s/entires/entries/ in local variable definition.

(andvar)

s/entires/entries/

(andvar)

regen

(nia)

More Xeon Scalable / Skylake-E devices

(nia)

White space KNF nits.

(simonb)

Add mutt_test to the test runs, this test tries to emulate a real
curses application to test the interaction of various curses calls.

(blymn)

doc/CHANGES: add indent(1), lint(1), make(1)

(rillig)

Update check files, changes due to bug fixes.

(blymn)

fix various typos, mainly s/trasfering/transferring/ and s/theese/these/.

(andvar)

Catch up with aarch64 TTBR0 handling in pmap_{,de}activate_efirt and
kpreempt_{en,dis}able.

(skrll)

Style. NFCI.

(skrll)

Sprinkle some KASSERT(kpreempt_disabled());

(skrll)

make: remove test scenarios for old GCC versions

These had been useful for NetBSD 8, which used GCC 5.

(rillig)

Catch up with Elftoolchain head: changes to support 'tools' builds
out of the box.

(jkoshy)

Fix vnode locking around VOP_OPEN.

(mlelstv)

libwsemul_vt100: Rumpified wscons vt100 emulator (wsvt25).

This library provides wscons vt100 emulation (aka TERM=wsvt25).  This
is internal module, not a device, so there's no RUMP_COMPONENT.  The
public interface provided by this module is wsemul_vt100_ops and it
needs the user to provide (in rump namespace) wsdisplay_emulinput and
wsdisplay_emulbell, provided in the real kernel by wsdisplay.

This subdirectory is not hooked into the build.

(uwe)

Regen.

(hannken)

Really depend the lockdebug checks on option VNODE_LOCKDEBUG -- were
enabled by accident.

(hannken)

libXcursor: Add missing MLINKS.

(uwe)

xorg-server: Add xorg.conf.d(5) MLINKS.

(uwe)

Minor: fix a typo.

(jkoshy)

Prepare the ground for <elfdefinitions.h> to be used by the 'tools'
build of 'libctf'.

(jkoshy)

Trailing whitespace

(jmcneill)

Regen.

(hannken)

Make option VNODE_LOCKDEBUG more functional.

- Add "LOCKED=EXCL" option for vnode operations needing exclusive locked nodes.

- As it is impossible to check for shared locks held by current thread avoid
  false assertions by testing for exclusive lock only in the "LOCKED=NO" case.

- New option "RUMP_VNODE_LOCKDEBUG" enables VNODE_LOCKDEBUG for the rump kernel.

(hannken)

Lock vnode for VOP_LOOKUP().

(hannken)

No IO_NODELOCKED for unlocked vnode.

(hannken)

Lock devvp for kauth KAUTH_REQ_SYSTEM_MOUNT_DEVICE.

(hannken)

Lock devvp for vinvalbuf().

(hannken)

Output changed due to bug fixes in curses.

(blymn)

* Don't redraw the background if the new background character is the
  same as the old one.  This prevents excessive redraws in some
  applications.

* Fix bug introduced when wbkgrndset was fixed, we cannot blindly
  replace any instance of the old background character with the new one
  because some of those characters were put there by the application
  leading to display corruption.  So flag characters as background when
  they are erased and only update the flagged characters when setting
  the background.

(blymn)

libXrandr: Add missing MLINKS.

(uwe)

if_wm.c: various spelling and grammar fixes

(gutteridge)

tap.4: trim introduction to match current reality

As of 10.0, tap(4) can no longer be used as a bridge(4) endpoint, so
avoid potentially confusing the reader, who'd later encounter the
caveats section, noting use of vether(4) is required instead.

(gutteridge)

Catch up with Elftoolchain upstream <r3980>.

(jkoshy)

Follow FreeBSD revision <0e15d9fbeba4> and simplify #ifdef'ed code.

This change avoids redefining 'SHN_SUNW_IGNORE' if the symbol is already
present.

(jkoshy)

explicitly set TZ=UTC so that we don't depend on the testing rig for it.

(christos)

Bug fix: use the correct target path in a dependency.

(jkoshy)

xterm: Try to make manpage substitutions more readable.

Use X11EXTRAMANTRANSFORMS to set most of them, so that the sed(1)
syntax noise doesn't drown out the keys/values.  Add comments.

Keep using X11EXTRAMANDEFS for __app_version__ - my make-fu is too
weak to figure out how to do that properly while keeping happy
multiple levels of quoting along the way.

Same output is generated.

(uwe)

Only print the appropriate PAR fields for PAR.F={0,1}

Group the fields in the header.

(skrll)

Add NetBSD version control markers.

(jkoshy)

gdb-12.1 out

(wiz)

fix non-MULTIPROCESSOR build

(ryo)

xterm(1): Add substitution for __double_buffer__

(uwe)

libXi: Fix typo in __libmansuffix_

(uwe)

bsd.x11.mk: Add xlocaledir substitution.

(uwe)

libX11: Install Compose(5) manual page.

(uwe)

Remove trailing whitespace from upstream revision strings.

No functional change.

(jkoshy)

Make the generated <sys/elfdefinitions.h> file visible when compiling
libdwarf.

(jkoshy)

Remove trailing whitespace from upstream revision strings to reduce
diffs against upstream.

No functional change.

(jkoshy)

libX11: Add missing MLINKS.

(uwe)

attach the other pciid for intel 100 series KT serial.

seen on a thinkpad 460s.

(mrg)

Rename variables to be more descriptive of their local meaning rather
than referring to external usage.

(pgoyette)

tun.4: some xrefs added and minor grammar fixes applied

(gutteridge)

lint: inline macro 'sflag'

Mark all places where lint's C90 mode is stricter than its C99 mode.

Most of the situations in which lint produces only warnings instead of
errors covered by the "Constraints" sections in the relevant standards.
This doesn't prevent a specific compiler from accepting it though.

No functional change.

(rillig)

lint: inline macro 'tflag'

The definition of the macro tested both allow_trad and allow_c90, but
there is only a single mode in which allow_c90 is false, therefore it
suffices to test only that.

While double-checking each occurrence of tflag individually, I learned
why lint performs lookups of struct members only by name, independently
of the struct in which they are declared.  See typeok_arrow for details.

No functional change.

(rillig)

lint: document why in GCC mode, C99 keywords are enabled

(rillig)

lint: document why C11 keywords are available in C99 mode as well

(rillig)

lint: move GCC attributes to the bottom of the list

These identifiers have a simpler definition than the other keywords
since they do not define when they were introduced.

No functional change.

(rillig)

lint: inline macro Sflag in cases of expected behavior

In the grammar rules 'enums_with_opt_comma' and 'block_item_list',
checking for allow_c99 was redundant since c99ism does not warn in C99
mode.

In the grammar rule 'designator', align the two actions structurally.

No functional change.

(rillig)

lint: inline macro c11flag

No functional change.

(rillig)

Don't put temporary files and directories into the atf_srcdir (that
might be read-only!).  Instead, put them in the test's working
directory what the atf framework will automatically clean up after
the test completes.

(pgoyette)

Mention the changes for /etc/rc.d/gpio

(brad)

add CPU_ID_APPLE_M1_ICESTORM and CPU_ID_APPLE_M1_FIRESTORM.

sync cpuids[] to sys/arch/aarch64/aarch64/cpu.c r1.69

(ryo)

Like /etc/rc.d/network, add the ability to put a '!' character as the
first character in /etc/gpio.conf and have the rest evaluated in a
shell.  It is useful to be able to do actions after setting up gpio
pins or attaching gpioiic(4).

(brad)

Handle warnings and errors from pcap_activate properly (from bch)

(christos)

Now, NULL is always defined correctly, as we include <sys/param.h> above.

(rin)

whitespace

(skrll)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

Use '\0' instead of NULL for NUL-character.

(rin)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

Do not use NULL for integer 0.

(rin)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

Use '\0' instead of NULL for NUL-character.

(rin)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

Use '\0' instead of NULL for NUL-character.

(rin)

lint: make grammar compatible with Bison

Bison and BSD Yacc differ in how they translate the grammar into C code.

Bison:
code above '%union' (YYDEBUG is either undefined or 1)
#include "cgram.h" (which defines YYDEBUG to 0 if undefined)
code from below '%union' until the first '%%'
actions
code below the second '%%'

BSD Yacc:
code above the first '%%' (YYDEBUG is either undefined or 1)
(no #include for "cgram.h")
define YYDEBUG to 0 if undefined
code below the second '%%' (YYDEBUG is either 0 or 1)
actions

The easiest way around these differences is to remove the 'defined(...)'
around the conditionals YYDEBUG, YYBISON and YYBYACC.

(rillig)

Add a new test for PR kern/56713 and set to expected_failure for now.

(pgoyette)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

SCANWAIT and PWAIT are renamed to SCANDELAY and SCANWAIT, respectively,
in order not to redefine PWAIT.

XXX
Too many copies for tgets.c...

(rin)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

Use '\0' instead of NULL for NUL-character.

(rin)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

SCANWAIT and PWAIT are renamed to SCANDELAY and SCANWAIT, respectively,
in order not to redefine PWAIT.

(rin)

Fix fallout from libsa change; stand.h includes <sys/param.h> now.

- Stop redefining roundup2().
- Compare kvamap->start (uint64_t) with 0 instead of NULL.
- Include <sys/param.h> explicitly.

Thanks gson@ for notification.

(rin)

lint: reorder grammar rule begin_type_typespec

This makes the rule structurally more similar to the other rules.  The
goal is that all rules whose name starts with 'begin_type_' actually
start with the rule 'begin_type'.  Currently, some rules parse their
type attributes before 'begin_type', which feels wrong, as type
attributes should be part of the type.

No functional change intended.

(rillig)

Enable SA_HARDCODED_SECSIZE to shrink loaders slightly.

(rin)

Re-introduce SA_HARDCODED_SECSIZE hack, by which hardcoded DEV_BSIZE is
used instead of secsize obtained by SAIOSECSIZE ioctl.

This hack avoids divdi3 and friends from being linked, in order to
support variable secsize for some archs.

Should be useful for ancient archs, for which secsize is fixed.

Thanks christos@ for comment.

(rin)

Remove unused; switched to usr.sbin/installboot/arch/amiga.c long ago.

(rin)

Stop twiddling when loading boot.amiga; it is small enough.

Slightly reduce insn's for primary loaders, and stop them to output garbage.

(rin)

Simplify relocation rules.

Now, both single- and triple-byte encodings in relocation table
represent addrdiff (instead of absolute file offset for the latter).

Shave off one insn for boot.amiga. Also, for elf2bb(1), relocation
overflow is mitigated and detectable at least even if inevitable
(it seems unlikely to happen although).

(rin)

lint: only allow GCC attribute in null statement

The only indented usage in that position is a fallthrough statement, as
a replacement for a /* FALLTHROUGH */ comment.  There is no point in
using _Noreturn or _Alignas as a statement.

Ideally, instead of gcc_attribute, the grammar would allow
gcc_attribute_list here (as always with GCC attributes), but that would
have introduced a reduce/reduce conflict, due to the attributes in the
middle of labeled_statement.

(rillig)

Comment that (buffer + 14) stands for reltab in bbstart.s.

(rin)

Now, cread.c can be built with -Os.

It was not a compiler bug actually; object code generated with
"-Os -Wa,-l" contains R_68K_16, which is not supported by elf2bb(1).

We should have fixed elf2bb(1), but the problem is gone at the moment.

(rin)

Compile boot.amiga without -l option for gas(1).

Fix build failure due to relocation overflows for R_68K_PC16.

(rin)

Fix boot.amiga when relocation table is located beyond 0x8000.

In this case, we cannot load reltab directly into %a2 by movw:

For m68k, movw is actually synonym for different instructions when
its destination is data or address register, respectively.

(1) For data register, it is for move.w, which loads source into lower
    half word with upper wharf word being kept untouched.

(2) For address register, it is movea.w, which loads source with
    *sign extension*.

XXX
We should really stop using movw, and use move.w or movea.w explicitly.
However, this will end up with BIG diff...

Thanks ryo@ and isaki@ for comments offlist.

(rin)

lint: do not allow type attributes after label

After a label, the type attributes _Alignas, __packed and _Noreturn do
not make sense, so don't allow them.  The only kind of attributes that
makes sense there syntactically is GCC __attribute__.  Allow multiple
such attributes, like everywhere in the GCC grammar.

(rillig)

lint: revert resolving grammar conflicts for labeled statements

Restore the grammar rule for labeled_statement as it was before cgram.y
1.400 from 2022-04-24.  This allows labels with attributes again.  Fix
the wrong interpretation in the tests; the attributes belong to the
label, not to the statement.

Today in the morning, when I thought that the change in cgram.y 1.400
were innocent, I accidentally ran lint only with the options '-Sw' but
forgot the option '-g' for GNU mode.  Without that option, the token
'__attribute__' is unknown, which unsurprisingly leads to lots of syntax
errors, and these didn't change with that commit.  The actual change was
only visible in GNU mode.

(rillig)

Make the thmap(9) used for params use sleepable allocations,
suggested by rmind@. Should fix PR 56802.

(martin)

Temporary hack to make PR 56802 (when it happens) tell us for sure that
it is caused by KM_NOSLEEP memory allocation failure.

(martin)

G/C

(skrll)

Add missing "show vmem[s]".

(msaitoh)

tests/lint: add test for GCC unused statement

Seen in xsrc/LRGB.c:799.

Even though it looks like this lint error may have been introduced by
cgram.y 1.400 from 2022-04-24 (since LRGB.c has been unchanged for a
month), earlier versions of lint produce the same parse errors.

(rillig)

Remove 'a' modifier from "show vmem". It was removed in sbur_vmem.c rev. 1.55.

(msaitoh)

be lazy about clearing the cursor - in most cases we're about to overwrite it
anyway. Same logic as with VCONS_DONT_READ.
visible speedup in things like systat and top

(macallan)

add locations where pkgsrc installs X11 fonts to the default font path

(macallan)

since pmap_activate_efirt() rewrites TTBR0, it is necessary to pmap_activate() again after pmap_deactivate_efirt() to restore the original TTBR0.

- Fix to do pmap_{de,}activate() before/after pmap_{,de}activate_efirt().
- moved kpreempt_{disable,enable}() to the caller since everything between
  arm_efirt_md_enter() and arm_efirt_md_exit() should be kpreempt disabled.

ok skrll@

(ryo)

I have been using this for a while and forgot to commit it.  This add
a number of KERNEL_LOCK() calls and at the very least prevents gpioiic
from panicing the system.  This may be a fix for kern/56438.

(brad)

After testing with a couple of more samples of the SHT3x sensor chip
it was found that the datasheet does not appear to provide enough
information to make use of the alarm and interrupt function work.  So
actually remove the dead code that was an attempt at making that all
work.  Adjust the man page to mention that this.

(brad)

Fix for PR 56243
clear the past EOL flags when moving the cursor - they are no longer
valid as move cannot put the cursor past the EOL.

(blymn)

As VOP_GETATTR() needs a shared lock at least move the preopattr lookup
inside nfs_namei() where we may lock the start directory without violating
the lock order.

(hannken)

Ticket #1739

(martin)

Pull up following revision(s) (requested by hannken in ticket #1739):

usr.sbin/quotaon/quotaon.c: revision 1.31
lib/libquota/quota_oldfiles.c: revision 1.10
sys/ufs/ufs/ufs_quota1.c: revision 1.25

Fix default quota file names, both user and group quota used the
same default name "QUOTAFILENAME, names[USRQUOTA])" resulting in
diagnostic assertion and possibly corrupted quota data.

Keep flag "UFS_QUOTA" set until the last quota is closed.
Prevents a live lock when dqrele() finds a struct with "dq_cnt == 1"
and flag "DQ_MOD" and cannot sync as flag UFS_QUOTA is unset.

As the quota type comes from the kernel and is only valid when
quota is on get the type before quota_off and after quota_on.

(martin)

Ticket #1440

(martin)

Pull up following revision(s) (requested by hannken in ticket #1440):

usr.sbin/quotaon/quotaon.c: revision 1.31
lib/libquota/quota_oldfiles.c: revision 1.10
sys/ufs/ufs/ufs_quota1.c: revision 1.25

Fix default quota file names, both user and group quota used the
same default name "QUOTAFILENAME, names[USRQUOTA])" resulting in
diagnostic assertion and possibly corrupted quota data.

Keep flag "UFS_QUOTA" set until the last quota is closed.
Prevents a live lock when dqrele() finds a struct with "dq_cnt == 1"
and flag "DQ_MOD" and cannot sync as flag UFS_QUOTA is unset.

As the quota type comes from the kernel and is only valid when
quota is on get the type before quota_off and after quota_on.

(martin)

Revert previous.

On 2022/04/27 21:10, Rin Okuyama wrote:
> Module Name: src
> Committed By: rin
> Date: Wed Apr 27 12:10:47 UTC 2022
>
> Modified Files:
> src/sys/arch/amiga/stand/bootblock/boot: Makefile
>
> Log Message:
> Sprinkle SA_HARDCODED_SECSIZE; now everything fits into
> +/-32KB range visible for PC relative addressing mode.
>
>
> To generate a diff of this commit:
> cvs rdiff -u -r1.57 -r1.58 src/sys/arch/amiga/stand/bootblock/boot/Makefile
>
> Please note that diffs are not public domain; they are subject to the
> copyright notices on the relevant files.

(rin)

Revert previous at the moment.

This is wrong reasoning; 68020 and above (incl. 040 and 060) support
32-bit displacements for PC relative addressing (via "fully extension
addressing mode" with null index register).

I've still not figured out what goes wrong with amiga/boot(8) when
compiled without -l option for gas(1)...

On 2022/04/27 20:48, Rin Okuyama wrote:
> Module Name: src
> Committed By: rin
> Date: Wed Apr 27 11:48:26 UTC 2022
>
> Modified Files:
> src/sys/lib/libsa: ext2fs.c minixfs3.c stand.h ufs.c
>
> Log Message:
> Introduce SA_HARDCODED_SECSIZE hack, by which hardcoded DEV_BSIZE is
> used instead of secsize obtained by SAIOSECSIZE ioctl.
>
> This hack avoids divdi3 and friends from being linked, in order to
> support variable secsize.
>
> This is useful for amiga/boot(8); it is loaded by firmware into
> unpredictable address, and therefore all symbols should be
> addressable by PC relative mode with only 16-bit displacements.
>
> See sys/arch/amiga/stand/bootblock/{boot/bbstart.s,elf2bb,txlt} for
> more details.
>
>
> To generate a diff of this commit:
> cvs rdiff -u -r1.31 -r1.32 src/sys/lib/libsa/ext2fs.c
> cvs rdiff -u -r1.10 -r1.11 src/sys/lib/libsa/minixfs3.c
> cvs rdiff -u -r1.83 -r1.84 src/sys/lib/libsa/stand.h src/sys/lib/libsa/ufs.c
>
> Please note that diffs are not public domain; they are subject to the
> copyright notices on the relevant files.

(rin)

Sprinkle SA_HARDCODED_SECSIZE; now everything fits into
+/-32KB range visible for PC relative addressing mode.

(rin)

Introduce SA_HARDCODED_SECSIZE hack, by which hardcoded DEV_BSIZE is
used instead of secsize obtained by SAIOSECSIZE ioctl.

This hack avoids divdi3 and friends from being linked, in order to
support variable secsize.

This is useful for amiga/boot(8); it is loaded by firmware into
unpredictable address, and therefore all symbols should be
addressable by PC relative mode with only 16-bit displacements.

See sys/arch/amiga/stand/bootblock/{boot/bbstart.s,elf2bb,txlt} for
more details.

(rin)

Add support for the mailbox that provides a communications channel with
additional cores integrated in Apple Silicon.

Also, add an implementation for the protocol to communicate with
coprocessors running firmware based Apple's RTKit OS.

From OpenBSD.

(skrll)

Sync with OpenBSD

- port initialisation and device power up. Latter requires the SMC
  GPIO controller which is WIP.

While here
- improve the MSI vector search algorithm; and
- spinkle some BITS(3)

(skrll)

Add a driver for the Apple GPIO controller found in Apple silicon.

Derived from the OpenBSD driver.

(skrll)

Add support for the power domain controller found in Apple silicon
machines.

(skrll)

Sync with OpenBSD

- Use subpage feature
- Catch up with new device tree bindings
- Skip locked DARTs
- Use bypass mode for USB3 DARTs

Other changes from me
- Sprinkle BITS(3)
- Improve interrupt error handling

(skrll)

Rename for AMD F15/6X device. No functional change.

(msaitoh)

Regen.

(msaitoh)

Update some AMD devices:

- Rename AMD 15h/6xh devices.
- Add some devices.
- Use OHCI for OHCI device.
- Use EHCI for EHCI device.
- and some minor changes.

(msaitoh)

Added new libcurses test.

(blymn)

Add test for wresize.

(blymn)

Fix for PR 55517.
Actually us x values to recalculate subwin columns.

(blymn)

As the quota type comes from the kernel and is only valid when
quota is on get the type before quota_off and after quota_on.

(hannken)

Keep flag "UFS_QUOTA" set until the last quota is closed.

Prevents a live lock when dqrele() finds a struct with "dq_cnt == 1"
and flag "DQ_MOD" and cannot sync as flag UFS_QUOTA is unset.

(hannken)

Fix default quota file names, both user and group quota used the
same default name "QUOTAFILENAME, names[USRQUOTA])" resulting in
diagnostic assertion and possibly corrupted quota data.

(hannken)

Avoid warning by clang of unused variable

While here, use the computated obj_size even though it's always the same as
its information length since we don't use extended attribute files or
stream directories.

(reinoud)

Fix endian bug also in commented out code

(reinoud)

Fix endian bug : Logblks recorded in fe/efe VAT is 64 bits

(reinoud)

Fix endian bug that resulted in UniqueIDs of 0 in file identifiers; note that
the the unique ID in a longad stored in a FID is 32 bit where in the file
descriptor (fe/efe) its 64 bits long.

(reinoud)

Fix endian bug in permission storage

(reinoud)

Fix endian bug with VAT 1.50 format

(reinoud)

No need to use ceilf() when ceil() is already used

(reinoud)

Fix endian bug in makefs for udf

(reinoud)

Fix typo. No funcitonal change.

(msaitoh)

Fix endian bug in logical volume creation for metadata/VAT partitions.

(reinoud)

One more style. No binary changes.

(rin)

Relax constraints on sequential media detection by checking for VAT format.
This allows fixing images of VAT formatted media too.

(reinoud)

Fix endian bug in descriptor CRC length on updating a FE/EFE

(reinoud)

libsa now needs ioctl to support media with large sectors. Provide
missing functions.

libsa defines files[] array itself, just reference it.
This also adds another global variable that tracks the root
filesystem module.

(mlelstv)

libsa now needs ioctl support to handle media with large sectors.
Provide dummy functions.

(mlelstv)

static-ify function bodies of usage() and eval().

No binary changes.

(rin)

One more style... No binary changes.

(rin)

Wrap looong lines, and misc style/cosmetic fixes.

No binary changes.

(rin)

Do not cast return value from malloc(3).

No binary changes.

(rin)

- static-ify usage(), intcmp(), and eval()
- remove prototype declaration for main()

NFCI; no binary changes observed for amd64 with -O0.

(rin)

Use htobe{16,32}(9) instead of be{16,32}toh(9) where appropriate.

No binary changes both for little and big endian machines.

(rin)

u_intN_t --> uintN_t

No binary changes.

(rin)

Ticket #1439

(martin)

Pull up following revision(s) (requested by nia in ticket #1439):

xsrc/external/mit/xinit/dist/startx.cpp: revision 1.9

Do not attempt to trap SIGKILL

(martin)

Add missing num_{tx,rx}_desc sysctl to sync with ixgbe.c.

(msaitoh)

Add missing num_tx_desc sysctl.

(msaitoh)

Limit {tx,rx}_process_limit sysctl values from 1 to num_{tx,rx}_desc.

(msaitoh)

Use cached rx_copy_len in ixgbe_rxeof().

(msaitoh)

Modify comment for consistency. No functional change.

(msaitoh)