+static void EmitBranchToAt(uint64_t At, uint64_t To, bool isCall, bool is64Bit){
+ intptr_t Offset = ((intptr_t)To - (intptr_t)At) >> 2;
+ unsigned *AtI = (unsigned*)(intptr_t)At;
+
+ if (Offset >= -(1 << 23) && Offset < (1 << 23)) { // In range?
+ AtI[0] = BUILD_B(Offset, isCall); // b/bl target
+ } else if (!is64Bit) {
+ AtI[0] = BUILD_LIS(12, To >> 16); // lis r12, hi16(address)
+ AtI[1] = BUILD_ORI(12, 12, To); // ori r12, r12, lo16(address)
+ AtI[2] = BUILD_MTCTR(12); // mtctr r12
+ AtI[3] = BUILD_BCTR(isCall); // bctr/bctrl
+ } else {
+ AtI[0] = BUILD_LIS(12, To >> 48); // lis r12, hi16(address)
+ AtI[1] = BUILD_ORI(12, 12, To >> 32); // ori r12, r12, lo16(address)
+ AtI[2] = BUILD_SLDI(12, 12, 32); // sldi r12, r12, 32
+ AtI[3] = BUILD_ORIS(12, 12, To >> 16); // oris r12, r12, hi16(address)
+ AtI[4] = BUILD_ORI(12, 12, To); // ori r12, r12, lo16(address)
+ AtI[5] = BUILD_MTCTR(12); // mtctr r12
+ AtI[6] = BUILD_BCTR(isCall); // bctr/bctrl
+ }
+}
+
+extern "C" void PPC32CompilationCallback();
+extern "C" void PPC64CompilationCallback();
+
+// The first clause of the preprocessor directive looks wrong, but it is
+// necessary when compiling this code on non-PowerPC hosts.
+#if (!defined(__ppc__) && !defined(__powerpc__)) || defined(__powerpc64__) || defined(__ppc64__)
+void PPC32CompilationCallback() {
+ llvm_unreachable("This is not a 32bit PowerPC, you can't execute this!");
+}
+#elif !defined(__ELF__)
+// CompilationCallback stub - We can't use a C function with inline assembly in
+// it, because we the prolog/epilog inserted by GCC won't work for us. Instead,
+// write our own wrapper, which does things our way, so we have complete control
+// over register saving and restoring.
+asm(
+ ".text\n"
+ ".align 2\n"
+ ".globl _PPC32CompilationCallback\n"
+"_PPC32CompilationCallback:\n"
+ // Make space for 8 ints r[3-10] and 13 doubles f[1-13] and the
+ // FIXME: need to save v[0-19] for altivec?
+ // FIXME: could shrink frame
+ // Set up a proper stack frame
+ // FIXME Layout
+ // PowerPC32 ABI linkage - 24 bytes
+ // parameters - 32 bytes
+ // 13 double registers - 104 bytes
+ // 8 int registers - 32 bytes
+ "mflr r0\n"
+ "stw r0, 8(r1)\n"
+ "stwu r1, -208(r1)\n"
+ // Save all int arg registers
+ "stw r10, 204(r1)\n" "stw r9, 200(r1)\n"
+ "stw r8, 196(r1)\n" "stw r7, 192(r1)\n"
+ "stw r6, 188(r1)\n" "stw r5, 184(r1)\n"
+ "stw r4, 180(r1)\n" "stw r3, 176(r1)\n"
+ // Save all call-clobbered FP regs.
+ "stfd f13, 168(r1)\n" "stfd f12, 160(r1)\n"
+ "stfd f11, 152(r1)\n" "stfd f10, 144(r1)\n"
+ "stfd f9, 136(r1)\n" "stfd f8, 128(r1)\n"
+ "stfd f7, 120(r1)\n" "stfd f6, 112(r1)\n"
+ "stfd f5, 104(r1)\n" "stfd f4, 96(r1)\n"
+ "stfd f3, 88(r1)\n" "stfd f2, 80(r1)\n"
+ "stfd f1, 72(r1)\n"
+ // Arguments to Compilation Callback:
+ // r3 - our lr (address of the call instruction in stub plus 4)
+ // r4 - stub's lr (address of instruction that called the stub plus 4)
+ // r5 - is64Bit - always 0.
+ "mr r3, r0\n"
+ "lwz r2, 208(r1)\n" // stub's frame
+ "lwz r4, 8(r2)\n" // stub's lr
+ "li r5, 0\n" // 0 == 32 bit
+ "bl _LLVMPPCCompilationCallback\n"
+ "mtctr r3\n"
+ // Restore all int arg registers
+ "lwz r10, 204(r1)\n" "lwz r9, 200(r1)\n"
+ "lwz r8, 196(r1)\n" "lwz r7, 192(r1)\n"
+ "lwz r6, 188(r1)\n" "lwz r5, 184(r1)\n"
+ "lwz r4, 180(r1)\n" "lwz r3, 176(r1)\n"
+ // Restore all FP arg registers
+ "lfd f13, 168(r1)\n" "lfd f12, 160(r1)\n"
+ "lfd f11, 152(r1)\n" "lfd f10, 144(r1)\n"
+ "lfd f9, 136(r1)\n" "lfd f8, 128(r1)\n"
+ "lfd f7, 120(r1)\n" "lfd f6, 112(r1)\n"
+ "lfd f5, 104(r1)\n" "lfd f4, 96(r1)\n"
+ "lfd f3, 88(r1)\n" "lfd f2, 80(r1)\n"
+ "lfd f1, 72(r1)\n"
+ // Pop 3 frames off the stack and branch to target
+ "lwz r1, 208(r1)\n"
+ "lwz r2, 8(r1)\n"
+ "mtlr r2\n"
+ "bctr\n"
+ );
+
+#else
+// ELF PPC 32 support
+
+// CompilationCallback stub - We can't use a C function with inline assembly in
+// it, because we the prolog/epilog inserted by GCC won't work for us. Instead,
+// write our own wrapper, which does things our way, so we have complete control
+// over register saving and restoring.
+asm(
+ ".text\n"
+ ".align 2\n"
+ ".globl PPC32CompilationCallback\n"
+"PPC32CompilationCallback:\n"
+ // Make space for 8 ints r[3-10] and 8 doubles f[1-8] and the
+ // FIXME: need to save v[0-19] for altivec?
+ // FIXME: could shrink frame
+ // Set up a proper stack frame
+ // FIXME Layout
+ // 8 double registers - 64 bytes
+ // 8 int registers - 32 bytes
+ "mflr 0\n"
+ "stw 0, 4(1)\n"
+ "stwu 1, -104(1)\n"
+ // Save all int arg registers
+ "stw 10, 100(1)\n" "stw 9, 96(1)\n"
+ "stw 8, 92(1)\n" "stw 7, 88(1)\n"
+ "stw 6, 84(1)\n" "stw 5, 80(1)\n"
+ "stw 4, 76(1)\n" "stw 3, 72(1)\n"
+ // Save all call-clobbered FP regs.
+ "stfd 8, 64(1)\n"
+ "stfd 7, 56(1)\n" "stfd 6, 48(1)\n"
+ "stfd 5, 40(1)\n" "stfd 4, 32(1)\n"
+ "stfd 3, 24(1)\n" "stfd 2, 16(1)\n"
+ "stfd 1, 8(1)\n"
+ // Arguments to Compilation Callback:
+ // r3 - our lr (address of the call instruction in stub plus 4)
+ // r4 - stub's lr (address of instruction that called the stub plus 4)
+ // r5 - is64Bit - always 0.
+ "mr 3, 0\n"
+ "lwz 5, 104(1)\n" // stub's frame
+ "lwz 4, 4(5)\n" // stub's lr
+ "li 5, 0\n" // 0 == 32 bit
+ "bl LLVMPPCCompilationCallback\n"
+ "mtctr 3\n"
+ // Restore all int arg registers
+ "lwz 10, 100(1)\n" "lwz 9, 96(1)\n"
+ "lwz 8, 92(1)\n" "lwz 7, 88(1)\n"
+ "lwz 6, 84(1)\n" "lwz 5, 80(1)\n"
+ "lwz 4, 76(1)\n" "lwz 3, 72(1)\n"
+ // Restore all FP arg registers
+ "lfd 8, 64(1)\n"
+ "lfd 7, 56(1)\n" "lfd 6, 48(1)\n"
+ "lfd 5, 40(1)\n" "lfd 4, 32(1)\n"
+ "lfd 3, 24(1)\n" "lfd 2, 16(1)\n"
+ "lfd 1, 8(1)\n"
+ // Pop 3 frames off the stack and branch to target
+ "lwz 1, 104(1)\n"
+ "lwz 0, 4(1)\n"
+ "mtlr 0\n"
+ "bctr\n"
+ );
+#endif
+
+#if !defined(__powerpc64__) && !defined(__ppc64__)
+void PPC64CompilationCallback() {
+ llvm_unreachable("This is not a 64bit PowerPC, you can't execute this!");