There's a lot of stuff wrong with the representation of these types:
TyFnDef doesn't actually uniquely identify a function, TyFnPtr is used to
represent method calls, TyFnDef in the sub-expression of a cast isn't
correctly reified, and probably some other stuff I haven't discovered yet.
Splitting them seems like the right first step, though.
This commit transitions the compiler to using the new exception handling
instructions in LLVM for implementing unwinding for MSVC. This affects both 32
and 64-bit MSVC as they're both now using SEH-based strategies. In terms of
standard library support, lots more details about how SEH unwinding is
implemented can be found in the commits.
In terms of trans, this change necessitated a few modifications:
* Branches were added to detect when the old landingpad instruction is used or
the new cleanuppad instruction is used to `trans::cleanup`.
* The return value from `cleanuppad` is not stored in an `alloca` (because it
cannot be).
* Each block in trans now has an `Option<LandingPad>` instead of `is_lpad: bool`
for indicating whether it's in a landing pad or not. The new exception
handling intrinsics require that on MSVC each `call` inside of a landing pad
is annotated with which landing pad that it's in. This change to the basic
block means that whenever a `call` or `invoke` instruction is generated we
know whether to annotate it as part of a cleanuppad or not.
* Lots of modifications were made to the instruction builders to construct the
new instructions as well as pass the tagging information for the call/invoke
instructions.
* The translation of the `try` intrinsics for MSVC has been overhauled to use
the new `catchpad` instruction. The filter function is now also a
rustc-generated function instead of a purely libstd-defined function. The
libstd definition still exists, it just has a stable ABI across architectures
and leaves some of the really weird implementation details to the compiler
(e.g. the `localescape` and `localrecover` intrinsics).
This brings some routine upgrades to the bundled LLVM that we're using, the most
notable of which is a bug fix to the way we handle range asserts when loading
the discriminant of an enum. This fix ended up being very similar to f9d4149c
where we basically can't have a range assert when loading a discriminant due to
filling drop, and appropriate flags were added to communicate this to
`trans::adt`.
DST fields, being of an unknown type, are not automatically aligned
properly, so a pointer to the field needs to be aligned using the
information in the vtable.
Fixes#26403 and a number of other DST-related bugs discovered while
implementing this.
Note: for now, this change only affects `-windows-gnu` builds.
So why was this `libgcc` dylib dependency needed in the first place?
The stack unwinder needs to know about locations of unwind tables of all the modules loaded in the current process. The easiest portable way of achieving this is to have each module register itself with the unwinder when loaded into the process. All modules compiled by GCC do this by calling the __register_frame_info() in their startup code (that's `crtbegin.o` and `crtend.o`, which are automatically linked into any gcc output).
Another important piece is that there should be only one copy of the unwinder (and thus unwind tables registry) in the process. This pretty much means that the unwinder must be in a shared library (unless everything is statically linked).
Now, Rust compiler tries very hard to make sure that any given Rust crate appears in the final output just once. So if we link the unwinder statically to one of Rust's crates, everything should be fine.
Unfortunately, GCC startup objects are built under assumption that `libgcc` is the one true place for the unwind info registry, so I couldn't find any better way than to replace them. So out go `crtbegin`/`crtend`, in come `rsbegin`/`rsend`!
A side benefit of this change is that rustc is now more in control of the command line that goes to the linker, so we could stop using `gcc` as the linker driver and just invoke `ld` directly.
Rather than injecting a local `_Unwind_Resume` into the current translation unit,
just replace `resume` instruction with a direct call the the `eh_unwind_resume` lang item.
This is likely to be more robust in the face of future LLVM changes, and also allows us to delegate
work back to libgcc's `_Unwind_Resume`.
This PR turns statically known erroneous code (e.g. numeric overflow) into a warning and continues normal code-generation to emit the same code that would have been generated without `check_const` detecting that the result can be computed at compile-time.
<del>It's not done yet, as I don't know how to properly emit a lint from trans. I can't seem to extract the real lint level of the item the erroneous expression is in.</del> It's an unconditional warning now.
r? @pnkfelix
cc @nikomatsakis
* [RFC 1229 text](https://github.com/rust-lang/rfcs/blob/master/text/1229-compile-time-asserts.md)
* RFC PR: rust-lang/rfcs#1229
* tracking issue: https://github.com/rust-lang/rust/issues/28238
This also involved adding `[TYPE;N]` syntax and aggregate indexing
support to the generator script: it's the only way to be able to have a
parameterised intrinsic that returns an aggregate, since one can't refer
to previous elements of the current aggregate (and that was harder to
implement).
This adds a new Python script (compatible with 2.7 and 3.x) that will consume some JSON files that define a platform's intrinsics. It can output a file that defines the intrinsics in the compiler, or an `extern` block that will import them.
The complexity of the generator is to be DRY: platforms (especially ARM and AArch64) have a lot of repetition with their intrinsics, for different versions with different types, so being able to write it once is nice.
This adds support for flattened intrinsics, which are called in Rust
with tuples but in LLVM without them (e.g. `foo((a, b))` becomes `foo(a,
b)`). Unflattened ones could be supported, but are not yet.
This is necessary to reflect the ARM APIs accurately, since some
functions explicitly take an unsigned parameter and a signed one, of the
same integer shape, so the no-duplicates check will fail unless we
distinguish.
Combining them seemed like a good idea at the time, but turns out that
handling lifetimes separately makes it somewhat easier to handle cases
where we don't want the intrinsics, and let's you see more easily where
the start/end pairs are.
The functions is useful for all kinds of fat pointers, but get_len()
just feels so wrong for trait object fat pointers. Let's use get_meta()
because that's rather neutral.
We're currently possibly introducing an unneeded temporary, make use of
InsertValue which is said to kick us off of FastISel and we generate
loads/stores of first class aggregates, which is bad as well. Let's not
do all these things.
