This adds an implementation of thread local storage for the
`wasm32-unknown-unknown` target when the `atomics` feature is
implemented. This, however, comes with a notable caveat of that it
requires a new feature of the standard library, `wasm-bindgen-threads`,
to be enabled.
Thread local storage for wasm (when `atomics` are enabled and there's
actually more than one thread) is powered by the assumption that an
external entity can fill in some information for us. It's not currently
clear who will fill in this information nor whose responsibility it
should be long-term. In the meantime there's a strategy being gamed out
in the `wasm-bindgen` project specifically, and the hope is that we can
continue to test and iterate on the standard library without committing
to a particular strategy yet.
As to the details of `wasm-bindgen`'s strategy, LLVM doesn't currently
have the ability to emit custom `global` values (thread locals in a
`WebAssembly.Module`) so we leverage the `wasm-bindgen` CLI tool to do
it for us. To that end we have a few intrinsics, assuming two global values:
* `__wbindgen_current_id` - gets the current thread id as a 32-bit
integer. It's `wasm-bindgen`'s responsibility to initialize this
per-thread and then inform libstd of the id. Currently `wasm-bindgen`
performs this initialization as part of the `start` function.
* `__wbindgen_tcb_{get,set}` - in addition to a thread id it's assumed
that there's a global available for simply storing a pointer's worth
of information (a thread control block, which currently only contains
thread local storage). This would ideally be a native `global`
injected by LLVM, but we don't have a great way to support that right
now.
To reiterate, this is all intended to be unstable and purely intended
for testing out Rust on the web with threads. The story is very likely
to change in the future and we want to make sure that we're able to do
that!
This commit applies a few code size optimizations for the wasm target to
the standard library, namely around panics. We notably know that in most
configurations it's impossible for us to print anything in
wasm32-unknown-unknown so we can skip larger portions of panicking that
are otherwise simply informative. This allows us to get quite a nice
size reduction.
Finally we can also tweak where the allocation happens for the
`Box<Any>` that we panic with. By only allocating once unwinding starts
we can reduce the size of a panicking wasm module from 44k to 350 bytes.
support pub(restricted) in thread_local! (round 2)
Resurrected #40984 now that the issue blocking it was fixed. Original description:
`pub(restricted)` was stabilized in #40556 so let's go!
Here is a [playground](https://play.rust-lang.org/?gist=f55f32f164a6ed18c219fec8f8293b98&version=nightly&backtrace=1).
I changed the interface of `__thread_local_inner!`, which is supposedly unstable but this is not checked for macros (#34097 cc @petrochenkov @jseyfried), so this may be an issue.
I think LLVM has had support for quite some time now for this, we just never got
around to testing it out and binding it. We've had some trouble landing this in
the past I believe, but it's time to try again!
This commit flags the `#[thread_local]` attribute as being available for Windows
targets and adds an implementation of `register_dtor` in the `thread::local`
module to ensure we can destroy these keys. The same functionality is
implemented in clang via a function called `__tlregdtor` (presumably provided in
some Windows runtime somewhere), but this function unfortunately does not take a
data pointer (just a thunk) which means we can't easily call it. For now
destructors are just run in the same way the Linux fallback is implemented,
which is just keeping track via a single OS-based TLS key.
This test has been deadlocking and causing problems on the bots basically since
its inception. Some memory safety issues were fixed in 987dc84b, but the
deadlocks remained afterwards unfortunately.
After some investigation, I've concluded that this is just a situation where OSX
is not guaranteed to run destructors. The fix in 987dc84b observed that OSX was
rewriting the backing TLS memory to its initial state during destruction while
we weren't looking, and this would have the effect of canceling the destructors
of any other initialized TLS slots.
While very difficult to pin down, this is basically what I assume is happening
here, so there doesn't seem to really be anythig we can do to ensure the test
robustly passes on OSX, so just ignore it for now.
This transitions the standard library's `thread_local!` macro to use the
freshly-added and gated `#[cfg(target_thread_local)]` attribute. This greatly
simplifies the `#[cfg]` logic in play here, but requires that the standard
library expose both the OS and ELF TLS implementation modules as unstable
implementation details.
The implementation details were shuffled around a bit but end up generally
compiling to the same thing.
Closes#26581 (this supersedes the need for the option)
Closes#27057 (this also starts ignoring the option)
This commit removes the injection of `std::env::args()` from `--test` expanded
code, relying on the test runner itself to call this funciton. This is more
hygienic because we can't assume that `std` exists at the top layer all the
time, and it meaks the injected test module entirely self contained.
This commit is an implementation of [RFC 1184][rfc] which tweaks the behavior of
the `#![no_std]` attribute and adds a new `#![no_core]` attribute. The
`#![no_std]` attribute now injects `extern crate core` at the top of the crate
as well as the libcore prelude into all modules (in the same manner as the
standard library's prelude). The `#![no_core]` attribute disables both std and
core injection.
[rfc]: https://github.com/rust-lang/rfcs/pull/1184
This isn't actually necessary any more with the advent of `$crate` and changes
in the compiler to expand macros to `::core::$foo` in the context of a
`#![no_std]` crate.
The libcore inner module was also trimmed down a bit to the bare bones.
Makes the lint a bit more accurate, and improves the quality of the diagnostic
messages by explicitly returning an error message.
The new lint is also a little more aggressive: specifically, it now
rejects tuples, and it recurses into function pointers.
TLS tests have been deadlocking on the OSX bots for quite some time now and this
commit is the result of the investigation into what's going on. It turns out
that a value in TLS which is being destroyed (e.g. the destructor is run) can be
reset back to the initial state **while the destructor is running** if TLS is
re-accessed.
To fix this we stop calling drop_in_place on OSX and instead move the data to a
temporary location on the stack.
