[8u] RFR: aarch32: array copy stub copies wrong data count

[Anton Kozlov]

Hi, All!

Current implementation of aarch32 array_copy_stub relies on word size to be 8 bytes, as it is on aarch64; it leads to less bytes copied than it should. Problem arises with copy of longs and copy of small number of elements. Test below demonstrates the issue:

=== Test ===
import java.lang.System;

class TestArrayCopy {
        static void test1() {
                long a0[] = { 0, 0 };
                long a1[] = { 1, 1 };
                System.arraycopy(a1, 0, a0, 0, 2); // should be inlined
                System.out.print(a0[1] == 1 ? "OK\n" : "FAIL\n");
        }

        static void test2() {
                int a10[] = { 0, 0, 0 };
                int a11[] = { 1, 1, 1 };
                System.arraycopy(a11, 0, a10, 0, 3); // should be inlined
                System.out.print(a10[2] == 1 ? "OK\n" : "FAIL\n");
        }

        public static void main(String[] args) {
                int a0[] = { 0 };
                int a1[] = { 0 };
                System.arraycopy(a1, 0, a0, 0, 1); // loads class

// System class should be loaded at the moment,
// test methods should not be inlined
                test1();
                test2();
        }
}

=== End Of Test ===

Patch in attach resolves the issue.

Thanks,
Anton
-------------- next part --------------
A non-text attachment was scrubbed...
Name: array_copy_stub.patch
Type: text/x-patch
Size: 2197 bytes
Desc: array_copy_stub.patch
URL: <http://mail.openjdk.java.net/pipermail/aarch32-port-dev/attachments/20160310/3868cf92/array_copy_stub.patch>