On x86_64 GCC implements va_list as an array. This makes the syntax
for taking a pointer to it break & crash. The workaround / solution is
to create a copy. Since va_list is a tiny struct referencing the
actual varargs, this is little overhead (especially compared to
va_args itself)
We have had these for quite a while, but we didn't compile them, and
used GCC's version instead. Clang does not come with these, so we have
to provide our own implementation.
Our implementation follows what `musl` and `FreeBSD` do, so this should
work fine, even if documentation can hardly be found for them.
This is required to make SSE instructions work when building with
Clang. Apparently Clang uses SSE instructions where GCC didn't so
we didn't previously run into this problem.
It's a single expression, no do/while needed. This makes assert() work
with the comma operator (assert(foo), assert(bar), assert(baz)).
Found because exactly this is being used somewhere in the guts of LLVM.
The System V ABI for both x86 and x86_64 requires that the stack pointer
is 16-byte aligned on entry. Previously we did not align the stack
pointer properly.
As far as "main" was concerned the stack alignment was correct even
without this patch due to how the C++ _start function and the kernel
interacted, i.e. the kernel misaligned the stack as far as the ABI
was concerned but that misalignment (read: it was properly aligned for
a regular function call - but misaligned in terms of what the ABI
dictates) was actually expected by our _start function.
As we've opted to make these inline functions and not macros, let's at
least make sure that the users don't *observe* multiple definitions of
these functions.
- Use a simple pthread_mutex_t instead of bringing in headers from
LibThreading just to get a mutex.
- Use a normal mutex instead of a recursive one.
- Remove redundant locking in realloc().
This fixes a build issue where `compiler-rt` tried to declare its own
version of the `FE_*` macros, and included `float.h` in order to get the
constants. `compiler-rt` tried to declare these as an enum, and failed
with a syntax error when the constant's literal values were substituted.
Without these attributes present on these custom formatting functions,
Clang would warn use about nonliteral format strings. As an added bonus,
we now get type checking on these.
This was an old SerenityOS-specific syscall for donating the remainder
of the calling thread's time-slice to another thread within the same
process.
Now that Threading::Lock uses a pthread_mutex_t internally, we no
longer need this syscall, which allows us to get rid of a surprising
amount of unnecessary scheduler logic. :^)
This implementation features a fast path for pthread_cond_signal() and
pthread_cond_broadcast() for the case there's no thread waiting, and
does not exhibit the "thundering herd" issue in
pthread_cond_broadcast().
Fixes https://github.com/SerenityOS/serenity/issues/8432
This is a private function that locks the lock much like the regular
pthread_mutex_lock(), but causes the corresponding unlock operation to
always assume there may be other waiters. This is useful in case some
waiters are made to wait on the mutex's futex directly, without going
through pthread_mutex_lock(). This is going to be used by the condition
variable implementation in the next commit.
pthread_mutex is now an actual "sleeping" mutex, and not just a
spinlock! It still has a fast path that only uses atomics and (in the
successful case) returns immediately without sleeping. In case of
contention, it calls futex_wait(), which lets the kernel scheduler put
this thread to sleep, *and* lets it know exactly when to consider
scheduling it again.
These are convinient wrappers over the most used futex operations.
futex_wait() also does some smarts for timeout and clock handling.
Use the new futex_wait() instead of a similar private helper in
LibPthread.
Depending on the driver, the second buffer may not be located right
after the first, e.g. it may be page aligned. This removes this
assumption and queries the driver for the appropriate offset.
Some devices may require DMA transfers to flush the updated buffer
areas prior to flipping. For those devices we track the areas that
require flushing prior to the next flip. For devices that do not
support flipping, but require flushing, we'll simply flush after
updating the front buffer.
This also adds a small optimization that skips these steps entirely for
a screen that doesn't have any updates that need to be rendered.
These were preventing some AK classes from using the AK Concepts header
due to the non-strictly namespaced ConversionSpecifier::Unsigned, and
are not used as their underlying value, so enum classes are more
appropriate anyways.
When creating uninitialized storage for variables, we need to make sure
that the alignment is correct. Fixes a KUBSAN failure when running
kernels compiled with Clang.
In `Syscalls/socket.cpp`, we can simply use local variables, as
`sockaddr_un` is a POD type.
Along with moving the `alignas` specifier to the correct member,
`AK::Optional`'s internal buffer has been made non-zeroed by default.
GCC emitted bogus uninitialized memory access warnings, so we now use
`__builtin_launder` to tell the compiler that we know what we are doing.
This might disable some optimizations, but judging by how GCC failed to
notice that the memory's initialization is dependent on `m_has_value`,
I'm not sure that's a bad thing.
We regularily need to flush many rectangles, so instead of making many
expensive ioctl() calls to the framebuffer driver, collect the
rectangles and only make one call. And if we have too many rectangles
then it may be cheaper to just update the entire region, in which case
we simply convert them all into a union and just flush that one
rectangle instead.
