diff --git a/libc/arch-arm/bionic/memcpy.S b/libc/arch-arm/bionic/memcpy.S
index 0dc86d5..f25b3e3 100644
--- a/libc/arch-arm/bionic/memcpy.S
+++ b/libc/arch-arm/bionic/memcpy.S
@@ -352,9 +352,9 @@
 
         // preload the destination because we'll align it to a cache line
         // with small writes. Also start the source "pump".
-        PLD         (r0, #0)
-        PLD         (r1, #0)
-        PLD         (r1, #32)
+        pld         [r0, #0]
+        pld         [r1, #0]
+        pld         [r1, #32]
 
 		/* it simplifies things to take care of len<4 early */
 		cmp			r2, #4
@@ -442,7 +442,7 @@
         add         r12, r12, #64
 
 1:      ldmia       r1!, { r4-r11 }
-        PLD         (r12, #64)
+        pld         [r12, #64]
         subs        r2, r2, #32
 
         // NOTE: if r12 is more than 64 ahead of r1, the following ldrhi
@@ -563,7 +563,7 @@
         ldr         r12, [r1], #4
 1:      mov         r4, r12
 		ldmia		r1!, {   r5,r6,r7,  r8,r9,r10,r11}
-        PLD         (r1, #64)
+        pld         [r1, #64]
         subs        r2, r2, #32
         ldrhs       r12, [r1], #4
 		orr			r3, r3, r4,		lsl #16
@@ -590,7 +590,7 @@
         ldr         r12, [r1], #4
 1:      mov         r4, r12
 		ldmia		r1!, {   r5,r6,r7,  r8,r9,r10,r11}
-        PLD         (r1, #64)
+        pld         [r1, #64]
 		subs		r2, r2, #32
         ldrhs       r12, [r1], #4
 		orr			r3, r3, r4,		lsl #24
@@ -617,7 +617,7 @@
         ldr         r12, [r1], #4
 1:      mov         r4, r12
 		ldmia		r1!, {   r5,r6,r7,  r8,r9,r10,r11}
-        PLD         (r1, #64)
+        pld         [r1, #64]
 		subs		r2, r2, #32
         ldrhs       r12, [r1], #4
 		orr			r3, r3, r4,		lsl #8