diff --git a/arch/tile/include/asm/Kbuild b/arch/tile/include/asm/Kbuild
index 2d1f563..20f2ba6 100644
--- a/arch/tile/include/asm/Kbuild
+++ b/arch/tile/include/asm/Kbuild
@@ -5,7 +5,6 @@
 generic-y += bugs.h
 generic-y += clkdev.h
 generic-y += cputime.h
-generic-y += div64.h
 generic-y += emergency-restart.h
 generic-y += errno.h
 generic-y += exec.h
diff --git a/arch/tile/include/asm/div64.h b/arch/tile/include/asm/div64.h
new file mode 100644
index 0000000..bf61619
--- /dev/null
+++ b/arch/tile/include/asm/div64.h
@@ -0,0 +1,14 @@
+#ifndef _ASM_TILE_DIV64_H
+#define _ASM_TILE_DIV64_H
+
+#ifdef __tilegx__
+static inline u64 mul_u32_u32(u32 a, u32 b)
+{
+	return __insn_mul_lu_lu(a, b);
+}
+#define mul_u32_u32 mul_u32_u32
+#endif
+
+#include <asm-generic/div64.h>
+
+#endif /* _ASM_TILE_DIV64_H */
diff --git a/arch/x86/include/asm/div64.h b/arch/x86/include/asm/div64.h
index ced283a..af95c47 100644
--- a/arch/x86/include/asm/div64.h
+++ b/arch/x86/include/asm/div64.h
@@ -59,6 +59,17 @@
 }
 #define div_u64_rem	div_u64_rem
 
+static inline u64 mul_u32_u32(u32 a, u32 b)
+{
+	u32 high, low;
+
+	asm ("mull %[b]" : "=a" (low), "=d" (high)
+			 : [a] "a" (a), [b] "rm" (b) );
+
+	return low | ((u64)high) << 32;
+}
+#define mul_u32_u32 mul_u32_u32
+
 #else
 # include <asm-generic/div64.h>
 #endif /* CONFIG_X86_32 */
diff --git a/include/linux/math64.h b/include/linux/math64.h
index 6e8b5b2..80690c9 100644
--- a/include/linux/math64.h
+++ b/include/linux/math64.h
@@ -133,6 +133,16 @@
 	return ret;
 }
 
+#ifndef mul_u32_u32
+/*
+ * Many a GCC version messes this up and generates a 64x64 mult :-(
+ */
+static inline u64 mul_u32_u32(u32 a, u32 b)
+{
+	return (u64)a * b;
+}
+#endif
+
 #if defined(CONFIG_ARCH_SUPPORTS_INT128) && defined(__SIZEOF_INT128__)
 
 #ifndef mul_u64_u32_shr
@@ -160,9 +170,9 @@
 	al = a;
 	ah = a >> 32;
 
-	ret = ((u64)al * mul) >> shift;
+	ret = mul_u32_u32(al, mul) >> shift;
 	if (ah)
-		ret += ((u64)ah * mul) << (32 - shift);
+		ret += mul_u32_u32(ah, mul) << (32 - shift);
 
 	return ret;
 }
@@ -186,10 +196,10 @@
 	a0.ll = a;
 	b0.ll = b;
 
-	rl.ll = (u64)a0.l.low * b0.l.low;
-	rm.ll = (u64)a0.l.low * b0.l.high;
-	rn.ll = (u64)a0.l.high * b0.l.low;
-	rh.ll = (u64)a0.l.high * b0.l.high;
+	rl.ll = mul_u32_u32(a0.l.low, b0.l.low);
+	rm.ll = mul_u32_u32(a0.l.low, b0.l.high);
+	rn.ll = mul_u32_u32(a0.l.high, b0.l.low);
+	rh.ll = mul_u32_u32(a0.l.high, b0.l.high);
 
 	/*
 	 * Each of these lines computes a 64-bit intermediate result into "c",
@@ -229,8 +239,8 @@
 	} u, rl, rh;
 
 	u.ll = a;
-	rl.ll = (u64)u.l.low * mul;
-	rh.ll = (u64)u.l.high * mul + rl.l.high;
+	rl.ll = mul_u32_u32(u.l.low, mul);
+	rh.ll = mul_u32_u32(u.l.high, mul) + rl.l.high;
 
 	/* Bits 32-63 of the result will be in rh.l.low. */
 	rl.l.high = do_div(rh.ll, divisor);