Enable file-system MFS. sysinst(8) assumes it's default.

(tsutsui)

Disable module autoload so that it won't foil our module unload test.

(pooka)

Fix pasto in comment.

(tsutsui)

Fix pasto in comment.

(tsutsui)

fix !USE_ATF

(pooka)

Return correct value from union_islocked.
(XXX: anything asserting LK_SHARED will fail because union doesn't
support shared locks)

(pooka)

fix !USE_ATF case

(pooka)

Now that we have a big devsw_conv0, need at least as many entries
in devsw as the highest device number in the conv table.  Do a
"good for a hundred years" guess now, fix properly later.

(pooka)

Note that the return type sould arguably be intmax_t instead of long long.

(jruoho)

Remove the (incorrect) note about inconsistent return values. Name the exit
code as EXIT_FAILURE instead of integer 1.

(jruoho)

Add more bugs. Namely, besides the return values that are confusing, no
function in a modern standard C library, exposed by <stdlib.h>, should be as
easy to (buffer) overflow as strsuftollx(3).

In addition, improvements to wording and markup.

(jruoho)

Improvements to wording and markup.

In addition, list more bugs and caveats. Namely, the NetBSD implementation
(like the FreeBSD one where this was ported to) requires that the comparison
keys are allocated dynamically, and that hdestroy() will try to free(3) each
key. This obviously severely limits the portability, given that other
implementations (for example, the Linux one) make no such assumptions. Both
approaches are legitimate, and thus the real bug is in the ambiguity of the
standard.

(jruoho)

in evbmips_iointr() for RMI, where we ack the EIRR,
replace (relatively expensive) splhigh()/splx()
protection with (more efficient) EIMR-based disable/restore.

(cliff)

remove some development debug prints,
in cpu_setup_trampoline_common()
and cpu_setup_trampoline_callback()

(cliff)

Update uebayasi-xip's base.

(uebayasi)

Following precedent in the i386 GENERIC kernel and in similar amd64
kernels, comment a few *VERBOSE options that have questionable value.

In follow up to PR 38579 and PR 38581.

(jakllsch)

IPL_VM is the highest interrupt priority where alldevs is read/written,
and acquiring alldevs_mtx already blocks those interrupts, so delete the
splhigh()/splx() in config_alldevs_lock()/_unlock().

Release alldevs_mtx while freeing memory with kmem_free(9); according to
new documentation, kmem_free(9) can sleep! :-) Thanks to rmind@ for the
tip.

Next step: use finer-grained locking, probably by adding a mutex to
    cfdriver_t.

And after that: make sure that all threads of execution are out of the
    device_t and/or softc before releasing their memory.

(dyoung)

silence "boot -z"

(hubertf)

Remove bpf modload code.  With the current state of affairs, the
kernel module either be autoloaded from the host or the kernel
module ABI is not supported on the given arch.

(pooka)

Add an exponential back-off to the loop that tries to lock the
PCI configuration-access registers, to avoid excessive cacheline
ping-ponging.  Thanks to rmind@ for the tip.

(dyoung)

Autogenerate /dev nodes.  Use (recreate) the naming policy in
MAKEDEV. -- Not the famous irrational file system devfs, but an
incredible simulation.

(pooka)

Include devsw_conv0 from an i386 kernel compilation (no, we don't
care about the arch as long as all the devices we care about are
present).  The file should be autogenerated, but that requires some
more changes to config(1).

(pooka)

For the simple cases, augment device-major with information on how
a driver expects /dev/node -> minor mappings to go and include that
information in devsw_conv.
(no, I didn't plow through all the MD majors files)

(pooka)

compress whitespace.  no functional change.

(pooka)

genfs_eopnotsuppify symlink, otherwise unlocky things happen!

(pooka)

add some missing kpreempt calls.  Fix an assert.

(matt)

Define mcontext_o32_t if !O32

(matt)

Syslog if login_tty fails

(matt)

Semantic change: Use _REG_x with gregs, no binary difference

(matt)

Add per process cache values to the total name translations count.

(njoly)

When .set reorder, the move goes before the jump, not after.

(matt)

No reason to set R_ZERO.

(matt)

Report alignment faults as SIGBUS, not SIGSEGV.

(matt)

In N32 kernel, cpu_setmcontext needs to deal O32 mcontext

(matt)

Enable PA2.0 CPUs, elroy, astro, ssio and attachments.

(skrll)

Note the PA of pages when allocated.

Simplify mapping of pages by traversing segs rather than the pageq and
calling VM_PAGE_TO_PHYS.

Only use PMAP_NOCACHE when hppa_cpu_hastlbu_p.

(skrll)

Add two DPRINTFs

(skrll)

Add HPPA_FTRS_TLBU to 7100LC and 7300LC CPUs.

(skrll)

Sync with HEAD.

(uebayasi)

Sync with HEAD.

(uebayasi)

Sync with HEAD.

(uebayasi)

Re-formatting.

(skrll)

For N32/N64, GP is callee-saved

(matt)

Fix rb_tree_iterate to get the min/max element of the tree for the
RBSMALL case.

(joerg)

In vmstat display, add one more column to output memory values. Avoid
overflow with more than 100GB.

(njoly)

Raise the blade, make the change, don't quite rearrange to where
it's sane.  But improves the non-monolithic kernel case anyway.

(pooka)

Rewrite the HISTORY section.

(jruoho)

Put the period outside the quotation marks.

(jruoho)

Add STANDARDS and CAVEATS. The latter notes the ambiguity of tdelete() when
deleting the root node.

(jruoho)

Add translation from vtype to dirent type.  Convert rumpfs now.
I'll convert the rest of the file servers in need after the next
version bump to avoid the coding module crisis.

(pooka)

Add a CTASSERT to make sure the cexp and ldavg arrays are kept in sync

(martin)

Get better results if initialize a field before calling strlen().
(can't believe that worked last night.  the stars must've been in
the "lottery, stupid" position)

(pooka)

Remove the note about eaccess(); no such thing in NetBSD.

(jruoho)

Also: in the example EOVERFLOW is probably better than ENOMEM.

(jruoho)

As err(3) instructs, use EXIT_FAILURE in the example.

(jruoho)

Steal the "malloc() vs. calloc()" -idiom from the OpenBSD's malloc(3).

While it may be debated how useful this is, good idiomatic usage examples
are exactly the kind of thing one would hope to see more in manual pages.

(jruoho)

Wording.

(wiz)

Sort errors.

(wiz)

Merge two entries for the same error.

(wiz)

Use a SECURITY CONSIDERATIONS instead of the BUGS section, and steal this
section from FreeBSD (with minor modifications).

(jruoho)

Put SECURITY CONSIDERATIONS as the last section.

(jruoho)

Fix the mess in the sections; someone went to sleep in the middle of a
sentence...

(jruoho)

They've finally made gets(3) obsolete (in POSIX, at least).

(jruoho)

Add HISTORY.

(jruoho)

Marked as legacy already in SUSv2; removed in POSIX 2001.

(jruoho)

Fix MKDEBUG=yes library builds for sparc64 X11R7

(lukem)

Deprecated in POSIX (rev. 2008).

(jruoho)

Also tmpnam(3) and tempnam(3) are now obsolete in POSIX.

(jruoho)

Interestingly, the committees plan to remove also isascii(3) and toascii(3)
in the future.

(jruoho)

More obsolete things. These are all already deprecated in NetBSD, but it is
worth noting that also the standard has marked these as legacy.

(jruoho)

Reference timeradd(3).

(jruoho)

Add STANDARDS section. Note that both functions were marked as obsolescent
in POSIX (2008); "timer_gettime(2) and timer_settime(2) should be used".

(jruoho)

Small improvements to wording and markup.

(jruoho)

.Nm -> .Fn.

(jruoho)

add mpt.

(chs)

fflush stdout, before writing to stderr.

(sjg)

p2k asserts get upset if VOP_ISLOCKED() doesn't reflect lock status.
So let genfs do its thing.

(pooka)

Add rump_au-naturel, which instead of mounting a file system inside
the rump kernel just attaches to / in said rump kernel instance.
This is mostly useful in case wanting to see what nodes components
create (see Makefile).

not built without manual intervention

(pooka)

regen

(chs)

enable TMPFS in all GENERICs that have MFS enabled.

(chs)

If vfstype is MOUNT_RUMPFS, assume we don't actually want to mount
a file system, but attach to the root in the fs namespace of a rump
kernel.

(pooka)

add simple VOP_READDIR for rumpfs

(pooka)

Wording.

(wiz)

Rename -w to -c by popular demand
(allow file 'c'reation - 'w'riting was always available, with some twist;
plus there's already such a feature elsewhere, so let's try to not
diverge needlessly)

(hubertf)

Fix date strings.

(wiz)

Don't mention "umac-64@openssh.com" MAC algorithm which is not enabled
in our SSH client and daemon as it causes crashes on architectures which
strict aligment requirements (e.g. NetBSD/sparc64).

This fixes PR bin/43221 by myself.

(tron)

It was nice to expect fts(3) to appear in the '88 revision. But perhaps
people had too high hopes.

(jruoho)

No more usleep(3) in POSIX; "use nanosleep(2) instead".

(jruoho)

Fix markup.

(jruoho)

Note that also POSIX has deprecated utime(3).

(jruoho)

Fix the standards: utimes(2) is still in SUSv4, although marked as legacy.

(jruoho)

Update this, mechanically, to match the new functions and their prototypes.

XXX: Someone more familiar with the code should proofread the page and
    evaluate how well it reflects the reality in 2010.

(jruoho)

Remove gsignal(9) and kgsignal(). Add links for kpgsignal(9) and kpsignal(9).

(jruoho)

Remove gsignal() and kgsignal() that do not prototype anything any more.

(jruoho)

Only use complete prototypes in the SYNOPSIS.

(jruoho)

Remove sched_psignal(9) and sigmasked(9). From Tetsuya Isaki in PR #40125.

(jruoho)

Fix MKDEBUG=yes for libraries on sparc64 for base.

(lukem)

devsw_conv is not used outside _KERNEL

(pooka)

Note the ambiguity of the return value in the BUGS section.

From Christian Biere <christianbiere@gmx.de> in PR #39392.

(jruoho)

Don't mix opening and closing macros of different types.

(joerg)

Note POSIX compliance.

(jruoho)

Sort errors.

(wiz)

Add punctuation.

(wiz)

Improve a sentence.

(wiz)

Note that utimes(2) no longer enjoys the blessing of POSIX.

(jruoho)

Sort options in SYNOPSIS; improve wording.

(wiz)

Call malloc(M_ZERO) instead of malloc() and memset().

(kiyohara)

Add some STANDARDS lingo, mainly to note the functions that have been
removed from POSIX (2008).

(jruoho)

Remove unnecessary cast.

(kiyohara)

+ Fix typo.
+ Call aprint_error_dev() instead of aprint_error().

(kiyohara)

+ Use structs scsipi_inquiry_data/scsi_sense_data instead of scsi3_inquiry_data/
  scsi3_sense_data.
+ s/switch()/switch ()/.
+ Change args for printf().  ("FUNCNAME ...") -> ("%s ...\n", __func__, ...)
+ Indent.
+ Remove unnecessary cast.

(kiyohara)

Adds hw quirk for SUN firewire controllers.  From FreeBSD.
  Thanks for cegger@.

(kiyohara)

Switch to the less restrictive binary detection algorithm from zip 3.0e.

(joerg)

Fix build of rump.

(uebayasi)

Don't specify anything in "int flags" of bus_space_physload_device(9)
for now.

(uebayasi)

Properly mark the standard that "deprecated" all of these.

(jruoho)

Note the recent changes.

(jruoho)

fix indent.

(nonaka)

Adapt to the new uvm_page_physload_device(9).

(uebayasi)

"int free_list" (VM_FREELIST_*) is specific to struct vm_page (memory
page).  Handle it only in memory physseg parts.

Record device page's properties in struct vm_physseg for future uses.
For example, framebuffers that is capable of some accelarated bus access
(e.g. write-combining) should register its capability through "int
flags".

(uebayasi)

Fix a thinko in Rev. 1.153.2.30.

(uebayasi)

Revert previous; unintended changes mixed.

(uebayasi)

Fix thinko in previous.

(uebayasi)

Fold long lines.

(uebayasi)

FIx an off-by-one in my new code.

(uebayasi)

use platid_match().

(nonaka)

WS003SH is normally reseted.

(nonaka)

Bump kernel version for bus_space_tag_t changes on x86.

(dyoung)

add byte-swapping so that this works on BE platforms.
the logic is based on the current freebsd driver.
fixes PR 42870.

(chs)

tftpd(8): Add -w so files can be uploaded without requiring them
to be created before the upload. See the section on security
considerations before enabling. [hubertf 20100429]

Addresses PR bin/43164.

(hubertf)

Silence for boot -z

(hubertf)

To reduce differences between ath(4)'s PCI attachment and its CardBus
attachment, delete several unnecessary #includes, re-order subroutines,
move around line breaks, change to different names for the same
constant, et cetera.  Comparing this version of if_ath_pci.c with my
uncommitted version of if_ath_cardbus.c, only about five lines are
different.

(dyoung)

Provide an x86 implementation of pci_chipset_tag_create(9) and
pci_chipset_tag_destroy(9).

(dyoung)

Document pci_chipset_tag_create(9) and pci_chipset_tag_destroy(9).

(dyoung)

ACPICA 20100428 is out.

(jruoho)

Add data types, function prototypes, and stub implementations
for pci_chipset_tag_create() and pci_chipset_tag_destroy().  On
architectures that support it, an MI PCI bus driver can override the
architecture's default pci(9) and pci_intr(9) implementation.

Coming up next: documentation.
After that: x86 implementation.
Last but not least: make cbb(4) use MI PCI overrides.

(dyoung)

#include <sys/bus.h> instead of <machine/bus.h> here to get all of the
MI prototypes.

(dyoung)

Fix 'tags' target.

(dyoung)

Cosmetic: delete whitespace at end of line.

(dyoung)

Two small grammar fixes.

(jruoho)

On x86, change the bus_space_tag_t to a pointer to a struct
bus_space_tag.  For now, bus_space_tag's only member is
bst_type, the type of space, which is either X86_BUS_SPACE_IO
or X86_BUS_SPACE_MEM.  In the future, new bus_space_tag members
will refer to override-functions installed by a new function,
bus_space_tag_create(9).

Add pointers to constant struct bus_space_tag, x86_bus_space_io and
x86_bus_space_mem.  Use them to replace most uses of X86_BUS_SPACE_IO
and X86_BUS_SPACE_MEM.

Add an x86-specific bus_space_is_equal(9) implementation that compares
the two tags' bst_type.

(dyoung)

rename pollard_pminus1->pollard_rho for consistency

(drochner)

Add ASUS CUV4X-D. From Pierre Pronchery in PR #41351.

(jruoho)

Fix warnings for -DRBSMALL

(joerg)

Instead of causing a catastrophic failure when uvm_loanuobjpages()
is called, just pretend it hit a wired page and return EBUSY.  This
should cause callers to use a non-loaning access path instead.

Fixes file read path of stock nfs kernel module (previously it was
easy to just frob the nfsd_use_loan boolean into the other position
by simple value assingment, but now that nfsserver can be autoloaded
at runtime into the rump kernel, that approach is a little more
challenging since the variable isn't available in application
linkage.  yes, things like adding a sysctl for the variable would
work, but now everything works out-of-the-box).

(pooka)

When mounting a block device as XIP, pass registered struct vm_physseg
* as a cookie from the block device to the caller (== mount code).
struct vm_physseg * will be passed to XIP vnode pager
(genfs_do_getpages_xip()), then converted back to paddr_t.

(My future plan is to pass struct vm_physseg * back to the fault handler,
and to pmap_enter() as is.)

(uebayasi)

kill stray free. from Sverre Froyen.

(christos)

Initial implemention of bus_space_physload(9) for NetBSD/arm (only i.MX31
for now).

(uebayasi)

Uncrack my brain, buggy let me !be.

In other words, make previous commit to this file actually work.
Funny how it's possible to fix things to not dump core and still
have them not work.

(pooka)

update python bindings to work with newer library interface

(agc)

Set l_sysent so that syscall autoload actually works instead of autocores.

(pooka)

Note that the ucontext(2) et. al. were first moved from X/OPEN extensions to
the base POSIX-2001 and later marked as obsolete in POSIX-2008.

(jruoho)

refresh the perl bindings to catch up with the changes in the library

(agc)

Clean up gt and peripherals.
  This change tested compile only.

(kiyohara)

At device attachment, register the given bus address region as possible
"managed" region using the new bus_space_physload_device(9) API.

(uebayasi)

Initial support of uvm_page_physunload(9) and uvm_page_physunload_device(9).
Note that callers of these functions are responsible to ensure that the
segment is not used.

(uebayasi)

Improvements to markup.

(jruoho)

If RUMP_NCPU is set, use that for virtual CPU count instead of
host CPU count.

(pooka)

Fix MKDEBUG=yes library builds for amd64

(lukem)

Manage struct vm_physseg as a list, which means that struct vm_physseg
objects don't move when a segment is added / removed.

(uebayasi)

Always use struct vm_physseg *vm_physmem_ptrs[] in MD code.

(uebayasi)