Fix padding in jvmtiCapabilities structure

[EricYangIBM]

JvmtiCapabilities struct cannot have padding between fields.

Signed-off-by: Eric Yang eric.yang@ibm.com

[EricYangIBM]

Jdk19 before and after:

	unsigned int can_generate_resource_exhaustion_threads_events : 1;
	unsigned int can_generate_early_vmstart : 1;
	unsigned int can_generate_early_class_hook_events : 1;
	unsigned int can_generate_sampled_object_alloc_events : 1;
	unsigned int : 4;
	unsigned int can_support_virtual_threads : 1;
	unsigned int : 15;
	unsigned int : 16;
	unsigned int : 16;
	unsigned int : 16;
	unsigned int : 16;
} jvmtiCapabilities;

	unsigned int can_generate_resource_exhaustion_threads_events : 1;
	unsigned int can_generate_early_vmstart : 1;
	unsigned int can_generate_early_class_hook_events : 1;
	unsigned int can_generate_sampled_object_alloc_events : 1;
	unsigned int can_support_virtual_threads : 1;
	unsigned int : 3;
	unsigned int : 16;
	unsigned int : 16;
	unsigned int : 16;
	unsigned int : 16;
	unsigned int : 16;
} jvmtiCapabilities;

[EricYangIBM]

@tajila @babsingh I could also change it to use compound conditions.
Currently:

if (JAVA_SPEC_VERSION >= 9) {
	unsigned int can_generate_early_vmstart : 1;
	unsigned int can_generate_early_class_hook_events : 1;
	if (JAVA_SPEC_VERSION >= 11) {
		unsigned int can_generate_sampled_object_alloc_events : 1;
		if (JAVA_SPEC_VERSION >= 19) {
			unsigned int can_support_virtual_threads : 1;
			unsigned int : 3;
		} else {
			unsigned int : 4;
		}
	} else {
		unsigned int : 5;
	}
} else {
	unsigned int : 7;
}

Alternative:

if (JAVA_SPEC_VERSION >= 9 && JAVA_SPEC_VERSION < 11) {
	unsigned int can_generate_early_vmstart : 1;
	unsigned int can_generate_early_class_hook_events : 1;
	unsigned int : 5;
} else if (JAVA_SPEC_VERSION >= 11 && JAVA_SPEC_VERSION < 19) {
	unsigned int can_generate_early_vmstart : 1;
	unsigned int can_generate_early_class_hook_events : 1;
	unsigned int can_generate_sampled_object_alloc_events : 1;
	unsigned int : 4;
} else if (JAVA_SPEC_VERSION >= 19) {
	unsigned int can_generate_early_vmstart : 1;
	unsigned int can_generate_early_class_hook_events : 1;
	unsigned int can_generate_sampled_object_alloc_events : 1;
	unsigned int can_support_virtual_threads : 1;
	unsigned int : 3;
} else {
	unsigned int : 7;
}

Is this better?

[babsingh]

re #15451 (comment):

IMO, the alternative is easier to understand. So, +1 for the alternative. The current nested version can lead to errors due to lack of simplicity. Will the below also work; it avoids duplicates?

if (JAVA_SPEC_VERSION >= 9) {
	unsigned int can_generate_early_vmstart : 1;
	unsigned int can_generate_early_class_hook_events : 1;
}
if (JAVA_SPEC_VERSION >= 11 ) {
	unsigned int can_generate_sampled_object_alloc_events : 1;
}
if (JAVA_SPEC_VERSION >= 19 ) {
	unsigned int can_support_virtual_threads : 1;
}
/* Add a comment about PADDING:
 * There will be errors if PADDING is wrong, and
 * it needs to be updated whenever a new field is added.
 */
if (JAVA_SPEC_VERSION >= 9 && JAVA_SPEC_VERSION < 11) {
	unsigned int : 5;
} else if (JAVA_SPEC_VERSION >= 11 && JAVA_SPEC_VERSION < 19) {
	unsigned int : 4;
} else if (JAVA_SPEC_VERSION >= 19) {
	unsigned int : 3;
} else {
	/* PADDING for Java 8. */
	unsigned int : 7;
}

[EricYangIBM]

Will the below also work; it avoids duplicates?

I think this is more confusing than the first alternative
Edit - the above updated version looks good

[babsingh]

In the first alternative, the duplicates are a drawback. @tajila What are your thoughts?

[tajila]

I generally like to avoid duplication, so I dont mind the nesting too much

[EricYangIBM]

So this pr is fine to be merged?

[tajila]

jenkins test sanity amac jdk19

[tajila]

jenkins compile win jdk19

[tajila]

only test failure is known issue