When the element is comptime-known, we can check if it has a repeated
byte representation. In this case, `@memset` can be lowered with the
LLVM intrinsic rather than with a loop.
Pointer comparisons were triggering `-Wcompare-distinct-pointer-types`
before this fix, which adds `(void*)` casts if the lhs type and rhs type
do not match pointer sizeness.
store:
The value to store may be undefined, in which case the destination
memory region has undefined bytes after this instruction is
evaluated. In such case ignoring this instruction is legal
lowering.
store_safe:
Same as `store`, except if the value to store is undefined, the
memory region should be filled with 0xaa bytes, and any other
safety metadata such as Valgrind integrations should be notified of
this memory region being undefined.
Previously, this code casted the array pointer to u8 pointer, but I
removed that in a different commit. This commit restores the cast, but
instead of hard-coding u8, it uses the destination element pointer,
since memset now supports arbitrary element types.
* Sema: upgrade operands to array pointers if possible when emitting
AIR.
* Implement safety checks for length mismatch and aliasing.
* AIR: make ptrtoint support slice operands. Implement in LLVM backend.
* C backend: implement new `@memset` semantics. `@memcpy` is not done
yet.
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
* Disable 128-bit atomics for x86_64 generic (currently also baseline)
because they require heavy abi agreement to correctly lower.
** This is a breaking change **
* Enable 128-bit atomics for aarch64 in Sema since it just works.
Since the Zig language documentation claims support for `.Min` and
`.Max` in `@atomicRmw` with floats, allow in Sema and implement for both
the llvm and C backends.
* CompileStep: Avoid calling producesPdbFile() to determine whether the
option should be respected. If the user asks for it, put it on the
command line and let the Zig CLI deal with it appropriately.
* Make the namespace of `std.dwarf.Format.dwarf32` no longer have a
redundant "dwarf" in it.
* Add `zig cc` integration for `-gdwarf32` and `-gdwarf64`.
* Toss in a bonus bug fix for `-gdwarf-2`, `-gdwarf-3`, etc.
* Avoid using default init values for struct fields unnecessarily.
* Add missing cache hash addition for the new option.
This commit enables producing 64-bit DWARF format for Zig executables
that are produced through the LLVM backend. This is achieved by exposing
both command-line flags and CompileStep flags. The production of the
64-bit format only affects binaries that use the DWARF format and it is
disabled on MacOS due to it being problematic. This commit, despite
generating the interface for the Zig user to be able to tell the compile
which format is wanted, is just implemented for the LLVM backend, so
clang and the self-hosted backends will need this to be implemented in a
future commit.
This is an effort to work around #7962, since the emission of the 64-bit
format automatically produces 64-bit relocations. Further investigation
will be needed to make DWARF 32-bit format to emit bigger relocations
when needed and not make the linker angry.
Backends want to avoid emitting unused instructions which do not have
side effects: to that end, they all have `Liveness.isUnused` checks for
many instructions. However, checking this in the backends avoids a lot
of potential optimizations. For instance, if a nested field is loaded,
then the first field access would still be emitted, since its result is
used by the next access (which is then unreferenced).
To elide more instructions, Liveness can track this data instead. For
operands which do not have to be lowered (i.e. are not side effecting
and are not something special like `arg), Liveness can ignore their
operand usages, and push the unused information further up, potentially
marking many more instructions as unreferenced.
In doing this, I also uncovered a bug in the LLVM backend relating to
discarding the result of `@cVaArg`, which this change fixes. A behaviour
test has been added to cover it.
This is a partial rewrite of Liveness, so has some other notable changes:
- A proper multi-pass system to prevent code duplication
- Better logging
- Minor bugfixes
It seems that some implementations may have problems with these right now,
like Intel and Rusticl. In theory, these attributes should be superficial
on the pointer type, as alignment guarantees are also added via the
alignment option of the OpLoad and OpStore instructions. Therefore, get rid
of them for now.
AmdgpuKernel and NvptxKernel are unified into a Kernel calling convention.
There is really no reason for these to be separate; no backend is allowed to
emit the calling convention of the other. This is in the same spirit as the
.Interrupt calling convention lowering to different LLVM calling conventions,
and opens the way for SPIR-V kernels to be exported using the Kernel calling
convention.
Entry points need to be attributed with a complete list of
global variables that they use. To that end, the global dependencies
mechanism is extended to also allow functions - when flushing the
module, the list of dependencies is examined to generate this
list of global variable result-ids.
SPIR-V cannot represent function pointers without extensions
that no vendor implements. For the time being, generate a test
kernel for each error, so that we can at least run SOME tests.
In the future we may be able to emulate function pointers in some
way, but that is not today.
