[lworld] RFR: 8276812: [lworld] [lw3] JVM_ACC_FIELD_INLINED flag creates a conflict with an existing flag

[Mandy Chung]

On Wed, 10 Nov 2021 19:57:42 GMT, Frederic Parain <fparain at openjdk.org> wrote:

> Please review this changeset removing the JVM_ACC_FIELD_INLINED flag from AccessFlags which was conflicting with an existing JVMTI flag.
> The flag was use by Unsafe and reflection.
> Thanks to Mandy for her fix to the MemberName code.
> 
> Tested locally with hotspot_valhalla and jdk_valhalla test suites on x86.
> Testing in progress with Mach5 tiers 1 to 3 (no failure related to these changes yet).
> 
> Thank you,
> 
> Fred

Other than the null check needed in `Unsafe::isFlattened`, it looks good.

src/java.base/share/classes/jdk/internal/misc/Unsafe.java line 185:

> 183:      */
> 184:     public boolean isFlattened(Field f) {
> 185:         return isFlattenedField0(f);

It should do a null check before calling the native entry:


    if (f == null) {
         throw new NullPointerException();
    }

-------------

Marked as reviewed by mchung (Committer).

PR: https://git.openjdk.java.net/valhalla/pull/580