When -Z profile is passed, the GCDAProfiling LLVM pass is added
to the pipeline, which uses debug information to instrument the IR.
After compiling with -Z profile, the $(OUT_DIR)/$(CRATE_NAME).gcno
file is created, containing initial profiling information.
After running the program built, the $(OUT_DIR)/$(CRATE_NAME).gcda
file is created, containing branch counters.
The created *.gcno and *.gcda files can be processed using
the "llvm-cov gcov" and "lcov" tools. The profiling data LLVM
generates does not faithfully follow the GCC's format for *.gcno
and *.gcda files, and so it will probably not work with other tools
(such as gcov itself) that consume these files.
[MIR] SwitchInt Everywhere
Something I've been meaning to do for a very long while. This PR essentially gets rid of 3 kinds of conditional branching and only keeps the most general one - `SwitchInt`. Primary benefits are such that dealing with MIR now does not involve dealing with 3 different ways to do conditional control flow. On the other hand, constructing a `SwitchInt` currently requires more code than what previously was necessary to build an equivalent `If` terminator. Something trivially "fixable" with some constructor methods somewhere (MIR needs stuff like that badly in general).
Some timings (tl;dr: slightly faster^1 (unexpected), but also uses slightly more memory at peak (expected)):
^1: Not sure if the speed benefits are because of LLVM liking the generated code better or the compiler itself getting compiled better. Either way, its a net benefit. The CORE and SYNTAX timings done for compilation without optimisation.
```
AFTER:
Building stage1 std artifacts (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
Finished release [optimized] target(s) in 31.50 secs
Finished release [optimized] target(s) in 31.42 secs
Building stage1 compiler artifacts (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
Finished release [optimized] target(s) in 439.56 secs
Finished release [optimized] target(s) in 435.15 secs
CORE: 99% (24.81 real, 0.13 kernel, 24.57 user); 358536k resident
CORE: 99% (24.56 real, 0.15 kernel, 24.36 user); 359168k resident
SYNTAX: 99% (49.98 real, 0.48 kernel, 49.42 user); 653416k resident
SYNTAX: 99% (50.07 real, 0.58 kernel, 49.43 user); 653604k resident
BEFORE:
Building stage1 std artifacts (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
Finished release [optimized] target(s) in 31.84 secs
Building stage1 compiler artifacts (x86_64-unknown-linux-gnu -> x86_64-unknown-linux-gnu)
Finished release [optimized] target(s) in 451.17 secs
CORE: 99% (24.66 real, 0.20 kernel, 24.38 user); 351096k resident
CORE: 99% (24.36 real, 0.17 kernel, 24.18 user); 352284k resident
SYNTAX: 99% (52.24 real, 0.56 kernel, 51.66 user); 645544k resident
SYNTAX: 99% (51.55 real, 0.48 kernel, 50.99 user); 646428k resident
```
cc @nikomatsakis @eddyb
Emit DW_AT_main_subprogram
This changes rustc to emit DW_AT_main_subprogram on the "main" program.
This lets gdb suitably stop at the user's main in response to
"start" (rather than the library's main, which is what happens
currently).
Fixes#32620
r? michaelwoerister
This changes rustc to emit DW_AT_main_subprogram on the "main" program.
This lets gdb suitably stop at the user's main in response to
"start" (rather than the library's main, which is what happens
currently).
Fixes#32620
r? michaelwoerister
Instead of directly creating a 'DIGlobalVariable', we now have to create
a 'DIGlobalVariableExpression' which itself contains a reference to a
'DIGlobalVariable'.
This is a straightforward change.
In the future, we should rename 'DIGlobalVariable' in the FFI
bindings, assuming we will only refer to 'DIGlobalVariableExpression'
and not 'DIGlobalVariable'.
LLVM Core C bindings provide this function for all the versions back to what we support (3.7), and
helps to avoid this unnecessary builder->function transition every time. Also a negative diff.
These switch statements cover all possible values, so the default case
is dead code (it contains an llvm_unreachable anyway), triggering a
-Wcovered-switch-default warning. Moving the unreachable after the
switch resolves these warnings. This keeps the build output clean.
Improve naming style in rustllvm.
As per the LLVM style guide, use CamelCase for all locals and classes,
and camelCase for all non-FFI functions.
Also, make names of variables of commonly used types more consistent.
Fixes#38688.
r? @rkruppe
i128 and u128 support
Brings i128 and u128 support to nightly rust, behind a feature flag. The goal of this PR is to do the bulk of the work for 128 bit integer support. Smaller but just as tricky features needed for stabilisation like 128 bit enum discriminants are left for future PRs.
Rebased version of #37900, which in turn was a rebase + improvement of #35954 . Sadly I couldn't reopen#37900 due to github. There goes my premium position in the homu queue...
[plugin-breaking-change]
cc #35118 (tracking issue)
As per the LLVM style guide, use CamelCase for all locals and classes,
and camelCase for all non-FFI functions.
Also, make names of variables of commonly used types more consistent.
Fixes#38688.
Fixes rebase fallout, makes code correct in presence of 128-bit constants.
This commit includes manual merge conflict resolution changes from a rebase by @est31.
This teaches Rust about an LLVM 4.0 API change for creating debug info
for global variables.
This change was made in upstream LLVM patch https://reviews.llvm.org/D20147
This is almost 1:1 copy of how clang did it in http://reviews.llvm.org/D20415
Alignment was removed from createBasicType and moved to
- createGlobalVariable
- createAutoVariable
- createStaticMemberType (unused in Rust)
- createTempGlobalVariableFwdDecl (unused in Rust)
e69c459a6e
In LLVM 4.0, this enum becomes an actual type-safe enum, which breaks
all of the interfaces. Introduce our own copy of the bitflags that we
can then safely convert to the LLVM one.
StringRefs have a length and their contents are not usually null-terminated.
The solution is to either copy the string data (in rustc_llvm::diagnostic) or take the size into account (in LLVMRustPrintPasses).
I couldn't trigger a bug caused by this (apparently all the strings returned in practice are actually null-terminated) but this is more correct and more future-proof.
The librustc_llvm API remains mostly unchanged, except that llvm::Attribute is no longer a bitflag but represents only a *single* attribute.
The ability to store many attributes in a small number of bits and modify them without interacting with LLVM is only used in rustc_trans::abi and closely related modules, and only attributes for function arguments are considered there.
Thus rustc_trans::abi now has its own bit-packed representation of argument attributes, which are translated to rustc_llvm::Attribute when applying the attributes.
The `Linkage` enum in librustc_llvm got out of sync with the version in LLVM and it caused two variants of the #[linkage=""] attribute to break.
This adds the functions `LLVMRustGetLinkage` and `LLVMRustSetLinkage` which convert between the Rust Linkage enum and the LLVM one, which should stop this from breaking every time LLVM changes it.
Fixes#33992
Macro expansions produce code tagged with debug locations that are completely different from the surrounding expressions. This wrecks havoc on debugger's ability the step over source lines.
In order to have a good line stepping behavior in debugger, we overwrite debug locations of macro expansions with that of the outermost expansion site.
When reuing a definition across codegen units, we obviously cannot use
internal linkage, but using external linkage means that we can end up
with multiple conflicting definitions of a single symbol across
multiple crates. Since the definitions should all be equal
semantically, we can use weak_odr linkage to resolve the situation.
Fixes#32518
We use a 64bit integer to pass the set of attributes that is to be
removed, but the called C function expects a 32bit integer. On most
platforms this doesn't cause any problems other than being unable to
unset some attributes, but on ARM even the lower 32bit aren't handled
correctly because the 64bit value is passed in different registers, so
the C function actually sees random garbage.
So we need to fix the relevant functions to use 32bit integers instead.
Additionally we need an implementation that actually accepts 64bit
integers because some attributes can only be unset that way.
Fixes#32360
`fast` a.k.a UnsafeAlgebra is the flag for enabling all "unsafe"
(according to llvm) float optimizations.
See LangRef for more information http://llvm.org/docs/LangRef.html#fast-math-flags
Providing these operations with less precise associativity rules (for
example) is useful to numerical applications.
For example, the summation loop:
let sum = 0.;
for element in data {
sum += *element;
}
Using the default floating point semantics, this loop expresses the
floats must be added in a sequence, one after another. This constraint
is usually completely unintended, and it means that no autovectorization
is possible.
whitequark
A.J. Gardner
Tim Neumann
bors
Matt Ickstadt
Simonas Kazlauskas
Corey Farwell
Jorge Aparicio
Tom Tromey
Dylan McKay
Ruud van Asseldonk
Ian Kerins
karpinski
Ivan Molodetskikh
Jake Goulding
Michael Woerister
Robin Kruppe
Seo Sanghyeon
Vadim Chugunov
Ariel Ben-Yehuda
Jan-Erik Rediger
Björn Steinbrink
Ulrik Sverdrup