BSD Commit Log Search

Firewall: Aliases: Add target to getAliasSource() (#9985)

Firewall: NAT: One-to-One NAT: add missing log statement

[NFC][analyzer] Update some incorrect doc-comments (#186852)

These were incorrectly copy-pasted more than ten years ago.

[ValueTracking] fadd never produces subnormal with no underflow (#186985)

In the cases where the fadd does not underflow, the result is both
non-zero and non-subnormal.

alive2 results for the added positive testcases:

testcase 1: https://alive2.llvm.org/ce/z/Mxjott

testcase 2: https://alive2.llvm.org/ce/z/Q-_A-v

testcase 3: https://alive2.llvm.org/ce/z/Y3XpSk

testcase 4: https://alive2.llvm.org/ce/z/34n8MZ

Fix #186975.

[Bazel] Port 429e9717 (#187019)

Fix deadlock in systemd plugin

refactor and pull dbus class into its own file

[MLIR][Interfaces] Make `getMutableSuccessorOperands` overridable on `ReturnLike` ops (#186832)

Move the `getMutableSuccessorOperands` implementation from `ReturnLike`
trait to the `RegionBranchTerminatorOpInterface` to allow overriding of
the implementation. This allows to have the trait on operations that are
not a return of all of their operands. This can be used, for example, to
implement custom `ReturnLike` terminator that consumes non-returned
operands in combination with `func.func`.

The `RegionBranchTerminatorOpInterface` now provides a default
implementation for the `getMutableSuccessorOperands` method that returns
all of the operands.

[lldb] Add pointer arithmetics for addition and subtraction to DIL (#184652)

[sancov] add -fsanitize-coverage=trace-pc-entry-exit (#185972)

Add a SanCov flag for calling dedicated hook functions on function entry
and exit. This flag can be used either in combination with
-fsanitize-coverage=trace-pc (in which case this patch changes which
hook is called for the entry BB, and generates an additional hook call
before return), or it can be used by itself (in which case only the
dedicated entry/exit callbacks are invoked).

This can be used to track the call stack throughout a sancov trace.

cc @vitalybuka @dvyukov

cache global dbus router connection

[libc++] Build the library with C++26 (#181021)

All supported compilers support C++26. This allows simplifying some of
the upcoming <text_encoding> implementation.

[NFC][analyzer] Improve computeObjectUnderConstruction (#186186)

Previously the method `ExprEngine::computeObjectUnderConstruction` took
a `NodeBuilderContext` parameter which was only used to call its
`blockCount()` method; this commit replaces this with directly taking
`NumVisitedCaller` (= number of times the caller was visited, the
`blockCount`) as an unsigned value.

In `CallEvent::getReturnValueUnderConstruction` this method is invoked
with `getNumVisitedCurrent()`, the visitation count of the _current_
`LocationContext` and `Block`; instead of calling `getNumVisited()` on
the `LocationContext` and `Block` corresponding to the `CallEvent`
instance (available through its data members). This is logically
incorrect, but (at least within the lit testsuite) there is no situation
where it leads to actually incorrect behavior. This is currently marked
with a FIXME comment; it will be fixed in a follow-up commit.

Merge branch 'main' into users/kparzysz/e10-check-depth

[SPIR-V] Handle spirv.MemoryModel metadata (#186138)

   Update to libgweather4-4.6.0.

[Flang][OpenMP] Add semantic support for OpenMP Loop Interchange and permutation clause in Flang (#183435)

This patch adds semantics for the `omp interchange` directive in flang
and the permutation clause, as specified in OpenMP 6.0.
Relevant tests have been added in every step.

[SPIR-V] Handle undef aggregate initializers for global variables (#186785)

Expand undef aggregate global initializers into per-element spv_undef
intrinsics

[LifetimeSafety] Exclude basic_string::insert from capturing methods (#186989)

Fixes https://github.com/llvm/llvm-project/issues/186817

   use fast blits whenever possible - much faster scrolling

Formatting fix

Reintroduce barrier-phase-with usage

[mlir][tosa] Remove 'Pure' trait from operations that are not speculatable (#185700)

This commit removes the 'Pure' trait from a number of TOSA operations.
Instead of marking most ops as pure by default, the trait is now opt-in
for operations that are provably side-effect free and speculatable.

Several operations were previously marked as pure unintentionally.

The following operations have had 'Pure' removed (reason in brackets):
- ARGMAX (out-of-range index)
- AVG_POOL2D (accumulator overflow/underflow)
- AVG_POOL2D_ADAPTIVE (same as above)
- CONV2D (accumulator overflow/underflow)
- CONV2D_BLOCK_SCALED (accumulator overflow/underflow)
- CONV3D (accumulator overflow/underflow)
- DEPTHWISE_CONV2D (accumulator overflow/underflow)
- MATMUL (accumulator overflow/underflow)
- MATMUL_T_BLOCK_SCALED (accumulator overflow/underflow)
- TRANSPOSE_CONV2D (accumulator overflow/underflow)

    [25 lines not shown]

[lldb][windows] fix TestReplaceDLL.py reruns (#187002)

[AMDGPU] Remove AMDGPUCallLowering dependency on AMDGPUTargetLowering. NFC. (#187008)

review: chnage design to track uniform values

[X86] Improve handling of i512 SHL(-1,Amt) + SRL(-1,Amt) "mask shifts" (#186806)

An extension of the existing one-bit shift patterns - perform an initial
select to handle 'allones/allzeros' elements and then insert the element
that has a partial mask on top of it.

Often turns up in bit manipulation patterns

[X86][APX] Enable NDD tunings (#186049)

For latest Intel processors with APX, mem form of all NDD instructions
(except for RIP based addressing) and imm form of NDD add/sub (include
inc/dec) instructions needed to be turned off by default for optimal
hardware performance.

Two new tunings enable-ndd-mem and enable-ndd-imm was added then
disabled by default. The new isa attributes is adopted for different
alternatives for NDD related patterns to control the generation of
mem/imm form.

dns/ddclient: remove references to ddclient_flags

ddclient_flags is not implemented, please sure the configuration file.

While here, portfmt.

PR:             293762

   Update to libgsf-1.14.56.