This patch adds the `USING_AK_GLOBALLY` macro which is enabled by
default, but can be overridden by build flags.
This is a step towards integrating Jakt and AK types.
Instead of having two separate implementations of AK::RefCounted, one
for userspace and one for kernelspace, there is now RefCounted and
AtomicRefCounted.
This mechanism was unsafe to use in any multithreaded context, since
the hook function was invoked on a raw pointer *after* decrementing
the local ref count.
Since we don't use it for anything anymore, let's just get rid of it.
cert-dcl50-cpp: No variadic functions, suppressed in RefCounted and
ThreadSafeRefCounted for implementing the magic one_ref_left and
will_be_destroyed functions.
cert-dcl58-cpp: No opening ::std, suppressed in the places we put names
in ::std to aid tools (move, forward, nullptr_t, align_val_t, etc).
Some time ago, automatic locking was added to the AK smart pointers to
paper over various race conditions in the kernel. Until we've actually
solved the issues in the kernel, we're stuck with the locking.
However, we don't need to punish single-threaded userspace programs with
the high cost of locking. This patch moves the thread-safe variants of
RefPtr, NonnullRefPtr, WeakPtr and RefCounted into Kernel/Library/.
Previously we'd incur the costs for a function call via the PLT even
for the most trivial ref-count actions like increasing/decreasing the
reference count.
By moving the code to the header file we allow the compiler to inline
this code into the caller's function.
SPDX License Identifiers are a more compact / standardized
way of representing file license information.
See: https://spdx.dev/resources/use/#identifiers
This was done with the `ambr` search and replace tool.
ambr --no-parent-ignore --key-from-file --rep-from-file key.txt rep.txt *
(...and ASSERT_NOT_REACHED => VERIFY_NOT_REACHED)
Since all of these checks are done in release builds as well,
let's rename them to VERIFY to prevent confusion, as everyone is
used to assertions being compiled out in release.
We can introduce a new ASSERT macro that is specifically for debug
checks, but I'm doing this wholesale conversion first since we've
accumulated thousands of these already, and it's not immediately
obvious which ones are suitable for ASSERT.
Problem:
- Many constructors are defined as `{}` rather than using the ` =
default` compiler-provided constructor.
- Some types provide an implicit conversion operator from `nullptr_t`
instead of requiring the caller to default construct. This violates
the C++ Core Guidelines suggestion to declare single-argument
constructors explicit
(https://isocpp.github.io/CppCoreGuidelines/CppCoreGuidelines#c46-by-default-declare-single-argument-constructors-explicit).
Solution:
- Change default constructors to use the compiler-provided default
constructor.
- Remove implicit conversion operators from `nullptr_t` and change
usage to enforce type consistency without conversion.
Since RefPtr<T> decrements the ref counter to 0 and after that starts
destructing the object, there is a window where the ref count is 0
and the weak references have not been revoked.
Also change WeakLink to be able to obtain a strong reference
concurrently and block revoking instead, which should happen a lot
less often.
Fixes a problem observed in #4621
Fixes a regression introduced by 5c1b3ce. The commit description there
asserts that the changes allow calling will_be_destroyed and
one_ref_left, which are not required to be const qualified. The
implementation in fact does require the methods to be const qualified,
because we forgot to add the const_cast inside the decltypes :^)
Problem:
- `typedef` is a keyword which comes from C and carries with it old
syntax that is hard to read.
- Creating type aliases with the `using` keyword allows for easier
future maintenance because it supports template syntax.
- There is inconsistent use of `typedef` vs `using`.
Solution:
- Use `clang-tidy`'s checker called `modernize-use-using` to update
the syntax to use the newer syntax.
- Remove unused functions to make `clang-tidy` happy.
- This results in consistency within the codebase.
This makes most operations thread safe, especially so that they
can safely be used in the Kernel. This includes obtaining a strong
reference from a weak reference, which now requires an explicit
call to WeakPtr::strong_ref(). Another major change is that
Weakable::make_weak_ref() may require the explicit target type.
Previously we used reinterpret_cast in WeakPtr, assuming that it
can be properly converted. But WeakPtr does not necessarily have
the knowledge to be able to do this. Instead, we now ask the class
itself to deliver a WeakPtr to the type that we want.
Also, WeakLink is no longer specific to a target type. The reason
for this is that we want to be able to safely convert e.g. WeakPtr<T>
to WeakPtr<U>, and before this we just reinterpret_cast the internal
WeakLink<T> to WeakLink<U>, which is a bold assumption that it would
actually produce the correct code. Instead, WeakLink now operates
on just a raw pointer and we only make those constructors/operators
available if we can verify that it can be safely cast.
In order to guarantee thread safety, we now use the least significant
bit in the pointer for locking purposes. This also means that only
properly aligned pointers can be used.
This looks at three things:
- if the type has a typedef `AllowOwnPtr', respect that
- if not, disallow construction if both of `ref()' and `unref()' are
present.
Note that in the second case, if a type only defines `ref()' or only
defines `unref()', an OwnPtr can be created, as a RefPtr of that type
would be ill-formed.
Also marks a `Performance' to explicitly allow OwnPtrs.
Before, we had about these occurrence counts:
COPY: 13 without, 33 with
MOVE: 12 without, 28 with
Clearly, 'with' was the preferred way. However, this introduced double-semicolons
all over the place, and caused some warnings to trigger.
This patch *forces* the usage of a semi-colon when calling the macro,
by removing the semi-colon within the macro. (And thus also gets rid
of the double-semicolon.)
Before this, it has been possible to assign a RefCounted object to another
RefCounted object. Hilariosly (or sadly), that copied the refcount among
the other fields, meaning the target value ended up with a wrong refcount.
Ensure this never happens by disallowing copies and moves for RefCounted types.
This fixes all sorts of race conditions, primarily in the kernel, where till
now it's been possible to obtain either double free or use-after-free by
exploiting refcounting races.
If these methods get inlined, the compiler is able to statically eliminate most
of the assertions. Alas, it doesn't realize this, and believes inlining them to
be too expensive. So give it a strong hint that it's not the case.
This *decreases* the kernel binary size.
We allow the ref-counting parts of an object to be mutated even when the
object itself is a const.
An important detail is that we allow invoking 'will_be_destroyed' and
'one_ref_left', which are not required to be const qualified, on const
objects.
As suggested by Joshua, this commit adds the 2-clause BSD license as a
comment block to the top of every source file.
For the first pass, I've just added myself for simplicity. I encourage
everyone to add themselves as copyright holders of any file they've
added or modified in some significant way. If I've added myself in
error somewhere, feel free to replace it with the appropriate copyright
holder instead.
Going forward, all new source files should include a license header.
