HardenedBSD/src cd2100astand/powerpc/ofw cas.c, sys/dev/dwc if_dwc.c dwc1000_dma.c

Merge branch 'freebsd/current/main' into hardened/current/master
DeltaFile
+10-4sys/dev/dwc/if_dwc.c
+3-3sys/dev/dwc/dwc1000_dma.c
+1-1stand/powerpc/ofw/cas.c
+1-1sys/dev/dwc/dwc1000_core.c
+1-1tests/sys/netpfil/pf/pflog.sh
+16-105 files

LLVM/project 38be580.github/workflows release-binaries.yml

workflows/release-binaries: Fix attestation artifact name (#176417)

We were contructing the attestation artifact name using the arch and the
OS of the current runner instead of using the runner that the builds
were done on. This led to a conflict in artifact names between all the
release binary jobs.
DeltaFile
+3-1.github/workflows/release-binaries.yml
+3-11 files

NetBSD/pkgsrc lkyUf9Zlang/rust options.mk

   lang/rust: Fix build under NetBSD 9
VersionDeltaFile
1.51+5-1lang/rust/options.mk
+5-11 files

LLVM/project 1529433clang/lib/CodeGen CodeGenModule.cpp, clang/test/CodeGenOpenCL amdgpu-features.cl amdgpu-features-default-delta.cl

[RFC][Clang][AMDGPU] Emit only delta target-features to reduce IR bloat

Currently, AMDGPU functions have `target-features` attribute populated with all default features for the target GPU. This is redundant because the backend can derive these defaults from the `target-cpu` attribute via `AMDGPUTargetMachine::getFeatureString()`.

In this PR, for AMDGPU targets only:

- Functions without explicit target attributes no longer emit `target-features`
- Functions with `__attribute__((target(...)))` or `-target-feature` emit only features that differ from the target's defaults (delta)

The backend already handles missing `target-features` correctly by falling back to the TargetMachine's defaults.

A new cc1 flag `-famdgpu-emit-full-target-features` is added to emit full features when needed.

Example:

Before:

```llvm
attributes #0 = { "target-cpu"="gfx90a" "target-features"="+16-bit-insts,+atomic-buffer-global-pk-add-f16-insts,+atomic-fadd-rtn-insts,+ci-insts,+dl-insts,+dot1-insts,+dot2-insts,..." }

    [13 lines not shown]
DeltaFile
+341-12clang/test/OpenMP/amdgcn-attributes.cpp
+53-53clang/test/CodeGenOpenCL/amdgpu-features.cl
+70-0clang/test/CodeGenOpenCL/amdgpu-features-default-delta.cl
+43-5clang/lib/CodeGen/CodeGenModule.cpp
+4-6clang/test/CodeGenOpenCL/amdgpu-readonly-features-written-with-no-target.cl
+4-4clang/test/CodeGenOpenCL/amdgpu-enqueue-kernel.cl
+515-804 files not shown
+528-8310 files

NetBSD/src H2zQDmodistrib/luna68k Makefile, distrib/luna68k/liveimage Makefile diskproto.in

   distrib/luna68k: add "build.sh live-image" support for luna68k

   Using recently added "USE_MDBOOTPART" in Makefile.bootimage with MD rules.
   A generated live-image is confirmed working on nono 1.6.4 with LUNA's ROM:
   ---
   % nono-cli -V vmtype=luna -V prom-image=LUNAROM.DAT -V dipsw-autoboot=no -V dipsw-serial=yes -V hostcom-driver=stdio -V spc0-id6-image=hd,NetBSD-11.99.5-luna68k-live.img -V spc0-id6-writeignore=1
   (PROM) Unknown PROM version
   (SCSIHD6) write is ignored
   Diagnostic Start....  11.549'836'096 805b XP(HD647180) WriteInternalIO 53H(DERA) <- $07 (NOT IMPLEMENTED)
     11.549'839'519 8060 XP(HD647180) WriteInternalIO 60H(IODRA) <- $50 (NOT IMPLEMENTED)
     11.549'842'942 8065 XP(HD647180) WriteInternalIO 70H(DDRA) <- $70 (NOT IMPLEMENTED)
     11.650'402'532 8528 XP(HD647180) WriteInternalIO 60H(IODRA) <- $20 (NOT IMPLEMENTED)
     11.653'401'080 8bd1 XP(HD647180) ReadInternalIO 48H(T2CSR1) -> $ff (NOT IMPLEMENTED)
     11.653'404'340 8bd6 XP(HD647180) WriteInternalIO 48H(T2CSR1) <- $ff (NOT IMPLEMENTED)
     11.653'406'785 8bd9 XP(HD647180) ReadInternalIO 49H(T2CSR2) -> $ff (NOT IMPLEMENTED)
     11.653'410'045 8bde XP(HD647180) WriteInternalIO 49H(T2CSR2) <- $ff (NOT IMPLEMENTED)
   No problems in hardware
   OMRON WS Monitor Ver4.25 (Thu May 16 11:54:02 1991) - 0x01000000 bytes of memory
   "BSD Monitor-mode"

    [38 lines not shown]
VersionDeltaFile
1.1+51-0distrib/luna68k/liveimage/Makefile
1.1+25-0distrib/luna68k/liveimage/diskproto.in
1.1+7-0distrib/luna68k/liveimage/fstab.in
1.3+5-1distrib/luna68k/Makefile
1.1+2-0distrib/luna68k/liveimage/spec.boot.in
1.1+2-0distrib/luna68k/liveimage/spec.in
+92-16 files

NetBSD/src VPhjFPBdistrib/common/bootimage Makefile.bootimage

   bootimage: allow MD Makefiles to add a port-specific boot partition

   Some ports require an additional port-specific boot partition even
   for live images so that firmware ROMs can load a bootloader.
   For example, luna68k ROMs require old 4.3-BSD style FFS, and hp300
   ROMs require an HP LIF format diretory entries.

   Add a USE_MDBOOTPART switch and adjust size/offset calculations and
   image concatination so MD liveimage Makefiles can provide own rules
   to build and populate the boot-partition image.
VersionDeltaFile
1.50+41-5distrib/common/bootimage/Makefile.bootimage
+41-51 files

NetBSD/pkgsrc-wip d49c79apy-identify distinfo COMMIT_MSG

py-identify: update to 2.6.16
DeltaFile
+3-3py-identify/distinfo
+1-1py-identify/COMMIT_MSG
+1-1py-identify/Makefile
+5-53 files

NetBSD/pkgsrc-wip 42e103aiosevka-curly-slab-ttf distinfo

iosevka-curly-slab-ttf: Correct distinfo hashes
DeltaFile
+2-2iosevka-curly-slab-ttf/distinfo
+2-21 files

FreeBSD/ports 5980368audio/ocp distinfo Makefile

audio/ocp: Update to 3.1.1
DeltaFile
+3-3audio/ocp/distinfo
+1-4audio/ocp/Makefile
+4-72 files

NetBSD/src NiZs7s0tools/gcc mknative-gcc Makefile

   prepare for /usr/bin/gfortran.

   in the Makefile, add fortran to the list of languages to configure.
   add configure-target-libquadmath and configure-target-libgfortran
   to the listof mknative targets to invoke.

   while here also remove some HAVE_GCC < 12 support.

   in mknative-gcc, remove ancient dead get_libg2c(), grab a couple
   more libgomp variables (will be used to replace some hard coded
   things and also adds the gomp/fortran portions), add new get
   functions for libquadmath and libgfortran, and also fetch the rest
   of the necessary variables.
VersionDeltaFile
1.123+44-22tools/gcc/mknative-gcc
1.119+12-5tools/gcc/Makefile
+56-272 files

LLVM/project 739e997libc/src/__support/GPU allocator.cpp

[libc] Remove ballot on slab find (#176606)

Summary:
This negatively impacts performance, while the other changes in the
initial PR slightly improved it. This was originally done to make Volta
independent thread scheduling work, but that doesn't seem to work
correctly all the time either so we should make this faster.
DeltaFile
+9-8libc/src/__support/GPU/allocator.cpp
+9-81 files

HardenedBSD/src 964d91etests/sys/netpfil/pf pflog.sh

pflog: tests: Fix rdr_action_head()

Fix a typo in the rdr_action_head() test.

Fixes:          685fb4253819 ("pf: Log the intended action when a NAT rule matches a packet")
MFC after:      1 week
DeltaFile
+1-1tests/sys/netpfil/pf/pflog.sh
+1-11 files

FreeBSD/src 964d91etests/sys/netpfil/pf pflog.sh

pflog: tests: Fix rdr_action_head()

Fix a typo in the rdr_action_head() test.

Fixes:          685fb4253819 ("pf: Log the intended action when a NAT rule matches a packet")
MFC after:      1 week
DeltaFile
+1-1tests/sys/netpfil/pf/pflog.sh
+1-11 files

LLVM/project 49cd842mlir/utils/vscode package-lock.json

[mlir][vscode] Remove resolved from lock file (#176611)

DeltaFile
+457-4,789mlir/utils/vscode/package-lock.json
+457-4,7891 files

LLVM/project 032eb06clang/test/CodeGenOpenCL amdgpu-features-illegal.cl, clang/test/SemaOpenCL builtins-amdgcn-error-wave32.cl builtins-amdgcn-wave32-func-attr.cl

[Clang][AMDGPU] Handle `wavefrontsize32` and `wavefrontsize64` features more robustly

We should also not allow `-wavefrontsize32` and `-wavefrontsize64` to be specified at the same time.
DeltaFile
+30-9llvm/lib/TargetParser/TargetParser.cpp
+14-7clang/test/CodeGenOpenCL/amdgpu-features-illegal.cl
+6-2flang/test/Driver/target-cpu-features-invalid.f90
+2-2clang/test/SemaOpenCL/builtins-amdgcn-error-wave32.cl
+2-2clang/test/SemaOpenCL/builtins-amdgcn-wave32-func-attr.cl
+54-225 files

LLVM/project d3dad64llvm/utils/gn/secondary/clang/lib/Analysis BUILD.gn

[gn build] Port 17ff9b3c67ab
DeltaFile
+0-1llvm/utils/gn/secondary/clang/lib/Analysis/BUILD.gn
+0-11 files

LLVM/project b8e3276mlir/utils/vscode package-lock.json

[mlir][vscode] Update lock file to match (#176608)

DeltaFile
+5,759-343mlir/utils/vscode/package-lock.json
+5,759-3431 files

NetBSD/pkgsrc oH1Rvs2doc CHANGES-2026

   Updated devel/p5-App-Cmd to 0.338
VersionDeltaFile
1.393+2-1doc/CHANGES-2026
+2-11 files

NetBSD/pkgsrc 1X91uFjdevel/p5-App-Cmd distinfo Makefile

   Update to 0.338

   Upstream changes:
   0.338     2025-10-03 15:56:06-04:00 America/New_York
           - add experimental --for-completion option to "commands" to help make
             shell completion; THIS MAY CHANGE WEIRDLY
VersionDeltaFile
1.33+4-4devel/p5-App-Cmd/distinfo
1.54+2-3devel/p5-App-Cmd/Makefile
+6-72 files

LLVM/project 676b292mlir/utils/vscode package.json

[mlir][vscode] Update engine (#176605)

DeltaFile
+1-1mlir/utils/vscode/package.json
+1-11 files

LLVM/project 462673amlir/utils/vscode package.json

[mlir][vscode] Update dev dependencies (#176604)

Update to match vscode-clangd.
DeltaFile
+4-4mlir/utils/vscode/package.json
+4-41 files

LLVM/project 17ff9b3clang-tools-extra/clang-tidy/bugprone UnsafeFunctionsCheck.cpp, clang/include/clang/Analysis AnnexKDetection.h

Revert "[clang][analyzer] Add ReportInC99AndEarlier option to DeprecatedOrUnsafeBuf…" (#176603)

Reverts llvm/llvm-project#168704
Checking what causes the clang-bolt buildbot failure.
DeltaFile
+12-76clang/lib/StaticAnalyzer/Checkers/CheckSecuritySyntaxOnly.cpp
+0-43clang/lib/Analysis/AnnexKDetection.cpp
+0-40clang/include/clang/Analysis/AnnexKDetection.h
+0-40clang/test/Analysis/security-deprecated-buffer-handling-report-modes.c
+6-19clang/include/clang/StaticAnalyzer/Checkers/Checkers.td
+20-2clang-tools-extra/clang-tidy/bugprone/UnsafeFunctionsCheck.cpp
+38-2204 files not shown
+38-23610 files

LLVM/project d5e14afmlir/utils/vscode language-configuration.json package.json

[mlir][vscode] Add angle bracket support to MLIR language configuration (#176602)

Add angle brackets (<>) to brackets, autoClosingPairs, and
surroundingPairs for better editing of types like tensor<3xf32>. Also
add colorizedBracketPairs for visual distinction between nested bracket
types.
DeltaFile
+10-1mlir/utils/vscode/language-configuration.json
+1-1mlir/utils/vscode/package.json
+11-22 files

NetBSD/pkgsrc V9SHniawww/w3m/patches patch-etc.c

   w3m: remove merged patch
VersionDeltaFile
1.2+1-1www/w3m/patches/patch-etc.c
+1-11 files

NetBSD/pkgsrc XOIYvlmdoc CHANGES-2026

   doc: Updated www/w3m-img to 0.5.5
VersionDeltaFile
1.392+2-1doc/CHANGES-2026
+2-11 files

NetBSD/pkgsrc g4suz4Jwww/w3m-img Makefile

   w3m-img: update to 0.5.5.

   See w3m for details.
VersionDeltaFile
1.36+2-3www/w3m-img/Makefile
+2-31 files

LLVM/project f64b69bmlir/utils/vscode pdll-grammar.json

[mlir][vscode] Fix PDLL grammar character class regex (#176601)

The character class [aA-zZ_0-9] incorrectly matches characters between
ASCII 90-97 (Z-a range), which includes: [ \ ] ^ _ `. This should be
[a-zA-Z_0-9] for proper identifier matching.
DeltaFile
+8-8mlir/utils/vscode/pdll-grammar.json
+8-81 files

NetBSD/pkgsrc OFaifegdoc CHANGES-2026

   doc: Updated www/w3m to 0.5.5
VersionDeltaFile
1.391+2-1doc/CHANGES-2026
+2-11 files

NetBSD/pkgsrc fwslfsJwww/w3m Makefile.common distinfo, www/w3m/patches patch-file.c patch-fm.h

   w3m: update to 0.5.5.

   w3m 0.5.5 - 2025-08-20

   * new feature
     - Use gettext everywhere, allow for more localization

   * bug fixes
     - Prevent segfault when user aborts session store command
     - Fix cookie sync logic

   w3m 0.5.4 - 2025-08-03

   * deprecated
     - option -H Users should use -o high-intensity=true instead
     - ChangeLog file, use the git log for all the details

   * new features
     - Add command to save the history (SAVE_HISTORY)

    [49 lines not shown]
VersionDeltaFile
1.2+16-17www/w3m/patches/patch-file.c
1.82+9-19www/w3m/Makefile.common
1.43+8-9www/w3m/distinfo
1.2+6-6www/w3m/patches/patch-fm.h
1.2+5-5www/w3m/patches/patch-Makefile.in
1.2+3-3www/w3m/patches/patch-Str.c
+47-592 files not shown
+49-648 files

LLVM/project 727f8f9clang/test/CodeGenOpenCL amdgpu-features-illegal.cl, clang/test/SemaOpenCL builtins-amdgcn-error-wave32.cl

[Clang][AMDGPU] Handle `wavefrontsize32` and `wavefrontsize64` features more robustly

We should also not allow `-wavefrontsize32` and `-wavefrontsize64` to be specified at the same time.
DeltaFile
+18-7llvm/lib/TargetParser/TargetParser.cpp
+8-5clang/test/CodeGenOpenCL/amdgpu-features-illegal.cl
+1-1clang/test/SemaOpenCL/builtins-amdgcn-error-wave32.cl
+27-133 files