Use pci_intr_establish_xname().

(msaitoh)

Remove extra newline in SPMI table.

(msaitoh)

s/conver_to/convert_to/. No functional change.

(msaitoh)

fix build error without DDB

(ryo)

for some reason my 6800 Ultra stalls when drawing *some* off-screen rectangles
as a workaround, draw them by software until I figure out what's going on

(macallan)

Back out sftchg/fstcnt deadlock workaround

The change did prevent some system freeze, but caused spurious
unmount failures reporter by bouyer@.

hannken@ is working on the right fix, see kern/53624

(manu)

cleanup locore, and changed the way to map memories during boot.
- add functions bootpage_enter() and bootpage_alloc() to adapt various layout
  of physical memory map. especially for 64bit physical memory layout.
  pmapboot_alloc() allocates pagetable pages from _end[].
- changed to map only the required amount for PA=VA identity mapping
  (kernel image, UART device, and FDT blob) with L2_BLOCK(2Mbyte).
- changing page permission for kernel image, and making KSEG mapping are done
  at cpu_kernel_vm_init() instead of at locore.
- optimize PTE entries with PTE Contiguous bit. it is enabled on devmap only for now.

reviewed by skrll@, thanks.

(ryo)

relax the NOERROR + C_CHAOS, just rerun the xfer instead of panic, to handle
another variant of bad/missing AHCI error recovery, this time under Parallels

(jdolecek)

curses: once __slk_init is called, reset slk_fmt

This allows soft label keys to be used on more than one terminal.

(roy)

curses: declare SLK format as invalid until slk_init(3) called

If we don't, then when __slk_init with a valid format of 0 it sets
up labels to be drawn without a window.

(roy)

further channel locking pass for reset

(jdolecek)

don't add +1 to retry on error

(jdolecek)

Remove duplicate #include

(skrll)

Add missing format string annotation.

(joerg)

Fix an obvious botch in the previous rev, found by martin@

(kre)

Ticket #1047

(martin)

Pull up following revision(s) (requested by roy in ticket #1047):

lib/libcurses/curses_private.h: revision 1.64
lib/libcurses/curses_private.h: revision 1.65
lib/libcurses/initscr.c: revision 1.33
lib/libcurses/curses_screen.3: revision 1.24
lib/libcurses/curses_screen.3: revision 1.25
lib/libcurses/newwin.c: revision 1.52
lib/libcurses/fileio.c: revision 1.6
lib/libcurses/ripoffline.c: revision 1.4
lib/libcurses/resize.c: revision 1.28
lib/libcurses/ripoffline.c: revision 1.5
lib/libcurses/resize.c: revision 1.29
lib/libcurses/screen.c: revision 1.34
lib/libcurses/setterm.c: revision 1.67

curses: fix ripoffline

When creating stdscr, ensure it's placed and sized in accordance with
lines ripped off.

LINES is no longer adjusted for lines ripped off.
POSIX makes no mention that it should be adjusted.
Bottom lines are now placed correctly.
Lines ripped off are now displayed after calling initscr.

ok kamil@

Fixes PR #53635

-

curses: resize ripped off windows
The application must still redraw them though.

-

Fix typos in curses_screen.3
Fn -> Ft in the resize_term(3) prototype
funcion -> function

(martin)

* define LX_BLKPAG_{OS,ATTR}_* for OS dependent PTE attributes in pmap.h
* cleanup macros

(ryo)

fix to boot APs of RPI3 with GENERIC64.
if no psci, try other methods.

(ryo)

Don't print extra newline if the PCI extended configuration space doesn't exist.

(msaitoh)

disassemblable sha512 insns

(ryo)

remove XXX delay to attach cpus in order

(ryo)

don't call plat->ap_early_putchar if not defined EARLYCONS

(ryo)

Also in the UNSET_HOOK2 case, release the mutex before destroying it.

(pgoyette)

Add missing file to the build

(pgoyette)

Grab the mutex for UNSET_HOOK2 (fixed previously for UNSET_HOOK)

(pgoyette)

Don't include common compat_80 code in the raid-speific module

(pgoyette)

Better diagnostic message

(pgoyette)

There are version-specific raid modules so don't include their code in
the "generic" version-specific module.

(pgoyette)

Don't define if_43_{init,fini}() in multiple .h files.

(pgoyette)

Make if_43_{init,fini} return an int, for consistency with all the
related *_{init,fini} routines.

(pgoyette)

change channel reset and drive reset for all ATA controllers to always
run via thread, and with channel lock held the whole time; the queue is
frozen while reset is pending

for this repurpose ata_reset_channel() into new ata_thread_run()

also adjust some device printfs to not leak xfer pointer, and avoid
aprint_* for non-autoconf messages

(jdolecek)

Tickets #1043 - #1046

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1046):

sys/net/if_bridge.c: revision 1.157
sys/net/if_bridge.c: revision 1.158
sys/net/if_bridge.c: revision 1.159

  Fix a bug that bridge_enqueue() incorrectly cleared outgoing packet's offload
flags. bridge_enqueue() is called from bridge_output() when a packet is
spontaneous. Clear csum_flags before calling brige_enqueue() in
bridge_forward() or bridge_broadcast() instead of in the beginning of
bridge_enqueue().

Note that this change doesn't fix a problem on the following configuration:

A bridge has two or more interfaces.
An address is assigned to an bridge member interface and
some offload flags are set.
Another interface has no address and has no any offload flag.

XXX pullup-[78]

- Fix bridge_enqueue() which was broken by last commit. Use correct mbuf
  pointer.
- Modify comment.

Micro optimization. m_copym(M_COPYALL) -> m_copypacket().

(martin)

Pull up following revision(s) (requested by maxv in ticket #1045):

sys/netinet/ip_reass.c: revision 1.19

Hold ip_off and ip_len in the fragment entry, instead of always reading
the associated mbuf (and converting to host order). This reduces the
cache/TLB misses when processing long lists.

(martin)

Pull up following revision(s) (requested by nakayama in ticket #1044):

sys/dev/pad/pad.c: revision 1.58

pad(4) mixer has only 1 channel, so return EINVAL in the case other than 1.

This fixes the following strange output of mixerctl(1):

        outputs.master=255,0
        inputs.dac=255,0

(martin)

Pull up following revision(s) (requested by roy in ticket #1043):

lib/libcurses/resize.c: revision 1.27
lib/libcurses/get_wch.c: revision 1.19
lib/libcurses/getch.c: revision 1.70

curses: resizeterm(3) should always send KEY_RESIZE

Fixes #53636

OK kamil@

(martin)

Another space that hurts Jared's eyes.

(skrll)

Add libpuffs (fixes sun2 build)

(jmcneill)

curses: resize ripped off windows

The application must still redraw them though.

(roy)

Normalize the required lists.  Each compat_netbsd32_xx needs the
non-32bit module of the same version _and_ the 32-bit module for
the next higher version.

(pgoyette)

Add support for Panasonic N5HBZ0000055, from David H. Gutteridge
in PR kern/53647.

(martin)

Regen

(martin)

David H. Gutteridge in PR kern/53647: add Panasonic N5HBZ0000055
WiFi device.

(martin)

Fix some product names and details as suggested by jmcneill

(skrll)

- Decode DMAR's ANDD (ACPI Name Space).
- Add X2APIC_MODE flag of DMAR. Note that Intel's VT-d's document says the
  flag's name is DMA_CTRL_PLATFORM_OPT_IN but actbl1.h says it's
  DMAR_X2APIC_MODE.

(msaitoh)

Add ACPI LPIT (Low Power Idle Table) from FreeBSD r336185.

(msaitoh)

Embryonic Cavium ThunderX support

(skrll)

Make sure that each version-specific compat_netbsd32_xx module depends
on all higher-version-specific modules, similar to the non-32bit
compat_xx modules.

(pgoyette)

- Don't print TPH requester's ST Table Size if the ST table location field
  is not PCI_TPH_REQ_STTBLLOC_TPHREQ because the size field is only applicaple
  for PCI_TPH_REQ_STTBLLOC_TPHREQ case.
- Add comment.

(msaitoh)

Sort the SOC_ list

(skrll)

Identify some Cavium ThunderX CPUs

(skrll)

Add commented out Cavium CN88XX ThunderX EARLYCONS and CONSADDR

(skrll)

Add some Cavium CPU_IDs (implementor and primary part number only aka
CPU_PARTMASK)

(skrll)

Advice from Bob Morris
https://www.ieee-security.org/Cipher/ConfReports/conf-rep-Crypto95.html

(sevan)

Don't treat mostly connection closed events as filtering events.
There a a failed to negotiate instance too, but I don't want to generate
more diff.

(christos)

Undo previous change. It made filtering a lot more aggressive.

(christos)

Update

(pgoyette)

Remove conditional compilation based on COMPAT_SUNOS - just update the
sf_scp field regardless.  We've previously removed the comparison that
checks if a lwp's emul is emul_sunos, so now we can also remove the
include for "opt_sunos.h"

(pgoyette)

PR/53646: Thomas Reim: Incorrect detection of the packet complete code in
fragment list check.

While the fix in https://launchpad.net/~rdratlos/+archive/ubuntu/racoon

- if (i > last_frag) /* It is complete */
+ if (i >= last_frag) /* It is complete */

has the correct behavior, it violates the test for successful
completion of the invariant of the loop:

    for (i = 1; i <= last_frag; i++) {
if (!check_fragment_index())
    break;
    }
    if (i > last_frag)
return ok;

It is better to move the check for NULL in the loop earlier, so that
the final iteration is done and the test is kept the same. It makes
the code easier to understand and preserves the original intent.

XXX: pullup-8

(christos)

fix diagnostic build.

XXX: pullup-7, pullup-8.

(mrg)

curses: fix ripoffline

When creating stdscr, ensure it's placed and sized in accordance with
lines ripped off.
LINES is no longer adjusted for lines ripped off.
POSIX makes no mention that it should be adjusted.
Bottom lines are now placed correctly.
Lines ripped off are now displayed after calling initscr.

ok kamil@

Fixes PR #53635

(roy)

Use hex string syntax to prevent overflow warnings for character constants
>= 128.

(christos)

Use a hook callback to allow sparc fpu code to determine if a process
is running under sunos emulation (in which case, fpu cleanup uses a
different set of fpu_codes[]).

(pgoyette)

Revert most of previous.  Leave the original dependencies, but still
remove the duplicate attribute definition to make config(1) happier.

(pgoyette)

Implement dummy netbsd32_compat_{13,16} routines.  aarch64 doesn't have
compat going that far back, but the build infrastructure expects to see
these sources and *_{init,fini} symbols.

(pgoyette)

Comment out printing L2CTLR_EL1 as it is implementation specific.

OK ryo

(skrll)

Make gzip on some crunched install media support xz decompression
(just like our base version does).

(martin)

Sync with version on Brendan's site
http://www.brendangregg.com/DTrace/iosnoop

(sevan)

Use more .Ql (quoted literal) for variable values.
Misc markup fixes.

(uwe)

Redo xz vs. gzip selection via conditional make expressions, suggested
by joerg.

(martin)

Expand code covered by sc->sc_mtx, to make sure an interrupt would not be
handled before the cv_timedwait_sig() call, or while polling.
Seems to fix "sunxitwi0: send STOP failed" messages frequently seen related
to axp20x0 sensors.

(bouyer)

Fix debug sets for mount_qemufwcfg

(jmcneill)

Sync with changes from HEAD: no more ATM

(pgoyette)

Use more markup.

(wiz)

atm is gone

(pgoyette)

Annotate format functions and fix format errors.

(christos)

Add qemufwcfg

(jmcneill)

Make more ld device nodes.

(jmcneill)

Install mount_qemufwcfg everywhere instead of just x86.

(jmcneill)

Suppress "not configured" message when no device is present

(jmcneill)

Add plrtc

(jmcneill)

Avoid reference of APbus workarea that can't be read via kvm(3).

(tsutsui)

Account APbus interrupts by evcnt(9).

(tsutsui)

Enable DHCP boot for modern servers rather than ancient BOOTPARAM.

(tsutsui)

Switch bootloader to using DHCP rather than BOOTPARAM.

Also bump version.

(tsutsui)

Set NS access bit when writing ICC_PMR_EL1

(jmcneill)

curses.h: Fix typo in a comment of the word 'handling'

No functional change intended.

(kamil)

Discussed some years ago but never commited: add an option to have a
single tmpfs (on /tmp) and use that for /var/shm as well (via a symlink
created after the tmpfs on /tmp has been mounted)

(martin)

If we're going to allow pl0 to access virtual and physical counters then
we should allow it from all CPUs.

kern/53630 (openssl fallout on arm)

(skrll)

Fix addchnstr() macro in curses.h

Add a missing argument 'n'.

(kamil)

remove hardcoded bullshit, probably fixes PR/53644

(maxv)

Newer gcc seems to use __ARM_ARCH_ISA_THUMB to tell use we have thumb
instructions available

(martin)

Fix news5000 panic during boot on APbus attach. PR port-newsmips/53626

Discussed with ryo@ (the original author of news5000 support).
Should be pulled up to netbsd-8.

(tsutsui)

mention  only the version with have in current.

(christos)

Note issue with compat_sunos vs sparc64's compat_netbsd32

(pgoyette)

Ssync with HEAD

(pgoyette)

add gcc 4.5 files.  updating a m68k system gave me:

rmdir: /usr/include/gcc-4.5: Directory not empty
        Can't remove obsolete directory /usr/include/gcc-4.5

because these two files were still present.

(mrg)

Split sparc64/netbsd32_machdep.c into pieces specific to versions _13
and _16

XXX Somehow this has broken compat_sunos - must investigate further

(pgoyette)

Fix the documented function prototype of slk_attr_set(3)

Note the missing argument short pair.

(kamil)

new ntpd

(christos)

merge conflicts

(christos)

Fix function prototypes in curses_underscore.3

Correct function prototype of wunderscore(3) and wunderend(3).

(kamil)

In MODULE_HOOK_CALL_DECL we don't need to provide the actual argument
list for calling the hook function, nor do we need to provide the
default value (for when the hook has not been set).

(pgoyette)

If the total size of the DTB exceeds our static storage, attempt to pack
it in place to make it fit.

(jmcneill)

Regen

(martin)

Upstream does not support armv4 any more - but it is trivial to add
add regen time. Make the "regen" target here do it.

(martin)

Fix comment markers.
https://github.com/opendtrace/toolkit/issues/11

(sevan)

Mention that NPF was converted to use libnv.

(rmind)

Stop searching for redistributors in a region after we find a redistributor
with the Last bit set in GICR_TYPER.

(jmcneill)

npf_return_tcp: fix no-INET6 case.

(rmind)

Set load offset to +2M to make room for QEMU's bootloader stub

(jmcneill)

Add virtio

(jmcneill)

Fix 64-bit build

(jmcneill)

- re-enable lint (otherwise the sets break)
- no space after -I (for consistency)

(christos)

Increase storage space for DTB (for QEMU)

(jmcneill)

No need to cover boot ROM with devmap

(jmcneill)

be less aggressive about blocking connections from disconnected sessions.

(christos)

NPF: Major rework -- migrate NPF to the libnv library.
- This conversion significantly simplifies the code and moves NPF to
  a binary serialisation format (replacing the XML-like format).
- Fix some memory/reference leaks and possibly use-after-free bugs.
- Bump NPF_VERSION as this change makes libnpf incompatible with the
  previous versions.  Also, different serialisation format means NPF
  connection/config saving and loading is not compatible with the
  previous versions either.

Thanks to christos@ for extra testing.

(rmind)

For i386 XEN3PAE_DOM0, use the "native" idt registration
infrastructure by removing the #ifndef XEN clause.

This will hopefully be the last commit to "fix" boot
breakage of XEN3PAE_DOM0

Thanks to bouyer@ to focussed bug reports with

# xl dmesg
and relevant ddb> bt

(cherry)

Correct function prototypes in curses_background(3)

Fix prototypes for: bkgdset(3), wbkgd(3) and wbkgdset(3).

(kamil)

Remove redundant definition of machine_arch32 and machine32

(pgoyette)

A timeout of 480 seconds is not enough for the t_pubkey:ec test case
on qemu/sparc.

(gson)

The default timeout of 300 seconds is not enough for the t_ciphers:evp
test case on qemu/sparc since the import of openssl 1.1.0i on 2018-08-18.

(gson)

Make netbsd32_sendsig_sigino() non-static since we need to call it from
the _md_16 code.

While here, use __func__ in a diagnostic message rather than hardcoding
the function name.

(pgoyette)

Use proper names for netbsd32_machdep_md_16_{init,fini}() routines

(pgoyette)

Add prototype for the default netbsd32_sendsig_siginfo()

(pgoyette)

Here there is no 32-bit specific sendsig_sigcontext() so just use the
native one.

(pgoyette)

Put netbsd32_sendsig_siginfo() here, where it belongs.

(pgoyette)

Remove netbsd32_sendsig_sigcontext() - it belongs in netbsd32_machdep.c

(pgoyette)

Define the netbsd32_sendsig_hook and create its accessor function.

(pgoyette)

Add a routine to actually invoke the hook for netbsd32_sendsig()

(pgoyette)

Keep only one copy of the netbsd32_compat_16 code.

Remove unnecessary #ifdef COMPAT_13 since the source file will only be
processed if COMPAT_13 is already defined.

(pgoyette)

Make sure that the netbsd32_compat_16 code is available to compat_netbsd_13

(pgoyette)

Don't fall off the end of a non-void function.  (It used to be void,
but now returns an int to conform with requirements of the MP-safe
module_hook stuff.)

(pgoyette)

Remove unnecesasry #ifdef COMPAT_16 (this file won't get selected unless
COMPAT_16 is defined).

Rename some functions for consistency with other architectures' machdep
code.

Provide some prototypes/declarations.

Use __func__ in a diagnostic message rather than hardcoding it in the
message string.

(pgoyette)

Allocate the netbsd32_sendsig_hook in the main compat_netbsd32 code
(where it call be called) and not in a subsidiary module which may
not even exist.

(pgoyette)

Wrap long line, KNF

(pgoyette)

Fix "use before init" of *idt.
This should fix boot time pagefaulting and panic(9)
on XEN kernels due to the boottime idt API reorg.

(cherry)

#include the module hook stuff
add prototype for netbsd32_sendsig_sigcontext

(pgoyette)

Add defaults and documentation for USE_XZ_SETS - a new variable
controlling whether an architecture will use .tgz or tar.xz
sets.

(martin)

Another USE_XZ_SETS which probably should be fixed elsewhere
in a better way

(kre)

Add glue for netbsd32 compat_13 and _16 modules

(pgoyette)

Remove some spurious declarations.

(pgoyette)

Clarify comment, pointed out by maya: we do not expect bootloaders
to understand xz compression, so the kernel.gz stay as they are, only
the sets are moving to .tar.xz.

(martin)

Regen

(pgoyette)

Only keep the compat syscall entry;  the main entry needs to remain
conditional.

(pgoyette)

Regen

(pgoyette)

Don't #ifdef out the NTP compat calls.

The syscall number needs to be defined (for the autoload table) whether
or not the compat code is built in.  The compat code will be included
in the module in all cases, but it makes explicit checks to ensure that
NTP code exists before executing.

(pgoyette)

Regen

(pgoyette)

Conditionalize the compat_10 sysv ipc syscalls on both the COMPAT_10
and SYSVxxx attributes

(pgoyette)

Regen

(pgoyette)

Conditionalize compat_09 syscalls

(pgoyette)

Regen

(pgoyette)