[llvm-dev] LLVM Weekly - #351, September 21st 2020

Mon Sep 21 10:49:17 PDT 2020

Would it please be possible to fix the HTTPS version of the site? :)

Roman

On Mon, Sep 21, 2020 at 8:43 PM Alex Bradbury via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
>
> LLVM Weekly - #351, September 21st 2020
> =======================================
>
> If you prefer, you can read a HTML version of this email at
> <http://llvmweekly.org/issue/351>.
>
> Welcome to the three hundred and fifty-first 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
>
> Slides from CppCon 2020 are [now
> available](https://github.com/CppCon/CppCon2020).
>
> If you're considering a round table at the virtual LLVM Dev Meeting, please
> [submit a proposal by September
> 25th](http://lists.llvm.org/pipermail/llvm-dev/2020-September/145186.html).
> As well as these scheduled round table discussions, impromptu sessions will
> also be possible.
>
>
> ## On the mailing lists
>
> * Mitch Philipps started a thread to discuss [tagging of globals with Arm's
> Memory Tagging
> Extension](http://lists.llvm.org/pipermail/llvm-dev/2020-September/145173.html),
> and is seeking feedback on the proposed schemes.
>
> * Walter Erquinigo shared an RFC on [processor trace support in
> LLDB](http://lists.llvm.org/pipermail/lldb-dev/2020-September/016437.html).
>
> * Tanya Lattner is [seeking
> moderators](http://lists.llvm.org/pipermail/llvm-dev/2020-September/145205.html)
> for the 2020 LLVM Dev Meeting.
>
> * Arthur Eubanks proposes [removing the
> DeadInstEliminationPass](http://lists.llvm.org/pipermail/llvm-dev/2020-September/145187.html)
> on the basis that DeadCodeEliminationPass is more featureful and
> DeadInstEliminationPass hasn't been ported to the new pass manager.
>
> * Artur Pilipenko
> [proposes](http://lists.llvm.org/pipermail/llvm-dev/2020-September/145192.html)
> adding a new requires-statement call attribute to the
> llvm.memcpy/memmov.element.unordered.atomic calls to make it possible to have
> a GC safepoint during the copy operation.
>
>
> ## LLVM commits
>
> * Documentation was added on writing passes for the new pass manager.
> [c2590de](https://reviews.llvm.org/rGc2590de30df).
>
> * An initial constraint elimination pass was added. This pass aims to add a
> more flexible way to eliminate compares based on more complex constraints
> collected from dominating conditions.
> [3d42d54](https://reviews.llvm.org/rG3d42d549554).
>
> * A sequence of 19 DWARF64 support patches were committed.
> [380e746](https://reviews.llvm.org/rG380e746bcca),
> [982b31f](https://reviews.llvm.org/rG982b31fad29), and 17 more.
>
> * `-basic-block-sections=labels` will now emit basic block metadata in a new
> `.bb_addr_map` section. [7841e21](https://reviews.llvm.org/rG7841e21c984).
>
> * The PowerPC backend now supports thread local storage for local exec.
> [c0f199e](https://reviews.llvm.org/rGc0f199e5667).
>
> * `update_llc_test_checks` was updated to allow `$` in function names.
> [ce0eb81](https://reviews.llvm.org/rGce0eb81c727).
>
> * A new IRSimilarityIdentifier analysis was added.
> [7e4c6fb](https://reviews.llvm.org/rG7e4c6fb8546).
>
> * The WebAssembly backend gained initial support for dwarf fission.
> [0ff28fa](https://reviews.llvm.org/rG0ff28fa6a75).
>
> * The implicit null check transformation is now supported on AArch64.
> [b04c181](https://reviews.llvm.org/rGb04c181ed77).
>
> * The RISC-V backend now supports a shadow call stack.
> [1c46647](https://reviews.llvm.org/rG1c466477ad4).
>
> * `update_cc_test_checks` gained a new `--include-generated-funcs` option so
> that any functions created by the compiler that don't exist in the source will
> also be checked. [7c8bb40](https://reviews.llvm.org/rG7c8bb409f31).
>
>
> ## Clang commits
>
> * clang-format gained a new option allowing unknown format options to be
> tolerated (only giving a warning rather than an error).
> [f64903f](https://reviews.llvm.org/rGf64903fd817).
>
> * Clang now checks if builtin/library declarations have the correct type
> and builtin-ness is tracked per declaration using BuiltinAttr.
> Previously it was simply assumed based on the name; for example
> incompatible redeclarations in the non-global scope would result in
> crashes. This work also fixed a long-standing bug where builtins aren't
> recognized if they aren't directly inside extern "C". Thanks to Raul Tambre
> for emailing in a summary of this patch.
> [e09107a](https://reviews.llvm.org/rGe09107ab80d).
>
> * A new command line flag was added for the machine function splitter.
> [f1a3ab9](https://reviews.llvm.org/rGf1a3ab90443).
>
> * Implicit casts are now supported between GNU and SVE vectors.
> [9218f92](https://reviews.llvm.org/rG9218f928380).
>
> * Initial work has started on a random forest model for code completion in
> clangd. [9b6765e](https://reviews.llvm.org/rG9b6765e784b).
>
>
> ## Other project commits
>
> * Documentation was added on getting involved in Flang.
> [fe395ae](https://reviews.llvm.org/rGfe395aecd9e).
>
> * libunwind gained support for leaf function unwinding.
> [22b615a](https://reviews.llvm.org/rG22b615a9659).
>
> * LLVM's libc gained an implementation of hypotf.
> [f55963d](https://reviews.llvm.org/rGf55963d501e).
>
> * LLD's `--warn-backrefs` was documented.
> [f6f3402](https://reviews.llvm.org/rGf6f34024e9a).
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev