alpha/htons.S ~ arch/alpha/memcpy.S
~ arch/alpha/memmove.S ~ arch/alpha/ovbcopy.S
~ arch/amd64/index.S ~ arch/amd64/rindex.S
~ arch/arm/ovbcopy.S ~ arch/hppa/memcpy.S
~ arch/hppa/memmove.S ~ arch/hppa/ovbcopy.S
~ arch/i386/bcopy.S ~ arch/i386/memcpy.S
~ arch/i386/memmove.S ~ arch/i386/ovbcopy.S
~ arch/m68k/bcopy.S ~ arch/m68k/memcpy.S
~ arch/m68k/memmove.S ~ arch/m68k/ovbcopy.S
~ arch/mips64/memcpy.S ~ arch/mips64/memmove.S
~ arch/mips64/ovbcopy.S ~ arch/powerpc/memmove.S
~ arch/powerpc/ovbcopy.S ~ arch/sh/ovbcopy.S
~ arch/sparc/bcopy.S ~ arch/sparc/memcpy.S
~ arch/sparc/memmove.S ~ arch/sparc/ovbcopy.S
~ arch/sparc/random.S ~ arch/sparc64/bcopy.S
~ arch/sparc64/bzero.S ~ arch/sparc64/memcpy.S
~ arch/sparc64/memmove.S ~ arch/sparc64/memset.S
~ arch/sparc64/ovbcopy.S ~ arch/sparc64/random.S
~ arch/vax/bcmp.S ~ arch/vax/bcopy.S
~ arch/vax/bzero.S ~ arch/vax/memmove.S
~ arch/vax/ovbcopy.S
> place a comment as to why these are otherwise empty (deraadt@)
nfs
~ nfs_subs.c
> fill the va_blocksize with something a litte saner,
> make stat(2) return st_blocksize thats a bit closer
> to reality.
> From NetBSD;
> OK art@, beck@, tedu@ (thib@)
sys
~ syscall.h ~ syscallargs.h
> sync (deraadt@)
~ workq.h
> remove the DIRECTOK flag. the caller semantics were too weird, and it's a
> bad idea to have functions that react unpredictably. (tedu@)
~ mbuf.h
> move protos to right place (deraadt@)
~ proc.h
> Add msleep. This is identical to tsleep but it takes a mutex as a
> parameter. The mutex is unlocked just before sleep and relocked after
> unless P_NORELOCK is in flags, in which case it is left unlocked.
> ok art at . (oga@)
~ conf.h ~ malloc.h
> Initial import of the DRM (direct rendering manager).
> This is the kernel part necessary for DRI support in X. Disabled for now
> because it still has a few bugs, but now I can work on it in tree. Also
> requires the requisite bits in X, which are currently under discussion
> on how to deal with them with privsep. ported from a combination of the
> free and netbsd implementations.
> Known bugs:
> 1) only the first occurence of X in any session will have dri, after
> that something prevents it working.
> 2) if the machine does not have a dri capable card, the kernel panics.
> Something's up in one of the probe functions. I haven't been able to
> find it though.
> 3) radeon cards need to be forced to use PCI mode otherwise they get
> into an infinite loop.
> This is known to at least kinda work with SiS, radeons in pci mode and
> intel cards.
> ok deraadt, kinda ok art, a few other people had a quick look. (oga@)
ufs/mfs
~ mfs_extern.h ~ mfs_vfsops.c
~ mfs_vnops.c
> g/c mfs_mountroot() and mfs_initminiroot(). We don't support root on
> mfs. OK deraadt@ (millert@)
== usr.bin =========================================================== 06/07 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin
ftp
~ ftp.1
> expand the description of -C; from martynas and myself (jmc@)
~ ftp.1 ~ main.c
> - [-C] and [-c cookie] do not belong in the first synopsis/usage();
> after some discussion with martynas
> - tweak the description of -C a little more, for readability (jmc@)
make
~ targ.c
> PHONY targets never correspond to real files. (espie@)
~ make.c
> grrr (espie@)
~ defines.h ~ job.c
~ main.c ~ make.1
~ make.c ~ make.h
> debug scaffolding: allows the insertion of a random delay before firing up
> jobs in parallel mode. (espie@)
== usr.sbin ========================================================== 07/07 ==
http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin
hoststated
~ hoststated.c
> unref the label when free'ing a protocol node (reyk@)
~ relay.c
> typos (reyk@)
~ relay.c
> bump the current file descriptor resource limit (openfiles-cur) to the
> maximum number of file descriptors for this login class
> (openfiles-max) of the relay child processes. this will allow 1024
> instead of just 128 open file descriptors in the default configuration
> (class daemon), use the openfiles-max capability and the sysctl
> kern.maxfiles to adjust the value.
> ok gilles@ pyr@ (reyk@)
procmap
~ procmap.1
> this is openbsd, therefore the native display mode is called openbsd
> (tedu@)
pstat
~ pstat.8 ~ pstat.c
> add a -d argument, which can be used to peek at values in the kernel.
> this is easier than using the ddb console. ok deraadt (tedu@)
~ pstat.8
> tweak previous; (jmc@)
~ pstat.8 ~ pstat.c
> Be more careful with printf-style formats, and fix a few other niggles
> ok tedu (deraadt@)
~ pstat.c
> refine format string, support hh and c for real. ok deraadt (tedu@)
~ pstat.8
> - bring back a tweak that got reverted in previous
> - Xr printf 3 whilst here (jmc@)
sensorsd
~ sensorsd.8 ~ sensorsd.c
> add options to specify the check and report intervals. ok deraadt (tedu@)
~ sensorsd.8
> sort options list; (jmc@)
===============================================================================
More information about the odc
mailing list