LIBISPRIVATE=yes

(christos)

Set LIBISPRIVATE=yes

(christos)

Arrange to create an _pic.a library too.

(christos)

1. if ${LIBISPRIVATE} == "pic", create _pic.a so that we can embed the library
  to another shared object
2. Don't compare ${LIBISPRIVATE} to "yes", because there are 3 places in
  Makefiles which set it to empty (this was a bug)
3. For private libraries, don't create .so* files

(christos)

lwp_thread_cleanup(): Remove overly-aggressive assertion.

(thorpej)

Fix extra whitespace added by accident.

(uwe)

In the interests of reliability simplify waiter handling more and redo
condvars to manage the list of waiters with atomic ops.

(ad)

Adjust to "npfctl debug" command line changes, from rmind@.

(martin)

Remove some unnecessary bit shifts by using appropriate read size

(skrll)

even if the line size of Icache and Dcache is different, it was fixed to work correctly.

- MAX(IcacheShift,DcacheShift) is wrong and should be MIN(IcacheShift,DcacheShift).
  Dcache and Icache are now done in independent loops instead of in the same loop.
- simplify the handling of cache_handle_range() macro arguments.
- cache_handle_range macro doesn't include "ret" anymore.

(ryo)

Add 0xa5 and 0xa6 for Comet Lake.

(msaitoh)

Add some definitions from the latest Intel SDM plus small fix:

- Add CPUID leaf 6 %eax bit 19 for HW_FEEDBACK* and IA32_PACKAGE_TERM* MSRs.
- Add CPUID leaf 7 %ecx bit 31 for Protection Keys.
- Add definition of Load only TLB and Store only TLB.
- Add IF_PSCHANGE_MC_NO bit of IA32_ARCH_CAPABILITIES
- Fix HWP_IGNIDL.

(msaitoh)

Typo in error message

(martin)

Example of a kernel module that registers DDB commands.

(uwe)

Make -f description actually use the word "force" so that it can be
found when searched for.

(uwe)

no need to make the PTE writable to do icache_sync, enough to accessible.

(ryo)

Use the default motd for stable builds

(jmcneill)

Avoid redefining _REENTRANT under sanitizers

Switch away from -Wno-macro-redefined which was Clang specific.

(kamil)

Avoid redefining _REENTRANT under sanitizers

Switch away from -Wno-macro-redefined which was Clang/LLVM specific.

(kamil)

Avoid redefining _REENTRANT under sanitizers

Switch away from -Wno-macro-redefined which was Clang specific.

(kamil)

Oops, missed a line in the last change.

(thorpej)

Remove superfluous checking for a "disable" property in the device_t
properties dictionary.

(thorpej)

Remove superfluous checking for a "disable" property in the device_t
properties dictionary.

(thorpej)

Switch from kmem_intr_alloc(sz, KM_NOSLEEP) to kmem_alloc(sz, KM_SLEEP).

Clearly document these functions are *not* for DDB session, but for
permanent data storage when initializing DDB.

(rin)

use strlcpy() instead of strncpy() for gcc happiness

(christos)

Stop allocating buffers dynamically in a DDB session, in order not to
disturb on-going debugged state of kernel datastructures.

Since DDB is running on 1 CPU at a time, static buffers are enough.

Increase in BSS section is:
  52552 for amd64 (LP64)
  9152 for m68k  (ILP32)

Requested by thorpej@ and mrg@.
Also suggested by ryo@.
Thanks!

(rin)

Reset ud_ifaces and ud_cdesc to NULL, to prevent use-after-free in
usb_free_device().

Reported-by: syzbot+c7e74d0ae89e9f08f863@syzkaller.appspotmail.com

(maxv)

also set ifc->ui_endpoints to NULL in usbd_free_iface_data() when the value
is freed, to make it impossible to re-enter this by mistake

very likely has no effect for the syzbot problem, but good to do nevetheless

Reported-by: syzbot+c555801d6bc0d768f402@syzkaller.appspotmail.com

(jdolecek)

If we failed because we didn't encounter an endpoint, do not attempt to
read 'ed', because its value is past the end of the buffer, and we thus
perform out-of-bounds accesses.

Detected thanks to vHCI+KASAN. First bug found by USB fuzzing.

Reported-by: syzbot+59e7f6b3f353584ac810@syzkaller.appspotmail.com

(maxv)

sync with regular socket code (no effect)

(christos)

Mask NULL + 0 LLVM UBSan reports in the ATF test: t_pslist.c

Pass -fno-delete-null-pointer-checks for Clang for the
MKSANITIZER/MKLIBCSANITIZER build

(kamil)

npf-params.7: s/filer/filter/

(tnn)

npf-params(7): fix the state.key defaults.

(rmind)

Finish rename of all Cavium Octeon device driver functions, structs etc from
"octeon_foo" to "octfoo" (missed octeon_uart + entries in iobus config).

(simonb)

postinstall: Move files out of dhcpcd chroot

(roy)

Revert prior - mark dirs as obsolete.

Thanks wiz@

(roy)

We don't install any chroot subdirs for dhcpcd anymore

(roy)

dhcpcd: Empty the chroot

While here, set correct optional hooks.

(roy)

Sync

(roy)

Fix typo in comment.

(rin)

Add m68k assembler version of __muldi3().

This is intended for 68060:
  - GCC does not emit __muldi3() for 68020-40, that have 32 * 32 --> 64 mulul
  - mulsl (and moveml), used in this code, are not implemented for 68010

In comparison with that from compiler_rt, this version saves:
  - 12% of processing time
  - 12 bytes of stack
  - 50 bytes of code size
Also, slightly faster, memory saving, and smaller than libgcc version.

By examining with evcnt(9), __muldi3() is invoked more than 1000 times per
sec by kernel, which should justify to introduce assembler version of this
function.

(rin)

Avoid using excessive kernel stack.
XXX maybe we should unite all "static proc_t" and "static lwp_t" in ddb
into one global each?

(martin)

make bwfm_usb_bus_ops static and const

(jdolecek)

Stop returning while PMAP_LOCK() (= KERNEL_LOCK(1, NULL)) is held.

Kernel freeze with heavy load is significantly mitigated (fixed?),
in which I could not even enter DDB from console.

XXX
There is still inconsistency in usage of two PVO pools.
I will send-pr later.

XXX
pullup to netbsd-[987]

(rin)

Tickets #1551 - #1553

(martin)

Pull up following revision(s) (requested by nat in ticket #1553):

usr.sbin/bta2dpd/bta2dpd/bta2dpd.c: revision 1.7

Avoid running of the end of the array if a file cannot be opened.
Found by plunky@.

(martin)

Tickets #933 - #937

(martin)

Pull up following revision(s) (requested by nat in ticket #937):

usr.sbin/bta2dpd/bta2dpd/bta2dpd.c: revision 1.7

Avoid running of the end of the array if a file cannot be opened.
Found by plunky@.

(martin)

Pull up following revision(s) (requested by roy in ticket #1552):

lib/libterminfo/curterm.c: revision 1.14

terminfo: test strlcpy result against space free, not string length

riastradh@ yep, looks good.

(martin)

Pull up following revision(s) (requested by roy in ticket #936):

lib/libterminfo/curterm.c: revision 1.14

terminfo: test strlcpy result against space free, not string length

riastradh@ yep, looks good.

(martin)

Pull up following revision(s) (requested by bouyer in ticket #935):

sys/arch/xen/x86/x86_xpmap.c: revision 1.89
sys/arch/x86/include/pmap.h: revision 1.121
sys/arch/xen/xen/privcmd.c: revision 1.58
sys/external/mit/xen-include-public/dist/xen/include/public/memory.h: revision 1.2
sys/arch/xen/include/xenpmap.h: revision 1.44
sys/arch/xen/include/xenio.h: revision 1.12
sys/arch/x86/x86/pmap.c: revision 1.394
(all via patch)

Ajust pmap_enter_ma() for upcoming new Xen privcmd ioctl:
pass flags to xpq_update_foreign()

Introduce a pmap MD flag: PMAP_MD_XEN_NOTR, which cause xpq_update_foreign()
to use the MMU_PT_UPDATE_NO_TRANSLATE flag.
make xpq_update_foreign() return the raw Xen error. This will cause
pmap_enter_ma() to return a negative error number in this case, but the
only user of this code path is privcmd.c and it can deal with it.

Add pmap_enter_gnt()m which maps a set of Xen grant entries at the
specified va in the specified pmap. Use the hooks implemented for EPT to
keep track of mapped grand entries in the pmap, and unmap them
when pmap_remove() is called. This requires pmap_remove() to be split
into a pmap_remove_locked(), to be called from pmap_remove_gnt().

Implement new ioctl, needed by Xen 4.13:
IOCTL_PRIVCMD_MMAPBATCH_V2
IOCTL_PRIVCMD_MMAP_RESOURCE
IOCTL_GNTDEV_MMAP_GRANT_REF
IOCTL_GNTDEV_ALLOC_GRANT_REF

Always enable declarations needed by privcmd.c

(martin)

Pull up following revision(s) (requested by skrll in ticket #1551):

sys/dev/usb/usb.c: revision 1.187

Don't allow open of /dev/usb if there are no attached busses.
PR kern/55303 mutex_vector_enter,512: uninitialized lock

(martin)

Pull up following revision(s) (requested by skrll in ticket #934):

sys/dev/usb/usb.c: revision 1.187

Don't allow open of /dev/usb if there are no attached busses.
PR kern/55303 mutex_vector_enter,512: uninitialized lock

(martin)

Pull up following revision(s) (requested by maya in ticket #933):

distrib/amd64/uefi-installimage/spec.inst: file removal
distrib/amd64/uefi-installimage/etc.rc: file removal
distrib/amd64/uefi-installimage/Makefile.bootimage: file removal
distrib/amd64/uefi-installimage/Makefile.installimage: file removal
distrib/amd64/uefi-installimage/install.sh: file removal
distrib/amd64/uefi-installimage/etc.ttys: file removal
distrib/amd64/Makefile: revision 1.14
distrib/amd64/uefi-installimage/boot.cfg.in: file removal
distrib/amd64/uefi-installimage/Makefile: file removal
distrib/amd64/installimage/Makefile: revision 1.11

rename amd64 uefi-installimage to be just installimage.

Works for both purposes, no confusing name.

(martin)

Switch to db_alloc() from kmem_intr_alloc(9).

Fix build failure as a part of crash(8).
Noticed by tnn@, thanks!

(rin)

Revert introduction of db_alloc.h.

db_alloc() and friends are already provided in crash.c.
Sorry for confusing you...

(rin)

XXX
DDB can be running in the interrupt context, e.g., when activated from
console. Therefore, we use kmem_intr_alloc(9) and friends in order to
avoid assertion failure.

(rin)

Introduce db_alloc.h.

Provide db_alloc() and friends to userland, i.e., crash(8).

(rin)

#ifdef out debug garbage.

(rin)

For wdc_obio_match(), struct wdc_regs is too large for stack.
Use kmem_alloc(9) instead.

Frame size becomes:
    3088 --> 0

(rin)

struct statvfs is too large for stack. Use malloc(9) instead.

XXX
Switch to kmem(9) for entire this file.

Frame size, e.g. for m68k, becomes:
    3292 --> 12

(rin)

db_show_callout(): struct callout_cpu and cpu_info are too much for stack.

XXX
DDB can be running in the interrupt context, e.g., when activated from
console. Therefore, use kmem_intr_alloc(9) instead of kmem_alloc(9).

Frame size, e.g. for m68k, becomes:
    9212 (oops!) --> 0

(rin)

sc_statuspend is allocated with kmem_zalloc, so no need to memset it.

(maxv)

Add comments.

(maxv)

Rename all Cavium Octeon device driver functions, structs etc from
"octeon_foo" to "octfoo", except "octeon_eth" becomes "cnmac".

(simonb)

Remove unused Timer Unit register definitions and stubs.  Can be
resurrected from the attic in the unlikely event we'll ever have
a driver for this device.

(simonb)

Avoid running of the end of the array if a file cannot be opened.

Found by plunky@.

(nat)

Clean up Cavium Octeon device names.  Rename devices from "octeon_foo"
to "octfoo" - this follows the naming conventions used by many other
MIPS CPUs.

(simonb)

Disable read of the ACPI timer during idle, fixing performance degradation
observed with acpicpu.  C1 (MWAIT) is the deepest sleep currently enabled so
there should be no functional change.  Also, the computed nap time is run
through hztoms() which seems incorrect as the ACPI timer is running in the
MHz range.

(ad)

Don't define psl as common symbol, move it into the only file using it.

(joerg)

Use device_xname() to access dv_xname

(sevan)

Use device_t for sc_dev

(sevan)

Switch from printf to aprintf_error/normal

(sevan)

Initialise sc->sc_dev, otherwise unnamed interfaces show up and things go
downhill from there.

(sevan)

npftest -- npf_test_init(): add a workaround for NetBSD.

(rmind)

npf_worker_sys{init,fini}: initialize/destroy the exit_cv condvar.

(rmind)

Create pic lib and don't install pkgconfig file.

(christos)

Make libuv private, requested by joerg@

(christos)

Fix merge error - adjust assertions.

(ad)

Regen.

(ad)

A couple of small changes to lookup that cut 5-10% system time from
"build.sh release" on my test system:

- Crossing mount points during lookup is slow because the set up for, and
  act of doing VFS_ROOT() is quite involved.  Use the name cache to help
  with this.  Cache an "impossible" zero-length name with covered vnodes,
  that points to the root of the file system mounted there.  Use it to cross
  mounts.  When cache_purge() is called on either of the vnodes involved the
  cache entry will disappear.  All of the needed calls for that are already
  in place (vnode reclaim, unmount, etc).

- In lookup_fastforward(), if the the last component has been found and the
  parent directory (searchdir) is not going to be returned, then don't get a
  reference to it.

(ad)

Allow LIBISPRIVATE to build a pic version of itself for embedding into
other shared libraries.

(christos)

catch up with changes - need bwfmreg.h before bwfmvar.h

(jdolecek)

Fix a lock order reversal that caused hangs.

(ad)

Merge updates from upstream to reduce stack usage of SHA3_Selftest.

(riastradh)

cache_lookup_linked():

- If the lookup fails return with the same lock held.  There's no bug here
  because either parent or child's lock is sufficient to prevent both from
  disappearing into thin air, but may as well be correct.

- if FSCRED is passed in then skip the auth check.

(ad)

sctlr_el1 and ctr_el0 are 64-bit registers

(jmcneill)

aubtfwl(4): Use kmem_asprintf() for firmware file name creation.

A MAXPATHLEN+1 array of char on stack is "too much".

(jakllsch)

slightly improve sentence, no need to 'Note' when in NOTES section

(jdolecek)

remove M_SOFTDEP from list of supported malloc types, it's long gone

also remove M_PCB - while it is used by SCTP, it's actually not defined
anywhere

in fact add new NOTES section explaining that the malloc types are
actually not used on NetBSD

(jdolecek)

gem_pci_attach(): avoid allocating a 2K buffer on the stack.

(thorpej)

coredump_note_elf{32,64}(): avoid allcating register state buffers on
the stack; they could be large.

(thorpej)

terminfo: test strlcpy result against space free, not string length

riastradh@ yep, looks good.

(roy)

discussed with thorpej@, and it seems it's better to allocate the bss
alignment buffer statically (as part of softc) just to be sure to not
affect performance

(jdolecek)

Major NPF improvements (merge from upstream):

- Switch to the C11-style atomic primitives using atomic_loadstore(9).

- npfkern: introduce the 'state.key.interface' and 'state.key.direction'
  settings.  Users can now choose whether the connection state should be
  strictly per-interface or global at the configuration level.  Keep NAT
  logic to be always per-interface, though.

- npfkern: rewrite the G/C worker logic and make it self-tuning.

- npfkern and libnpf: multiple bug fixes; add param exporting; introduce
  more parameters.  Remove npf_nvlist_{copyin,copyout}() functions and
  refactor npfctl_load_nvlist() with others; add npfctl_run_op() to have
  a single entry point for operations.  Introduce npf_flow_t and clean up
  some code.

- npfctl: lots of fixes for the 'npfctl show' logic; make 'npfctl list'
  more informative; misc usability improvements and more user-friendly
  error messages.

- Amend and improve the manual pages.

(rmind)

reduce stack usage in genfb_calc_hsize()

(jdolecek)

reduce stack usage in bwfm_rx_event_cb(), can use KM_SLEEP because
this is called from workqueue handler (i.e. thread context)

(jdolecek)

constify, and make static where possible

(jdolecek)

add bwfm* at pci?, which is present in GENERIC

(jdolecek)

fix wdcprobe_with_reset() to avoid allocating big structures on stack

(jdolecek)

allocate memory for dtv_ts_section using kmem_alloc() in dtv_demux_read(),
instead of on-stack

XXX compile-tested only

(jdolecek)

Make room for debug sets

(jmcneill)

change pci_conf_print() to allocate memory for the regs dynamically
instead of on-stack

(jdolecek)

remove #if 0'ed pci_conf_print() call

(jdolecek)

Avoid passing file paths in panic strings, this results in extra long
output that is annoying and that syzbot classifies as independent reports
due to the instances having different build paths.

(maxv)

Introduce PTRACE_REGS_ALIGN, and on x86, enforce a 16-byte alignment, due
to fpregs having fxsave which requires 16-byte alignment.

Reported-by: syzbot+f44d47e617ebf7fda081@syzkaller.appspotmail.com

(maxv)

Revert 0(sp) -> CALLFRAME_S0(sp) changes in rev 1.4.

0(sp) is safe here even on O32 because the only callee function
mips_cp0_cause_read() doesn't take arguments, but it's caller's
responsibility to choose proper stackframe region and the name
of CALLFRAME_S0 is just confusing.  No binary change.

(tsutsui)

Avoid use of CALLFRAME_S0 on calling ARC BIOS functions.

On O32 and O64, 0(sp) could be clobbered by callee functions
to save argument a0, so we can't use it to save caller registers.
Use CALLFRAME_SP instead because a frame pointer is not used here.
Maybe all other references of CALLFRAME_S0 might also be problematic
on O32.  Discussed with soda@.

(tsutsui)

CN70XX iobdma limit appears to be 128 words, so gather that many samples
as a time.
Gather the full 512 bytes of samples and process in a single call to
rnd_add_data_sync() - about 10% faster than 4 calls to rnd_add_data_sync().
Put sample buffer in the softc to save some stack usage.

(simonb)

Explicitly use MIPS_CURLWP macro.  No binary change.

(tsutsui)

Initialize puffs9p::server (though it's not used anywhere).

(uwe)

Don't forget to save MIPS_CURLWP (t8) register before restoring it.

Maybe harmless because the only ARC BIOS function that takes
more than four arguments is currently unused.

(tsutsui)

Provide list of options for ease of reference.

(uwe)

Fix vio9p device name in a comment.

(uwe)

Try to shrink kernels to smaller than ~6MB due to ARC BIOS restriction.

Currently NetBSD/arc kernels are loaded at 0x80200000
(on some machines ARC BIOS uses regions before it) and
ARC BIOS also uses around 0x807f0000, so bootloader fails
on loading >6MB kernels.

Note GXemul doesn't have this restriction because it loads
a kernel directly via a command line argument.

Tested by installing using sysinst on RAMDISK kernel on netbsd-9.

Should be pulled up to netbsd-9.

(tsutsui)

Shrink ramdisk to make RAMDISK kernel smaller for ARC BIOS restriction.

(tsutsui)

Make sure to preserve MIPS_CURLWP (t8) register from ARC BIOS calls.

Fixes silent hangup right after starting a kernel on my
Express5800/230 R4400 PCI (finally I've repaired its PSU and RTC).
Reviewed by soda@.  Maybe other ARC machines are also affected.

Note GXemul was not affected by this problem because it doesn't
have actual ARC BIOS ROM and emulates ARC functions internally.

Should be pulled up to at least netbsd-9.

(tsutsui)

Reported-by: syzbot+fd9be59aa613bbf4eba8@syzkaller.appspotmail.com
Reported-by: syzbot+15dd4dbac6ed159faa4a@syzkaller.appspotmail.com
Reported-by: syzbot+38fa02d3b0e46e57c156@syzkaller.appspotmail.com

pmap_remove_all(): need to drain PV pages only after the PTEs are unmapped,
otherwise there can be a context switch with them mapped in.  XXX amd64
should use the direct map.

(ad)

make it compile again

(christos)

Add more printf annotations

(christos)

put back the warn flags

(christos)

fix printf format warning with format_arg

(christos)

Add explicit casts

(christos)

For struct timecounter, use C99 initializers.
Compile tested. No functional changes intended.

(rin)

delete previous and fix bug instead.

(christos)

Fix printf format error.

(christos)

fix WORDS_BIGENDIAN

(christos)

Fix the ntpd build with Clang/LLVM

Set -Wno-format-nonliteral for ntp_refclock.c

(kamil)

Fix incorrect merge.

(christos)

Fix the libntp build with Clang/LLVM

Set -Wno-error=implicit-int-float-conversion

(kamil)

DTB are no longer part of kernel set

(jmcneill)

Specify -Wno-format-nonlitera for scp.c (OpenSSH)

Fixes build with Clang/LLVM.

(kamil)

Make this compile again. Not tested.

(rin)

Use __BITS consistently and inline with specification. NFCI.

(skrll)

For clk_timecounter:
- Use C99 initializer.
- Leave .tc_name NULL initially.
No functional changes.

(rin)

u_int64_t --> uint64_t. No binary changes.

(rin)

Fix suspend/resume.
- Revert temporary usage of sc_[pr]busy during suspend.  These indicate
  whether the mixer needs to be restarted or not.
- Avoid timeout error when about to suspend.

(isaki)

remove useless MACHINE_ARCH == sh3 check for GCC 8.

(mrg)

alpha, vax, and sh3 have switched.  (sh3 swithed a month ago by
accident when i wrote ${MACHINE_ARCH} != "sh3", but this should
have been MACHINE_CPU.)  this means m68k is the only one left.

update various other parts and note some of the m68k issues.

(mrg)

switch vax to GCC 8.  testing in simh shows the same set of bugs.

(mrg)

Set -fno-delete-null-pointer-checks for RUMP/MKSANITIZER with Clang

NULL + 0 arithmetic raises LLVM UBSan warnings, specially in sys/pslist.h
in the type-safe macros _PSLIST_VALIDATE_PTRS and _PSLIST_VALIDATE_CONTAINER.

Proposes on <tech-userlevel> without objections.

(kamil)

Allocate proper storage for the event counter group names.

Can't use a stack buffer for these because the evcnt remembers the
pointer!

(riastradh)

Accept ioctl(RNDADDDATA) estimates at securelevel 1 (but not 2).

securelevel=1 is supposed to be a reasonable default for normal
computers.  This got in the way of ever getting entropy from a seed
on a machine with no HWRNG -- e.g., from another machine, or by
making the executive decision that what has been sampled is good
enough and issuing `head -c 32 < /dev/urandom > /dev/random'.

(riastradh)

evbarm: Add install media for earmv7hf and aarch64.

(jmcneill)

At least panic with a useful message if there are too many CPUs.

(ad)

Remove "write a BSD-licensed web browser" from the list of suggested smaller projects.

(jmcneill)

PR port-i386/55314: i386 no longer boots

Oops, EDI and ESI are callee saved on i386.

(ad)

Shrink the evbarm install image by not including the live images on the install media

(jmcneill)

Set HAVE_INSTALL_IMAGE for evbarm

(jmcneill)

Remove "If you booted from a floppy" message

(jmcneill)

new OpenSSH

(christos)

bump openssh

(christos)

Merge conflicts

(christos)

Remove support for evbarm cdrom images. Not sure that any of them can even
be used.

(jmcneill)

Build install images for aarch64 and earmv7hf.

(jmcneill)

Only install gptmbr.bin if USE_GPTMBR=yes

(jmcneill)

Default USE_MBR to no

(jmcneill)

Do not send dummy Tauth, it's completely unnecessary.

The best way to not use authentication is to not use authentication.
Cf. http://man.cat-v.org/plan_9/5/attach

(uwe)

Add some new CTR_EL0 bits

(skrll)

Create a separate Arm64 image that uses MBR partitioning. Amlogic SoCs
require the bootloader to be installed on SD cards at LBA1 which makes
them incompatible with GPT images.

(jmcneill)

Allow config file to override hostname

(jmcneill)

Add AMD MSR_DE_CFG's bit 1 as DE_CFG_LFENCE_SERIALIZE.
This bit makes lfence instruction serializing.

(msaitoh)