Noratrieb/blog
1
0
Fork 0
mirror of https://github.com/Noratrieb/blog.git synced 2026-01-14 20:35:02 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions

deploy: 750b012b77

Browse source
This commit is contained in:
Nilstrieb 2023-03-17 14:22:01 +00:00
parent 1e20001c22
commit 11964435c0
8 changed files with 13 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

10
posts/box-is-a-unique-type/index.html
View file

@ -44,7 +44,7 @@ reveals the solution: (severely shortened to only show the relevant parts)</p><p
</code></pre><p>See the little attribute on the first parameter called <code>noalias</code>? That&rsquo;s what&rsquo;s doing the magic here.
<code>noalias</code> is an LLVM attribute on pointers that allows for various optimizations. If there are two pointers,
and at least one of them is <code>noalias</code>, there are some restrictions around the two. Approximately:</p><ul><li>If one of them writes, they must not point to the same value (alias each other)</li><li>If neither of them writes, they can alias just fine.
Therefore, we also apply <code>noalias</code> to <code>&mut T</code> and <code>&T</code> (if it doesn&rsquo;t contain interior mutability through
Therefore, we also apply <code>noalias</code> to <code>&amp;mut T</code> and <code>&amp;T</code> (if it doesn&rsquo;t contain interior mutability through
<code>UnsafeCell&lt;T></code>), since they uphold these rules.</li></ul><p>For more info on <code>noalias</code>, see <a href=https://llvm.org/docs/LangRef.html#parameter-attributes>LLVMs LangRef</a>.</p><p>This might sound familiar to you if you&rsquo;re a viewer of <a href=https://twitter.com/jonhoo>Jon Gjengset</a>&rsquo;s content (which I can highly recommend). Jon has made an entire video about this before since his crate <code>left-right</code>
was affected by this (<a href=https://youtu.be/EY7Wi9fV5bk)>https://youtu.be/EY7Wi9fV5bk)</a>.</p><p>If you&rsquo;re looking for <em>any</em> hint that using box emits <code>noalias</code>, you have to look no further than the documentation
for <a href=https://doc.rust-lang.org/nightly/std/boxed/index.html#considerations-for-unsafe-code><code>std::boxed</code></a>. Well, the nightly or beta docs, because I only added this section very recently. For years, this behaviour was not really documented, and you had to
@ -96,7 +96,7 @@ If box didn&rsquo;t invalidate pointers on move and instead behaved like a norma
the heap, and therefore moving it should invalidate pointers, since moving <code>T</code> definitely has to invalidate pointers to it,
this comparison doesn&rsquo;t make sense to me. While <code>Box&lt;T></code> usually behaves like a <code>T</code>, it&rsquo;s just a pointer. Writers of unsafe
code <em>know</em> that box is just a pointer and will abuse that knowledge, accidentally causing UB with it. While this can be
mitigated with better docs and teaching, like how no one questions the uniqueness of <code>&mut T</code> (maybe that&rsquo;s also because that
mitigated with better docs and teaching, like how no one questions the uniqueness of <code>&amp;mut T</code> (maybe that&rsquo;s also because that
one makes intuitive sense, &ldquo;shared xor mutable&rdquo; is a simple concept), I think it will always be a problem,
because in my opinion, box being unique and invalidating pointers on move is simply not intuitive.</p><p>When a box is moved, the pointer bytes change their location in memory. But the bytes the box points to stay the same. They don&rsquo;t
move in memory. This is the fundamental missing intuition about the box behaviour.</p><p>There are also other reasons why the box behaviour is not desirable. Even people who know about the behaviour of box will want
@ -114,10 +114,10 @@ provide an aliasable version of <code>Box&lt;T></code>, which is used by the sel
So if box stayed unique, people could also just pick up that crate as a dependency and use the aliasable box from there instead of
having to write their own. Interestingly, this crate also provides a <code>Vec&lt;T></code>, even though <code>Vec&lt;T></code> can currently be aliased in practice and
in the current version of stacked borrows just fine, although it&rsquo;s also not clear whether we want to keep it like this, but I
don&rsquo;t think this can reasonable be changed.</p><blockquote><p>One thing was just pointed out to me after releasing the post: Mutation usually goes through <code>&mut T</code> anyways, even when the value
don&rsquo;t think this can reasonable be changed.</p><blockquote><p>One thing was just pointed out to me after releasing the post: Mutation usually goes through <code>&amp;mut T</code> anyways, even when the value
is stored as a <code>Box&lt;T></code>. Therefore, all the guarantees of uniqueness are already present when mutating boxes, making the uniqueness
of box even less important.</p></blockquote><h1 id=noalias-noslow>noalias, noslow<a href=#noalias-noslow class=hanchor arialabel=Anchor>&#8983;</a></h1><p>There is one clear potential benefit from this box behaviour: ✨Optimizations✨. <code>noalias</code> doesn&rsquo;t exist for fun, it&rsquo;s something
that can bring clear performance wins (for <code>noalias</code> on <code>&mut T</code>, those were measureable). So the only question remains:
that can bring clear performance wins (for <code>noalias</code> on <code>&amp;mut T</code>, those were measureable). So the only question remains:
<strong>How much performance does <code>noalias</code> on <code>Box&lt;T></code> give us now, and how many potential performance improvements could we get in the
future?</strong> For the latter, there is no simple answer. For the former, there is. <code>rustc</code> has <a href=https://github.com/rust-lang/rust/pull/99527><em>no</em> performance improvements</a>
from being compiled with <code>noalias</code> on <code>Box&lt;T></code>, but this isn&rsquo;t really representative since rustc mostly uses arenas instead of box internally.</p><p>I have also benchmarked a few crates from the ecosystem with and without noalias on box, and the <a href=https://gist.github.com/Nilstrieb/9a0751fb9fd1044a30ab55cef9a7d335>results</a>
@ -131,6 +131,6 @@ are very strong, and still giving code an option to obtain these seems useful. O
<code>&'static mut T</code> that is unleaked for drop, but the semantics of this are still <a href=https://github.com/rust-lang/unsafe-code-guidelines/issues/316>unclear</a>.
If that is not possible, exposing <code>std::ptr::Unique</code> (with it getting boxes aliasing semantics) could be desirable. For this, all existing usages of <code>Unique</code>
inside the standard library would have to be removed. We could also offer a <code>std::boxed::UniqueBox</code> that keeps the current semantics, but this would also bring direct aliasing
decisions more towards safe code, which I am not a huge fan of. Ownership is enough already.</p><p>I guess what I am wishing for are some good and flexible raw pointer types. But that&rsquo;s still in the stars&mldr;</p><p>For more information about this topic, see <a href=https://github.com/rust-lang/unsafe-code-guidelines/issues/326>https://github.com/rust-lang/unsafe-code-guidelines/issues/326</a></p><p><em>Thanks to the nice people on the Rust Community Discord for their feedback on the draft of this post!</em></p></div></div></div></div><footer class=footer><div class=footer__inner><div class=copyright><span>© 2022 Powered by <a href=http://gohugo.io>Hugo</a></span>
decisions more towards safe code, which I am not a huge fan of. Ownership is enough already.</p><p>I guess what I am wishing for are some good and flexible raw pointer types. But that&rsquo;s still in the stars&mldr;</p><p>For more information about this topic, see <a href=https://github.com/rust-lang/unsafe-code-guidelines/issues/326>https://github.com/rust-lang/unsafe-code-guidelines/issues/326</a></p><p><em>Thanks to the nice people on the Rust Community Discord for their feedback on the draft of this post!</em></p></div></div></div></div><footer class=footer><div class=footer__inner><div class=copyright><span>© 2023 Powered by <a href=http://gohugo.io>Hugo</a></span>
<span>:: Theme made by <a href=https://twitter.com/panr>panr</a></span></div></div></footer><script src=/assets/main.js></script>
<script src=/assets/prism.js></script></div></body></html>

2
posts/index.html
View file

@ -1,5 +1,5 @@
<!doctype html><html lang=en><head><title>Posts :: nilstriebs blog</title><meta charset=utf-8><meta name=viewport content="width=device-width,initial-scale=1"><meta name=description content><meta name=keywords content><meta name=robots content="noodp"><link rel=canonical href=/posts/><link rel=stylesheet href=/assets/style.css><link rel=apple-touch-icon href=/img/apple-touch-icon-192x192.png><link rel="shortcut icon" href=/img/favicon/orange.png><meta name=twitter:card content="summary"><meta property="og:locale" content="en"><meta property="og:type" content="website"><meta property="og:title" content="Posts"><meta property="og:description" content><meta property="og:url" content="/posts/"><meta property="og:site_name" content="nilstriebs blog"><meta property="og:image" content="/img/favicon/orange.png"><meta property="og:image:width" content="2048"><meta property="og:image:height" content="1024"><link href=/posts/index.xml rel=alternate type=application/rss+xml title="nilstriebs blog"></head><body class=orange><div class="container center headings--one-size"><header class=header><div class=header__inner><div class=header__logo><a href=/><div class=logo>nilstriebs blog</div></a></div></div></header><div class=content><div class=posts><div class="post on-list"><h1 class=post-title><a href=/posts/box-is-a-unique-type/>Box Is a Unique Type</a></h1><div class=post-meta><span class=post-date>2022-07-23</span>
<span class=post-author>:: Nilstrieb</span></div><span class=post-tags>#<a href=/tags/rust/>rust</a>&nbsp;
#<a href=/tags/unsafe-code/>unsafe code</a>&nbsp;</span><div class=post-content>About better aliasing semantics for <code>Box&lt;T></code></div><div><a class="read-more button" href=/posts/box-is-a-unique-type/>→</a></div></div><div class=pagination><div class=pagination__buttons></div></div></div></div><footer class=footer><div class=footer__inner><div class=copyright><span>© 2022 Powered by <a href=http://gohugo.io>Hugo</a></span>
#<a href=/tags/unsafe-code/>unsafe code</a>&nbsp;</span><div class=post-content>About better aliasing semantics for <code>Box&lt;T></code></div><div><a class="read-more button" href=/posts/box-is-a-unique-type/>→</a></div></div><div class=pagination><div class=pagination__buttons></div></div></div></div><footer class=footer><div class=footer__inner><div class=copyright><span>© 2023 Powered by <a href=http://gohugo.io>Hugo</a></span>
<span>:: Theme made by <a href=https://twitter.com/panr>panr</a></span></div></div></footer><script src=/assets/main.js></script>
<script src=/assets/prism.js></script></div></body></html>