When performing a "fat" LTO the compiler has a whole mess of codegen
units that it links together. To do this it needs to select one module
as a "base" module and then link everything else into this module.
Previously LTO passes assume that there's at least one module in-memory
to link into, but nowadays that's not always true! With incremental
compilation modules may actually largely be cached and it may be
possible that there's no in-memory modules to work with.
This commit updates the logic of the LTO backend to handle modules a bit
more uniformly during a fat LTO. This commit immediately splits them
into two lists, one serialized and one in-memory. The in-memory list is
then searched for the largest module and failing that we simply
deserialize the first serialized module and link into that. This
refactoring avoids juggling three lists, two of which are serialized
modules and one of which is half serialized and half in-memory.
Closes#63349
libtest: add --show-output flag to print stdout of successful tests
This pull request adds a new flag `--show-output` for tests to show the output of successful tests. For most formatters this was already supported just not exposed via the CLI (apparently only used by `librustdoc`). I've also added support for this option in the JSON formatter.
This kind of fixes https://github.com/rust-lang/rust/issues/54669 which wants `--format json` to work with `--nocapture`, which is... well, impossible. What this issue really calls for is `--show-output` as implemented here.
I think this was left out by accident from the "convert everything to
rlibs" commit, there's no need for this to be a dylib just as everything
else doesn't need to be a dylib!
So far it is assumed that using a DLL as a -l parameter argument is ok,
but the assumption doesn't hold when compiling the native code with
llvm.
In which case, an import library is required, so let's build one
This also requires the cargo counterpart to add the import library in
the stamp files, at least when compiling libstd. Otherwise, the files
don't get uplifted
Now that procedural macros no longer link transitively to libsyntax,
this shouldn't be needed any more! This commit is an experiment in
removing all dynamic libraries from rustc except for librustc_driver
itself. Let's see how far we can get with that!
Refactor C FFI variadics to more closely match their C counterparts, and add Clone implementation
We had to make some changes to expose `va_copy` and `va_end` directly to users (mainly for C2Rust, but not exclusively):
- redefine the Rust variadic structures to more closely correspond to C: `VaList` now matches `va_list`, and `VaListImpl` matches `__va_list_tag`
- add `Clone` for `VaListImpl`
- add explicit `as_va_list()` conversion function from `VaListImpl` to `VaList`
- add deref coercion from `VaList` to `VaListImpl`
- add support for the `asmjs` target
All these changes were needed for use cases like:
```Rust
let mut ap2 = va_copy(ap);
vprintf(fmt, ap2);
va_end(&mut ap2);
```
Emit error when trying to use PGO in conjunction with unwinding on Windows.
This PR makes `rustc` emit an error when trying use PGO in conjunction with `-Cpanic=unwind` on Windows, isn't supported by LLVM yet. The error messages points to https://github.com/rust-lang/rust/issues/61002, which documents this known limitation.
Remove special case for *ios* builds in run-make-fulldeps/print-target-list Makefile
Previous `TODO` comment in this file mentions [an issue that was closed](https://github.com/rust-lang/rust/issues/29812), and I was able to confirm locally that provided code in that issue no longer produces an ICE. Discussion on that issue seems to indicate this code was no longer needed as of 1.12.0.
I removed the `*ios*` branch from this `case` statement as it may cause confusion, then removed the case statement entirely as it only had a wildcard branch.
Didn't think it was this particular about things, but I also should
have tested locally.
It makes sense, though---`\` followed by LF would eat it, so we'd have
`sysroot done` instead of `sysroot; done` as it is parsed. This should
pass now.
Signed-off-by: Kristofer Rye <kristofer.rye@gmail.com>
Tested-by: Kristofer Rye <kristofer.rye@gmail.com>
Based on the TODO, this case was added to short-circuit for ios builds,
which is no longer necessary.
The comment in this Makefile mentions rust-lang/rust#29812 as a
dependency, but that issue has been since closed, with a statement that
the ICE of concern was resolved circa 1.12.
Here we remove this case, and just run the same branch for all targets.
Signed-off-by: Kristofer Rye <kristofer.rye@gmail.com>
This commit separates metadata encoding (`tcx.encode_metadata`) from the
creation of the metadata module (which is now handled by
`write_compressed_metadata`, formerly `write_metadata`).
The metadata encoding now occurs slightly earlier in the pipeline, at
the very start of code generation within `start_codegen`.
Metadata *writing* still occurs near the end of compilation; that will
be moved forward in subsequent commits.
Add codegen test for PGO instrumentation.
This PR adds a codegen test that makes sure that LLVM actually generates instrumentation code when we enable PGO instrumentation in `rustc`.
The second commit updates a test case to the new commandline option syntax introduced in #59874. Without the fix the test still works, but it confusingly creates a directory called `test.profraw`, which usually is the name of the _file_ where profiling data is collected.
Disallow double trailing newlines in tidy
This wasn't done previously in https://github.com/rust-lang/rust/pull/47064#issuecomment-354533010 as it affected too many files, but I think it's best to fix it now so that the number of files with double trailing newlines doesn't keep increasing.
r? kennytm
