See https://reviews.llvm.org/D81809 for upstream description.
In summary this is ~10x improvement for small divisors and similar
performance for equal divisors.
Closes#13523.
The idea here is that there are two ways we can reference a function at runtime:
* Through a direct call, i.e. where the function is comptime-known
* Through a function pointer
This means we can easily perform a form of rudimentary escape analysis
on functions. If we ever see a `decl_ref` or `ref` of a function, we
have a function pointer, which could "leak" into runtime code, so we
emit the function; but for a plain `decl_val`, there's no need to.
This change means that `comptime { _ = f; }` no longer forces a function
to be emitted, which was used for some things (mainly tests). These use
sites have been replaced with `_ = &f;`, which still triggers analysis
of the function body, since you're taking a pointer to the function.
Resolves: #6256Resolves: #15353
Now they use slices or array pointers with any element type instead of
requiring byte pointers.
This is a breaking enhancement to the language.
The safety check for overlapping pointers will be implemented in a
future commit.
closes#14040
- Combine mulXi3 routines for follow-up cleanup.
- DRY up Dwords and Twords
- rename both to HalveInt and use instance
* Justification: Not all processors have word size 32 bit.
* remove test file from CMakeLists
* DRY things.
Justifications
- compiler_rt base routines are almost finished, so make 1:1
mapping of code and documentation.
- Make adjustments to code + documentation simpler to prevent technical or
documentation debt.
- Copy and adjust __divmodsi4 tests for __divmoddi4 and __divmodti4.
- Assuming d = a/b does not overflow (MIN/-1) or uses div by 0,
then tmp = (d * b) = (a/b * b) = a does not overflow.
=> Remove wraparound for remainder in applicable routines.
These are unused in LLVM, but in gcc used for example by rl78.
Move test files into common file, add tests for aforementioned routines
and missing ones for __lshrti3.
- In Integer Bit Operations, use zeros consistently and fix the
incorrect function descriptions
- In ucmp, fix the incorrect signedness of the operands
- In Integer Arithmetic, document the signedness of the operands in
div, udiv, mod, umod, udivmod and divmod as it is done in LLVM
documentation
- In Float Conversion, replace "int_to_float" to "int to float" and
"uint_to_float" to "uint to float"
This commit makes the text more dense with sources, noes the outstanding
audit and adds an explanation of the table before converting everything
to markdown tables.
For the status a checkmark or cross in utf8 encoding are used and the
input and output sizes of all operations are given inclusive comments for
understanding of the routines.
This should document all compiler_rt integer and float routines, but
does not include a documentation of existing ieee floating and math
routines.
- cbe: Implement linksection support, to support TLS when not linking libc
- cbe: Support under-aligned variables / struct fields
- cbe: Support packed structs (in the C definition of packed)
- windows: Fix regression with x86 _tls_array
- compiler_rt: Add 128-bit atomics to compiler_rt
- tests: Re-enable threadlocal tests on cbe+windows, and llvm+x86
- tests: Re-enable f80 tests that now pass
- ci: change windows ci to run the CBE behaviour tests with -lc, to match how the compiler is bootstrapped
- update zig1.wasm
When we're compiling compiler_rt for any WebAssembly target, we do
not want to expose all the compiler-rt functions to the host runtime.
By setting the visibility of all exports to `hidden`, we allow the
linker to resolve the symbols during linktime, while not expose the
functions to the host runtime. This also means the linker can
properly garbage collect any compiler-rt function that does not get
resolved. The symbol visibility for all target remains the same as
before: `default`.
These errdefer where never executed, while this didn't bother the stage1
compiler, it caused an error in stage2.
The fix is just removing those errdefer which doesn't change any
behaviour because they were never executed in the first place.
This is a partial revert of 0d533433e21621177fb291e2a4901bee11834501,
which regressed this behavior. The idea here is to avoid aliases, which
happens when the same function is exported with multiple names. The
problem with aliases is that weak aliases don't seem to work, causing
symbol collisions when multiple of the same symbol are provided, despite
the desired behavior that weak symbols are overridden.
In this case we export redundant functions with different names. Thanks
to -ffunction-sections, the unused functions will be garbage-collected
at link time. This leaves us with the best of both worlds: Zig's
compiler-rt will provide both sets of symbols, and it will be
binary-compatible with different compilers that expect different names,
while still resulting in binaries without garbage.
* Rely on libSystem when targeting macOS.
* Make tools/gen_outline_atomics.zig more idiomatic.
* Remove the CPU detection / auxval checking from compiler_rt. This
functionality belongs in a different component. Zig's compiler_rt
must not rely on constructors. Instead it will export a symbol for
setting the value, and start code can detect and activate it.
* Remove the separate logic for inline assembly when the target does or
does not have lse support. `.inst` works in both cases.
