[llvm-dev] LLVM Weekly - #360, November 23rd 2020

[Alex Bradbury via llvm-dev]

LLVM Weekly - #360, November 23rd 2020
======================================

If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/360>.

Welcome to the three hundred and sixtieth issue of LLVM Weekly, a weekly
newsletter (published every Monday) covering developments in LLVM, Clang, and
related projects. LLVM Weekly is brought to you by [Alex
Bradbury](https://www.linkedin.com/in/alex-bradbury/). Subscribe to future
issues at <http://llvmweekly.org> and pass it on to anyone else you think may
be interested. Please send any tips or feedback to <asb at asbradbury.org>, or
@llvmweekly or @asbradbury on Twitter.


## News and articles from around the web

Support for the Cranelift code generation backend has been [merged into
Rust](https://github.com/rust-lang/rust/pull/77975) as an optional alternative
to LLVM offering 20-30% faster build times.

Some previously missing videos from past LLVM Dev Meetings [have been uploaded
to the LLVM YouTube channel](https://www.youtube.com/c/LLVMPROJ/videos).


## On the mailing lists

* Paul C. Anagnostopoulos has posted the [second occasional TableGen
newsletter](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146649.html),
covering recent documentation updates and feature additions.

* Fāng-ruì Sòng
[proposes](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146676.html)
adding an `-fbinutils-verstion=` option to Clang.

* Geoffrey Martin-Noble [restarted discussion on contributing Bazel BUILD
files to
LLVM](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146670.html),
framing it in the context of the recently agreed [support
policy](http://llvm.org/docs/SupportPolicy.html).

* Paulo Matos has shared an [RFC on supporting multiple program address spaces
in LLVM](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146723.html).
This is motivated by the support of reference types in WebAssembly.

* Yuanke Luo authored an RFC on [supporting the Intel AMX programming model in
LLVM](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146770.html).

* Mike Edwards [linked to notes from the GitHub Pull Requests round table at
the LLVM Dev
Meeting](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146737.html),
which led to a sub-thread bout rebase-based CI setups.

* Rong Xu distributed an RFC on [control-flow sensitive
AutoFDO](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146694.html),
which aims to mantain more proced AutoFDO profile counts for duplicated
control flow.

* Tom Stellard [posted the proposed LLVM 11.0.1 release
schedule](http://lists.llvm.org/pipermail/llvm-dev/2020-November/146833.html).


## LLVM commits

* A new `llvm.pseudoprobe` IR intrinsic was introduced for pseudo-probe block
instrumentation. See the
[RFC](http://lists.llvm.org/pipermail/llvm-dev/2020-August/144101.html) for
more background. [f3c4456](https://reviews.llvm.org/rGf3c445697d2).

* Support was added for "out of line atomics" on AArch64.
[4d7df43](https://reviews.llvm.org/rG4d7df43ffdb).

* A default address space for globals can now be specified in DataLayout.
[3bc4157](https://reviews.llvm.org/rG3bc4157556b).

* A SchedMachineModel was added for the Cortex-M7.
[48b43c9](https://reviews.llvm.org/rG48b43c9d4f0).

* BasicAA's checks for whether two getelementptrs with different offsets alias
were generalised and unified.
[cd3c22c4](https://reviews.llvm.org/rGcd3c22c47e4).

* RISC-V bitmanip GREVI and GORCI are now lowered as custom nodes, allowing
more complex patterns to be matched.
[1ac9b54](https://reviews.llvm.org/rG1ac9b548310).


## Clang commits

* The OpenMP 5.1 feature status table was updated.
[7046be1](https://reviews.llvm.org/rG7046be17307)

* clangd gained an extension method to dump the AST.
[8adc4d1](https://reviews.llvm.org/rG8adc4d1ec76).


## Other project commits

* LLDB was extended to show tagged regions on AArch64 Linux when the memory
tagging extension (MTE) is enabled.
[3254168](https://reviews.llvm.org/rG32541685b2a).

* Sparse tensor compiler support was started in MLIR.
[eced4a8](https://reviews.llvm.org/rGeced4a8e6fe).

* New builtins were introduced to compiler-rt to convert between fp16 and
fp128. [1fb91fc](https://reviews.llvm.org/rG1fb91fcf9cf).

* LLVM's libc gained implementations of ldexp[f|l].
[bb8f258](https://reviews.llvm.org/rGbb8f2585c6e).

* LLVM's libcxx was ported to the NuttX RTOS.
[f0785c1](https://reviews.llvm.org/rGf0785c1f7ac).