ARM: net fix emit_udiv() for BPF_ALU | BPF_DIV | BPF_K intruction.

commit 19fc99d0 upstream.

In that case, emit_udiv() will be called with rn == ARM_R0 (r_scratch)
and loading rm first into ARM_R0 will result in jit_udiv() function
being called the same dividend and divisor. Fix that by loading rn
first into ARM_R1 and then rm into ARM_R0.
Signed-off-by: Nicolas Schichan <nschichan@freebox.fr>
Fixes: aee636c4 (bpf: do not use reciprocal divide)
Acked-by: Mircea Gherzan <mgherzan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

arch/arm/net/bpf_jit_32.c

@@ -449,10 +449,21 @@ static inline void emit_udiv(u8 rd, u8 rm, u8 rn, struct jit_ctx *ctx)
 		return;
 	}
 #endif
-	if (rm != ARM_R0)
-		emit(ARM_MOV_R(ARM_R0, rm), ctx);
+
+	/*
+	 * For BPF_ALU | BPF_DIV | BPF_K instructions, rm is ARM_R4
+	 * (r_A) and rn is ARM_R0 (r_scratch) so load rn first into
+	 * ARM_R1 to avoid accidentally overwriting ARM_R0 with rm
+	 * before using it as a source for ARM_R1.
+	 *
+	 * For BPF_ALU | BPF_DIV | BPF_X rm is ARM_R4 (r_A) and rn is
+	 * ARM_R5 (r_X) so there is no particular register overlap
+	 * issues.
+	 */
 	if (rn != ARM_R1)
 		emit(ARM_MOV_R(ARM_R1, rn), ctx);
+	if (rm != ARM_R0)
+		emit(ARM_MOV_R(ARM_R0, rm), ctx);
 
 	ctx->seen |= SEEN_CALL;
 	emit_mov_i(ARM_R3, (u32)jit_udiv, ctx);