Correction to previous.

(ad)

Cosmetic change to previous.

(ad)

- Rename cpu_lookup_byindex() to cpu_lookup(). The hardware ID isn't of
  interest to MI code. No functional change.
- Change /dev/cpu to operate on cpu index, not hardware ID. Now cpuctl
  shouldn't print confused output.

(ad)

Don't map cpu index to hardware id.

(ad)

Merge wrstuden-revivesa into HEAD. Catch some files that I forgot
to revive on HEAD in previous commit.

(wrstuden)

Merge wrstuden-revivesa into HEAD.

(wrstuden)

Don't install sys/sa.h and sys/savar.h.

(wrstuden)

Per discussion with core, bring this file back under sys/compat.
I chose compat/sa/compat_sa.c as there is only one file. Please
see sys/kern/kern_sa.c for revision history.

(wrstuden)

Remove some now-unneeded includes. Matches previous changes to i386 code.

(wrstuden)

Add sys/sa.h and sys/savar.h which I think are needed with
today's syscall change.

(wrstuden)

If we're doing a 2-byte read, return the entire 16-bit result.

(pgoyette)

- Adapt to changes on bsd.x11.mk
- Remove unnecessary targets that handled .man.pre files
- Add transformations for package-specific patterns in man page templates.

(cube)

Rework the way manual pages are handled.

- .man.pre files are treated just like .man files.
- Makefiles are offered the choice of using sed or cpp for the
  transformation.  MKXORG will default to sed, MKX11 to cpp.
- At least for now, versions of packages are not tracked individually.

(cube)

Some fixes to the TAU support code:
- set the ADC time according to clock speed, not some magic number
- move some mtspr()s which we really only need to do once
- make our envsys_data_t static, this should fix PR 39411

(macallan)

Adapt kern.no_sa_support so that it changes sa_system_disabled
and thus actually controls SA.

(wrstuden)

Add a new vp-private flag, SAVP_FLAG_DELIVERING. This flag indicates
that the vp is in the process of delivering a new-thread upcall.
This is a BLOCKED or NEWPROC upcall. Do NOT generate any BLOCKED
upcalls while this is set. Set it at the end of sa_switchcall()
and clear it at syscall entry. Sadly, I have found no other
way to handle this. The old SA interface has no way for
userland to acknowledge that it got the upcall, so use syscall
entry as the indicator that libpthread got it.

This addresses the issue seen on current-users with firefox
dying with a blockedgen != unbockedgen error. The problem is we
deliver a BLOCKED upcall, but between the end of userret()
and when the CPU indicates that it's running in user mode, we take
a page fault. Said page fault blocks, blocking upcall delivery.
The original block, however, was short-lived, and the UNBLOCK
happens shortly. It happens before libpthread actually notices
the BLOCKED upcall. Boom!

Unfortunately a LOCKDEBUG/DIAGNOSTIC kernel still dies before
firefox opens. However this new issue is with uvm code calling
mtsleep when not holding a correct lock; sa_switch() is exploding
at KASSERT(lwp_locked(l, NULL)); before it's done _anything_.

Also, while here, add sa_system_disabled. This is an integer
that we examine when starting SA for a proc. If it's non-zero,
we error out. Once this is hooked into sysctl, you'll be able to
disable SA w/o recompiling.

(wrstuden)

ehci_dump_itd and _sitd are unused

(jmcneill)

find the kmodule ldscript in the source tree again.

should fix a build issue reported by jmcneill.

(mrg)

kern/39727: race condition in ehci isoc abort xfer path

From Jeremy Morse:

  * Serialize access to the ehci intrlist.
  * Change the ehci intrlist to a tailq so xfers are not queued out of order.
  * In ehci_check_itd_intr, don't treat a transfer error as an indication
    that the xfer is no longer active.

(jmcneill)

Need to be __clz[sd]i2

(matt)

Give maximum level of network softinterrupts a symbolic constant
(which happened to get bumbed from 32 to 33 (AF_MAX) now).

(pooka)

Add FA526 support (compile tested only)

(matt)

do correct octal counting and use CPUID_APM_FLAGS in cpuctl

(cegger)

If no booted_device is found in find_root(), also check CD-ROM boot
with strategy bootloader does.  This allows one CD system with cd9660
root file system and mfs (like a restorecd for cobalt) using GENERIC.

No objection on port-i386, and no bad side effect on usual harddisk boot
or installation of GENERIC with miniroot module.

(tsutsui)

Regen autoconf stuff using tools version of autoconf (in src/tools/autoconf),
as per suggestion from apb@ on tech-toolchain.
Tested 'build.sh -U release iso-image' on NetBSD/i386 4.99.73.

(tsutsui)

Fix handling of lists with a single entry in list_for_each_safe();
this also plugs a memory leak.

(bjs)

add cpuid fn 80000007 %edx: AMD Power Management feature flags

(cegger)

PR kern/39726 Soekris 5501-60 boot/bootxx 120 second delay

Try enabling A20 using an extended BIOS call before playing around with
the keyboard controller or system config port.

(ad)

M_IOV is used all over the place, so stuff it in with malloc_stdtypes

(pooka)

+IPL_SOFTNET

(pooka)

Accept filters:

- Remove remaining #ifdef INET.
- Avoid holding locks so we don't need to do KM_NOSLEEP allocations.
- Use a rwlock to protect the accept filter list.
- Make it safe to unload accept filter modules.
- Minor KNF.

(ad)

Use designated initializers for struct accept_filter.

(ad)

fix output of 3DNOWPREFETCH feature flag

(cegger)

Add some more trivial emulations.

(pooka)

score some stubs

(pooka)

Typo fix and punctuation nit.

(wiz)

Correction to previous.

(ad)

Redo kern_resource.c:1.146 by pooka@: change ui_sbsize to a u_long.

(ad)

Use "?=", not "=", to set default values for the TOOL_* variables
added in the previous commit.

(apb)

Add a _clzsi2 for armv4 platforms.

(matt)

Add a simple example of how to use rump networking facilities.
This just contacts www.NetBSD.org, does GET / and displays the
result.

It either uses the host kernel socket provided PF_INET or the kernel
TCP/IP stack running in userspace (via if_virt and /dev/tap),
depending on the libraries it's linked with.

(not built yet, as I need to commit some more bits from my local
tree for it work)

(pooka)

Make the "something random" ethernet address ETHER_IS_LOCAL
(address assignment probably needs some more work if we want to
support multiple interfaces).

hat tip to Quentin Garnier

(pooka)

since the Makefile creates stuff in (objdir) ./ldscripts/ and then
later deletes it, move the kmodule ldscript into 'xldscripts' and
deal with it from there.

fixes build issues where src/gnu/usr.bin/binutils/ld/ldscripts went
missing as reported by at least 2 people so far.

(mrg)

Add tcp debugging as a build option.

(pooka)

Dear diary: an ethernet address with 0x01 set in the first octet
is an ethernet multicast address.  TCP does not like ethernet
multicast addresses.

(pooka)

print features4: cpuid fn80000001 %ecx on AMD CPUs.

(cegger)

Update lists of options (sync with i386/boot(8) man page).

(apb)

Update MI boot(8) man page from i386/boot(8)

(apb)

Update i386/boot(8) man page:
* Explain that the native boot protocol sets flags in boothowto,
  while multiboot passes args as strings.
* Add missing option letters.
* For each option letter, say which flag it sets in boothowto.
* Describe the "load" and "multiboot" commands.
* Xr multiboot(8).
* Some chnges in wording.

(apb)

POOL_INIT -> pool_init

(pooka)

award unlimited sbsize rlimit

(pooka)

Add cpuid 0x80000001 %ecx features flags. Rename CPUID_MASK4 to CPUID_INTEL_MASK4 for consistency with new CPUID_AMD_MASK4

(cegger)

tz{code,data}2008h out; tzdata2008h in.

(kleink)

Import tzdata2008h.

Status:

Vendor Tag: ADO
Release Tags: TZDATA2008H

(kleink)

Add missing TOOL_* variables to bsd.sys.mk and document them in
bsd.README.  Previously, several of these variables were present in
bsd.own.mk but not in bsd.sys.mk or bsd.README.

(apb)

Remove references to TOOL_BEBOXELF2PEF and TOOL_BEBOXMKBOOTIMAGE,
which were obsoleted in May 2008.

(apb)

regen: properly create weak references to sys_foo()

(pooka)

rump_enosys() is autogenerated into rump_syscalls (makes things
actually work)

(pooka)

Fix previous: weak symbols work so much better if they are exported
from the object as weak symbols...

There's also something to be said about "cosmetic" tweaks after
testing, but I do think I'll bother.

(pooka)

Do not descend into docs/ unless MKDOC != no.

(dyoung)

Per uwe@'s suggestion, build nbgroff under wider circumstances.
I choose to build it if MKDOC != no or if MKHTML != no.  It used
to be built only when MKMAN != no.

(dyoung)

Add uid_init() I managed to miss earlier.

(pooka)

Jumbo frames is supported.
Note new default values for interrupt mitigation counters.

(bouyer)

Add support for jumbo frames.
Some performances tweaks (including better defaults for interrupt mitigation)

(bouyer)

Treat .man.pre files just as regular .man files, that's what is done in the
package itself anyway.

(cube)

regen - weak aliasing for default syscall

(pooka)

Weak alias all rump system calls to rump_enosys()

(pooka)

add rump_enosys()

(pooka)

In bsd.sys.mk and bsd.own.mk, sort lists of TOOL_* variables.
In bsd.README, document all TOOL_* variables that are set in bsd.sys.mk.

There are several TOOL_* variables that are set in bsd.own.mk, but not
set in bsd.sys.mk and not documented in bsd.README.

(apb)

Intel Tempest can use ichlpcib as well.

(joerg)

Fix wrong memset; PR bin/39733 from Henning Petersen.

(dholland)

Fix reversed test in connection with combining -i and -l. If some
packets have already sent because of the preload count handling, we
shouldn't start off by sending another one right away. The reversed
test causes ping6 to wait one interval before sending anything at all
if you don't use -l. PR bin/39732.

(dholland)

Ensure optional MTRR sections are built if MTRR is enabled (missing #include
"opt_mtrr.h").
Fix build due to changes in revision 1.4 of sys/arch/x86/include/mtrr.h

(sborrill)

Fix up the bus-independant probe code.  Patch from njoly@

(pgoyette)

DMA fixes thanks to rtr & skrll
Don't set tsf in auth, this stops firmware errors on WPA renegotiation
Various other tweaks that the linux driver seems to have that we did not

(blymn)

Correct beacon structure
Add scan abort command
Add TX status codes from FreeBSD driver

(blymn)

Avoid local include.

(nakayama)

Remove unnecessary includes.

(nakayama)

Update to add i2c bus attachment

(pgoyette)

Add lm0 at iic?

(pgoyette)

Remove duplicated copy of the file - no idea where it came from.

(pgoyette)

Provide psignal() so that we don't get linked against psignal(3).
As there is no real process model currently, just "trap" the signals
appropriately.

(pooka)

add reachover for xf86-video-r128

(rtr)

add -I${DESTDIR}${X11INCDIR}/X11/dri to CPPFLAGS if DRI is enabled it
is needed but not added where DRI=yes is set so do it here

add ${X11FLAGS.EXTENSION} to CPPFLAGS since many of the drivers build
options depending on extensions enabled.

(rtr)

Bump date for previous.

(wiz)

Print actual maximum amount of MTRRs configured

(sborrill)

The firmware URL changed.

(is)

Load firmware from .../firmware/stuirda, not ../uirda, to avoid confusion.

(is)

Sort sections. Improve grammar. Short NAME and lengthen DESCRIPTION.
Use \*[Lt] and \*[Gt] for better HTML output.

(wiz)

Drop trailing whitespace. Sort supported hardware.
Put RCS Id at top. Sort SEE ALSO.
Comment out loader.conf(5) section.

(wiz)

Bump date for previous, use Fx for FreeBSD.

(wiz)

XXX: r200_tcl.c and radeon_tcl.c need -Wno-error on sparc64.

(mrg)

catch up with share/ldscripts -> libdata/ldscripts.

(mrg)

reallocate the correct amount. From Anon Ymous

(christos)

add X11SRCDIR.xf86-video-r128

(mrg)

initial import of xf86-video-r128-6.8.0

Status:

Vendor Tag: xorg
Release Tags: xf86-video-r128-6-8-0

(mrg)

Consistent use of softc member name - fixes build breakage from previous.

(pgoyette)

- install ldscripts on sparc64
- move ldscripts to /usr/libdata/ldscripts

(mrg)

Add some shameful kludges for the vax build.  Eventually I'd like
to use the native kernel machine directory, so don't waste too much
effort fighting this now.

(pooka)

__cpu_simple_lock_t changed to unsigned char on arm kernels almost
two years ago.  Catchup here also, although it seems nothing uses
acorn32 kernels with MULTIPROCESSOR.

(pooka)

Don't use unicode in usage.
Noted by Anon Ymous.

(wiz)

For libXaw backward compat links, replace "hardlink to symlink" by "symlink
to symlink" and thus make building on both NetBSD and Linux work again, see
my postings to current-users@ on the problem[1][2].

[1] http://mail-index.netbsd.org/current-users/2008/09/30/msg004940.html
[2] http://mail-index.netbsd.org/current-users/2008/09/30/msg004941.html

(hubertf)

Move printing of the "polling" message at attach time out of the wrong if
scope, so that it is actually printed even in non-error cases.
Suggested by jmcneill.

(martin)

Add callout inits which I managed to not commit earlier with the
rest of the callout stuff.

problem noticed by Arnaud Ysmal

(pooka)

Try to improve warning messages when stage2 bootstrap is not found
in the root of the file systrem in which the installation is being
performed.

Inspired by PR 39728 by der Mouse

(apb)

Try to improve documentation of the fact that the primary bootstrap is
specified using a file name on the running system, while the secondary
bootstrap is specified using a file name relative to the root of the
file systrem in which the installation is being performed.

Inspired by PR 39728 by der Mouse

(apb)

ip6_pcbopts() is called with the socket lock held, use M_NOWAIT

(plunky)

ip6_pcbopt() is in the ctloutput path, we should not
sleep here because socket lock is held. use M_NOWAIT

(plunky)

Fix kernels without accf_data or accf_http.  I'll modularize this better
on Monday, but this will get people building again while still removing
the silly dependency on INET.

(tls)

Add an i2c attachment for the LM78 family of temp sensor/fan control

Tested by Martin Husemann

(pgoyette)

Move the {read,write}reg() routine addresses into the dbcool_softc so
that we can use different routines for macppc's ki2c attachment.  Fixes
problems reported by Robert Fritzsche on current-users@ list.

(pgoyette)

convert ip6_[sg]etmoptions() to use sockopt(9) API
should be no functional change

(plunky)

Fix reading MAC address from eeprom: don't forget to read the last
descriptor.
While there, white space fixes.

(bouyer)

update ip_pcbopts() to use sockopt(9) API.

cleans up function and one small fix is that we now stop copying user
options to the mbuf when the _EOL is given, previously this function
would continue to copy options.

(plunky)

Add some required explicit scopes to fix ambiguity errors on
NetBSD/current.

(plunky)

do not sleep while allocating memory, socket lock is held
(use ENOBUFS for failure)

(plunky)

do not sleep while allocating memory here as socket lock is held

(plunky)

switch to per-arch multilib.h since we now use them.

now, given a correctly populated /usr/lib/{i386,sparc} directory, you
can "cc -m32" on amd64/sparc64 and have it be useful.

note that a trivial copy of 32 bit /usr/lib will not work, since you
end up with symlinks to the 64 bit /lib.  work on that part is in
progress...

(mrg)

update to clarify that
sockopt_init() may sleep
sockopt_set() will not sleep

(plunky)

fix problem pointed out by ad where sockopt may end up sleeping
inappropriately with the socket lock held.

    sockopt_init() may sleep

    sockopt_set() will not sleep

    sockopt_getmbuf() for legacy code will not sleep

(plunky)

Add sequencer and vcoda.

(tsutsui)

add reachover build for xf86-video-nvxbox

(rtr)

Put a workaround for PR toolchain/39528:
Adjust a definition of MKDIR_P to invoke $ac_install_sh script
via $(SHELL) so that it will work even if executable bits are not
preserved in the source archives.

No objection on tech-toolchain.

(tsutsui)

Fix mangled attach output. Noted by Jari Kuittinen.

(dholland)

output.c output.h: expose OUTPUT_ERR (flag for an exposed flags variable)
bltin.h: support ferror()
echo.c: use ferror() to fail on output write errors

Another piece of PR bin/39574.

(dholland)

jme(4), a driver for JMicron Technologies JME250 Gigabit Ethernet and
JME260 Fast Ethernet PCI Express controllers.
Written with a lot of cut-n-paste from the FreeBSD jme(4) driver.
No support for jumbo ethernet frames yet (but should come soon).
Thanks to JMicron Technologies for providing me sample boards and
documentation for this work.

(bouyer)

Support 64bit DMA on PCI busses.

(bouyer)

- use aprint_debug_dev(9) for debug info
- use aprint_normal_dev(9) rather than aprint_normal(9) + device_xname()

(tsutsui)

Print a message on two previously silent error paths in auvia_attach.

(dholland)

eval arguments one more time so that quoting works.

(christos)

Make SIOCSIFCAP work again on vlan interfaces: first check that the
capability is enabled on parent, then call ifioctl_common().

(bouyer)

Now the cobalt kernel can detect boot device properly by boot strings
passed from the firmware, so change root file system type from
explicit nfs to generic '?' so that the INSTALL kernel can also be used
as a rescue kernel on localdisk.

(tsutsui)

Address problems with accept filters noted by ad in his source-changes
mail: http://mail-index.netbsd.org/source-changes/2008/10/10/msg211109.html

* Scary-looking socket locking stubs (changed to KASSERT of locked)

* depends on INET inappropriately (though now you must add new
  accept filter names to the uipc_accf.c line in conf/files if
  you aren't using dataready or httpready)

* New code uses MALLOC/FREE -- changed to kmem_alloc/kmem_free;
  could be pool_cache, these are all fixed-size allocations.

We need to verify that this works as expected with protocols with per-socket
locking, like PF_LOCAL.  I'm a little concerned about the case where the
lock on the listen socket isn't the same lock as on the eventual connected
socket.

(tls)

Include <machine/cpu> for the benefit of archs where it doesn't happen
accidentally automagically.

Pointed out by Havard.

(pooka)

Update cross-references for addition of dbcool(4) and removal of adt7463c(4)
and adt7467c(4).

(pgoyette)

Mention multiboot support for Xen.

(joerg)

Make cirpoll work.

(jmcneill)

fix compile with options TRAPDEBUG, ok by matt

(hans)

cirkqfilter is not implemented, so remove any references to it so this
file will compile again

(jmcneill)

Provide time_uptime and push it forward every now and then if we happen
to have the timer thread.  rump uptime - woohoo.

(pooka)

* implement uvm_km_alloc/free_poolpage()
* add silly stubs for uvm_lwp_hold/rele()

(pooka)

reorganize routines a bit better.  no functional change

(pooka)

remember to call percpu_init()

(pooka)

Add a simple percpu implementation (which isn't actually percpu at
all, since we don't currently have the notion of "real" cpu in
rump...but that's beyond the point).

(pooka)

fix brainfart: kthread_create() sets curlwp

(pooka)

use kmem_alloc/kmem_free rather than malloc() for sockokpt

(plunky)

Add code to adjust how we generate upcalls in sa_yield(). The
key problem is that we can only return from sys_sa_yield() iff
there are upcalls, and in some cases we can only tell that there
are upcalls to deliver after we have run userret().

So if we get woken up in a way that might not have generated upcalls
(timers, signals), try delivering upcalls. If we delivered something,
exit.

Add a flag to the vp & set it if we do this. In userret(), clear
the flag if set. Also, clear it in sa_upcall_userret() iff we get
into there.

Add comments describing the above so that the next person looking at
the code can understand it.

(wrstuden)

Add audio and midi devices, per report from Mike Hebel on port-cobalt.

(tsutsui)

Add u3g(4) driver from FreeBSD. This driver provides better support for
3G datacards than ugensa and will replace the latter for the supported
devices.

(joerg)

Regen.

(joerg)

Add entries for Vodafone Mobile Connect 3G datacard (from FreeBSD)
and for the Novatel 950D.

(joerg)

Support callouts and call callout_hardclock() from the timer
interrupt thread.

The sleepq implementation required for callouts is horrible, kludged
only for callouts, and generally unacceptable.  It needs revisiting,
but I'm not sure yet should rump or kern_timeout be improved.  It's
almost untested as of now, but committing this will give me some
maneuvering space while letting application compile.

(pooka)

sysctl doesn't like names that start with digits, nor does it like names
with embedded '.' so change the names to be better-conforming.  Reported
by njoly

(pgoyette)

Update CALLOUT_INVOKING correctly, seems to have been lost.

(ad)

Redo 1.169 correctly (dsl's fix doesn't do what I intended, either!).

(ad)

tcp_close: rearrange sequence of events slightly to make this atomic.
It was possible for a half-destroyed tcpcb to be visble, as softnet_lock
was being dropped.

(ad)

tcp_delack: test for TF_DELACK.

(ad)

include prerequisite headers

(pooka)