0
Fork 0
mirror of https://codeberg.org/ziglang/zig synced 2026-11-18 07:05:05 +00:00
Commit graph

79 commits

Author SHA1 Message Date
Matthew Lugg
82d389279c
compiler: default to Elf2 instead of LLD on sparc64
`Elf2` is now a more complete linker implementation than LLD when
targeting sparc64 on ELF. For instance, LLD is essentially incapable of
building code which links against glibc on `sparc64-linux`, while `Elf2`
can handle that just fine.
2026-09-22 12:31:23 +01:00
Andrew Kelley
ae4e835a66 std.lang: rename OptimizeMode to Optimize
and remove "release" from the enum tag names.

no functional change, however, despite the addition of
backwards-compatibile declarations in this patch, it is breaking because
expressions that use `==` or `!=` operators will not able to use the
deprecated names.

I predict these names to survive until the 1.0 tag without any more
breakage.
2026-07-20 17:26:50 -07:00
Krzysztof Wolicki
8219c86100 std.Target: add ashetos tag (#35954)
Reviewed-on: https://codeberg.org/ziglang/zig/pulls/35954
2026-07-07 15:19:20 +02:00
Andrew Kelley
b384a7f858 compiler: import Module instead of Package
it's almost like all these references to Package actually wanted to be
references to Module instead...
2026-06-29 23:50:19 -07:00
Matthew Lugg
4a2a9e9d24
Compilation.Config: rename an error
The old error name didn't align with the actual message we emitted, and
kept confusing me.
2026-06-04 09:57:29 +01:00
Matthew Lugg
72f0ae7ec4
Compilation.Config: improve error reporting for '-fnew-linker'
...and allow using `Elf2` with backends other than self-hosted x86_64.
The new ELF linker is currently opt-in, so there's no harm in allowing
backends which are not yet "officially" supported. (And also I'm going
to make the LLVM backend work with `Elf2` in the next commit because
that seems pretty trivial.)
2026-06-04 07:27:24 +01:00
Matthew Lugg
95eb37a198
Compilation.Config: do not default to '-flld' if '-fnew-linker' given
This causes a confusing CLI error.
2026-06-04 07:27:24 +01:00
pentuppup
71950f8c34 fix error when trying to use LLD with self-hosted backend 2026-05-31 09:42:38 +02:00
Zhalkhas
f04799dc9c implicitly enable LLVM when TSan is enabled (#35311)
FIxes #32127

Reviewed-on: https://codeberg.org/ziglang/zig/pulls/35311
2026-05-28 04:32:23 +02:00
Alex Rønne Petersen
e919a0d175
compiler: don't allow statically linking executables on Haiku
Haiku always loads executables through its dynamic linker.
2026-05-24 21:38:04 +02:00
mlugg
7eb79daffb Merge pull request 'std.builtin -> std.lang migration progress' (#32182) from compiler-std-lang into master
Reviewed-on: https://codeberg.org/ziglang/zig/pulls/32182
2026-05-06 10:10:44 +02:00
johan0A
6a37df00ec Config.resolve: don't error on explicit dynamic linker for Lib/Obj outputs 2026-05-05 17:17:15 +02:00
Matthew Lugg
4c330e053b
compiler: use 'std.lang' instead of 'std.builtin' 2026-05-03 12:23:30 +01:00
Alex Rønne Petersen
5464392e11
std.os: add APIs to determine whether Zig std requires libc for a target
This is distinct from the question of whether the target formally considers libc
to be the only stable syscall interface; for example, FreeBSD and NetBSD have
stable syscalls, but we don't yet have syscall layers for them in std.os.
2026-04-17 12:07:57 +02:00
Alex Rønne Petersen
cef16b5130
std.Target: don't claim that OpenBSD requires libc
https://codeberg.org/ziglang/zig/issues/30982
2026-03-26 05:50:51 +01:00
Alex Rønne Petersen
4219037faf
compiler: default to PIC if the target defaults to PIE
Otherwise, by default and with no special flags, we produce objects and static
libraries that cannot be linked into executables.
2026-03-26 05:50:51 +01:00
Alex Rønne Petersen
909159ad8e
compiler: don't enforce PIC for x86-windows and thumb-windows
Only x86_64-windows and aarch64-windows actually require PIC.
2026-01-23 21:16:01 +01:00
Alex Rønne Petersen
00e6895bde
compiler: don't error on explicit link_libc=false on requiresLibC() targets
There are various reasons why one might want to still create libc-less
compilations on these targets. Case in point: Compiling our bundled crt0 for
OpenBSD.

We will still default to linking libc on these targets, though.
2026-01-05 16:50:46 +01:00
Alex Rønne Petersen
9ab7eec23e represent Mac Catalyst as aarch64-maccatalyst-none rather than aarch64-ios-macabi
Apple's own headers and tbd files prefer to think of Mac Catalyst as a distinct
OS target. Earlier, when DriverKit support was added to LLVM, it was represented
a distinct OS. So why Apple decided to only represent Mac Catalyst as an ABI in
the target triple is beyond me. But this isn't the first time they've ignored
established target triple norms (see: armv7k and aarch64_32) and it probably
won't be the last.

While doing this, I also audited all Darwin OS prongs throughout the codebase
and made sure they cover all the tags.
2025-11-14 11:33:35 +01:00
Jacob Young
0834e696f7 Elf2: start implementing dynamic linking 2025-10-29 18:15:09 -04:00
Alex Rønne Petersen
a7119d4269 remove all IBM AIX and z/OS support
As with Solaris (dba1bf9353), we have no way to
actually audit contributions for these OSs. IBM also makes it even harder than
Oracle to actually obtain these OSs.

closes #23695
closes #23694
closes #3655
closes #23693
2025-10-29 14:25:51 +01:00
Jacob Young
561e556aaf target: check for backend support for the new linker
Closes #25343
2025-09-24 16:08:26 -07:00
Jacob Young
f58200e3f2 Elf2: create a new linker from scratch
This iteration already has significantly better incremental support.

Closes #24110
2025-09-21 14:09:14 -07:00
Jacob Young
917640810e Target: pass and use locals by pointer instead of by value
This struct is larger than 256 bytes and code that copies it
consistently shows up in profiles of the compiler.
2025-06-19 11:45:06 -04:00
Jacob Young
0bf8617d96 x86_64: add support for pie executables 2025-06-06 23:42:14 -07:00
Andrew Kelley
e19886a898 Compilation.Config: prefer_llvm depends on pie 2025-06-06 23:42:14 -07:00
Andrew Kelley
d6b1ff7533 Compilation.Config: eliminate the only variable from this function 2025-06-06 23:42:14 -07:00
Andrew Kelley
2387305b23 Compilation.Config: "can_use_llvm" -> "prefer_llvm" 2025-06-06 23:42:14 -07:00
Jacob Young
5986bdf868 Compilation: enable the x86_64 backend by default for debug builds
Closes #22257
2025-06-06 23:42:14 -07:00
Alex Rønne Petersen
9d534790eb std.Target: Introduce Cpu convenience functions for feature tests.
Before:

* std.Target.arm.featureSetHas(target.cpu.features, .has_v7)
* std.Target.x86.featureSetHasAny(target.cpu.features, .{ .sse, .avx, .cmov })
* std.Target.wasm.featureSetHasAll(target.cpu.features, .{ .atomics, .bulk_memory })

After:

* target.cpu.has(.arm, .has_v7)
* target.cpu.hasAny(.x86, &.{ .sse, .avx, .cmov })
* target.cpu.hasAll(.wasm, &.{ .atomics, .bulk_memory })
2025-06-05 06:12:00 +02:00
Alex Rønne Petersen
0ccd2b0c5c
compiler: Always dynamically link executables for Fuchsia.
Fuchsia only supports PIE executables, specifically ET_DYN.

https://fuchsia.dev/fuchsia-src/concepts/process/program_loading
2025-06-04 06:54:10 +02:00
Alex Rønne Petersen
a81fb5fb76
compiler: Rework PIE option logic.
To my knowledge, the only platforms that actually *require* PIE are Fuchsia and
Android, and the latter *only* when building a dynamically-linked executable.
OpenBSD and macOS both strongly encourage using PIE by default, but it isn't
technically required. So for the latter platforms, we enable it by default but
don't enforce it.

Also, importantly, if we're building an object file or a static library, and the
user hasn't explicitly told us whether to build PIE or non-PIE code (and the
target doesn't require PIE), we should *not* default to PIE. Doing so produces
code that cannot be linked into non-PIE output. In other words, building an
object file or a static library as PIE is an optimization only to be done when
the user knows that it'll end up in a PIE executable in the end.

Closes #21837.
2025-06-04 06:48:19 +02:00
Alex Rønne Petersen
c5e669ff76
compiler: Link libc by default when targeting NetBSD.
We don't yet have a direct syscall layer in std.os.netbsd.
2025-05-17 04:41:27 +02:00
Alex Rønne Petersen
d29ba75c84
Compilation.Config: Default to dynamic linking with NetBSD libc. 2025-05-17 04:41:27 +02:00
Alex Rønne Petersen
833d4c9ce4
Merge pull request #23835 from alexrp/freebsd-libc
Support dynamically-linked FreeBSD libc when cross-compiling
2025-05-12 01:19:23 +02:00
Alex Rønne Petersen
f3e851dbd0
compiler: Link libc by default when targeting FreeBSD.
We don't yet have a direct syscall layer in std.os.freebsd.
2025-05-11 11:15:23 +02:00
Alex Rønne Petersen
e8992af7f0
Compilation.Config: Default to dynamic linking with FreeBSD libc. 2025-05-11 11:15:09 +02:00
Alex Rønne Petersen
837e0f9c37 std.Target: Remove ObjectFormat.nvptx (and associated linker code).
Textual PTX is just assembly language like any other. And if we do ever add
support for emitting PTX object files after reverse engineering the bytecode
format, we'd be emitting ELF files like the CUDA toolchain. So there's really no
need for a special ObjectFormat tag here, nor linker code that treats it as a
distinct format.
2025-05-10 12:21:57 +02:00
Alex Rønne Petersen
d2f92e1797
compiler: Link libunwind when linking glibc statically.
glibc's libc.a depends on the functions provided by libunwind.
2025-05-03 10:54:36 +02:00
Alex Rønne Petersen
f6476e9cae
compiler: Allow linking native glibc statically.
This is generally ill-advised, but can be useful in some niche situations where
the caveats don't apply. It might also be useful when providing a libc.txt that
points to Eyra.
2025-05-03 10:54:33 +02:00
Alex Rønne Petersen
ae38575b42
compiler: Rename misleading libcNeedsLibUnwind() function.
It's about libc++, not libc.
2025-05-03 10:31:31 +02:00
Alex Rønne Petersen
b3537d0f4a compiler: Allow configuring UBSan mode at the module level.
* Accept -fsanitize-c=trap|full in addition to the existing form.
* Accept -f(no-)sanitize-trap=undefined in zig cc.
* Change type of std.Build.Module.sanitize_c to std.zig.SanitizeC.
* Add some missing Compilation.Config fields to the cache.

Closes #23216.
2025-04-26 22:54:34 +02:00
imreallybadatnames™️
7733b5dbe6
Merge pull request #23501 from imreallybadatnames/master
Step.Compile: use LtoMode enum for lto option
2025-04-09 05:16:36 +00:00
David Rubin
931178494f Compilation: correct when to include ubsan 2025-02-25 11:22:33 -08:00
Alex Rønne Petersen
b60e39fe8f
Compilation: Disable LTO by default.
LLD has too many LTO bugs, and we're dropping the LLD dependency soon anyway.
2025-01-25 14:59:02 +01:00
Alex Rønne Petersen
41185d297f
Package.Module: Make create() fall back on options.global.root_optimize_mode.
As is done for root_strip and root_error_tracing.
2025-01-23 23:22:38 +00:00
Alex Rønne Petersen
ef4d7f01a5
compiler: Fix computation of Compilation.Config.any_unwind_tables.
This moves the default value logic to Package.Module.create() instead and makes
it so that Compilation.Config.any_unwind_tables is computed similarly to
any_sanitize_thread, any_fuzz, etc. It turns out that for any_unwind_tables, we
only actually care if unwind tables are enabled at all, not at what level.
2025-01-23 23:22:38 +00:00
Travis Lange
82e7f23c49 Added support for thin lto 2025-01-05 18:08:11 +01:00
Alex Rønne Petersen
8af82621d7
compiler: Improve the handling of unwind table levels.
The goal here is to support both levels of unwind tables (sync and async) in
zig cc and zig build. Previously, the LLVM backend always used async tables
while zig cc was partially influenced by whatever was Clang's default.
2024-12-11 00:10:15 +01:00
Alex Rønne Petersen
94be75a94f Compilation: Re-enable LTO for RISC-V. 2024-11-19 12:04:42 +01:00