NetBSD 9.99.58 - ubc_uiomove needs to see new flag

(ad)

Fix userland gprof profiling on aarch64.

Adjusts _PROF_PROLOGUE to match OpenBSD; reworks our MCOUNT to retrieve
frompc placed on stack by the prologue, and to streamline sp manipulation
when preserving argument registers.

(jakllsch)

cache_lookup_linked(): We can't use the name to decide how to lock the dir,
since the name refers to the child (found object) not the parent (the thing
that's being locked).

Fix it by always doing rw_tryenter().  There's not much to be won by
optimising for the contended case, and were this routine doing lockless
lookups (the eventual goal) it wouldn't be hanging around waiting for
changes either.

(ad)

Enable ubc_direct by default, but only on systems with no more than 2 CPUs
for now.

(ad)

PR kern/54759 (vm.ubc_direct deadlock when read()/write() into mapping of itself)

- Add new flag UBC_ISMAPPED which tells ubc_uiomove() the object is mmap()ed
  somewhere.  Use it to decide whether to do direct-mapped copy, rather than
  poking around directly in the vnode in ubc_uiomove(), which is ugly and
  doesn't work for tmpfs.  It would be nicer to contain all this in UVM but
  the filesystem provides the needed locking here (VV_MAPPED) and to
  reinvent that would suck more.

- Rename UBC_UNMAP_FLAG() to UBC_VNODE_FLAGS().  Pass in UBC_ISMAPPED where
  appropriate.

(ad)

- Install HPET based DELAY() before going multiuser then recalibrate the TSC.
  Idea from joerg@.

- Take overhead into account when computing CPU frequency.

- Don't flush cache before computing TSC skew.

(ad)

ubc_direct_release(): unbusy the pages directly since pg->interlock is
being taken.

(ad)

When computing TSC skew make 8 measurements and use the average.

(ad)

Implement a HPET based DELAY().

(ad)

Swirch arm to the new binutils

(christos)

Typo in comment

(skrll)

fix inverted logic in NETBSD32 user signal stack handling (PR evbarm/55200)

(tnn)

use shortcut variables for readability.

(christos)

Add the system binary paths too since tests use them.

(christos)

fix tyop

(jdolecek)

elaborate on why we don't support hardware checksum offloading, and probably
never will

(jdolecek)

also let backend know that feature-ipv6-csum-offload is supported

(jdolecek)

g/c unused macro

(jdolecek)

Ticket #1728

(martin)

Apply patch, requested by he in ticket #1728:

src/external/bsd/bind/include/config.h (apply patch)

Fix BIND configuration for big-endian hosts.
This should make DNSSEC work on such hosts as well.

(martin)

Tickets #1537 - #1539

(martin)

Apply patch, requested by he in ticket #1539:

external/bsd/bind/include/config.h (apply patch)

Fix BIND configuration for big-endian hosts.
This should make DNSSEC work on such hosts as well.

(martin)

Pull up following revision(s) (requested by kim in ticket #1538):

usr.sbin/rtsold/rtsold.8: revision 1.37
usr.sbin/rtsold/rtsold.8: revision 1.38
usr.sbin/rtsold/rtsold.8: revision 1.39
share/man/man5/ifconfig.if.5: revision 1.20
etc/rc.d/network: revision 1.77
etc/rc.d/network: revision 1.78 (plus patch)
usr.sbin/rtsold/rtsold.8: revision 1.40
usr.sbin/rtsold/rtsold.8: revision 1.41
usr.sbin/rtsold/rtsold.8: revision 1.42
share/man/man5/ifconfig.if.5: revision 1.19

Add rtsol to the NAME section as well

rc.d/network: improve wording of waiting for DAD to finish

Revert unrelated changes to prior

Use .Dl for one line literal display.  Add SEE ALSO.

Remove trailing dot in SEE ALSO.

Add an "rtsol" keyword to ifconfig.if for enabling IPv6 RS/RA

Sort SEE ALSO.

Update date

(martin)

Pull up following revision(s) (requested by kim in ticket #1537):

share/man/man5/rc.conf.5: revision 1.172
etc/rc.d/network: revision 1.75

Allow rc.conf to setup resolv.conf via resolvconf(8).

This allows all static network config to be in rc.conf rather than
spread across files.

(martin)

Apply previous change ("Don't mix sign and unsigned operands. Just use
size_t for the loop.") to another loop variable.

(simonb)

Tickets #844 and #845

(martin)

Pull up following revision(s) (requested by roy in ticket #845):

sys/netinet6/nd6_nbr.c: revision 1.178

inet6: nd6_na_input() now considers ln_state <= ND6_LLINFO_INCOMPLETE

Otherwise if ln_state != ND6_LLINFO_INCOMPLETE and the is no lladdr
and this message was solicited then ln_state is set to ND6_LLINFO_REACHABLE
which could then cause a panic in nd6_resolve().

If ln_state > ND6_LLINFO_INCOMPLETE then it's assumed we have a lladdr.
Potentially this could have been triggered by the introduction of
ND6_LLINFO_PURGE in nd6.c r1.143 but also by the re-introduction of
ND6_LLINFO_INCOMPLETE in nd6.c r1.263.

Depending on the timing, it's technically possible to receive such
a message after the llentry is created with ND6_LLINFO_NOSTATE.

(martin)

make this work properly:
- get rid of cargo-culted register assignments, I found the right ones by
  experiment, now both light sensors report sane values
- keyboard brightness seems to max out at 16, act like it
- do what MacOS does and fade keyboard brightness instead of just switching
- add sysctls to configure keyboard brightness and environmental light
  thresholds
- don't poll the chip more often than once a second

(macallan)

Apply patch, requested by roy in ticket #844:

external/bsd/dhcpcd/dist/src/defs.h       (apply patch)
external/bsd/dhcpcd/dist/src/dhcp.c       (apply patch)
external/bsd/dhcpcd/dist/src/dhcp6.c       (apply patch)
external/bsd/dhcpcd/dist/src/ipv6nd.c       (apply patch)
doc/3RDPARTY       (apply patch)

Update to dhcpcd-8.1.9 which fixes alignment issues.

(martin)

Add -unknown- to "./netbsd" in modules/mi to fix "build.sh syspkgs" failure

(uki)

fix obvious typo in error handling - buf error should go to b_error

(jdolecek)

drop openfirm.h include

(macallan)

pass b_flags B_PHYS and B_RAW when setting up the buf for underlying device

should fix misfired KASSERT() in xbd(4)

(jdolecek)

If we're having Thai units, make them correct.  A tarangwah is a square wah,
which is ~4m^2 but not exactly (just as a wah is not exactly 2m, but close).
(wah apparently comes from the width of a person with hands outstretched...)

(kre)

fix too many zeroes in date entry

(jdolecek)

xbdback(4) is MP-safe

(jdolecek)

make xbdback actually MPSAFE and stop using KERNEL_LOCK()

remove no longer necessary atomics, the counters are now always
updated with held mutex

(jdolecek)

Stop forcing the -e option off in the subshell createds for a command
substitution.  This was inherited in the big "-e" fixup patch set (rev 1.50)
of Jan 2000, which came from dash.  dash no longer acts this way.

(kre)

allocate xbdback_io instances as part of struct xbdback_instance and
stop using pool_cache(9), to remove possibility for the allocation to sleep

(jdolecek)

Add '-t' option for tcsh-style time output.

(simonb)

g/c no longer used xbdi_same_page and xbdi_lastfragio_time

(jdolecek)

g/c no longer needed xbdi_io structure member, just pass it as continuation
parameter

(jdolecek)

sort product table, and tabify

(ryo)

On FIBRE devices, there are times when linkstat interrupt doesn't occur?
reported from Andrius V. thanks.

- use polling instead of linkstat interrupt when FIBRE
- add AQ_FORCE_POLL_LINKSTAT options (not by default)

(ryo)

make umount NAME= work

(christos)

Regen.

(msaitoh)

Remove duplicated entry.

(msaitoh)

Fix outdated comment. We've enabled crashme for non-DEBUG kernels.

(rin)

Remove "#ifndef __mc68010__ #error" I added when these {,u}divsi3.S and
{,u}modsi3.S were moved here (It was better to split the commit, sorry);
bootloader for sun3 still depends on them.

(rin)

For m68k, do not compile in {,u}divsi3 and {,u}modsi3 since codes
generated by gcc do not depend on them.

We should not have unused functions in libkern for MODULAR.

(rin)

Avoid using common symbol definitions.

(joerg)

rump doesn't own pnbuf_cache, externalize it

(joerg)

Replace noatf global with conditional compilation

(joerg)

saved_output is owned by director.c

(joerg)

lineno, states and allocated should be owned by the parser

(joerg)

npftest_mbufops and npftest_ifops are owned by npf_mbuf_subr.c

(joerg)

Avoid common symbol definitions.

(joerg)

sockb is only used locally, so move it into the function

(joerg)

Avoid global scope for variables only used locally

(joerg)

Rename debug to mountd_debug to avoid overlap in rump tests

(joerg)

Avoid common symbol declarations

(joerg)

bootfile is owned by libsa

(joerg)

Uses LFS boot code

(joerg)

rootpath is owned by libsa

(joerg)

Don't play #define games with extern, just have a separate list to avoid commons

(joerg)

loop.c owns iflist

(joerg)

Let loop.c own iflist

(joerg)

Mark _rtld_invalid_paths static as ldd can end up with two copies

(joerg)

Avoid common symbol definitions

(joerg)

Externalize yytext, it belongs to the lexer

(joerg)

_rpcpmstart is defined by rpcgen, so externalize it here

(joerg)

tempstr is only used in one file, make it static

(joerg)

Don't depend on common symbol definitions.

(joerg)

Externalize variables owned logically by pf_ruleset.c

(joerg)

buffnum is shared between subs.c and fancy.c

(joerg)

Explicitly use -fcommon for globals shared between libc and CSU.

(joerg)

regen

(christos)

Add AQUANTIA AQC100 from Andrius V.

(christos)

lookup_fastforward(): handle dotdot lookups and give up less often in
the union mount case.

(ad)

Remove unit tests for the never-exposed _lwp_gettid(2).

(thorpej)

Remove man page for the never-exposed _lwp_gettid(2) call.

(thorpej)

Regen for removal of _lwp_gettid(2).

(thorpej)

Remove _lwp_gettid(2) system call.  This problem is going to be solved
another way.  (Note: this call was never exposed in libc, so we can just
recycle the syscall number.)

(thorpej)

Don't try to attach vcpu on !XENPV
remove debug printf and other misc cosmetic changes

(bouyer)

inet6: nd6_na_input() now considers ln_state <= ND6_LLINFO_INCOMPLETE

Otherwise if ln_state != ND6_LLINFO_INCOMPLETE and the is no lladdr
and this message was solicited then ln_state is set to ND6_LLINFO_REACHABLE
which could then cause a panic in nd6_resolve().
If ln_state > ND6_LLINFO_INCOMPLETE then it's assumed we have a lladdr.

Potentially this could have been triggered by the introduction of
ND6_LLINFO_PURGE in nd6.c r1.143 but also by the re-introduction of
ND6_LLINFO_INCOMPLETE in nd6.c r1.263.
Depending on the timing, it's technically possible to receive such
a message after the llentry is created with ND6_LLINFO_NOSTATE.

(roy)

Ticket #843

(martin)

Pull up following revision(s) (requested by mrg in ticket #843):

external/gpl3/gcc/lib/libsupc++/Makefile: revision 1.13

pull in atomicity.h (atomicity.cc) on sparc.  3 of the objects
use these interfaces and this avoids link errors

fixes PR#54660.

(martin)

Tickets #837 - #841

(martin)

Tickets #1533 - #1535

(martin)

Pull up following revision(s) (requested by gdt in ticket #1534):

sys/kern/vfs_syscalls.c: revision 1.544
lib/libc/sys/fdatasync.2: revision 1.17

Relax fdatasync restriction that fd be writable

The restriction that a fd passed to fdatasync(2) must be writable was
added in 2003 in order to comply with POSIX.  Since then, POSIX has
removed that requirement, and POSIX-valid programs have been therefore
encountering errors on NetBSD.

Patch by Paul Ripke after discussion on netbsd-users.  Issue
discovered with pkgsrc/databases/mongodb3 as used by pkgsrc/net/unifi.

(martin)

Pull up following revision(s) (requested by maxv in ticket #1535):

sys/compat/ossaudio/ossaudio.c: revision 1.83

ossaudio: Avoid giving userland uninitialized memory. Noticed by maxv.

The uninitalized field in this structure is `fillers`, an array that
simply reserves space for later changes in OSSv4, which this version
of the OSS compat layer (specifically for Linux applications) makes no
effort to implement.

(martin)

Pull up following revision(s) (requested by maxv in ticket #841):

sys/compat/ossaudio/ossaudio.c: revision 1.83

ossaudio: Avoid giving userland uninitialized memory. Noticed by maxv.

The uninitalized field in this structure is `fillers`, an array that
simply reserves space for later changes in OSSv4, which this version
of the OSS compat layer (specifically for Linux applications) makes no
effort to implement.

(martin)

Pull up following revision(s) (requested by gdt in ticket #840):

sys/kern/vfs_syscalls.c: revision 1.544
lib/libc/sys/fdatasync.2: revision 1.17

Relax fdatasync restriction that fd be writable

The restriction that a fd passed to fdatasync(2) must be writable was
added in 2003 in order to comply with POSIX.  Since then, POSIX has
removed that requirement, and POSIX-valid programs have been therefore
encountering errors on NetBSD.

Patch by Paul Ripke after discussion on netbsd-users.  Issue
discovered with pkgsrc/databases/mongodb3 as used by pkgsrc/net/unifi.

(martin)

Pull up following revision(s) (requested by gson in ticket #839):

sys/kern/vfs_mount.c: revision 1.79

Destroy anonymous device vnodes on reboot once the last file system
got unmounted and the mount list is empty.

PR kern/54969: Disk cache is no longer flushed on shutdown

(martin)

Pull up following revision(s) (requested by kim in ticket #838):

share/man/man5/ifconfig.if.5: revision 1.20
etc/rc.d/network: revision 1.78 (plus patch)
usr.sbin/rtsold/rtsold.8: revision 1.42

Add an "rtsol" keyword to ifconfig.if for enabling IPv6 RS/RA

Sort SEE ALSO.
Update date

(martin)

Add troyounce

(scole)

Remove HZ=20, it was to testing only.

(bouyer)

On second thought, don't include more than one spelling for thai units

(scole)

We have USER_LDT tests in ATF, remove the ones from regress.

(maxv)

Add more Thai length, area and weight units

(scole)

Add missing break's

(christos)

make variable constant since it is assigned from a string literal

(christos)

regen

(prlw1)

Add an eclectic mix of
GeForce GTX 680, RTS522A PCI-E Card Reader, intel wifi 9560
+ whitespace

(prlw1)

Add RFI, RFP

(ginsbach)

show the 'interrupting at ...' info on boot

(jdolecek)

Restrict usage of m68k assembler versions of {,u}divsi3 and {,u}divsi3 to
kernel and bootloader for 68010.

They requires a special calling convention to udivsi3, and cannot to be
mixed up in normal routines provided by libgcc or compiler_rt. Although,
there's no problem for using in a controlled situation, i.e., kernel and
standalone programs.

Note that this does not affect m68k ports other than sun2, since codes
generated by gcc do not call these routines.

Assembler files are moved from common/lib/libc/arch/m68k/gen to
sys/lib/libkern/arch/m68k in order not to be compiled in libc.

Revert hack introduced to lib/libc/compiler_rt/Makefile.inc rev 1.37:
http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libc/compiler_rt/Makefile.inc#rev1.37

Proposed on port-sun2@ with no response...
(Again, this does not affect m68k ports other than sun2.)
http://mail-index.netbsd.org/port-sun2/2020/03/10/msg000102.html

(rin)

For NO_KERNEL_RCSIDS, strip __RCSID() also for src/common/*.

(rin)

At the moment, we need kern/uipc_mbufdebug.c only if DDB is enabled.

(rin)

Make crypto/rijindael optional again as cprng_strong does no longer
depend on it. Dependency is explicitly declared in files.foo if a
component requires it.

(rin)

Use __mc68010__ to distinguish m68000 (sun2) from other m68k ports;
__mc68000__ is defined both for m68000 and m68k.

(rin)

Fix previous for libhack, where _REENTRANT is not defined;
arenas_map_key is used only when NO_TLS && _REENTRANT.

(rin)

Add Whiskey Lake U and Amber Lake Y support.

(msaitoh)

Regen.

(msaitoh)

Add Intel Whiskey Lake U and Amber Lake Y devices.

(msaitoh)

Add Aquantia AQC100, AQC100S and D100.

(msaitoh)

Drop obsolete special case for clang/ppc.

(joerg)

Switch to using TLS in old jemalloc for everywhere but VAX and sun2.

(joerg)

NetBSD 9.99.57 - cwdinfo changed

(ad)

- Give cwdi_lock its own cache line, for multithreaded processes.
- Comments.

(ad)

Revert the changes made in February to make cwdinfo use mostly lockless,
which relied on taking extra vnode refs.

Having benchmarked various experimental changes over the past few months it
seems that it's better to avoid vnode refs as much as possible.  cwdi_lock
as a RW lock already did that to some extent for getcwd() and will permit
the same for namei() too.

(ad)

Use -Wno-error=address-of-packed-member for a number of more boot
loaders.

(joerg)

two more files to convert to newer HYPERVISOR_physdev_op() interface

(jdolecek)

convert to newer HYPERVISOR_physdev_op() interface, now command and the
arg are separate arguments - this is needed for newer physdev_op commands

remove code for PHYSDEVOP_IRQ_UNMASK_NOTIFY, it is obsolete since
interface version 0x00030202 and is unsupported by newer versions of Xen

confirmed working on amd64 Dom0, i386 compile-tested only

(jdolecek)

Ooops, restore accidently removed files from merge mishap

(martin)

adjust so that this at least compiles and links with __HAVE_PCI_MSI_MSIX

(jdolecek)

Sync with HEAD

(martin)

XEN_IPI_KICK was replaced by XEN_IPI_AST.

(ad)

Remove spurious reference to XEN_IPI_KICK - it represents the absence of
a specific IPI type.

(ad)

Follow convention and put entire predicate inside __predict_false()

(ad)

vmem_list_lock needs to be an adaptive lock.  Fixes LOCKDEBUG panic.

(ad)

Pull up following revision(s) (requested by rin in ticket #1533):

sys/arch/macppc/stand/ofwboot/Locore.c: revision 1.34 (via patch)
sys/arch/powerpc/oea/ofw_consinit.c: revision 1.18

When boot up with auto-boot? == true, some systems do not provide stdout
if the normal output is screen.

Open screen explicitly as stdout in this case, both for ofwboot and kernel,
which fixes problems with auto-boot? == true for Mac mini G4:
- messages from ofwboot do not appear
- kernel freeze during early boot stage

Taken from OpenBSD:
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/arch/macppc/macppc/ofw_machdep.c#rev1.3
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/arch/macppc/stand/Locore.c#rev1.3

XXX
pullup to netbsd-9 and -8

(martin)

It is possible to use grant_v2 with HVM guest; but the status table has
to be mapped using XENMEM_add_to_physmap.

(bouyer)

Pull up following revision(s) (requested by rin in ticket #837):

sys/arch/macppc/stand/ofwboot/Locore.c: revision 1.34
sys/arch/powerpc/oea/ofw_consinit.c: revision 1.18

When boot up with auto-boot? == true, some systems do not provide stdout
if the normal output is screen.

Open screen explicitly as stdout in this case, both for ofwboot and kernel,
which fixes problems with auto-boot? == true for Mac mini G4:
- messages from ofwboot do not appear
- kernel freeze during early boot stage

Taken from OpenBSD:
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/arch/macppc/macppc/ofw_machdep.c#rev1.3
http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/arch/macppc/stand/Locore.c#rev1.3

XXX
pullup to netbsd-9 and -8

(martin)

use %s/__func__ so that the strings can be shared.

(christos)

add blkif_x86_{32,64}_request_indirect types

(jdolecek)

also note the 64KB MAXPHYS for xbd(4)

(jdolecek)

note indirect segment support for xbd(4) and xbdback(4)

(jdolecek)

PR/55193: Jose Luis Duran: Fix wrong location for the db file.

(christos)

add support for indirect segments, which makes it possible to pass
up to MAXPHYS (implementation limit, interface allows more) using
single request

request using indirect segment requires 1 extra copy hypercall per
request, but saves 2 shared memory hypercalls (map_grant/unmap_grant),
so should be net performance boost due to less TLB flushing

this also effectively doubles disk queue size for xbd(4)

(jdolecek)

new openssl

(sevan)

adjust the condition for max indirect segments to make it clear what
is really needed (enough to map MAXPHYS + 1 page); no functional difference

(jdolecek)

Remove dst variable that was only set but not used.

(This broke the build for me.)

(wiz)

Update getopt string.

Remove -M. Add -C (the code is there and documented, even if noone could
activate it...).
Sync usage with manpage.

(wiz)

Remove -M from SYNOPSIS as well.

(wiz)

rtadvd: Remove disabled use of SIOCGIFPREFIX_IN6

It's not been enabled since the functionality was added to ifconfig(8)
many many years ago.

(roy)

rtadvd: Remove router renumbering - it's been disable for years

(roy)

LLVM is a bit stricter and wants the ARM variant of %progbits.

(joerg)

Note newer dhcpcd

(roy)

Sync

(roy)

postfix-3.5.1 out.

(wiz)

Fix garbages in dmesg caused by uninitialized variables slipped in r1.228.

Noticed in HP9000/362 dmesg:
https://dmesgd.nycbug.org/index.cgi?do=view&id=5459

Should be pulled up to netbsd-8 and netbsd-9.

(tsutsui)

Print "Hz".

(msaitoh)

There is no fast spoon^Wfast path in pmap_clearbit.

PR port-arm/55186: tests crash arm pmap

Tested by martin@

(skrll)

pull in atomicity.h (atomicity.cc) on sparc.  3 of the objects
use these interfaces and this avoids link errors

fixes PR#54660.

(mrg)

Whitespace fix. No functional change.

(msaitoh)

Get TSC frequency from CPUID 0x15 and/or x16 for newer Intel processors.

- If the max CPUID leaf is >= 0x15, take TSC value from CPUID. Some processors
  can take TSC/core crystal clock ratio but core crystal clock frequency
  can't be taken. Intel SDM give us the values for some processors.
- It also required to change lapic_per_second to make LAPIC timer correctly.
- Add new file x86/x86/identcpu_subr.c to share common subroutines between
  kernel and userland. Some code in x86/x86/identcpu.c and cpuctl/arch/i386.c
  will be moved to this file in future.
- Add comment to clarify.

(msaitoh)

when sorting all_extensions_by_on[] be sure to keep the NULL
entry last.  correct the sizeof() to use the right variable
(they were the same size at least.)  as-was, with the NULL
entry first, all loops across this array were no-ops.

this should fix PR#55158's arm64 side.

(mrg)

Rename buf_syncwait() to vfs_syncwait(), and have it wait on v_numoutput
rather than BC_BUSY.  Removes the dependency on bufhash.

(ad)

channel %d -> chan %d, for the benefit of 'systat vm'

(bouyer)

Pull from bouyer-xenpvh:
Bad news:
* grant table v2 is not supported for HVM guests on 4.11 at last.
* see xen/arch/x86/hvm/hypercall.c in Xen sources (missing
* GNTTABOP_get_status_frames)

So restore grant table v1 for !XENPV

(bouyer)

Misc fixes after merge

(bouyer)