remove now dated comment.

(mrg)

turn on MKCOMPAT for riscv64.

(mrg)

Zero out the ifreq struct for SIOCGIFCONF to avoid up to 127 bytes of stack
disclosure. From Andy Nguyen, many thanks! This is the compat code part
pointed out by ozaki-r@

(christos)

Follow arm64 and provide two sizes of fonts, so a larger one is chosen
for larger displays.

(maya)

Yes, we support FFS_EI but not DISKLABEL_EI... How does that make sense?

(christos)

no tls for sun2 either.

(christos)

all we are saying is, give lint a chance

(christos)

Fix the bell period.

(isaki)

White space and indent fix.

(isaki)

Note restrictions on return values from tzgetname() and tzgmtoff()

(pgoyette)

We use pthread functions, link against pthread.

(maya)

if HAVE_XORG_GLAMOR != no, set define PLATFORM_HAS_EGL=1.

(mrg)

on NetBSD, don't define PLATFORM_HAS_EGL if it is already defined

(mrg)

- use the same set of emulations for riscv* on netbsd as linux does
- for riscv64-*netbsd* hosts and 32 bit target, set LIB_PATH
- build the other target emulations for riscv32 on riscv64
- regen riscv64 mknative for ld updates

(mrg)

- various updates for GCC 7 riscv definitions: -mabi and -march changed
- add missing LIBGCC_MACHINE_ARCH
- add the new ldscripts for riscv64 MKCOMPAT.

(mrg)

- undef STARTFILE_PREFIX_SPEC, it is wrong for netbsd
- make LINK_SPEC more like other multi-line string defines
- update multilib stuff for GCC 7 riscv -- -m32/-m64 gone, -mabi and
  -march now decide between several options

(mrg)

rump: add missing sanity checks at the end of syscalls

(ozaki-r)

tests: dump kernel stats on cleanup

(ozaki-r)

tests: make utility funtions easy to use for tests that don't use the framework

(ozaki-r)

delete files no longer present in GCC 7 port of riscv.

(mrg)

Add missing ./var/db/dhcpcd

(rin)

OpenSSH 8.0 is out

(sevan)

Note import of dhcpcd-7.2.0

(roy)

Remove non distributed file.

(roy)

Sync

(roy)

L2CAP

(sevan)

Get a large enough buffer for the LIF DIR

(skrll)

Summary: Ensure default TCP flags are applied to rules like 'pass stateful all'

CVS: ----------------------------------------------------------------------
CVS: CVSROOT  cvs.NetBSD.org:/cvsroot
CVS: please use "PR category/123" to have the commitmsg appended to PR 123
CVS:
CVS: Please evaluate your changes and consider the following.
CVS: Abort checkin if you answer no.
CVS: => For all changes:
CVS: Do the changed files compile?
CVS: Has the change been tested?
CVS: => If you are not completely familiar with the changed components:
CVS: Has the change been posted for review?
CVS: Have you allowed enough time for feedback?
CVS: => If the change is major:
CVS: => If the change adds files to, or removes files from $DESTDIR:
CVS: => If you are changing a library or kernel interface:
CVS: Have you successfully run "./build.sh release"?

The documented default "flags S/SAFR" for stateful rules that affect
TCP packets but don't specify any flags, doesn't actually get applied
to a rule like "pass stateful out all". The big problem with this is
that when you then do a "block return-rst" for an incoming packet, the
generated RST packet will create state for the connection attempt it's
blocking, so that a second attempt from the same source will pass.

This change makes the default flags actually apply to such simple
rules.  It also fixes a related bug in the code generation for the
flag matching, where part of the action could erroneously be omitted.

Reviewed by <rmind>
Closes PR bin/54124
Pullup to NetBSD 8

(tih)

NVMM

(sevan)

Pick up the latest matching (most recent) entry instead of the first one.

This fixes:
env TZ=Australia/Melbourne date
printing
Thu Apr 18 03:32:32 LMT 2019
instead of:
Thu Apr 18 03:32:32 AEST 2019

XXX: To be strictly correct, the tzgetname() call should probably take a time
reference point in order to return the appropriate zone name for the time
given.

(christos)

BCH DSF HFS ISL LIF ODE VXFS VXVM

(sevan)

Ticket #1231

(martin)

Pull up following revision(s) (requested by nonaka in ticket #1231):

sys/arch/i386/stand/efiboot/efidisk.c: revision 1.7

efiboot: Don't panic when BLOCK_IO_PROTOCOL is not found.

It has been reported that there is a machine where BLOCK_IO_PROTOCOL
can not be found when network boot without disk.

(martin)

Fix Clang/LLVM build in t_ptrace_wait.c

Do not compile functions that might be unused.

(kamil)

Make the illegal instruction macro safe, so the compiler does not fear it
would be an illegal instruction when compiling with certain cpu options.

(martin)

LZSS NC-SI NCSI ODM VPD

(sevan)

fix typo

(maya)

fix for riscv32.

(mrg)

- define __riscv__ again
- define *intptr_t in terms of 'long int' always
- move fixtfdi.c, fixunstfdi.c, floatditf.c, and floatunditf.c into t-netbsd64
- remove duplicated riscv case in config.gcc, reducing upstream diff as well
- regen riscv64 mknative

(mrg)

mknative-gcc and mknative-binutils for riscv32.  mostly builds.

(mrg)

KNF. No functional change.

(msaitoh)

fake mknative, mostly enough to build real for mknative.

(mrg)

fix MKCOMPAT=no builds.

(mrg)

Remove unused inclusion.

(msaitoh)

also want s_nexttoward.c on riscv.  now tests/lib/libm links.

(mrg)

add missing atomic_and_16_nv_cas.c atomic_and_8_nv_cas.c.

(mrg)

add missing aliases for _atomic_cas_32_ni and atomic_cas_uint_ni.

(mrg)

Tabify. No functional change.

(msaitoh)

efiboot: Don't panic when BLOCK_IO_PROTOCOL is not found.

It has been reported that there is a machine where BLOCK_IO_PROTOCOL
can not be found when network boot without disk.

(nonaka)

Don't check pserialize_not_in_read_section after panic

It can cause a false positive assertion failure.

(ozaki-r)

Bogus number police

(sevan)

- enable EGL on x86/arm64 (XXX also do evbarm32)
- enable gbm/glamor in the server.

(mrg)

enable EGL and glamor.  partly from maya@.

for x86 and evbarm:
- install headers and libEGL
- install xorg-server glamoregl component
- link xorg drivers against gbm/egl

bonus fix:
- fix some wrongly marked compatx11file files

build tested on shark, sgimips, evbarm64-el, amd64 and i386.
run tested on radeon, intel and nvidia on amd64, including
'modesetting' driver on amd64.  however, my systems disable
it due to llvmpipe so i'm not sure what is happening.

(mrg)

enable EGL and glamor.  partly from maya@.

for x86 and evbarm:
- install headers and libEGL
- install xorg-server glamoregl component
- link xorg drivers against gbm/egl

bonus fix:
- fix some wrongly marked compatx11file files

build tested on shark, sgimips, evbarm64-el, amd64 and i386.
run tested on radeon, intel and nvidia on amd64, including
'modesetting' driver on amd64.  however, my systems disable
it due to llvmpipe so i'm not sure what is happening.

(mrg)

If arg3 is zero then fill it with the value of 'end'.  Helps booting
with qemu.

(skrll)

s/evbarm64/evbarm

(sevan)

Tag evbarm related changes separately to separate out the entries into the
specific section on the generated webpage.
Others the entries show up in the section of each arm based port.

(sevan)

typo - via leot

(sevan)

Add manual categories to generate links to man.n.o on the webpage

(sevan)

Ran out of patience playing stack protector bingo.

(christos)

one tab to start an item, and two tabs to indent the content

(sevan)

Trailing whitespace

(skrll)

Trailing whitespace

(skrll)

Add PTRACE_ILLEGAL_ASM

(martin)

Indent to improve readability.
Add a description for log event.

(sevan)

fix various problems i've seen where cv_*wait*() return ERESTART,
which is -3 in netbsd, which we have mapped linux ERESTARTSYS to.

this has a problem because linux code often returns errors and
pointers in the same value, and pointer values between -4095 and
-1 are considered as error returns, but -3 ends up as 3 and thus
is not considered an error, and mayhem ensues.

with this in place my kabylake system seems actually stable, i
have not triggered any of my prior issues in almost 4 weeks now.

Taylor asked me to write up a description and then wrote most of
the text below for me :-)

In Linux code, we always work with ERESTARTSYS so the code meaning
start over is a positive NetBSD errno safe for PTR_ERR/ERR_PTR.
To achieve this:
1. adapt all cv_waits that return to Linux so they map ERESTART to
  ERESTARTSYS, and
2. adapt all returns to userland so they convert ERESTARTSYS to
  ERESTART.
Leave EINTR and all other error codes alone.

(mrg)

Ticket #1673

(martin)

Pull up following revision(s) (requested by dholland in ticket #1673):

sys/external/bsd/drm2/pci/drm_pci.c: revision 1.32 (via patch)

Don't try to interpret the second half of a 64-bit BAR as another one.
From msaitoh@.

(martin)

fix -I paths for powerpc*.  convert amd64 specific hack.

we've been building dp-bit.c and fp-bit.c backwards, but as they
use the same input source, they just had opposite objects and
the right total was provided.  oops.

(mrg)

revert previously. we want riscv32 to have 128 bit long double...
shall fix the compiler instead.

(mrg)

mknative-gcc for riscv64.

(mrg)

list a bunch of soft-fp as present in the current risc64 libgcc defs.mk.
this makes mknative-gcc work without breaking maya's manual work.

(mrg)

RISC-V ELF psABI says ILP32 also defaults to 128bit long double.

(maya)

regen mknative-gcc for GCC 7 and ppc64.  libgcc still has problems.

(mrg)

avoid differences to linux64.h that aren't needed.
merge the rs6000_abi_name check for netbsd into the linux/freebsd section.

(mrg)

no clone() on netbsd.

(mrg)

risc32 doesn't want long double specific versions

(mrg)

We're now using gcc netbsd-stdint.h instead of our own definitions, so
match those with the format types

XXX wrong for 32bit.
XXX unclear if changing the fast types was the right call

(maya)

Fix a comment

(skrll)

It's not required (and can't do) to convert OSIOCSIFMEDIA in ifmedia_ioct()
because the conversiosn is done in doifioctl().

(msaitoh)

Re-arrange dino_softc members to improve alignment

(skrll)

dino depends on gedoens

(skrll)

regen for mknative-binutils for riscv64.

(mrg)

Whitespace

(skrll)

Rename ifreqo2n() and ifreqo2n() to IFREQO2N_43() and IFREQN2O_43():
- ifreqo2n() and ifreqn2o() are for COMPAT_43, so add _43 to the name.
- Uppercase to make it clear those are macros.

(msaitoh)

The path of SOICSIFMEDA or TAPGIFNAME calls are as follows:

doifioctl()

    pre-convert (if_cvtcmd_43_hook & ifreqo2n)

    (*ifp->if_ioctl)(ifp, cmd, data);

    post-convert (ifreqn2o)

so it's not required to check OSIOCSIFMEDIA and OTAPGIFNAME in if_tap.c.
Those two command is converted to new command in if_cvtcmd_43_hook and
always new commands are seen in tap_ioctl().

OK'd by pgoyette.

(msaitoh)

Ticket #1688.

(msaitoh)

Ticket #1688.

(msaitoh)

Ticket #1688.

(msaitoh)

Pull up following revision(s) (requested by martin in ticket #1688):
sys/kern/sys_mqueue.c: revision 1.44
mq_send1: fix argument validation and reject too large lengths early.
Discovered by Andy Nguyen.

(msaitoh)

Pull up following revision(s) (requested by martin in ticket #1688):
sys/kern/sys_mqueue.c: revision 1.44
mq_send1: fix argument validation and reject too large lengths early.
Discovered by Andy Nguyen.

(msaitoh)

Pull up following revision(s) (requested by martin in ticket #1688):
sys/kern/sys_mqueue.c: revision 1.44
mq_send1: fix argument validation and reject too large lengths early.
Discovered by Andy Nguyen.

(msaitoh)

ticket #1230

(msaitoh)

Pull up following revision(s) (requested by martin in ticket #1230):
sys/kern/sys_mqueue.c: revision 1.44
mq_send1: fix argument validation and reject too large lengths early.
Discovered by Andy Nguyen.

(msaitoh)

fix merge botch - remove duplicated case statement (not in configure)

(mrg)

Sync gcc.old GetPcSpBp() with GCC9 (and recent LLVM)

Unify all NetBSD ports in a single ifdef.

(kamil)

Sync GetPcSpBp() with GCC9 (and recent LLVM)

Unify all NetBSD ports in a single ifdef.

(kamil)

mq_send1: fix argument validation and reject too large lengths early.
Discovered by Andy Nguyen.

(martin)

Re-arrange the ufetchstore tests to look like the other ones.

(christos)

spelling

(sevan)

npf_boot and /etc/defaults/npf.boot.conf

(sevan)

Provide a simpler config for a host which permits any traffic from the host out,
and small subset of traffic in (DHCP (v4 and v6), All ICMPv6, ICMP echo
requests, traceroute, mDNS).

(sevan)

cpuctl.8: update the URL for AMD microcode, the previous site listed is
apparently defunct.

(gutteridge)

Remove unused definition.

(maya)

Prevent a NULL pointer dereference when the local endpoint is not defined.
From Andy Nguyen, many thanks.

(christos)

Zero out the ifreq struct for SIOCGIFCONF to avoid up to 127 bytes of stack
disclosure. From Andy Nguyen, many thanks!

(christos)

Whitespace

(skrll)

Trailing whitespace

(skrll)

Ignore atomic alignment error for clang

(christos)

Trailing whitespace

(skrll)

Trailing whitespace

(skrll)

-compact must come last

(uwe)

use long long and volatile consistently.

(christos)

fix TESTSDIR.
http://releng.netbsd.org/builds/HEAD/201904150550Z/evbarm-earm.build.failed

(christos)

r_symndx is no longer declared, use symnum and initialize it before use.

Completes the changes in the commit:
"Replace COMBREL with just-in-time check in _rtld_relocate_nonplt_objects."
http://mail-index.netbsd.org/source-changes/2017/06/19/msg085496.html

(maya)

Re-do previous:
emulate the addition of NETBSD_SUBTARGET_EXTRA_SPECS usually done by
SUBTARGET_EXTRA_SPECS (not used in risc-v) with EXTRA_SPECS.
Go back to using %(netbsd_link_spec) in our link spec, now that it's
a valid string.

Try to mimic other riscv OSes in our LINK_SPEC. they all start with
"-melf" XLEN_SPEC "lriscv" LD_EMUL_SUFFIX

and a definition of LD_EMUL_SUFFIX,
let's copy the LD_EMUL_SUFFIX linux uses.

Now we can link executables.

(maya)

we need rintl (long double) to link binaries

(maya)

Cargo cult follow the other 128bit long double architecture (sparc64).

(maya)

Temporarily ifdef out PTRACE_VFORK and PTRACE_VFORKDONE tests

It's not reliable on all ports. sparc and evbarm are known to hang.

PR kern/54111 by Martin Husemann

(kamil)

Provide riscv implementation.

XXX this is MI, why not use it for everyone?

(maya)

Don't build compat libraries for riscv, for now.

(maya)

more fake mknative

(maya)

don't redefine types.
UINTMAX_TYPE / INTMAX_TYPE are mostly not redefined on netbsd.
These choices contradict typical 64bit archs, perhaps we don't want them.

XXX is the SIZE_TYPE default for riscv reasonable?
(long unsigned int for 64bit abi, unsigned int for 32bit abi)

(maya)

Avoid -Wconversion warnings

(maya)

Fake mknative

(maya)

Fake mknative

(maya)

Add reference counting to alias states to prevent them disappearing
while still in use.

(hannken)

Add support for fifos, character and block devices.

(hannken)

Clean up this mess and simplify, so that all the socket options get
handled correctly whether or not the compat_50 module is loaded.

(pgoyette)

$ORIGIN support is now enabled.

Thanks martin for the heads up.

(maya)

Regen.

(msaitoh)

Avoid duplication.

(msaitoh)

Regen.

(msaitoh)

- Add some Intel Xeon Scalable processor's devices.
- Add Some Qlogic devices from OpenBSD.
- Rename Qlogic 0x3022's name from ISP3022 to ISP4022.

(msaitoh)

Some KNF

(skrll)

Typo: s/a indirection/an indirection/

(abhinav)

Trailing whitespace

(skrll)

Protect ether_multi list

Fix this driver because it can set IFEF_MPSAFE.

(ozaki-r)

If the compat code successfully handled an option, don't return an error.

(pgoyette)

Actually update the timeout value for the compability sockops

(pgoyette)

Split the COMPAT_50 socket-timeout stuff out of kern/uipc_socket.c
and into its own source file, which is now included in the compat_50
module.

(Not sure how this got missed during the original [pgoyette-compat] work)

(pgoyette)

Use more Dv.

(wiz)

Note that we now search DT_RUNPATH paths as well.

(maya)

Provide a definition of quiet NaN for RISC-V, and a reference to where
it comes from.

(maya)

undo previous unlock/relock; it is not needed now that we added the tls model
declaration to all externs in jemalloc (suggested by joerg@)

(christos)

Declare the tls model in external declarations. Pointed out by joerg.

(christos)

PR/53990, PR/52380, PR/52102: UFS2 cylinder group inode allocation botch

Fix rare allocation botch in ffs_nodealloccg().

Conditions:
  a) less than
        #_of_initialized_inodes(cg->cg_initediblk)
        - inodes_per_filesystem_block
      are allocated in the cylinder group
  b) cg->cg_irotor points to a uninterupted run of
      allocated inodes in the inode bitmap up to the
      end of dynamically initialized inodes
      (cg->cg_initediblk)

In this case the next inode after this run was returned
without initializing the respective inode block. As the
block is not initialized these inodes could trigger panics
on inode consistency due to old (uninitialized) disk data.

In very rare cases data loss could occur when
the uninitialized inode block is initialized via the
normal mechanism.
Further conditions to occur after the above:
  c) no panic
  d) no (forced) fsck
  e) and more than cg->cg_initediblk - inodes_per_filesystem_block
      allocated inodes.

Fix:
Always insure allocation always in initialized inode range
extending the initialized inode range as needed.

Add KASSERTMSG() safeguards.

ok hannken@

(kardel)

Only need to #include sys/vnode.h once

(pgoyette)

Add more checks, if the values are negative we hit a KASSERT later in the
timeout.

Reported-by: syzbot+662dbeb526303f458255@syzkaller.appspotmail.com

(maxv)

Trailing whitespace

(skrll)

Set the snippet_length field of the callback_args

Because of this field not being set, apropos was failing to show snippet when piped to a pager
or when used with -p argument.

(abhinav)

Create a RPI compatible image which cn be used with

kernel_address=0x10000040

in config.txt.

XXX alignment requirement needs fixing

(skrll)

Fix build error.  size locater is gone.
Thanks mrg@.

(isaki)

ifix uninialized pid

(christos)

Remove unused declaration of __BIT and __BITS
it's defined already, tested with #error.

(maya)

Omit mention of old vnd(4) bug, now resolved.

Via elge, thanks.

(maya)

If an enable-method property is present on the cpu node and not supported
by the kernel, try to use the enable-method from the cpus node instead.

(jmcneill)

add missing files

(christos)

regen

(christos)

fix typo, unused variable

(christos)