* goals
- zig as linker for object files generated by other compilers
- zig-specific runtime features for eventual standardisation
* changes
- missing routines are marked with `missing`
- structure inspired by libgcc docs, but improved order and wording
- rename misspelled functions
- reorder and rephrase compiler_rt.zig to reflect documentation
- potential decimal float or fixed-point arithmetic support:
* 'Decimal float library routines' ca. 120 functions
* 'Fixed-point fractional library routines' ca. 300 functions
thanks to @Vexu for multiple reviews and @scheibo for review
Non-Arm CPUs use u16 as the parameter to __extendhfxf2 and the return value of
__truncxfhf2, so insert appropriate bitcasts in gen_soft_f80_widen_or_shorten.
Otherwise, LLVM might crash because the functions are called in a different way
than its compiler-rt definition.
This fixes stage1 build on SPARCv9, and possibly other non-x86, non-Arm CPUs.
Comment reproduced here:
Note the following u64 alignments:
x86-linux: 4
x86-windows: 8
LLVM makes x86_fp80 have the following alignment and sizes regardless
of operating system:
x86_64: size=16, align=16
x86: size=12, align=4
However in Zig we override x86-windows to have size=16, align=16
in order for the property to hold that u80 and f80 have the same ABI size.
Fixes "error: destination type 'f80' has size 12 but source type 'u80'
has size 16" when trying to bitcast between f80 and u80 on i386-windows.
Get rid of `std.math.F80Repr`. Instead of trying to match the memory
layout of f80, we treat it as a value, same as the other floating point
types. The functions `make_f80` and `break_f80` are introduced to
compose an f80 value out of its parts, and the inverse operation.
stage2 LLVM backend: fix pointer to zero length array tripping LLVM
assertion. It now checks for when the element type is a zero-bit type
and lowers such thing the same way that pointers to other zero-bit types
are lowered.
Both stage1 and stage2 LLVM backends are adjusted so that f80 is lowered
as x86_fp80 on x86_64 and i386 architectures, and identical to a u80 on
others. LLVM constants are lowered in a less hacky way now that #10860
is fixed, by using the expression `(exp << 64) | fraction` using llvm
constants.
Sema is improved to handle c_longdouble by recursively handling it
correctly for whatever the float bit width is. In both stage1 and
stage2.
* F80Repr extern struct needs no explicit padding; let's match the
target padding.
* stage2: fix lowering of f80 constants.
* stage1: decide ABI size and alignment of f80 based on alignment of
u64. x86 has alignof u64 equal to 4 but arm has it as 8.
* stage2: fix Value.floatReadFromMemory to use F80Repr
* pass air_tag instead of zir_tag
* also pass eval function so that the branch only happens once and the
body of zirUnaryMath is simplified
* Value.sqrt: update to handle f80 and f128 in the normalized way that
includes handling c_longdouble.
Semi-related change: fix incorrect sqrt builtin name for f80 in stage1.
Currently Zig lowers `@intToFloat` for f80 incorrectly on non-x86
targets:
```
broken LLVM module found:
UIToFP result must be FP or FP vector
%62 = uitofp i64 %61 to i128
SIToFP result must be FP or FP vector
%66 = sitofp i64 %65 to i128
```
This happens because on such targets, we use i128 instead of x86_fp80 in
order to avoid "LLVM ERROR: Cannot select". `@intToFloat` must be
lowered differently to account for this difference as well.
These options were removed in 5e63baae8 (CLI: remove --verbose-ast and
--verbose-tokenize, 2021-06-09) but some remainders were left in.
Signed-off-by: Johannes Löthberg <johannes@kyriasis.com>
stage1: fix issue with to_twos_complement that caused a compile error when performing wrapping addition on two signed ints, both of which have the minimum possible value
LLVM bitcast wants integers that match the number of bits. So the const
bitcast has to use an i80, not an i128.
This commit makes the behavior tests fail for me, so it seems I did not
correctly construct the type. But it gets rid of the LLVM segfault.
I noticed that the strategy of memcpy the buf worked if I simply did an
LLVMConstTrunc() on the i128 to make it into an i80 before the
LLVMConstBitCast().
But is that correct in the face of different endianness? I'm not sure.
Instead use the standarized option for communicating the
zig compiler backend at comptime, which is `zig_backend`. This was
introduced in commit 1c24ef0d0b09a12a1fe98056f2fc04de78a82df3.
This exposes a function from stage2 to stage1 to append symbols to automatically export them.
This happends under the following conditions:
- Target is wasm
- User has not provided --export/--rdynamic flags themselves.
Small packed structs weren't included in this resolution so their
c_abi_type would be NULL when attempting usage later, leading to a
compiler crash.
Resolves#10431.
This saves on comptime format string parsing, as the compiler caches
comptime calls. The catch here, is that parsePlaceHolder cannot take the
placeholder string as a slice. It must take it as an array by value for
the caching to occure.
There is also some logic in here that ensures that the specifier_arg is
always them same slice when the items they contain are the same. This
makes the compiler stamp out less copies of formatType.
Make `@returnAddress()` return for the BPF target, as the BPF target for
the time being does not support probing for the return address. Stack
traces for the general purpose allocator for the BPF target is also set
to not be captured.
