Also explicitly specify `-mstrict-align` (The current default `-mcpu`
on gcc doesn't support unaligned accesses, so aligned memory accesses
are already implicitly required).
The `-Wcast-align` warning seems to oversensitive as it flags code like
this: https://godbolt.org/z/c8481o8aa
This used to be added for aarch64 in a473cfd71b, but was later removed
in 11896868d6.
This flag makes the linker bind default-visibility functions locally, so
that calls to them do not have to go through the PLT. This makes it
impossible to override them by preloading a DSO. This was already the
case partially due to `-fno-semantic-interposition`, however that flag
is only able to optimize call sites that are in the same Translation
Unit as the function definitions.
This removes 80% of the PLT relocations in `libjs.so.serenity`.
Obsoletes #20877
This makes CMake pass `-fpie` instead of `-fpic` to the compiler when
building the Kernel and userland *executables*. This allows the compiler
to make certain optimizations based on the fact that the code will be
used in an executable, such as not having to emit `.localalias` symbols.
This leads to a 450 KiB decrease in the size of the Kernel binary.
We currently only use frame pointer-based backtrace generation.
This option is necessary for RISC-V as otherwise the compiler doesn't
save `fp` most of the time.
I made this flag not riscv64 exclusive, as we should do everything
to make that kind of backtrace generation work.
(1148020960)
Two non-functional changes:
- Remove pointless `-latomic` flag. It was specified via
`add_compile_options`, which only affects compilation and not linking,
so the library was never actually linked into the kernel. In fact, we
do not even build `libatomic` for our toolchain.
- Do not disable `-Wnonnull`. The warning-causing code was fixed at some
point.
This commit also removes `-mstrict-align` from the userland. Our target
AArch64 hardware natively supports unaligned accesses without a
significant performance penalty. Allowing the compiler to insert
unaligned accesses into aligned-as-written code allows for some
performance optimizations in fact. We keep this option turned on in the
kernel to preserve correctness for MMIO, as that might be sensitive to
alignment.
For the most part no behavior change, except that we now pass
-Wno-implicit-const-int-float-conversion and -Wno-literal-suffix
only to clang and gcc each in both lagom and serenity builds,
while we previously passed them to both in lagom builds (and
passed them to one each in serenity builds). The former is
a clang flag, the latter a gcc flag, but since we also use
-Wno-unknown-warning-option it doesn't really matter.
The aarch64 processor is set up to trap on unaligned memory accesses, so
to enforce that the compiler correctly generates aligned accesses, the
-mstrict-align flag is needed. We also need the -Wno-cast-align as there
are some files in AK that don't build without the flag.
Setting the DWARF version after having selected which level of debug
information to generate apparently undoes some settings again.
Doing the reverse apparently keeps both the version and the debug level
setting, resulting in a significantly smaller disk image size.
We currently don't support DWARF revision 5 and LLVM/Clang might create
such debug info into our binaries in x86_64, which will lead to a crash
in CrashReporter that is unable to parse that information correctly.
The shared parts are now firmly compiled into LibC instead of being
defined as a static library and then being copied over manually.
The non-shared ("local") parts are kept as a static library that is
linked into each binary on demand.
This finally allows us to support linking with the -fstack-protector
flag, which now replaces the `ssp` target being linked into each binary
accidentally via CMake.
