Remove unused fss_softc member sc_bs_size.

(hannken)

Follow christos' suggestions, and make ks_active a u_short, and
also only use 16 u_shorts instead of 32 ints.  Also add panic()
calls for under- and overflow of the ks_active members under
DIAGNOSTIC.  The MAXBUCKET constant ended up in sys/mallocvar.h
and not sys/param.h, as the latter caused build problems.

Ride the kernel revision bump of my previous change.

(he)

Sort ERRORS section.

(wiz)

Fix date.

(wiz)

Push the bpf_ops usage back into bpf.h. Push the common ifp->if_bpf
check into the inline functions as well the fourth argument for
bpf_attach.

(joerg)

Extend struct malloc_type to count the number of active allocations
per size, and make vmstat report this information under the "Memory
statistics by type" display, which is only printed when the kernel
has been compiled with KMEMSTATS defined, like this:

Memory statistics by type                                Type  Kern
          Type InUse  MemUse HighUse  Limit  Requests Limit Limit Size(s)
          wapbl    15  4192K  4192K  78644K    376426    0    0 32:0,256:3,512:6,131072:1,262144:2,524288:3

Since struct malloc_type is user-visible and is changed, bump kernel
revision to 5.99.26.

While it is true that malloc(9) is in general on the path of slowly
being replaced by kmem(9) (kmem_alloc/kmem_free), there remains a
lot of points of usage of malloc/free, and this could aid in finding
any leaks.  (It helped finding the leak fixed in PR#42661.)

This was discussed with and somewhat hestitantly OKed by rmind@

(he)

Avoid use of LOOKUP_BOOTINFO, since LOOKUP_BOOTINFO causes panic
when BTINFO_XXX is not passed from ofwboot.

This makes -current kernel bootable with old ofwboot.

(nakayama)

Ticket 1361.

(snj)

Cherry-pick the /libexec content for (non miniroot) CDs and add the sysinst
specific dhcpcd hook scripts

(martin)

Pull up following revision(s) (requested by mrg in ticket #1361):
external/mit/xorg-server/dist/render/glyph.c: revision 1.2
import fdo git change 185185eeb44a277c324be0f58a4b4a469b56b69b, which
i've seen twice in the last two days:
Fix crash when all glyphs of a given depth are freed, but not all glyphsets
This is how the crash can be triggered with only two clients on the system:
Client A: (already running)
Client B: Connect
Client B: CreateGlyphSet(depthN)
Client A: Disconnect
Server: free globalGlyphs(depthN)
Client B: AddGlyphs(depthN)
Server: SEGV
This crash was introduced with the FindGlyphsByHash function
in 516b96387b0e57b524a37a96da22dbeeeb041712. Before that revision,
ResizeGlyphSet was always called before FindGlyphRef, which would
re-create globalGlyphs(depthN) if necessary.
X.Org Bug 20718 <http://bugs.freedesktop.org/show_bug.cgi?id=20718>
XXX: needs netbsd-5 pullup.

(snj)

Remove bogus workaround hiding a real bug: /tmp/dhcpcd-lease should be
created by the sysinst specific dhcpcd hook

(martin)

Turn a printf into a DPRINTF(), since we will very likely recover from
the condition (and it happens on stock install CDs)

(martin)

cd_size: if we fake a size (and I realy have no idea why this would be a
good idea), at least set up all values to the fake values, as the caller
expects.
Should fix PR kern/39904, though if noone can find out why the fake value
would be needed, we should change it to just return 0 as suggested in the PR.

(martin)

FTBFS, POV, TOCTOU, and YAGNI.

(jruoho)

don't leak a vnode and don't call namei (implicitly) twice

(jnemeth)

Note Intel 82855GM AGP support in agp(4).

(jakllsch)

Add (non-IGP) AGP port support for Intel 82855GM.
agp_intel tested with radeondrm(4).  agp_i810 tested with i915drm(4).

(jakllsch)

ctime can return NULL

(christos)

ctime can return NULL

(christos)

handle ctime returning NULL.

(christos)

Fix the comments about cpuid flags, according cpuid documentation by
Intel and AMD.

(jym)

don't use NULL for 0

(christos)

relative .PATH

(christos)

remove useless double [[

(christos)

hack for amd64

(christos)

use relative .PATH, and remove one that is not needed.

(christos)

- use CPPFLAGS instead of CFLAGS
- use relative .PATH

(christos)

Only unregister the sensors if these were successfully installed;
sysmon_envsys_unregister() does not check for NULL pointers gracefully.

(jruoho)

Move most clock_getres syscall code, except for coypout call, to a new
clock_getres1() function which can be used by emulations. Adjust all
clock_getres syscalls to now make of use it.

(njoly)

Update the limits when a change from absent to present is detected.

(jruoho)

Mention the history of stat(2) and fix the history of open(2).

(jruoho)

The functions read(), wait(), and write() appeared already in v2 UNIX.

(jruoho)

Add a history section. From OpenBSD.

(jruoho)

An unlink() appeared already in v2 UNIX.

(jruoho)

Add some history.

(jruoho)

Some interesting additions to the history section. From OpenBSD.

(jruoho)

Trivia: one .Pp after .El, speak about NULL instead of nil pointers, etc.

(jruoho)

If we're detaching the device, make sure we unregister the sensors.

(pgoyette)

When unregistering a device, in addition to reducing the global count
of sensors, we need to adjust the sme_fsensor values for some devices.
Failure to do this leaves a sparesly-populated number-space and can
subsequently result in overlapping sensors number assignments.  (This
only affects the compatability-mode API, but still needs to be fixed.)

(pgoyette)

use -Wno-pointer-sign in the HAVE_PCC case also

(plunky)

- add atari's EtherNEC to SYNOPSIS
- also pull comments about MD device description from each kernel config
- note that RTL8019 is widely used on local busses other than ISA

(tsutsui)

Wrap long line.

(skrll)

Add md_astpending to struct mdlwp and use it.

(skrll)

Insert missing )

(pgoyette)

Use time_t (not long) to save time_second value.

(tsutsui)

Ooops, remove leftovers from the ramdisk /

(martin)

update portmaster and releng lists

(spz)

ctime can return NULL.

(christos)

check ctime.

(christos)

check ctime return

(christos)

Move install CDs over to / on cd9660 (instead of a memory disk image)

(martin)

regen

(jakllsch)

Add Intel 82855GM Host-AGP Bridge.

(jakllsch)

- int putbuf_func ARGS((const char *s, size_t len));
+ int (*putbuf_func) ARGS((const char *, size_t));

for pcc
- did not recognise pointer to function
- argument names shadowed other arguments

(plunky)

fix typo.

(christos)

Per-cpu "physical" trap save area.

(skrll)

fix printf format.

(christos)

starttime needs to be time_t (Izumi Tsutsui)

(christos)

Bump version

(martin)

Add support for /boot.cfg. Implement a single command for now: override
the boot partition, which will be used for bootable CDs. Add cd9660
support.

(martin)

Add a bootinfo node to allow the loader to override the boot device

(martin)

handle ctime returning NULL.

(christos)

remove bogus cast that hides bugs if tv_sec is non POSIX compliant.

(christos)

handle ctime returning NULL.

(christos)

handle ctime returning NULL.

(christos)

make it obvious what ctime it is used for.

(christos)

handle ctime returning NULL.

(christos)

Pull across pmap_remove fix from OpenBSD.

Fix user-after-free bug in pmap_remove().  Page table pages are freed as soon
as the last page table entry that was actually used is cleared.  So make sure
we check the page table page is still there for every page we remove.
Otherwise we will bring back the tlb entry and cache lines when we touch the
freed page, and we will create an illegal alias (non-equivalent mapping)
as soon as the page gets re-used.

(skrll)

make it obvious to grep that ctime is being checked.

(christos)

Change #ifdef DIAGNOSITC tests+panic to KASSERTs.

(skrll)

handle ctime returning NULL.

(christos)

make it obvious what ctime is used for.

(christos)

handle ctime returning NULL.

(christos)

handle ctime returning NULL.

(christos)

handle ctime returning NULL.

(christos)

make it obvious to grep that we are checking ctime.

(christos)

fix debugging printf.

(christos)

PR/43105: Peter Kerwien: Destroying a GPT partition with dd will cause gpt
destroy / create to fail. When reading the GPT label from the end of the disk
ignore errors if the GPT label at the beginning of the disk was not found.

(christos)

Delete ieee80211_setbasicrates().  It's buggy, and we drivers can get
along fine without it.

(dyoung)

handle ctime returning NULL

(christos)

Change booke_intr.h to intr.h since <powerpc/booke/intr.h> will be unique
enough.

(matt)

OpenPIC register defintions used by e500 booke.

(matt)

Mention Domex DMX-3191D support in nca(4).

(jakllsch)

Fix some TCR definitions.

(matt)

Mention EtherNEC support for atari.

(tsutsui)

Regen from GENERIC.in rev 1.85, HADES.in rev 1.10, and MILAN.in rev 1.24:
> Add support for the EtherNEC.

(tsutsui)

Add support for the EtherNEC.

The EtherNEC is yet another user designed device for Atari machines,
an NE2000 based ISA board connected to Atari's ROM cartridge slot with
simple but ingenious interface and cartridge port adapter, and
it's the most available network solution for the old Atari machines.

See following pages for more details about EtherNEC:
http://hardware.atari.org/ether/
http://home.arcor.de/thomas.redelberger/prj/atari/etherne/
http://www.freemint.org/ethernec/ethernec.html

(tsutsui)

Oops.  Fix LOCKDEBUG panic -- and spurious calls to tcp_output()! -- in
previous.  Be careful with that {}, Eugene.

(tls)

Add the "Statement for The NetBSD Foundation's Position on the GPLv3".

(reed)

Whitespace.

(skrll)

Permit creation of an event-monitor for alarm limits even if the limits
are not currently set (and, in case of battery capacity sensors, if there
is no value_max).  Ensure that such an event-monitor does not trigger the
delivery of any actual events.

Provide a mechanism for sensors to set their limits at time other than
system startup (for example, when a battery is inserted).

This allows us to boot a system with a battery missing, install the
battery some time later, and automatically monitor it without requiring
any user intervention to create the event-monitor.  (The actual changes
for battery sensors to use this new capability will come later.)

(pgoyette)

Small simplification

(skrll)

import fdo git change 185185eeb44a277c324be0f58a4b4a469b56b69b, which
i've seen twice in the last two days:

Fix crash when all glyphs of a given depth are freed, but not all glyphsets
This is how the crash can be triggered with only two clients on the system:
Client A: (already running)
Client B: Connect
Client B: CreateGlyphSet(depthN)
Client A: Disconnect
Server: free globalGlyphs(depthN)
Client B: AddGlyphs(depthN)
Server: SEGV

This crash was introduced with the FindGlyphsByHash function
in 516b96387b0e57b524a37a96da22dbeeeb041712. Before that revision,
ResizeGlyphSet was always called before FindGlyphRef, which would
re-create globalGlyphs(depthN) if necessary.

X.Org Bug 20718 <http://bugs.freedesktop.org/show_bug.cgi?id=20718>

XXX: needs netbsd-5 pullup.

(mrg)

fix a few pastos

(pooka)

Cosmetic: join a couple of lines.  Stop storing the value of a BAR in a
temporary variable with the name 'csr'.

(dyoung)

don't divide by zero. ever.

(macallan)

mention adm1031 support

(macallan)

mention ADM1031 support in dbcool

(macallan)

add some more ADM1031 support:
- additional temperature and fan speed sensors work now
- limits on should work on all temperature sensors
Controlling fan behaviour doesn't work yet, the ADM1031 uses the same
register for both fans which doesn't quite fit with the way the dbcool
driver works.

(macallan)

Add support for Domex 536 PCI SCSI controller to nca(4).
This truly remarkable chip is found on the Domex DMX-3191D SCSI adapter.

(jakllsch)

In case we create a file, reclaim the vnode, and lookup the file
without the directory timeout expiring, we get vattr_null as the
attributes for that file.  Ensure that we always report sane
attributes by getting them from the server if this is the case.
(also, sprinkle some const)

Fixes problem reported by dyoung.  But wait!  The bug's medallion
begins to glow!  The bug looks much better!  I crumble to dust.

There's probably another similar bug related to "lazy open".  It
will trigger if we reclaim a node before the response to the open
arrives.  Even the comments (typoless) know about this bug ...
But verifying it exists and fixing it will have to wait for another
day.

(pooka)

As suggested by at least 3 different people (the guilty parties know who
they are) avoid repeated kernel_lock/unlock by using an intrq on the stack.

About 5%-10% better from run to run, on my *very* simpleminded test.  Can't
possibly be worse.

(tls)

Fix the code for 5822 B0 errata. Use pci_conf_read() and access at
32bit alignment.

(msaitoh)

After discussion with ad@: it appears that KERNEL_LOCK also protects
the driver output path (that is, ifp->if_output()).  In the case of
entry through the socket code, we are fine, because pru_usrreq takes
KERNEL_LOCK.  However, there are a few other ways to cause output
which require protection:

1) direct calls to tcp_output() in tcp_input()
2) fast-forwarding code (ip_flow) -- protected elsewise
  against itself by the softnet lock.
3) *Possibly* the ARP code.  I have currently persuaded
  myself that it is safe because of how it's called.
4) Possibly the ICMP code.

This change addresses #1 and #2.

(tls)

Just ignore the errno of strtoimax, it doesn't provide much value.

(joerg)

Remove 2**31 limit on the number of lines or bytes to process.
For the byte operations, process in blocks of 64KB.

(joerg)

remove a now useless bus_space_barrier()

(macallan)

Fix copyrights.

(ad)

add necessary build bucatini for kernfs tests

(pooka)

Add some basic tests for kernfs (getdents & changing hostname).

Per a "mad \"scientist\"" experiment, on i386 load kernfs support
from the installed kernel module instead of linking the driver in.

(pooka)

KNF FTW

(ad)

Create kern.hostname sysctl node.  Using init_sysctl.c for creating
the nodes requires some weeding, so don't try to use it yet.

(pooka)

Use tp-glob keys when requesting for test pattern.

(njoly)

The Atffile tp/tp-glob keys does not allow more than one word. Put new
union value on its own line.

(njoly)

Add preliminary support for the ADM1031 found in the Sun Blade 2500 and similar
machines. For now we treat it like an ADM1030 - the only difference is the
device ID and a few previously reserved registers.

(macallan)

2 RCSIDs are enough for anyone.

(skrll)

Add the "Linkers and Loaders" to SEE ALSO.

(jruoho)

Make eval a local variable.

(joerg)

ANSIfy. Sprinkle static.

(joerg)

fix typo (at least it was in an error string instead of a comment)

(pooka)

regen

(pooka)

MAP_COPY is available only in _KERNEL .... uh, so don't use it here.

(pooka)

Add note saying that the manual page is autogenerated and all parts
may not apply to this particular server.

TODO: have three different texts, one for device backed file systems
(e.g. ffs), another for network backed (e.g. nfs), and a third for
fictional (e.g. kernfs)

(pooka)

generate

(pooka)

Add kernfs server.

Don't build & install it by default, since it represents the state
of the rump virtual kernel and not the host kernel.  However, it's
useful for debugging kernfs.

(pooka)

+rumpfs_kernfs

(pooka)

support kernfs

(pooka)

Misc KNF and cosmetics.

(tsutsui)

separate vfs-like components from actual drivers
(no functional change)

(pooka)

We don't have a real rootdev (by default at least), so set it to NODEV.

(pooka)

convert to parseargs

(pooka)

ANSIfy, some KNF.

(tsutsui)

initialize canon_dev

(pooka)

Make cpl, ipending, and intr_depth per CPU values.

(skrll)

Now that fw_port.h is gone, we need to directly include <sys/select.h>

Fixes build break reported by myself.

(pgoyette)

set kernel boottime

(pooka)

* initialize msgbuf
* for banner, use aprint_verbose instead of if (verbose) printf

(pooka)

Note import of dhcpcd-5.2.2

(roy)

Import dhcpcd-5.2.2 with the following changes:
* Don't send a maximum DHCP message size beyond what we can handle.
* ip_id is now randomized.
* IPTOS_LOWDELAY and IP_DF flags are no longer set.

Status:

Vendor Tag: roy
Release Tags: dhcpcd-5-2-2

(roy)

Don't hold kernel lock across call to ip_input() -- it blocked *all*
hardware interrupts for the length of time it took for all dequeued
packets to flow up the stack (on multiprocessors only).  Initial testing
shows performance impact is minimal -- since this temporary fix actually
means taking/releasing the kernel lock per-packet, that seems
acceptable.

Holding the kernel lock across the ip_input() call duplicated the
exclusion intended to be provided by the socket locks/softnet lock
(same lock, for INET/INET6 sockets) and could mask serious bugs.  Several
hours' testing didn't turn any up but I'd be surprised if some don't now
appear.

Damon Permezel noticed the problem.  Temporary fix suggested by matt@.

(tls)

Whitespace.

(skrll)

#define<tab>

(skrll)

Do as OpenSolaris does and read the status register after each write.
Now this driver works on my Blade 2500.

(macallan)

Kill stray t.

(joerg)

If msgbuf is not enabled, do not report the node in readdir.  That
way ls -l won't report funny errors because getattr for a readdir
result fails.

XXX: lookup for msgbuf still succeeds even if not enabled

(pooka)

Include more sources in the tags target.  Don't compute tags on some
non-existent assembly-language files.

(dyoung)

Do not bother to ask the user for CD details (like device name or path
on the CD) if we can find CDs (i.e. media, not drives) automagically
and check the standard paths on it.

If we find multiple CDs, offer a menu to select showing the volume
name. If something goes wrong, fall back to the manual input method.

In typical installs this makes us just proceed with extracting the sets
after selecting "install from CD/DVD", w/o any further interaction.

As a side effect fixes PR 43012.
XXX new messages need translations.

(martin)

Comments against #endifs.

(skrll)

Whitespace.

(skrll)

Whitespace.

(skrll)

Update for libarchive 2.8.2

(joerg)

yinterrupt -> interrupt

(tnn)

\\ -> \e

(joerg)

Fix \\ encoding

(joerg)

Kill non-existent macros.

(joerg)

Assume o umlaut is transliterated by nroff as needed.
\\ -> \e

(joerg)

\\ -> \e

(joerg)

move change from rev 1.232 to the initialization part.
This fixes the problem 'nbconfig: cannot create /GENERIC'
when KERNOBJDIR has a relative path and is defined as env variable.
Problem reported and fix proposed on tech-toolchain@.

'commit it then' christos@

(cegger)

Whitespace.

(skrll)

Correct comment.

(skrll)

make sure we pickup the powerpc majors file on powerpc64.

from dennis.c.ferguson@gmail.com in PR#43074.

(mrg)

install altivec.h in powerpc64.

from dennis.c.ferguson@gmail.com in PR#43075.

(mrg)

build elf2aout and fdformat on macppc64 (for now, it's a lot easier than
figuring out how to split the sets... but that will have to happen.)

fixes the first easy part of PR#43075.

(mrg)

for now, use the C versions of bn and des, not the (32 bit) versions,
on powerpc64.

from dennis.c.ferguson@gmail.com in PR#43073.

(mrg)

qft: quoted for truth

(lukem)

fix a KASSERT() in the previous.  add a couple of comments.

(mrg)

anti-pasto

(pooka)

add nullfs/umapfs tests to setlists etc

(pooka)

Add basic test for umapfs functionality.

XXX: the reverse mapping case (last subsubtest in t_basic) does
not make any sense, but apparently that how umapfs works.  I'm not
familiar enough with the code to determine if this is a wanted
feature or a pure and simple bug.

(pooka)

Basic test for nullfs, which just checks that the two views act
somewhat sensibly.

(pooka)

add nullfs and umapfs to setlists

(pooka)

support nullfs and umapfs

(pooka)