FreeBSD/src 56275b7usr.sbin/bsdinstall Makefile

bsdinstall: Ignore -p[0-9]+ in determining BRANCH

The patch level is not part of the branch per se and should not be
used in constructing the FreeBSD-base.conf file used by bsdinstall.

MFC after:      1 day

(cherry picked from commit d45816f369eb0099fd274fee23a3bf137ee794b1)
DeltaFile
+1-1usr.sbin/bsdinstall/Makefile
+1-11 files

LLVM/project bd643bcflang/include/flang/Optimizer/Transforms Passes.h Passes.td, flang/lib/Optimizer/Transforms FIRToSCF.cpp

[flang] Use default constructor for FIRToSCF pass (#169741)

DeltaFile
+2-4flang/lib/Optimizer/Transforms/FIRToSCF.cpp
+0-1flang/include/flang/Optimizer/Transforms/Passes.h
+0-1flang/include/flang/Optimizer/Transforms/Passes.td
+2-63 files

LLVM/project b028daclibcxx/include queue, libcxx/test/libcxx/diagnostics queue.nodiscard.verify.cpp

[libc++][queue] Applied `[[nodiscard]]` (#169469)

`[[nodiscard]]` should be applied to functions where discarding the
return value is most likely a correctness issue.

-
https://libcxx.llvm.org/CodingGuidelines.html#apply-nodiscard-where-relevant
DeltaFile
+19-7libcxx/test/libcxx/diagnostics/queue.nodiscard.verify.cpp
+9-7libcxx/include/queue
+28-142 files

LLVM/project 6910813mlir/lib/Transforms/Utils DialectConversion.cpp

address comments
DeltaFile
+4-3mlir/lib/Transforms/Utils/DialectConversion.cpp
+4-31 files

LLVM/project ceba82fllvm/lib/Transforms/Vectorize LoadStoreVectorizer.cpp, llvm/test/Transforms/LoadStoreVectorizer/AMDGPU vectorize-redund-loads.ll

[LoadStoreVectorizer] Fix one-element vector handling (#169671)

This is the followup of https://github.com/llvm/llvm-project/pull/168135
DeltaFile
+27-0llvm/test/Transforms/LoadStoreVectorizer/AMDGPU/vectorize-redund-loads.ll
+4-4llvm/lib/Transforms/Vectorize/LoadStoreVectorizer.cpp
+31-42 files

LLVM/project 1ff5c89mlir/test/Analysis/DataFlow test-liveness-analysis.mlir, mlir/test/lib/Analysis/DataFlow TestLivenessAnalysis.cpp

[mlir][dataflow] Add arguemnt print for test-liveness-analysis (#169625)

Add arguemnt print for test-liveness-analysis to better debug
remove-dead-values pass.

---------

Co-authored-by: Mehdi Amini <joker.eph at gmail.com>
DeltaFile
+13-1mlir/test/Analysis/DataFlow/test-liveness-analysis.mlir
+11-0mlir/test/lib/Analysis/DataFlow/TestLivenessAnalysis.cpp
+24-12 files

LLVM/project 48a9b07llvm/lib/Target/AMDGPU AMDGPUCodeGenPrepare.cpp

[AMDGPU] Remove unused functions isSigned. NFC (#169750)

These have been unused since
https://github.com/llvm/llvm-project/pull/145483.
DeltaFile
+0-18llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
+0-181 files

LLVM/project e2a29ecclang/lib/CodeGen CGExpr.cpp, clang/test/Driver fsanitize.c

[UBSan] Use -fsanitize-handler-preserve-all-regs in codegen



Pull Request: https://github.com/llvm/llvm-project/pull/168645
DeltaFile
+18-3compiler-rt/test/ubsan_minimal/TestCases/override-callback.c
+19-0llvm/test/Instrumentation/BoundsChecking/runtimes.ll
+14-4clang/test/Driver/fsanitize.c
+8-0clang/lib/CodeGen/CGExpr.cpp
+7-0llvm/lib/Passes/PassBuilder.cpp
+5-2llvm/include/llvm/Transforms/Instrumentation/BoundsChecking.h
+71-95 files not shown
+84-1411 files

LLVM/project 27a3e43llvm/test/CodeGen/AMDGPU amdgcn.bitcast.1024bit.ll amdgcn.bitcast.512bit.ll, llvm/test/tools/llvm-dwarfdump/X86 simplified-template-names.s

address review get() -> native_handle() and constexpr invalid value

Created using spr 1.3.7
DeltaFile
+45,267-48,746llvm/test/CodeGen/AMDGPU/amdgcn.bitcast.1024bit.ll
+11,954-11,000llvm/test/CodeGen/AMDGPU/amdgcn.bitcast.512bit.ll
+5,981-8,885llvm/test/CodeGen/AMDGPU/shufflevector.v4p0.v4p0.ll
+5,981-8,885llvm/test/CodeGen/AMDGPU/shufflevector.v4i64.v4i64.ll
+7,387-7,087llvm/test/tools/llvm-dwarfdump/X86/simplified-template-names.s
+5,060-5,874llvm/test/CodeGen/AMDGPU/amdgcn.bitcast.960bit.ll
+81,630-90,47712,517 files not shown
+674,509-546,02212,523 files

LLVM/project 9782413clang/lib/CodeGen CGExpr.cpp, clang/test/Driver fsanitize.c

[UBSan] Use -fsanitize-handler-preserve-all-regs in codegen (#168645)

DeltaFile
+18-3compiler-rt/test/ubsan_minimal/TestCases/override-callback.c
+19-0llvm/test/Instrumentation/BoundsChecking/runtimes.ll
+14-4clang/test/Driver/fsanitize.c
+8-0clang/lib/CodeGen/CGExpr.cpp
+6-1llvm/lib/Transforms/Instrumentation/BoundsChecking.cpp
+5-2llvm/include/llvm/Transforms/Instrumentation/BoundsChecking.h
+70-105 files not shown
+84-1411 files

LLVM/project 396f4f9mlir/lib/Conversion/ArithToAPFloat ArithToAPFloat.cpp, mlir/lib/ExecutionEngine APFloatWrappers.cpp

[mlir][arith] Add support for `cmpf` to `ArithToAPFloat`
x
DeltaFile
+147-5mlir/lib/Conversion/ArithToAPFloat/ArithToAPFloat.cpp
+15-0mlir/test/Conversion/ArithToApfloat/arith-to-apfloat.mlir
+11-0mlir/lib/ExecutionEngine/APFloatWrappers.cpp
+4-0mlir/test/Integration/Dialect/Arith/CPU/test-apfloat-emulation.mlir
+177-54 files

LLVM/project b7eb988lldb/source/Utility RegisterValue.cpp

[lldb] Use InlHostByteOrder in RegisterValue::SetValueFromData (#169624)

An existing code can be further simplified.

---------

Co-authored-by: Matej Košík <matej.kosik at codasip.com>
DeltaFile
+1-3lldb/source/Utility/RegisterValue.cpp
+1-31 files

LLVM/project 8cc0259llvm/include/llvm/ExecutionEngine/Orc WaitingOnGraph.h, llvm/unittests/ExecutionEngine/Orc WaitingOnGraphTest.cpp

[ORC] Clear stale ElemToPendingSN entries in WaitingOnGraph. (#169747)

WaitingOnGraph::processReadyOrFailed was not clearing stale entries from
the ElemToPendingSN map. If symbols were removed from the
ExecutionSession and then re-added this could lead to dependencies on
the stale entries, triggering a use-after-free bug.

https://github.com/llvm/llvm-project/issues/169135
DeltaFile
+46-0llvm/unittests/ExecutionEngine/Orc/WaitingOnGraphTest.cpp
+18-4llvm/include/llvm/ExecutionEngine/Orc/WaitingOnGraph.h
+64-42 files

HardenedBSD/src a12e9c0lib/libc/tests/stdlib strfmon_test.c, share/doc/papers/malloc implementation.ms performance.ms

Merge branch 'freebsd/current/main' into hardened/current/master
DeltaFile
+245-24usr.bin/mdo/mdo.1
+0-223share/doc/papers/malloc/implementation.ms
+0-111share/doc/papers/malloc/performance.ms
+57-23lib/libc/tests/stdlib/strfmon_test.c
+0-72share/doc/papers/malloc/intro.ms
+0-70share/doc/papers/malloc/malloc.ms
+302-52317 files not shown
+360-77723 files

HardenedBSD/src 8fef8a7sys/dev/sound/pcm channel.c vchan.c, usr.bin/mdo mdo.1

Merge branch 'freebsd/15-stable/main' into hardened/15-stable/main
DeltaFile
+245-24usr.bin/mdo/mdo.1
+22-48sys/dev/sound/pcm/channel.c
+0-18sys/dev/sound/pcm/vchan.c
+0-5sys/dev/sound/pcm/vchan.h
+267-954 files

HardenedBSD/ports e1ee377devel/rust-cbindgen distinfo Makefile.crates, net/haproxy33 pkg-plist Makefile

Merge branch 'freebsd/main' into hardenedbsd/main
DeltaFile
+316-295www/homepage/pkg-plist
+107-95devel/rust-cbindgen/distinfo
+130-0net/haproxy33/files/haproxy.in
+115-0net/haproxy33/pkg-plist
+52-46devel/rust-cbindgen/Makefile.crates
+96-0net/haproxy33/Makefile
+816-43635 files not shown
+914-63141 files

LLVM/project 227c15dmlir/lib/Dialect/XeGPU/Transforms XeGPUBlocking.cpp

simplify: Avoid calling getTileShape(Result&Operand) from getTileShape(op)
DeltaFile
+14-13mlir/lib/Dialect/XeGPU/Transforms/XeGPUBlocking.cpp
+14-131 files

LLVM/project 49516ballvm/tools/llvm-objdump SourcePrinter.cpp SourcePrinter.h

[llvm-objdump] Optimize live element tracking (#158763)

This patch significantly optimizes the LiveElementPrinter
by replacing a slow linear search with efficient hash map
lookups. It refactors the code to use a map-based system
for tracking live element addresses and managing column
assignments, leading to a major performance improvement
for large binaries.
DeltaFile
+203-56llvm/tools/llvm-objdump/SourcePrinter.cpp
+43-10llvm/tools/llvm-objdump/SourcePrinter.h
+5-5llvm/tools/llvm-objdump/llvm-objdump.cpp
+251-713 files

HardenedBSD/src 3cf8333sys/dev/sound/pcm vchan.c vchan.h

sound: Remove vchan_passthrough() and hw.snd.passthrough_verbose

Unused and confusing.

Sponsored by:   The FreeBSD Foundation
MFC after:      1 week

(cherry picked from commit 3612ef642f511a1bd9f759da87abeafe7d6ff110)
DeltaFile
+0-18sys/dev/sound/pcm/vchan.c
+0-5sys/dev/sound/pcm/vchan.h
+0-232 files

FreeBSD/src 3cf8333sys/dev/sound/pcm vchan.c vchan.h

sound: Remove vchan_passthrough() and hw.snd.passthrough_verbose

Unused and confusing.

Sponsored by:   The FreeBSD Foundation
MFC after:      1 week

(cherry picked from commit 3612ef642f511a1bd9f759da87abeafe7d6ff110)
DeltaFile
+0-18sys/dev/sound/pcm/vchan.c
+0-5sys/dev/sound/pcm/vchan.h
+0-232 files

HardenedBSD/src 068b20esys/dev/sound/pcm channel.c

sound: Fix KASSERT panics in chn_read() and chn_write()

INVARIANTS kernels may trigger a KASSERT panic from sndbuf_acquire(),
when fuzzing write(2) using stress2, because of a race in chn_write().

In the case of chn_write(), what sndbuf_acquire() does is extend the
ready-to-read area of the buffer by a specified amount of bytes. The
KASSERT in question makes sure the number of bytes we want to extend the
ready area by, is less than or equal to the number of free bytes in the
buffer. This makes sense, because we cannot extend the ready area to
something larger than what is available (i.e., free) in the first place.

What chn_write() currently does for every write is; calculate the
appropriate write size, let's say X, unlock the channel, uiomove() X
bytes to the channel's buffer, lock the channel, and call
sndbuf_acquire() to extend the ready area by X bytes. The problem with
this approach, however, is the following.

Suppose an empty channel buffer with a length of 1024 bytes, and 2

    [45 lines not shown]
DeltaFile
+22-48sys/dev/sound/pcm/channel.c
+22-481 files

FreeBSD/src 068b20esys/dev/sound/pcm channel.c

sound: Fix KASSERT panics in chn_read() and chn_write()

INVARIANTS kernels may trigger a KASSERT panic from sndbuf_acquire(),
when fuzzing write(2) using stress2, because of a race in chn_write().

In the case of chn_write(), what sndbuf_acquire() does is extend the
ready-to-read area of the buffer by a specified amount of bytes. The
KASSERT in question makes sure the number of bytes we want to extend the
ready area by, is less than or equal to the number of free bytes in the
buffer. This makes sense, because we cannot extend the ready area to
something larger than what is available (i.e., free) in the first place.

What chn_write() currently does for every write is; calculate the
appropriate write size, let's say X, unlock the channel, uiomove() X
bytes to the channel's buffer, lock the channel, and call
sndbuf_acquire() to extend the ready area by X bytes. The problem with
this approach, however, is the following.

Suppose an empty channel buffer with a length of 1024 bytes, and 2

    [45 lines not shown]
DeltaFile
+22-48sys/dev/sound/pcm/channel.c
+22-481 files

LLVM/project 31141b6llvm/docs LangRef.rst, llvm/lib/CodeGen PreISelIntrinsicLowering.cpp

Unsupport SW encoding

Created using spr 1.3.6-beta.1
DeltaFile
+4-60llvm/test/Transforms/PreISelIntrinsicLowering/protected-field-pointer.ll
+4-60llvm/test/Transforms/PreISelIntrinsicLowering/protected-field-pointer-addrspace1.ll
+3-23llvm/lib/CodeGen/PreISelIntrinsicLowering.cpp
+2-4llvm/docs/LangRef.rst
+13-1474 files

FreeBSD/src 35232d8share/doc/papers/malloc implementation.ms performance.ms

papers: Retire phkmalloc paper

It has not been installed since commit cdc37953165c ("In preparation for
the removal of the roff toolchain, disconnect the") and turned up in
a search for outdated MALLOC_OPTIONS settings.

The rendered paper is available at
https://papers.freebsd.org/1998/phk-malloc

PR:             287357
Reviewed by:    bapt
Event:          Kitchener-Waterloo Hackathon 202506
Sponsored by:   The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D50908
DeltaFile
+0-223share/doc/papers/malloc/implementation.ms
+0-111share/doc/papers/malloc/performance.ms
+0-72share/doc/papers/malloc/intro.ms
+0-70share/doc/papers/malloc/malloc.ms
+0-54share/doc/papers/malloc/kernel.ms
+0-52share/doc/papers/malloc/problems.ms
+0-5825 files not shown
+0-71211 files

HardenedBSD/src 35232d8share/doc/papers/malloc implementation.ms performance.ms

papers: Retire phkmalloc paper

It has not been installed since commit cdc37953165c ("In preparation for
the removal of the roff toolchain, disconnect the") and turned up in
a search for outdated MALLOC_OPTIONS settings.

The rendered paper is available at
https://papers.freebsd.org/1998/phk-malloc

PR:             287357
Reviewed by:    bapt
Event:          Kitchener-Waterloo Hackathon 202506
Sponsored by:   The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D50908
DeltaFile
+0-223share/doc/papers/malloc/implementation.ms
+0-111share/doc/papers/malloc/performance.ms
+0-72share/doc/papers/malloc/intro.ms
+0-70share/doc/papers/malloc/malloc.ms
+0-54share/doc/papers/malloc/kernel.ms
+0-52share/doc/papers/malloc/problems.ms
+0-5825 files not shown
+0-71211 files

LLVM/project 4795f24mlir/lib/Dialect/XeGPU/Transforms XeGPUBlocking.cpp, mlir/test/Dialect/XeGPU xegpu-blocking.mlir

simplify: load_nd/store_nd/prefetch_nd use anchor layout for blocking
DeltaFile
+39-39mlir/test/Dialect/XeGPU/xegpu-blocking.mlir
+6-5mlir/lib/Dialect/XeGPU/Transforms/XeGPUBlocking.cpp
+45-442 files

FreeBSD/src 5198c32sys/dev/vt vt_core.c

vt: Allow VT_SETMODE with frsig=0

Linux does not check that any of the signals in vt_mode VT_SETMODE ioctl
(relsig, acqsig, frsig) are valid, but FreeBSD required that all three
are valid.  frsig is unusued in both Linux and FreeBSD, and software
typically leaves it unset.  To improve portability, allow frsig to be
set to zero.

PR:             289812
Reported by:    Dušan Gvozdenović
Sponsored by:   The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D52835
DeltaFile
+3-3sys/dev/vt/vt_core.c
+3-31 files

HardenedBSD/src 5198c32sys/dev/vt vt_core.c

vt: Allow VT_SETMODE with frsig=0

Linux does not check that any of the signals in vt_mode VT_SETMODE ioctl
(relsig, acqsig, frsig) are valid, but FreeBSD required that all three
are valid.  frsig is unusued in both Linux and FreeBSD, and software
typically leaves it unset.  To improve portability, allow frsig to be
set to zero.

PR:             289812
Reported by:    Dušan Gvozdenović
Sponsored by:   The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D52835
DeltaFile
+3-3sys/dev/vt/vt_core.c
+3-31 files

LLVM/project 17d207aclang/lib/Analysis/FlowSensitive/Models UncheckedStatusOrAccessModel.cpp, clang/unittests/Analysis/FlowSensitive UncheckedStatusOrAccessModelTestFixture.cpp

[𝘀𝗽𝗿] initial version

Created using spr 1.3.7
DeltaFile
+102-0clang/lib/Analysis/FlowSensitive/Models/UncheckedStatusOrAccessModel.cpp
+84-0clang/unittests/Analysis/FlowSensitive/UncheckedStatusOrAccessModelTestFixture.cpp
+186-02 files

FreeBSD/src ddec420sys/netpfil/ipfilter/netinet fil.c ip_nat.c

ipfilter: fix broken build

Every commit earns me a dozen emails that LINT is broken. This should
stop that.

Fixes: eda1756d0454f ipfilter: Verify frentry on entry into kernel
Sponsored by: Netflix
DeltaFile
+1-3sys/netpfil/ipfilter/netinet/fil.c
+0-3sys/netpfil/ipfilter/netinet/ip_nat.c
+1-62 files