One more step towards fixing some builds.  Like my last attempt
this one also needs redoing properly...

(kre)

Register micro-UBSan in CHANGES.

There are 3 entries:
- Import of the clean-room reimplementation of the runtime.
- uUBSan (user-UBSan) integration with MKLIBCSANITIZER.
- kUBSan (kernel-UBSan) integration as KUBSAN.

(kamil)

Update TODO.sanitizers

Remove the following entries:
- kernel-ubsan is done and merged with src/.
- the no-ASRL boot.cfg option has been rejected.

(kamil)

Update TODO.ptrace

Drop the following entries:
- PaX MPROTECT tests - done, implemented.
- research of ipkdb(4) - it's gone, if it will be superseded it would be
  probably a networking interface for kgdb
- PTRACE_VFORK implemented and covered by tests, there are SMP bugs left

(kamil)

Build fix after recent KERN_{VTOPHYS,PHYSTOV} change

(skrll)

More whitespace

(skrll)

Refactor code to split aarch{32,64} kernel page tables and VM setup.  This
will help re-build the kernel page tables on aarch64 with correct section
mappings.

(skrll)

Expose kern_vtopdiff

(skrll)

Fix KERN_{VTOPHYS,PHYSTOV) fallout differently

(skrll)

Don't use $NetBSD$ in a sentence, refer to them as 'RCS identifiers'.

(alnsn)

README.import with step-by-step import instruction.

(alnsn)

Resolve conflicts.

(alnsn)

Fix KERN_VTOPHYS build fallout

(skrll)

Hack workaround to deal with KERN_VTOPHYS and KERN_PHYSTOV now
being defined in arm/arm32/machdep.h ... attempt to fix (some) earm
builds.  Even if builds succeed, resulting kernel might not boot,
and if it boots, could crash.  Someone with a clue, please use it!

(kre)

need to set _UC_TLSBASE of uc_flags for _lwp_makecontext().
pointed out from skrll@. thanks.

(ryo)

State Save:
  Scanning works, can see the broadcasts and stations.
  Association appears to work.  (open networks only, no encryption yet.)
  IP traffic doesn't flow yet, not sure where the error is.
  Added back more NetBSD ioctls not in the FreeBSD code.  Still have
      more to go.
  Updated my TESTWIFI kernel config.

(phil)

don't set lwp->l_private if no _UC_TLSBASE flag.
atf lib/libc/sys/t_swapcontext Passed.

(ryo)

set kernel text/rodata readonly when not defined DDB.
set readonly segment on 2Mbytes aligned. (kernel image is mapped with 2Mbytes L2 block)

(ryo)

Try to fix the evbppc-powerpc64 build

Avoid "comparison between signed and unsigned integer expressions" on
Big-Endian hosts.

(kamil)

Provide and use kern_vtopdiff in KERN_{VTOPHYS,PHYSTOV}.  A step towards
generic arm.

(skrll)

Subtract KERNEL_BASE_VOFFSET from all VAs to get PA.  How did this work
before?

(skrll)

Restrict -fno-sanitize=function to Clang/LLVM only

The base GCC version 6,x does not support this option.

(kamil)

Provide and use a evbarm/fdt/machdep.h

(skrll)

Fix build

(skrll)

Add a "gcc is stupid" comment to the previous change, as even
the most cursory analysis shows that the var ("eg") is not (cannot
be) used unitialialised, just gcc is too dumb to work it out.

In this case, the code could be rewritten easily enough to
appease even gcc, but that would cause unnecessary code churn,
and some minor duplication, so just put up with the nonsense init...

(kre)

s/DPRINTF/VPRINTF/ for consistency

(skrll)

Use a different psz for a different lock. Patch from riastradh, reviewed
by ozaki-r.

(jmcneill)

Allow for setting kernel breakpoints in our remote kgdb

(reinoud)

tun: fix locking against myself

filt_tunread is called with tun_lock held from tun_output (via tun_output =>
selnotify => knote), so we must not take tun_lock in filt_tunread.  The bug
is triggered only if a tun is used through kqueue.

Found by k-goda@IIJ

(ozaki-r)

Mark fdt_addr_r __attribute__((__section__(".data")))

(skrll)

Fix physical memory size issue. It will now happily run on more than one gb.

(reinoud)

Fix distribution lists with MKCATPAGES=yes

Add SCTP documentation.

(kamil)

Add some missing IFM_* entries.

(msaitoh)

Register kUBSan in the GENERIC amd64 kernel config

Tested with GCC.

(kamil)

Appease GCC in the openssh code when built with UBSan

Initialize eg to NULL in sshkey_ecdsa_key_to_nid().
The compiler warns that it might be uninitialized.

(kamil)

Add a support to build ubsan.c in libc

Under the condition of MKLIBCSANITIZER==yes link ubsan.c into libc.

This is a clean-room reimplementation from scratch of the Undefined
Behavior runtime called micro-UBSan (or uBSan - user-UBSan).

(kamil)

Register micro-UBSan ATF tests in the distribution

Populate distrib files, mtree lists and add the entry in Makefile to
include the new code.

(kamil)

Import micro-UBSan ATF tests

These tests are used only when a distribution is built without MKSANITIZER
and without MKLIBCSANITIZER. They build the ubsan.c micro-implementation
in userland as a standalone library (a .c file linked into tests).

The code generation part emits symbols that are handled by uUBSan.

Tested with Clang amd64 + i386 and GCC amd64 + i386.

There are two sets of tests: C code generation and C++ code generation.

There are various little differences that are handled with proper ifdefs.

Clang 7svn has significantly more checks for Undefined Behavior than GCC
5.x. Some of the tests are or might be amd64 and/or i386 specific and are
marked with proper ifdefs and comments according to the compiler
documentation.

(kamil)

Register a new directory in common/lib/libc/misc

Registe misc/ with ubsan.c.

(kamil)

Tidy up the comment in ubsan.c

As noted, style has no impact on the comparison of a similar code.
This version is a reimplementation from scratch with no code and no
algirithm (whenever possible) reuse.

Public symbols and struct layout must be kept in sync with the code
generation part. It casues violation of the style like with long filenames.

My previous comment was 'too perfect' and could trigger some unnecessary
attention.

(kamil)

Allow to overwrite sanitizer flags for userland

Always define -fsanitize=${USE_SANITIZER} and
-fsanitize=${USE_LIBCSANITIZER} before other flags defined with
SANITIZERFLAGS and LIBCSANITIZERFLAGS.

This allows to overwrite rules that are parsed from left to right.
With the current approach it wasn't possible to disable some checks, as
they were reenabled again by the "-fsanitize=" argument.

(kamil)

Disable sanitization of -fsanitize=function in libc

Generated code for the checks (in at least libunwind) depends on the C++
runtime feature of RTTI.

Pass -fno-sanitize=function to LIBCSANITIZERFLAGS.

(kamil)

Fix xf86-video-wsfb build with Clang/LLVM

The code uses non-orthodox types in format strings.
Add a compiler flag to disable a warning that is turned into a fatal error.

(kamil)

Fix libXfont2 build with Clang/LLVM on i386

The code uses non-orthodox types in format strings.
Add a compiler flag to disable a warning that is turned into a fatal error.

(kamil)

Fix libGLU build with Clang/LLVM

The register keyword is deprecated in newer C++ standards.
Add a compiler flag to disable a warning that is turned into a fatal error.

(kamil)

Fix dri7 build with Clang/LLVM

The register keyword is deprecated in newer C++ standards.
Add a compiler flag to disable a warning that is turned into a fatal error.

(kamil)

Fix build failure in dhcpcd under uUBSan

GCC with -fsanitize=undefiend detects a potential overflow in the code.
Cast the return value of ntohs(3) to size_t.

Fix merged upstream by <roy>.

(kamil)

Import micro-UBSan (ubsan.c)

This is a reimplementation of the Undefined Behavior Sanitizer with the
following properties:
- pure C implementation,
- no -fsanitize=vpts support, as it requires RTTI support and C++
  low-level routies to validate whether C++ objects are compatible
- designed to be used inside libc and known as uUBSan or user-UBSan
- designed to be shared with kernel and known as kUBSan or kernel-UBSan
- designed to be usable with ATF tests as a standalone runtime,
  reachable without any MK* switches
- designed to be safer for hardening as it does not have side effects on
  executables like writing to a selected location on demand
- controllable with environment variable LIBC_UBSAN with options:
  * a - abort on report
  * A - do not abort on a report (unless a failure is unrecoverable)
  * e - output report to stderr
  * E - do not output report on stderr
  * l - output report on syslog (LOG_DEBUG | LOG_USER)
  * L - do not output report on syslog
  * o - output report on stdout
  * O - do not output report on stdout
  The default options are: "AeLO".
- compatible with Clang (3.8, 7.x) and GCC (6.x) code generation
- all handlers (except =vptr) from Clang/LLVM up to 7svn are supported

This file does not follow the regular KNF style, due to potential licensing
concerns.

Tested with Clang amd64+i386 and GCC amd64+i386.

(kamil)

Mark two entries as done.

(maxv)

Add a "version" field in the prekern_args structure. The kernel checks it,
and if it's not happy it returns back to the prekern.

(maxv)

Don't forget to call init_slotspace when we're booted via the prekern.

(maxv)

Use Nm macro
Move history details to the HISTORY section
Spelling
Be less meta & clarify the website should be consulted.

(sevan)

Distribute GENERIC_KASLR on amd64.

(maxv)

Remove netbsd-INSTALL_XEN3_DOMU.gz (it doesn't exist anymore), and add
netbsd-XEN3PAE_DOM0.gz (has always existed, but was apparently forgotten).

(maxv)

Add missing SCTP manpage.

(rjs)

Add userland support for SCTP and manpages.

(rjs)

fix the sanitizer include path.

(mrg)

Fix usb_rem_task_wait API.

- Return whether it removed task from queue or not.
  . True if it was on the queue and we intercepted it before it ran.
  . False if we could not intercept it: either it wasn't queued,
    or it already ran.  (Up to caller to distinguish these cases.)
- Pass an optional interlock like callout_halt.

While here, simplify.

ok mrg@

(riastradh)

Clean up from sync-with-HEAD

(pgoyette)

Adjust alignment in m_pulldown().

IP6_EXTHDR_GET() and M_REGION_GET() do m_pulldown(). When m_pulldown() copies
data into M_TRAILINGSPACE, the alignment might be changed. There are a lot of
IP6_EXTHDR_GET() calls, so I think it's not good to check the alignment after
every IP6_EXTHDR_GET() call. This change fixes this problem in m_pulldown().
In this commit, the next mbuf are 4 byte aligned. For networking, I've never
heard that 64bit alignment is required, so I think it would be OK.

I don't know this is the best solution, but it's better than nothing.

OK'd by maxv@.

After committing this change, the workaround code for PR#50776 can be removed.

(msaitoh)

Mark _error() as printflikee

(roy)

Drop ISO/Xerox references which are now long gone.

(sevan)

merge GCC 6.4 into gcc.old.

(mrg)

Add ioctl(2) handler for kernel part of sctp_peeloff().

(rjs)

Add a comment to explain the purpose of POWER_IOC_GET_TYPE_WITH_LOSSAGE
and also define it only for the kernel, userland should never see that.

(kre)

Unreference IPF/PF from all the config files, and enable NPF instead when
wanted. This also fixes some inconsistencies I saw in several files (eg
IPF options while IPF was not compiled, IPF+PF enabled by default, etc).

(maxv)

fix long double NaN definition.

(ryo)

Build the umb(4) driver in the ALL kernels (amd64, i386)

As suggested by Robert Swindells; thank you!

(khorben)

Fix building umb(4) on NetBSD-current

Patch by Robert Swindells; thank you!

(khorben)

Move RCS Id to top. Fix date.

(wiz)

Various improvements to the man page.

(wiz)

strnlen(s, (size_t)-1) returned -1. it must return the length of s.

(ryo)

Unreference IPF/PF from the x86 config files (amd64, i386, xen), and enable
NPF instead when wanted.

(maxv)

Can't see a thing (which see)
Remove CLNP reference which is long gone.

(sevan)

Fix the alignment argument to usb_allocmem in dwc2_hc_setup_align_buf

(skrll)

s/_ARM32_BOOT_MACHDEP_H_/_ARM32_MACHDEP_H_/ in multiple inclusion
protection.

(skrll)

remove -DGHASH_ASM_X86; it is already defined.

(christos)

Some whitespace improvements.  NFC.

(skrll)

Note the removal of non-PAE-32bit-PV.

(maxv)

Xen is PAE, so remove ifdefs.

(maxv)

Add a bold note to say our PF is obsolete.

(maxv)

Allow kmem_alloc(9) to sleep when attaching

Without this, umb(4) may needlessly fail to attach, like when under memory
pressure.

Thanks skrll@ for the heads-up!

(khorben)

Avoid parentheses in return statements (KNF)

NFCI.

Thanks skrll@ for the heads-up!

(khorben)

Revert to working state

(reinoud)

Add missing defines:
https://github.com/openssl/openssl/pull/6828
When ghash-x86.S is generated with -DOPENSSL_IA32_SSE2 we need to compile
gcm128.c with the same flags.
Reported by manu@

(christos)

Remove yet another debug printf()

(reinoud)

Fix too long line

(reinoud)

Remove debugging printf()

(reinoud)

Add the kgdb meat

(reinoud)

Forgot the two header files

(reinoud)

Add preliminary KGDB support for NetBSD/usermode, currently only under amd64

(reinoud)

Add headers for support functions for  kgdb

(reinoud)

Max kernel address is end of kernel

(reinoud)

Oops, forgot a debug printf

(reinoud)

Allow the usermode pmaps to also contain the kernel itself. This is needed for
the kernel pmap.

(reinoud)

Note that mss clamping should be improved

(sborrill)

As described in comments, we treat non-spacing chars as single-width ones.

(rin)

Drop ISO references as it's no longer supported.

(sevan)

debug versions of libraries need minor version bump to match
the non-debug versions (bump for getsockopt2().)

Arguably this dependency ought to be handled by automation rather than
relying upon developers "just knowing" when the debug sets need updating.

(kre)

Remove ISO references

(sevan)

Remove references to XNS

(sevan)

regen.

(rjs)

Add getsockopt2().

(rjs)

Remove ISO & XNS references which are now long gone.

(sevan)

Suggest command line variables.

(darcy)

Add new item, about deprecating IPF/PF; I thought it was there, but
realized it wasn't.

(maxv)

Add man page for sctp(4).

(rjs)

Tickets #951, #952, #954, #955, #957

(martin)

Pull up following revision(s) (requested by nonaka in ticket #957):

usr.sbin/arp/arp.c: revision 1.63

arp(8): Fix the host output when arp -dav.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #955):

sys/dev/ppbus/if_plip.c: revision 1.29

KNF & fix typo. No functional change.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #954):

sys/sys/socketvar.h: revision 1.157
share/man/man4/ddb.4: revision 1.180
share/man/man4/ddb.4: revision 1.181
sys/kern/uipc_socket2.c: revision 1.131
sys/ddb/db_command.c: revision 1.154

Add "show socket" command written by Hiroki SUENAGA. It prints usage of
system's socket buffers.
Improve wording.

(martin)

Add a port of the umb(4) driver from OpenBSD

The umb(4) driver provides support for USB MBIM (Mobile Broadband
Interface Model) devices.

MBIM devices establish connections via cellular networks such as GPRS,
UMTS, and LTE. They appear as a regular point-to-point network interface, transporting raw IP frames.

Required configuration parameters like PIN and APN have to be set with
umbctl(8), a new tool specific to this driver. The IP address is configured
automatically; the default route and DNS server information have to be set
separately.

The driver is not fully functional yet, it is therefore still marked as
experimental and disabled by default. Any help welcome to complete it!

Tested on NetBSD/amd64, with a Sierra Wireless EM7345 LTE modem on a Lenovo
ThinkPad T440s. No functional change expected otherwise.

(khorben)

Enable SCTP sysctl nodes.

Rename auto asconf one to match FreeBSD.

(rjs)

Pull up following revision(s) (requested by msaitoh in ticket #952):

sys/dev/ic/malo.c: revision 1.11

bpf_mtap*() after ieee80211_encap() should be bpf_mtap3() rather than
bpf_mtap(). bpf_mtap3() is for raw bpf.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #951):

sys/dev/ic/rt2661.c: revision 1.38

bpf_mtap*() before ieee80211_encap() should be bpf_mtap() rather than
bpf_mtap3(). bpf_mtap3() is for raw bpf and be used after ieee80211_encap().

(martin)

Tickets #945 - #948

(martin)

Re-generate

(khorben)

Add the Sierra Wireless MC7304 and EM7455 USB modems

(khorben)

Pull up following revision(s) (requested by msaitoh in ticket #948):

sys/arch/sun3/dev/if_ie.c: revision 1.64

Remove bpf_tap() in iexmit() because ie_start() does bpf_mtap().

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #947):

sys/dev/pci/if_de.c: revision 1.153

Remove bpf_tap() call in rx. It's done in if_percpuq_softint().

(martin)

Pull up following revision(s) (requested by kre in ticket #946):

sbin/cgdconfig/cgdconfig.c: revision 1.42
sbin/cgdconfig/cgdconfig.c: revision 1.43
sbin/cgdconfig/cgdconfig.c: revision 1.44
sbin/cgdconfig/cgdconfig.c: revision 1.45
sbin/cgdconfig/cgdconfig.8: revision 1.37
sbin/cgdconfig/cgdconfig.8: revision 1.38

Check whether the cgd device selected is available to be
configured,that is, not already in use, before requesting
passwords from the user (or elsewhere).

Fix usage for rump.  Fixes test breakage caused by previous commit.
Check arg count in configure() at entry, rather than later.

This avoids the stupid null deref I added a couple of commits
ago (on bad usage) and also simplifies the rest of the routine
which no longer needs to check the arg count nearly as much.
Thanks to Alexander Nasonov for finding the null deref bug.

Fix missing -p in usage message (noted by Christoph Badura, thanks),
and update -l usage as well.
While here fix man page synopses and a few other odds and ends.
Add commas in enumeration.

(martin)

Pull up following revision(s) (requested by taca in ticket #945):

sys/arch/i386/i386/trap.c: revision 1.290

Don't use a const variable as an array size, because it is not a
constant expression and will cause the array to needlessly become a
variable-size one.  Fixes "error: stack protector not protecting local
variables: variable length buffer [-Werror=stack-protector]" when
building i386 with MKDEBUG=YES -V COPTS="-g".

(martin)

Change implementation of sctp_connectx() to use ioctl(2).

(rjs)

Add declaration of getsockopt2().

(rjs)

Bump version for addition of getsockopt2() syscall.

Welcome to NetBSD 8.99.23 !

(rjs)

Userland changes for getsockopt2().

(rjs)

regen.

(rjs)

Add getsockopt2() syscall.

(rjs)

#ifdef VERBOSE_INIT_ARM some format strings only used in that case.

(martin)

arp(8): Fix the host output when arp -dav.

(nonaka)

Initialize z_lockf for new znodes.

Ok: Chuck Silvers

(hannken)

Make jcl allocation per queue to reduce mutex spin. Tested by me and
knakahara.

(msaitoh)

Define and use VPRINTF

(skrll)

Sprinkle #include "opt_arm_debug.h" where VERBOSE_INIT_ARM is used

(skrll)

Avoid an infinite loop caused by a line accidentally dropped
in 1.22 (almost 30 months ago!)

[ Discovered by rhialto@ ]

(kre)

don't depend on clang code to backtrace. keep trapframe as framepointer if DDB.

(ryo)

fix copy{in,out}str to return ENAMETOOLONG if the string is longer than len bytes.

(ryo)

Add machine commands for aarch64.

(ryo)

Regen.

(msaitoh)

Add Toshiba's XG4 and XG5 NVMe SSD.

(msaitoh)

Regen.

(msaitoh)

Add some Intel devices. Some of them are found in Core i7-8550U.

(msaitoh)

Skip urtwn_chip_stop for 88E family chips as we do for 92EU.

Doing "ifconfig urtwn0 down" with an RTL8188EU on Pinebook causes the
device to disappear from the built-in GL850G USB 2.0 hub. Since this
happens at reboot, an unplug/replug cycle is the only way to make the
device re-appear. Bypassing urtwn_chip_stop works around this issue.

(jmcneill)

Remove files imported to a wrong directory.

(alnsn)

Fix mandoc warnings.

(wiz)

regen

(sevan)

Spelling mistake.
Reword MKINFO description.

(sevan)

Document "in" field in CPU section
Fix ps(1) Xr raised by mandoc -Tlint

(sevan)

mark satapmp_rescan() static, it's not to be used outside satapmp_subr.c

(jdolecek)

Reduce the confusion, rename a bunch of variables and reorg a little.
Tested on i386PAE-domU and amd64-dom0.

(maxv)