LLVM/project 1a35a3fllvm/lib/Target/AMDGPU SIInstructions.td, llvm/test/CodeGen/AMDGPU fneg-fabs-v2f32.ll packed-fp32.ll

[AMDGPU] Use two v_xor_b32 instructions in fneg v2f32 selection pattern (#195585)

'fneg x'  and 'fadd 0.0, -x' are not strictly identical. For example,
'fneg 0.0' may result in '-0.0' but 'fadd 0.0, -0.0' may not.
DeltaFile
+14-8llvm/test/CodeGen/AMDGPU/fneg-fabs-v2f32.ll
+5-3llvm/lib/Target/AMDGPU/SIInstructions.td
+4-2llvm/test/CodeGen/AMDGPU/packed-fp32.ll
+23-133 files

OPNSense/core 1bf1c69src/opnsense/mvc/app/controllers/OPNsense/CaptivePortal/Api ServiceController.php

Captive Portal: reload ipfw on captive portal reconfigure (#10253)
DeltaFile
+19-0src/opnsense/mvc/app/controllers/OPNsense/CaptivePortal/Api/ServiceController.php
+19-01 files

LLVM/project 4e33a3boffload/plugins-nextgen/level_zero/src L0Kernel.cpp

Respect group size limit for kernel (#195655)

The case with user-defined group sizes still needs to respect the group
size attribute for the kernel. Level Zero will return error otherwise.
DeltaFile
+5-3offload/plugins-nextgen/level_zero/src/L0Kernel.cpp
+5-31 files

LLVM/project 0fd3e26llvm/lib/CodeGen/SelectionDAG TargetLowering.cpp, llvm/lib/Target/ARM README.txt

[ARM] Optimize BSwap for ARM once again (#195347)

We can now use the optimal ARM sequence.
DeltaFile
+32-38llvm/test/CodeGen/ARM/load-combine-big-endian.ll
+24-28llvm/test/CodeGen/ARM/load-combine.ll
+0-26llvm/lib/Target/ARM/README.txt
+14-10llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
+70-1024 files

FreeNAS/freenas c1b54f3src/middlewared/middlewared/utils/filesystem directory.py, tests/api2 test_190_filesystem.py

NAS-140899 / 27.0.0-BETA.1 / Fix listdir contents for child datasets (#18892)

This commit fixes directory listing such that it includes child
filesystems by using the recently-added
include_mountpoints keyword argument for iter_filesystem_contents.
DeltaFile
+54-32src/middlewared/middlewared/utils/filesystem/directory.py
+25-4tests/api2/test_190_filesystem.py
+79-362 files

FreeNAS/freenas afd70f4src/middlewared/middlewared/plugins boot.py, src/middlewared/middlewared/plugins/system_advanced debug_kernel.py config.py

NAS-140884 / 27.0.0-BETA.1 / write debug kernel to initramfs file (#18886)

Same justification for doing this as outlined in
https://github.com/truenas/middleware/pull/18878
DeltaFile
+65-0src/middlewared/middlewared/plugins/system_advanced/debug_kernel.py
+0-9src/middlewared/middlewared/plugins/boot.py
+6-0src/middlewared/middlewared/plugins/system_advanced/config.py
+71-93 files

LLVM/project 665984fclang/lib/AST/ByteCode Interp.h, clang/test/AST/ByteCode vectors.cpp

[clang][bytecode] Check primtypes in CopyArray op (#195628)
DeltaFile
+9-0clang/test/AST/ByteCode/vectors.cpp
+6-0clang/lib/AST/ByteCode/Interp.h
+15-02 files

FreeNAS/freenas bbf248etests/unit/zfs test_zvol_block_cloning.py

NAS-140870 / 26.0.0-RC.1 / Add zvol block cloning test (by ixhamza) (#18893)

zvol block cloning depends on [custom TrueNAS kernel patches that add
blkdev_copy_offload() to the block
layer](https://github.com/truenas/linux/pull/203). The ZFS code path is
gated behind `HAVE_BLKDEV_COPY_OFFLOAD` and compiles out on stock
kernels, and there is no existing automated coverage of this path.

The test creates two zvols, issues `copy_file_range`, and asserts via
zdb that exactly 64 L0 blocks share identical DVAs between src and dst.
Identical DVAs prove the destination references the source's blocks
rather than holding an independent copy.

Original PR: https://github.com/truenas/middleware/pull/18877

Co-authored-by: Ameer Hamza <ahamza at ixsystems.com>
DeltaFile
+97-0tests/unit/zfs/test_zvol_block_cloning.py
+97-01 files

LLVM/project dd8be88libc/src/__support/FPUtil FEnvImpl.h, libc/src/__support/math log_bf16.h

bfloat16
DeltaFile
+211-7libc/test/shared/CMakeLists.txt
+156-3libc/test/shared/shared_math_constexpr_test.cpp
+3-2libc/src/__support/FPUtil/FEnvImpl.h
+2-1libc/test/shared/shared_math_test.cpp
+1-1libc/src/__support/math/log_bf16.h
+373-145 files

NetBSD/src 8fGcIMXshare/mk bsd.own.mk

   Pull up following revision(s) (requested by martin in ticket #275):
        share/mk/bsd.own.mk: 1.1481 via patch
   PR 58762: disable MKCOMPAT for earm*.
   If someone is interested in re-adding support for oabi compat library
   builds, they can figure out the missing bits. But for now, stop producing
   bogus compat32/debug32 sets on all evbarm builds.
VersionDeltaFile
1.1428.2.3+2-3share/mk/bsd.own.mk
+2-31 files

LLVM/project c252557flang-rt/lib/runtime extensions.cpp, flang-rt/unittests/Runtime Time.cpp

[flang-rt] Fix for TIMEF test case hanging on Windows (#195651)

The merging of PR https://github.com/llvm/llvm-project/pull/185377
caused a test case to hang on Windows. This PR fixes the same by adding
an upper bound on the number of iterations of the test-loop, post which
the test gracefully exits irrespective of whether the TIMEF
functionality is supported on Windows or not.
DeltaFile
+9-2flang-rt/unittests/Runtime/Time.cpp
+2-4flang-rt/lib/runtime/extensions.cpp
+11-62 files

FreeNAS/freenas 086ee81src/middlewared/middlewared/test/integration/assets iscsi.py, tests/sharing_protocols/iscsi test_264_iscsi_mode_compat.py test_265_iscsi_portal_binding.py

Skip LIO CI tests for now
DeltaFile
+34-0src/middlewared/middlewared/test/integration/assets/iscsi.py
+9-0tests/sharing_protocols/iscsi/test_264_iscsi_mode_compat.py
+9-0tests/sharing_protocols/iscsi/test_265_iscsi_portal_binding.py
+52-03 files

LLVM/project b07d2a9llvm/lib/IR Constants.cpp, llvm/test/Assembler aggregate-constant-values.ll

[RFC][IR] Support vector splats in `ConstantPointerNull`

This PR allows `ConstantPointerNull` to represent both scalar pointer nulls and
fixed or scalable vector splats of pointer nulls. This change first aligns with
the native splat behavior of `ConstantInt` and `ConstantFP`, and second, makes
it easier to eventually change the semantics of `ConstantPointerNull` to
represent a semantic null pointer instead of a zero value, which is what it
represents today.
DeltaFile
+31-31llvm/test/Transforms/RewriteStatepointsForGC/base-vector.ll
+30-30llvm/test/Transforms/LoopVectorize/X86/masked_load_store.ll
+49-4llvm/lib/IR/Constants.cpp
+28-0llvm/unittests/IR/ConstantsTest.cpp
+18-9llvm/test/Transforms/RewriteStatepointsForGC/base-inference.ll
+24-0llvm/test/Assembler/aggregate-constant-values.ll
+180-7463 files not shown
+351-22269 files

NetBSD/src D8NpTj1usr.bin/cksum cksum.c

   Pull up following revision(s) (requested by martin in ticket #274):
        usr.bin/cksum/cksum.c: 1.53
   PR 60154: do not print arbitrary control characters when printing
   file names while checking hashes with the -c option.
VersionDeltaFile
1.52.6.1+15-4usr.bin/cksum/cksum.c
+15-41 files

FreeBSD/ports 1d9bfe6devel/qtcreator distinfo Makefile

devel/qtcreator: Update to 18.0.2

PR:             294548
Submitted by:   Chad Jacob Milios

(cherry picked from commit 8c783c802a326d995674570180ffa39aa46d3ad6)
DeltaFile
+3-3devel/qtcreator/distinfo
+1-1devel/qtcreator/Makefile
+1-0devel/qtcreator/pkg-plist
+5-43 files

FreeBSD/ports 8c783c8devel/qtcreator distinfo Makefile

devel/qtcreator: Update to 18.0.2

PR:             294548
Submitted by:   Chad Jacob Milios
DeltaFile
+3-3devel/qtcreator/distinfo
+1-1devel/qtcreator/Makefile
+1-0devel/qtcreator/pkg-plist
+5-43 files

LLVM/project 3efe6d4llvm/lib/Transforms/Vectorize VPlanTransforms.cpp, llvm/test/Transforms/LoopVectorize multiple-early-exits.ll optimal-epilog-vectorization-liveout.ll

[LV] Remove redundant parameter in IV live-out optimization (#195036)

When optimizing induction variable live-out values, passing PredVPBB is
no longer necessary as the VPBasicBlock can be retrieved from the
incoming value of exit phis.

This patch ensures the escapee of induction variable is inserted before
the recipe that defines the incoming value, and removes the redundant
PredVPBB parameter passing.
DeltaFile
+17-17llvm/test/Transforms/LoopVectorize/X86/float-induction-x86.ll
+17-17llvm/test/Transforms/LoopVectorize/multiple-early-exits.ll
+12-12llvm/lib/Transforms/Vectorize/VPlanTransforms.cpp
+8-8llvm/test/Transforms/LoopVectorize/AArch64/epilog-iv-live-outs.ll
+8-8llvm/test/Transforms/LoopVectorize/optimal-epilog-vectorization-liveout.ll
+4-4llvm/test/Transforms/LoopVectorize/pointer-induction.ll
+66-664 files not shown
+76-7610 files

LLVM/project 8116195clang/include/clang/Options Options.td, clang/test/Driver cl-options.c

[clang] Expose -f(no-)math-errno as a clang-cl option (#195455)

Vaguely related:
https://developercommunity.visualstudio.com/t/Add-a-flag-similar-to--fno-math-errno-th/10710444
DeltaFile
+2-2clang/include/clang/Options/Options.td
+2-0clang/test/Driver/cl-options.c
+4-22 files

FreeNAS/freenas f8ae364tests/unit/zfs test_zvol_block_cloning.py

Add zvol block cloning test

zvol block cloning depends on custom TrueNAS kernel patches that add
blkdev_copy_offload() to the block layer. The ZFS code path is gated
behind HAVE_BLKDEV_COPY_OFFLOAD and compiles out on stock kernels, and
there is no existing automated coverage of this path.

The test creates two zvols, issues copy_file_range, and asserts via zdb
that exactly 64 L0 blocks share identical DVAs between src and dst.
Identical DVAs prove the destination references the source's blocks
rather than holding an independent copy.

(cherry picked from commit e15c42bf0d534d64c358814bec40712ed2be0bee)
DeltaFile
+97-0tests/unit/zfs/test_zvol_block_cloning.py
+97-01 files

FreeNAS/freenas ea1dad1tests/unit/zfs test_zvol_block_cloning.py

NAS-140870 / 27.0.0-BETA.1 / Add zvol block cloning test (#18877)

zvol block cloning depends on [custom TrueNAS kernel patches that add
blkdev_copy_offload() to the block
layer](https://github.com/truenas/linux/pull/203). The ZFS code path is
gated behind `HAVE_BLKDEV_COPY_OFFLOAD` and compiles out on stock
kernels, and there is no existing automated coverage of this path.

The test creates two zvols, issues `copy_file_range`, and asserts via
zdb that exactly 64 L0 blocks share identical DVAs between src and dst.
Identical DVAs prove the destination references the source's blocks
rather than holding an independent copy.
DeltaFile
+97-0tests/unit/zfs/test_zvol_block_cloning.py
+97-01 files

LLVM/project fca00ablibc/src/__support/FPUtil FEnvImpl.h, libc/src/__support/math log_bf16.h

bfloat16
DeltaFile
+211-7libc/test/shared/CMakeLists.txt
+156-3libc/test/shared/shared_math_constexpr_test.cpp
+2-2libc/src/__support/FPUtil/FEnvImpl.h
+2-1libc/test/shared/shared_math_test.cpp
+1-1libc/src/__support/math/log_bf16.h
+372-145 files

FreeBSD/ports a6a3edearchivers/7-zip distinfo Makefile, archivers/7-zip/files patch-C_CpuArch.h

archivers/7-zip: Update to 26.01

Drop stale patch.

Release notes:
https://sourceforge.net/p/sevenzip/discussion/45797/thread/555e132ba4/

(cherry picked from commit 9f60d78bf6d9120158108f9423bf736c3d1f3a7a)
DeltaFile
+0-10archivers/7-zip/files/patch-C_CpuArch.h
+3-3archivers/7-zip/distinfo
+1-2archivers/7-zip/Makefile
+4-153 files

FreeBSD/ports 9bba5dcnet-im/kaidan Makefile

net-im/kaidan: Rebuild after net-im/qxmpp update

(cherry picked from commit 55db57431bac29d569ee4a540ffd84f60777d280)
DeltaFile
+1-0net-im/kaidan/Makefile
+1-01 files

FreeBSD/ports a89c86bnet-im/qxmpp distinfo Makefile

net-im/qxmpp: Update to 1.15.1

(cherry picked from commit cb6dc1a6c62e36ef1445773a3f21d6b9753db667)
DeltaFile
+3-3net-im/qxmpp/distinfo
+2-2net-im/qxmpp/Makefile
+5-52 files

FreeBSD/ports 983f2b3sysutils/polkit-qt-1 pkg-plist distinfo

sysutils/polkit-qt-1: Update to 0.201.1

(cherry picked from commit 79dcbfc4c8f73b998d574ca73f4f871223063761)
DeltaFile
+3-3sysutils/polkit-qt-1/pkg-plist
+3-3sysutils/polkit-qt-1/distinfo
+1-1sysutils/polkit-qt-1/Makefile
+7-73 files

FreeBSD/ports c3827c3Mk/Uses mlt.mk, multimedia/mlt7 pkg-plist Makefile

multimedia/mlt7*: Update to 7.38.0

Release notes: https://github.com/mltframework/mlt/releases/tag/v7.38.0

multimedia/mlt7:
 - Add OPENFX option

PR:             294868
(cherry picked from commit e1202f0b10f5a0a44f40c611f250fb64acf623fc)
DeltaFile
+12-0multimedia/mlt7/pkg-plist
+5-2multimedia/mlt7/Makefile
+3-3multimedia/mlt7/distinfo
+1-1Mk/Uses/mlt.mk
+1-0multimedia/mlt7-qt/pkg-plist
+22-65 files

FreeBSD/ports caef895astro/kstars Makefile, astro/kstars/files patch-kstars_catalogsdb_sqlstatements.cpp

astro/kstars: fix string quoting for SQL queries

PR:             282980
Submitted by:   Raúl Zighelboim

(cherry picked from commit 97a1f0777a61844ae0410e1c243bd3abf5ca5cdd)
DeltaFile
+11-0astro/kstars/files/patch-kstars_catalogsdb_sqlstatements.cpp
+1-1astro/kstars/Makefile
+12-12 files

FreeNAS/freenas acfad9bsrc/middlewared/middlewared/test/integration/assets iscsi.py, tests/sharing_protocols/iscsi test_264_iscsi_mode_compat.py test_265_iscsi_portal_binding.py

Skip LIO CI tests for now
DeltaFile
+35-0src/middlewared/middlewared/test/integration/assets/iscsi.py
+9-0tests/sharing_protocols/iscsi/test_264_iscsi_mode_compat.py
+9-0tests/sharing_protocols/iscsi/test_265_iscsi_portal_binding.py
+53-03 files

OpenBSD/src EwOumB6regress/lib/libcrypto/certs README

   certs/README: fix previous: 14b should fail to verify
VersionDeltaFile
1.4+1-1regress/lib/libcrypto/certs/README
+1-11 files

LLVM/project 309948aclang/lib/CIR/CodeGen CIRGenBuiltinNVPTX.cpp

fix fmt
DeltaFile
+4-6clang/lib/CIR/CodeGen/CIRGenBuiltinNVPTX.cpp
+4-61 files