* `RB_OBJ_SET_SHAREABLE(obj)` makes obj shareable.
All of reachable objects from `obj` should be shareable.
* `RB_OBJ_SET_FROZEN_SHAREABLE(obj)` same as above
but freeze `obj` before making it shareable.
Also `rb_gc_verify_shareable(obj)` is introduced to check
the `obj` does not violate shareable rule (an shareable object
only refers shareable objects) strictly.
The rule has some exceptions (some shareable objects can refer to
unshareable objects, such as a Ractor object (which is a shareable
object) can refer to the Ractor local objects.
To handle such case, `check_shareable` flag is also introduced.
`STRICT_VERIFY_SHAREABLE` macro is also introduced to verify
the strict shareable rule at `SET_SHAREABLE`.
Ref: https://github.com/ruby/ruby/pull/14134#issuecomment-3207733725
We can't safely use low-bit pointer tagging anymore because `RTypedData.type`
lines up with `RData.dfree` and there is no aligment guarantee on function
pointers, as evidenced by `memcached` and `gpgme` gems.
We also can't use FL_USER* for this, because extensions may use these
for other purposes.
Using a general flag for this is a bit unfortunate, as general flags
are hard to come by, however I recently freed several of them, and
we still have two or three free ones left.
Also adds a static assertion to ensure the documented behavior stays
true, namely that the data field is at the same position in the RData
and RTypedData structs.
Similar to f3206cc79bec2fd852e81ec56de59f0a67ab32b7 but for TypedData.
It's quite common for TypedData objects to have a mix of reference in
their struct and some ivars.
Since we do happen to have 8B free in the RtypedData struct, we could
use it to keep a direct reference to the IMEMO/fields saving having
to synchronize the VM and lookup the `gen_fields_tbl` on every ivar
access.
For old school Data classes however, we don't have free space, but
this API is soft-deprecated and no longer very common.
This only adds the rbimpl_ version to include/ruby/atomic.h so that it
is not a new public interface.
We were already using RUBY_ATOMIC_VALUE_LOAD in a few locations. This
will allow us to use other memory orders internally when desired.
"store" is the terminology the C11 standard uses, which allows us to use
this as a fallback.
This only changes the private rbimpl_ version of the method,
RUBY_ATOMIC_SET et al. keep the same name.
* Add support for `cause:` argument to `Fiber#raise` and `Thread#raise`.
The implementation behaviour is consistent with `Kernel#raise` and
`Exception#initialize` methods, allowing the `cause:` argument to be
passed to `Fiber#raise` and `Thread#raise`. This change ensures that
the `cause:` argument is handled correctly, providing a more consistent
and expected behavior when raising exceptions in fibers and threads.
[Feature #21360]
* Shared specs for Fiber/Thread/Kernel raise.
---------
Co-authored-by: Samuel Williams <samuel.williams@shopify.com>
These branches are only active for 32-bit Windows and Solaris platforms,
codify the fact by changing `#else` to `#elif`'s that explicitly include
those targets and `#error`-out otherwise.
This should be a minimal C-API needed to deal with Set objects. It
supports creating the sets, checking whether an element is the set,
adding and removing elements, iterating over the elements, clearing
a set, and returning the size of the set.
Co-authored-by: Nobuyoshi Nakada <nobu.nakada@gmail.com>
Previously, any time we used FL_TEST or variations we would need to add
an additional test that it wasn't a T_NODE. I think that was only needed
historically and we can avoid generating that extra code.
Some GC implementations want to always know when an object is written to,
even if the written value is a special constant. Checking special constants
in rb_obj_written was a micro-optimization that made assumptions about
the GC implementation.
Whenever we run into an inline cache miss when we try to set
an ivar, we may need to take the global lock, just to be able to
lookup inside `shape->edges`.
To solve that, when we're in multi-ractor mode, we can treat
the `shape->edges` as immutable. When we need to add a new
edge, we first copy the table, and then replace it with
CAS.
This increases memory allocations, however we expect that
creating new transitions becomes increasingly rare over time.
```ruby
class A
def initialize(bool)
@a = 1
if bool
@b = 2
else
@c = 3
end
end
def test
@d = 4
end
end
def bench(iterations)
i = iterations
while i > 0
A.new(true).test
A.new(false).test
i -= 1
end
end
if ARGV.first == "ractor"
ractors = 8.times.map do
Ractor.new do
bench(20_000_000 / 8)
end
end
ractors.each(&:take)
else
bench(20_000_000)
end
```
The above benchmark takes 27 seconds in Ractor mode on Ruby 3.4,
and only 1.7s with this branch.
Co-Authored-By: Étienne Barrié <etienne.barrie@gmail.com>
Previously we used a flag to set whether a module was uninitialized.
When checked whether a class was initialized, we first had to check that
it had a non-zero superclass, as well as that it wasn't BasicObject.
With the advent of namespaces, RCLASS_SUPER is now an expensive
operation, and though we could just check for the prime superclass, we
might as well take this opportunity to use a flag so that we can perform
the initialized check with as few instructions as possible.
It's possible in the future that we could prevent uninitialized classes
from being available to the user, but currently there are a few ways to
do that.
This makes `RBobject` `4B` larger on 32 bit systems
but simplifies the implementation a lot.
[Feature #21353]
Co-authored-by: Jean Boussier <byroot@ruby-lang.org>
As reported in <https://bugs.ruby-lang.org/issues/21340>, older autoconf
have an AC_HEADER_STDBOOL that's incompatible with C23. Autoconf 2.72
fixed the macro, but also mentions that it's obsolescent since all
current compilers have this header.
Since we require C99 [1] and VS 2015 [2], we might actually be able take
that suggestion and include stdbool.h without a check. I want to try
this on rubyci.org and will revert if this cause any issues. Not
touching AC_HEADER_STDBOOL in configure.ac for now.
[1]: https://bugs.ruby-lang.org/issues/15347
[2]: https://bugs.ruby-lang.org/issues/19982
`struct RTypedData` was changed significantly in https://github.com/ruby/ruby/pull/13190
which breaks many extensions.
Bumping the ABI version might save some people from needlessly
investigating crashes.
If the object isn't shareable and already has a object_id
we can access it without a lock.
If we need to generate an ID, we may need to lock to find
the child shape.
We also generate the next `object_id` using atomics.
And get rid of the `obj_to_id_tbl`
It's no longer needed, the `object_id` is now stored inline
in the object alongside instance variables.
We still need the inverse table in case `_id2ref` is invoked, but
we lazily build it by walking the heap if that happens.
The `object_id` concern is also no longer a GC implementation
concern, but a generic implementation.
Co-Authored-By: Matt Valentine-House <matt@eightbitraptor.com>
Ivars will longer be the only thing stored inline
via shapes, so keeping the `iv_index` and `ivptr` names
would be confusing.
Instance variables won't be the only thing stored inline
via shapes, so keeping the `ivptr` name would be confusing.
`field` encompass anything that can be stored in a VALUE array.
Similarly, `gen_ivtbl` becomes `gen_fields_tbl`.
