Whitespace fixes. No functional change.

(msaitoh)

G/C ri_delta.

XXX
Bump kernel version after other changes for struct rasops_info.

(rin)

Switch to per-device stamp, and retire stamp_mutex.

XXX
Bump kernel version after other changes for struct rasops_info.

(rin)

Misc clean-up's:
- protect private stuff in rasops.h by _RASOPS_PRIVATE
- staticify rasops_copycols() and rasops_isgray[]
- G/C unused extern int cold

(rin)

Add a forward declaration for the specialized __timepunct<wchar_t> destructor
that the dragonfly code needs to avoid "instantiation before specialization",
and an empty implementation for the generic code.

(christos)

There is no specialized constructor for ctype<char> so the
destructor ends up trying to free uninitialized memory for
_M_c_locale_ctype.

(christos)

Arrange so that the Makefile works for both the generic and the dragonfly
locale code.

(christos)

Include pe_lwp member in 'struct ptrace_state'

Reviewed by kamil.

(mgorny)

Welcome to 9.0_BETA!

(martin)

Avoid undefined behavior when setting multicast address. found by kUBSan.

(msaitoh)

Avoid undefined behavior when reset the chip. found by kUBSan.

(msaitoh)

Welcome to NetBSD 9.99.1!

(martin)

- Avoid undefined behavior in bge_setmulti(). found by kUBSan.
- Avoid undefined behavior when setting the MAC address in bge_init().
  found by kUBSan.

(msaitoh)

Try to improve performance when shadow framebuffer is present;
Use block copy from shadow fb to real fb, instead of repeating
the same operations to two fb's.

(rin)

Support shadow framebuffer for 24bpp screen.

(rin)

Update libarchive entry for 3.4.0 import.

(joerg)

In rasops_copy{rows,cols}(), if src == dst, we have nothing to do.
Otherwise, we can use memcpy safely instead of memmove.

(rin)

Fix catastrophe when ri_emustride != ri_stride in rasops_copyrows().

(rin)

Treat highlighted and reversed text in the same manner to xterm.

(rin)

vmx(4) can select workqueue for packet processing like ixg(4).

(knakahara)

Fix overflow to calculate frequencies.

(hkenken)

Note import of dhcpcd-8.0.2

(roy)

Sync

(roy)

ixgbe_setup_low_power_mode() calls ixgbe_stop() in it. ixgbe_detach() first
called ixgbe_stop() and then called ixgbe_setup_low_power_mode() which resulted
in calling ixgbe_stop() twice. Call only ixgbe_setup_low_power_mode() first to
remove extra ixgbe_stop() call.

(msaitoh)

It's not required to call ixv_stop() twice in ixv_detach().

(msaitoh)

Add libpthread to DPADD and LDADD, as libarchive now references it

(martin)

Fix loop to calculate prescaler divisor.

(skrll)

Add IIC support.

(hkenken)

Whitespace

(skrll)

Remove fdtbus_gpio_release() from enet_phy_reset().

(hkenken)

- rnd_add_*() call must be serialized. Unserialized accesses broke the
  sc->rnd_source structure and caused off by one in kern_rndq.c's
  rnd_add_data_ts(). This problem was found by kUBSan. Make krndsource_t's
  data per queue-pair and call rnd_add_uint32() in wm_txrxintr_msix(). One of
  the reason why rnd_add_uint32() calls were moved from wm_{tx,rx}eof() is that
  rnd(9) manual page says "To allow the system to gather the timing information
  accurately, this call should be placed within the actual hardware interrupt
  service routine."
- Pass a better value to rnd_add_uint32() in wm_intr_legacy().

XXX pullup-8.

(msaitoh)

MAKEDEV.tmpl: Create nodes for 16 USB hubs

As proposed on current-users, but with better formatting.

(gdt)

Try to fix problem reported by ryoon:
http://mail-index.netbsd.org/source-changes-d/2019/07/29/msg011516.html

There are apparent bugs for shadow framebuffer. Sorry for the breakage.

(rin)

Hmmm, color was still strange for 24bpp on little endian machine,
only when font width is 12.

We need to use different devcmap for that case, if we wish to share
codes for other depths/font widths as possible as we can.

XXX
What should we do for big endian? I have no big endian machines with
24bpp framebuffer...

(rin)

Simplify and tiny optimization.

(rin)

Fix color on 24bpp screen for little endian.

(rin)

Enable anti-aliasing for depth 24.
Style.

(rin)

Support anti-aliasing for depth 24.

(rin)

Added missing efi_memory_probe() call.

(nonaka)

sys/dev/ic/mfi.c: Add missing break in switch

(The entire switch is guarded by MFI_DEBUG and is known not to build.)
Reported by Oskar.

(gdt)

Sync the output of memmap command to the output of stand/efiboot mem command.

(nonaka)

Added BTINFO_EFIMEMMAP compaction support to x86 efiboot.

(nonaka)

The CUBOX-I boards have Cortex-A9 cpus, so force -march=armv7-a
when compiling this kernel.

(martin)

Factor out putchar_aa functions into rasops_putchar_aa.h, which includes
the following fixes:

- stop using memset to framebuffer for depth 8
- correctly support non-standard positions/lengths of RGB bits in pixel

(rin)

make vmx(4)'s *_process_limit tunable.

(knakahara)

Remove duplicate options TPIDRPRW_IS_CURCPU

(skrll)

vmx(4) uses deferred interrupt handlering like ixg(4).

(knakahara)

Fix missing NULL check after softint_establish().

(knakahara)

Fix info leak: the padding after the header causes uninitialized heap
memory to be copied to userland in sys_recvmsg().

(maxv)

Fix warning in userdel -r.

If the 'rm -rf' works, then the home directory does not exist any more,
and rmdir() on it will return an error (ENOENT). Catch and ignore it.

Avoids:

# useradd -m foo
# userdel -r foo
userdel: Unable to remove all files in `/home/foo': No such file or directory

(wiz)

Avoid undefined behavior when converting unaligned be32 data to
host integer, found by kUBSan.

Pointed out by msaitoh.

(rin)

Remove extra paren

(christos)

remove extra paren

(christos)

Convert rasops2.c and rasops4.c to use rasops_putchar_width.h.
Style.

(rin)

Clean up for rasops1.c:
- dedup ugly copy-paste
- rewrite to factorize width-optimized putchar functions
- misc style

(rin)

Join Tx interrupt handler and Rx interrupt handler of vmx(4).

That can reduce interrupt resources.

(knakahara)

Fix missing underlines on mono screen.
Style.

(rin)

SGIs from the MIPS & ARM world :)

(sevan)

Factor out the uselocale() calls by providing the _l missing functions as
inlines in xlocale_port.h except in c_locale.{cc,h} because the header gets
installed.

(christos)

If available, show additional names of foreign partitions in the
size menu. Makes installing on machines with other OSes installed
a lot less confusing.

(martin)

When merging existing partitions into the planned installation partitions,
keep the separation between the original planned partitions and the
pre-exiting partitions.
Also do not call found partitions "outer" if they are not (but could be)

(martin)

regen

(mlelstv)

Add BCM2711 PCIe Host bridge for RPI4

(mlelstv)

typo

(sevan)

PR/54414: Valery Ushakov: c99 requires that wcsrtombs(3) updates the source
argument on conversion error

(christos)

PR/54414: Valery Ushakov: c99 requires that wcsrtombs(3) updates the source
argument on conversion error

(christos)

add new wcsrtombs test

(christos)

PR/54414: Valery Ushakov: add a test for wcsrtombs(3) doesn't update the
source argument on conversion error

(christos)

Deal with partition types we don't know anything about.

(martin)

Explicitly link against libpthread and liblzma to make static linking
happy. Newer libarchive supports LZMA compression in zip files.

(joerg)

Avoid memset to framebuffer.

XXX
There still remains memset in rasops8.c, which will be removed soon
when putchar_aa functions are factorized.

(rin)

Cast attr to uint32_t before right shift to avoid undefined behavior.

Also, misc style/cosmetic changes for clarity.

(rin)

Fix infinite loop in dwc_mmc_exec_command error path

(jmcneill)

Use the non-WIDE_CHAR version of libcurses from libhack, and also build
the included nvi w/o wide character support.

(martin)

Use the non-WIDE_CHAR libcurses version from libhack

(martin)

Provide a way for crunched programs to use a non-WIDE_CHAR version of
libcurses by setting HACK_CURSES=yes before including this makefile.

(martin)

Free kmem_alloc'd memory with kmem_free

(martin)

We (especially libcurses and nvi) use more multibyte character locale
related symbols nowadays. Update libhack to avoid pulling in full grown
multibyte locale support on small install media.

(martin)

Simplify and style. No functional changes intended.

(rin)

Set max link speed, remove assigned clocks

(jmcneill)

PR/54415: Ricky Zhou: libedit stats completions for non-file completions
Use the proper completion function and account for the character appended
by the function when computing the number of columns.

(christos)

Avoid undefined behavior in uao_pagein_page(). Found by kUBSan. OK'd by
riastradh. I think this is a real bug on amd64 at least.

(msaitoh)

Redo fix in rev 1.2
http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/dev/rasops/rasops_putchar_width.h#rev1.2
in a correct way.

For 24bpp framebuffer, we need to use ``stamp triplets'' even for fully
blank or filled rows.

Now, background-color for white space and underline are correctly drawn.

(rin)

Fix factorization fallout; pointer for shadow framebuffer should be
rolled back before drawing underline.

(rin)

Fix corruption of background-colored white space and underline when
font width is 8 or 16 on 2bpp screen.

(rin)

Fix underline for anti-aliased fonts on 15/16 depth screen.

(rin)

sys/dev/rasops routines support anti-aliasing for depths 15 and 16.

(rin)

For wsvt25, stop disabling color attributes for underlined text.

With this change, underlined text is drawn in correct colors if
MD backend of wscons is underline-capable.

As the removed comment said, we render underlined text in red, if
MD backend of wscons does not support underline. However, this is
done regardless of original color attributes; foreground color is
forcibly changed to red. So, we don't need to disable original
colors for underlined.

(rin)

__slk_set_finalise - fix length vs. width confusion

(uwe)

__slk_set - fix calculating "len" for HAVE_WCHAR

(uwe)

Ensure that percpu_arena_update is really dead if it shouldn't work and
really alive otherwise.

(joerg)

Whitespace

(maya)

Add option -D to run in the background using daemon(3).
Also changed the message output to using syslog(3).

(nakayama)

Ticket #1308

(martin)

Pull up following revision(s) (requested by jakllsch in ticket #1308):

sys/dev/pci/if_wpi.c: revision 1.86

Put back correct recieve ring allocation size which was lost just over
five years ago mae culpa.

Found by msaitoh@

Should fix PR kern/54320 once pulled up.

(martin)

nit

(sevan)

No need to check cold to add I2C_F_POLL; the i2c API does it for us now.

(thorpej)

list the wide variant of functions.

(sevan)

Remove unused needs-flag

(skrll)

Fixed compile error.

(hkenken)

Fix builds properly

(skrll)

sort some #includes

(skrll)

Note update hickup build issue for uefi boot.

(martin)

Need opt_console for CONSADDR, etc

(skrll)

Rename a register define to make it more obvious it's watchdog only

(skrll)

Put slk_color() undef #ifdef HAVE_WCHAR too.

(uwe)

Put wide character functions under #ifdef HAVE_WCHAR

(uwe)

Add missing ``.'' at the ends of sentences, split long lines,
and misc style unification.

(rin)

doc/CHANGES.prev: s/mfi(8)/mfi(4)/g

(cnst)

doc/CHANGES.prev: address 7 definite time-travelling entries (year-based)

(cnst)

doc/CHANGES: fix OpenSSH 8.0 date; same for imcsmb(4) per consistency audit

(cnst)

TTTR

(sevan)

leot wants the ability to start wpa_supplicant from dhcpcd
without having to start wpa_supplicant seperately, thus the example
hook is restored.

(roy)

PR/54410: fmoon: typos in mail(1)

(christos)

Pre-allocate memory for the kernel space at startup.

(nonaka)

Build and install WSFB and WSFBSBC.

(rin)

Added tftp support to x86 efiboot.

(nonaka)

Ticket #1307

(martin)

Pull up following revision(s) (requested by christos in ticket #1307):

sys/netinet6/nd6.c: revision 1.256

Decrease the reference count before freeing, so that the entries actually
get free'd. (Ryota Ozaki)

(martin)

Put back byte-wise copy to stop using memcpy, which
does not work for device memory on some platforms.

Pointed out by Jared, many thanks!

(rin)

Even though the spec says SRST should read back as 1 after being set, this
does not appear to be the case with QEMU when using single byte accesses.

Instead of printing an error and giving up, continue and try to leave
stream reset state even if setting SRST times out.

(jmcneill)

As dhcpcd no longer supports IPv4 address advertisement for SMALL builds,
remove the equivalent IPv6 functionality.
This shouldn't be an issue as this is only used for IPv6 address sharing,
which only the NetBSD kernel currently supports.

(roy)

Note genfb(4) for mac68k.

(rin)

Sort SEE ALSO.

(wiz)

Build dhcpcd without ARP support for SMALL builds.
The only functionality lost on modern NetBSD is the ability to
announce the preferred address for address sharing.
This shouldn't be an issue for installations.

(roy)

Add genfb(4) driver for mac68k grfbus.
1, 2, 4, 8, 15, and 32 color-depths are supported.

ANSI colors on console are functional for depth >= 4.

Graphic applications based on wsfb API should work, provided
they can correctly handle fbi_fboffset and your color depth.

wsfb driver of Xorg 1.20 and mlterm-fb (framebuffer version of
x11/mlterm) work fine for depth == 1 || depth >= 8.

For depth == 8 (256 colors), graphic applications require
colormap callback, which is currently supported for some
internal graphic adapters, and only Civic (found on Quadra
AV series) was tested. Register definition and its usage are
taken from Linux. You can easily support other adapters if
Linux supports that.

Have fun, and any problem reports are welcomed!

(rin)

Allow dhcpcd to be built without ARP support for SMALL builds.
This is fine because the kernel supports RFC 5227 which dhcpcd uses
instead to detect and act on address duplication.

(roy)

Fix set name in previous.

(wiz)

Replace hook example 10-wpa_supplicant with 29-lookup-hostname.
Fixes PR install/54351.

(roy)

Stop installing 10-wpa_supplicant.
Only install 50-ypbind when MKYP=yes.

Partially addresses PR install/54351.

(roy)

Decrease the reference count before freeing, so that the entries actually
get free'd. (Ryota Ozaki)

(christos)

- AMD CPUID Fn8000_0001d Cache Topology Information leaf is almost the same as
  Intel Deterministic Cache Parameter Leaf(0x04), so make new
  cpu_dcp_cacheinfo() and share it.
- AMD's L2 and L3's cache descriptor's definition is the same, so use one
  common definition.
- KNF.

XXX Split some common functions to new identcpu_subr.c or use #ifdef _KERNEK
... #endif in identcpu.c to share from both kernel and cpuctl?

(msaitoh)

Use aprint_debug_dev() instead of printf() in wm_ulp_disable().

(msaitoh)

Strip trailing / from last mounted strings.
No idea how they happen, but for cgd root (init.root = "/altroot")
they have been reported to exist.

(martin)

The mother of all demos

(sevan)

Remove outdated note about sysinst restrictions

(martin)

comment out WSDISPLAY_COMPAT_RAWKBD for now, it does not compile

(martin)

KNF on #includes

(skrll)

Fix builds

(skrll)

Fix 4-bpp stamp for big endian.

(rin)

Calculate the multicast hash table correctly to avoid undefined behavior.
Found by KUBSan.

(msaitoh)

Set kcpuset's bit correctly to avoid undefined behavior. Found by KUBSan.

(msaitoh)

Set sc_mask correctly in selsysinit() to avoid undefined behavior.
Found by KUBSan.

(msaitoh)

Replace manually unrolled loops with memcpy/memmove or simple loops.
Modern compilers are smart enough; there's no measurable changes in
performance even on MC68040 with optimization level -Os.

Also, convert loop of byte-wise copy into memset.

(rin)

Misc creen up for rasops.c:

- sort headers
- return error value instead of panic
- use uintptr_t for cast pointer to integer
- use macros appropriately
- use __func__
- some consistency check ifndef DEBUG
- try to avoid undefined behaviors related to shift
- convert malloc/free to kmem_alloc/kmem_free
- convert MIN to uimin
- style

(rin)

Set IVAR (Interrupt Vector Allocation Register) correctly for 82598 in
ixgbe_set_ivar(). Found by KUBSan.

(msaitoh)

Simplify code. No functional change intended.

(msaitoh)

Simplify and avoid undefined behavior. This is not a real bug because
the max number of queue is 2.

(msaitoh)

XXX
Add hack so that colorbars(6) works for 24-bpp.
The code before factorization did the same in a strange way.

(rin)

More Whitespace

(skrll)

Whitespace

(skrll)

Halt the old console on both cnattach and preattach

(jmcneill)

#ifdef HAVE_WCHAR around enhanced curses stuff.

(uwe)

Remove a duplicate psci* at fdt?

(skrll)

Gracefully get out of extended partitioning if the user did not create
any partitions at all.

(martin)

Do not free() an array embedded in another structure

(martin)

Partially undo previous so NETWALKER_INSTALL compiles

(skrll)

Factor out putchar and width-optimized putchar functions into
rasops_putchar.h and rasops_putchar_width.h, respectively.

XXX
Possibly, we can do the same for putchar_aa functions. But it is
currently missing for 24-bpp.

(rin)

Add width-optimized putchar functions for 32-bpp, that work fine on
both little- and big-endian machines.

WIP genfb(4) driver for mac68k becomes 1.5 times faster!

(rin)

Make register address unsigned to avoid undefined behavior. Found by KUBSan.

(msaitoh)

Avoid undefined behavior. Found by KUBSan.

(msaitoh)

Fix a few calloc() calls with wrong arg order (harmless, but for consistency)

(martin)

When re-discovering a disk where we have free the partitions, read
the current partitions again (they might have changed).

(martin)

Avoid undefined behavior. Found by KUBSan. Some of errors are still remaining.

(msaitoh)

Fix arm build

(skrll)

Use unsigned (RAL_RF_BUSY). Found by KUBSan.

(msaitoh)

Fix VLAN hardware filter initialization. Use unsigned. Found by KUBSan.

(msaitoh)

Fix VLAN filter setting. Found by KUBSan.

(msaitoh)

Note import of dhcpcd-8.0.1

(roy)

Ticket #1306

(martin)

Revert prior as dhcpcd-8.0.1 now builds small without IPv4LL again.

(roy)

Pull up following revision(s) (requested by ozaki-r in ticket #1306):

crypto/dist/ipsec-tools/src/setkey/parse.y: revision 1.23
sys/netipsec/key.c: revision 1.265
crypto/dist/ipsec-tools/src/setkey/token.l: revision 1.23
tests/net/ipsec/t_ipsec_misc.sh: revision 1.23

ipsec: fix a regression of the update API

The update API updates an SA by creating a new SA and removing an existing SA.
The previous change removed a newly added SA wrongly if an existing SA had been
created by the getspi API.

setkey: enable to use the getspi API

If a specified SPI is not zero, tell the kernel to use the SPI by using
SADB_EXT_SPIRANGE.  Otherwise, the kernel picks a random SPI.

It enables to mimic racoon.

tests: add tests for getspi and udpate

(martin)

Sync

(roy)