Hi,
Currently, the class `AdapterFingerPrint` only tracks the types of the
arguments in the calling convention. This does not work well with the inclusion
of value classes which need to be packed and unpacked when passed between
different tiers. This is because 2 types can have the same sequence of field
types, yet the fields are placed at different offsets. In those cases, the code
to pack those types cannot be the same, hence the adapter code cannot be
shared. Currently, a hack is implemented to cover the cases of inherited
fields. For example:
value class MyAbstract {
int x;
}
value class MyValue1 extends MyAbstract {}
value class MyValue2 {
int x;
}
In this case, since `MyAbstract` is an abstract class, we do not know the
complete field sequence of every class that has the field `MyAbstract.x`.
Moreover, we need to ensure that the payload of a value class is stable when it
is flattened, and it can be accessed atomically if atomicity is required. As a
result, `MyAbstract.x` must have the highest alignment, which is currently 8
because the largest atomic flat access is 8-byte. However, `MyValue2.x` does
not need such a large alignment.
The current hack put a pair of `T_METADATA`, `T_VOID` before an inherited
field. This is fine for the case above, but it clashes with another case:
value class MyValue3 {
@NullRestricted
Empty e;
int x;
}
The solution I propose is to add field offsets into `AdapterFingerPrint`, as
well as removing the current hack. Please take a look and leave your reviews,
thanks a lot.
-------------
Commit messages:
- Fix incorrect adapter sharing with value classes
Changes: https://git.openjdk.org/valhalla/pull/1679/files
Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=1679&range=00
Issue: https://bugs.openjdk.org/browse/JDK-8348547
Stats: 252 lines in 3 files changed: 139 ins; 59 del; 54 mod
Patch: https://git.openjdk.org/valhalla/pull/1679.diff
Fetch: git fetch https://git.openjdk.org/valhalla.git pull/1679/head:pull/1679
PR: https://git.openjdk.org/valhalla/pull/1679
