[llvm-dev] LLVM Weekly - #372, February 15th 2021

[Alex Bradbury via llvm-dev]

LLVM Weekly - #372, February 15th 2021
======================================

If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/372>.

Welcome to the three hundred and seventy-second 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

The LLVM Foundation have announced the first [virtual Community.o
Summit](https://llvm.swoogo.com/community-o-summit/). Taking place March 8th
to March 10th, this three day virtual event aims to provide an inclusive space
for underrepresented groups and newcomers. For more information on
Community.o, see the [recent LLVM blog
post](https://blog.llvm.org/posts/2021-02-10-community-dot-o/) on it.

The schedule for the 5th LLVM workshop at CGO [is
out](https://lists.llvm.org/pipermail/llvm-dev/2021-February/148434.html).

Stephen Kelly has blogged about [a new AST traversal
mode](https://steveire.wordpress.com/2021/02/14/ast-matchmaking-made-easy/)
available in Clang 12 that makes it easier to write AST matchers.


## On the mailing lists

* Geoffrey Martin-Noble has posted a [proposal for adding a Bazel build
configuration to the LLVM
monorepo](https://lists.llvm.org/pipermail/llvm-dev/2021-February/148437.html).

* Cassie Jones shares [Vim syntax hightling support for
MIR](https://lists.llvm.org/pipermail/llvm-dev/2021-February/148545.html).

* Ricky Taylor kicked off a discussion on ['code beads' for M68k machine code
generation](https://lists.llvm.org/pipermail/llvm-dev/2021-February/148408.html).


## LLVM commits

* `-print-changed=[diff | diff-quiet]` will show IR changes between passes in
a patch-like format. [4b661b4](https://reviews.llvm.org/rG4b661b4059b0).

* FileCheck now defaults to `allow-unused-prefixes=false`.
[87dbdd2](https://reviews.llvm.org/rG87dbdd2e3bb6).

* Scalable vector fused multiply-add (FMA) on RISC-V were made commutable to
improve register allocation.
[b7b4f4c](https://reviews.llvm.org/rGb7b4f4cbc3a6).

* Initial support as added for converting fixed vectors to scalable vectors in
the RISC-V backend. [a719b66](https://reviews.llvm.org/rGa719b667a979), and
more.

* opt's new `--print-passes` flag will print all available passes.
[5d960cb](https://reviews.llvm.org/rG5d960cba3468).

* Alignment attributes are now supported for `atomicrmw` and `cmpxchg`.
[d06ab79](https://reviews.llvm.org/rGd06ab7981678),
[17517f3](https://reviews.llvm.org/rG17517f3178b5).


## Clang commits

* The `-f[no-]finite-loops` options were added to control when Clang adds the
`mustprogress` attribute. [51bf4c0](https://reviews.llvm.org/rG51bf4c0e6d4c).

* Support was added for the `swift_async_error` attribute.
[1e8afba](https://reviews.llvm.org/rG1e8afba6f176).


## Other project commits

* lld-macho received initial ARM64 target support.
[87104fa](https://reviews.llvm.org/rG87104faac433).

* LLD gained support for big endian AArch64.
[7605a9a](https://reviews.llvm.org/rG7605a9a009b5).

* An initial alias analysis framework was added to MLIR.
[b9c876b](https://reviews.llvm.org/rGb9c876bd7e9a).

* The flang driver gained support for `-fopenmp` and `-fopenacc`.
[6d48a1a](https://reviews.llvm.org/rG6d48a1a53fab).

* Various C++20 concepts were added to libcxx.
[8caf835](https://reviews.llvm.org/rG8caf835337f7),
[2b2f36a](https://reviews.llvm.org/rG2b2f36a8b12b),
[2193e8b](https://reviews.llvm.org/rG2193e8be3efa), and more.

* The LLDB FreeBSD plugin gained support for mips64 and powerpc.
[8244fc5](https://reviews.llvm.org/rG8244fc505def),
[bd03f6d](https://reviews.llvm.org/rGbd03f6df51d1).

* Dialect interfaces were introduced to MLIR for translation to LLVM IR, and
the OpenMP to LLVM IR translation was migrated to use this infrastructure.
[b77bac0](https://reviews.llvm.org/rGb77bac057234),
[66900b3](https://reviews.llvm.org/rG66900b3eae96).