$NetBSD: patch-bu,v 1.1.1.1 2008/11/06 11:50:47 thomasklausner Exp $

--- xpcom/reflect/xptcall/src/md/unix/xptcinvoke_unixish_amd64.cpp.orig	2006-02-02 17:14:03.000000000 +0100
+++ xpcom/reflect/xptcall/src/md/unix/xptcinvoke_unixish_amd64.cpp
@@ -0,0 +1,174 @@
+/* -*- Mode: C; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
+// Platform specific code to invoke XPCOM methods on native objects
+
+#include "xptcprivate.h"
+
+// 6 integral parameters are passed in registers
+const PRUint32 GPR_COUNT = 6;
+
+// 8 floating point parameters are passed in SSE registers
+const PRUint32 FPR_COUNT = 8;
+
+// Remember that these 'words' are 64-bit long
+static inline void
+invoke_count_words(PRUint32 paramCount, nsXPTCVariant * s,
+                   PRUint32 & nr_gpr, PRUint32 & nr_fpr, PRUint32 & nr_stack)
+{
+    nr_gpr = 1; // skip one GP register for 'that'
+    nr_fpr = 0;
+    nr_stack = 0;
+
+    /* Compute number of eightbytes of class MEMORY.  */
+    for (uint32 i = 0; i < paramCount; i++, s++) {
+        if (!s->IsPtrData()
+            && (s->type == nsXPTType::T_FLOAT || s->type == nsXPTType::T_DOUBLE)) {
+            if (nr_fpr < FPR_COUNT)
+                nr_fpr++;
+            else
+                nr_stack++;
+        }
+        else {
+            if (nr_gpr < GPR_COUNT)
+                nr_gpr++;
+            else
+                nr_stack++;
+        }
+    }
+}
+
+static void
+invoke_copy_to_stack(PRUint64 * d, PRUint32 paramCount, nsXPTCVariant * s,
+                     PRUint64 * gpregs, double * fpregs)
+{
+    PRUint32 nr_gpr = 1; // skip one GP register for 'that'
+    PRUint32 nr_fpr = 0;
+    PRUint64 value;
+
+    for (uint32 i = 0; i < paramCount; i++, s++) {
+        if (s->IsPtrData())
+            value = (PRUint64) s->ptr;
+        else {
+            switch (s->type) {
+            case nsXPTType::T_FLOAT:                                break;
+            case nsXPTType::T_DOUBLE:                               break;
+            case nsXPTType::T_I8:     value = s->val.i8;            break;
+            case nsXPTType::T_I16:    value = s->val.i16;           break;
+            case nsXPTType::T_I32:    value = s->val.i32;           break;
+            case nsXPTType::T_I64:    value = s->val.i64;           break;
+            case nsXPTType::T_U8:     value = s->val.u8;            break;
+            case nsXPTType::T_U16:    value = s->val.u16;           break;
+            case nsXPTType::T_U32:    value = s->val.u32;           break;
+            case nsXPTType::T_U64:    value = s->val.u64;           break;
+            case nsXPTType::T_BOOL:   value = s->val.b;             break;
+            case nsXPTType::T_CHAR:   value = s->val.c;             break;
+            case nsXPTType::T_WCHAR:  value = s->val.wc;            break;
+            default:                  value = (PRUint64) s->val.p;  break;
+            }
+        }
+
+        if (!s->IsPtrData() && s->type == nsXPTType::T_DOUBLE) {
+            if (nr_fpr < FPR_COUNT)
+                fpregs[nr_fpr++] = s->val.d;
+            else {
+                *((double *)d) = s->val.d;
+                d++;
+            }
+        }
+        else if (!s->IsPtrData() && s->type == nsXPTType::T_FLOAT) {
+            if (nr_fpr < FPR_COUNT)
+                // The value in %xmm register is already prepared to
+                // be retrieved as a float. Therefore, we pass the
+                // value verbatim, as a double without conversion.
+                fpregs[nr_fpr++] = s->val.d;
+            else {
+                *((float *)d) = s->val.f;
+                d++;
+            }
+        }
+        else {
+            if (nr_gpr < GPR_COUNT)
+                gpregs[nr_gpr++] = value;
+            else
+                *d++ = value;
+        }
+    }
+}
+
+extern "C"
+XPTC_PUBLIC_API(nsresult)
+XPTC_InvokeByIndex(nsISupports * that, PRUint32 methodIndex,
+                   PRUint32 paramCount, nsXPTCVariant * params)
+{
+    PRUint32 nr_gpr, nr_fpr, nr_stack;
+    invoke_count_words(paramCount, params, nr_gpr, nr_fpr, nr_stack);
+    
+    // Stack, if used, must be 16-bytes aligned
+    if (nr_stack)
+        nr_stack = (nr_stack + 1) & ~1;
+
+    // Load parameters to stack, if necessary
+    PRUint64 *stack = (PRUint64 *) __builtin_alloca(nr_stack * 8);
+    PRUint64 gpregs[GPR_COUNT];
+    double fpregs[FPR_COUNT];
+    invoke_copy_to_stack(stack, paramCount, params, gpregs, fpregs);
+
+    // Load FPR registers from fpregs[]
+    register double d0 asm("xmm0");
+    register double d1 asm("xmm1");
+    register double d2 asm("xmm2");
+    register double d3 asm("xmm3");
+    register double d4 asm("xmm4");
+    register double d5 asm("xmm5");
+    register double d6 asm("xmm6");
+    register double d7 asm("xmm7");
+
+    switch (nr_fpr) {
+#define ARG_FPR(N) \
+    case N+1: d##N = fpregs[N];
+        ARG_FPR(7);
+        ARG_FPR(6);
+        ARG_FPR(5);
+        ARG_FPR(4);
+        ARG_FPR(3);
+        ARG_FPR(2);
+        ARG_FPR(1);
+        ARG_FPR(0);
+    case 0:;
+#undef ARG_FPR
+    }
+    
+    // Load GPR registers from gpregs[]
+    register PRUint64 a0 asm("rdi");
+    register PRUint64 a1 asm("rsi");
+    register PRUint64 a2 asm("rdx");
+    register PRUint64 a3 asm("rcx");
+    register PRUint64 a4 asm("r8");
+    register PRUint64 a5 asm("r9");
+    
+    switch (nr_gpr) {
+#define ARG_GPR(N) \
+    case N+1: a##N = gpregs[N];
+        ARG_GPR(5);
+        ARG_GPR(4);
+        ARG_GPR(3);
+        ARG_GPR(2);
+        ARG_GPR(1);
+    case 1: a0 = (PRUint64) that;
+    case 0:;
+#undef ARG_GPR
+    }
+
+    // Ensure that assignments to SSE registers won't be optimized away
+    asm("" ::
+        "x" (d0), "x" (d1), "x" (d2), "x" (d3),
+        "x" (d4), "x" (d5), "x" (d6), "x" (d7));
+    
+    // Get pointer to method
+    PRUint64 methodAddress = *((PRUint64 *)that);
+    methodAddress += 8 * methodIndex;
+    methodAddress = *((PRUint64 *)methodAddress);
+    
+    typedef PRUint32 (*Method)(PRUint64, PRUint64, PRUint64, PRUint64, PRUint64, PRUint64);
+    PRUint32 result = ((Method)methodAddress)(a0, a1, a2, a3, a4, a5);
+    return result;
+}