arm32, bpf: add support for sign-extension mov instruction

The cpuv4 added a new BPF_MOVSX instruction that sign extends the src
before moving it to the destination.

BPF_ALU | BPF_MOVSX sign extends 8-bit and 16-bit operands into 32-bit
operands, and zeroes the remaining upper 32 bits.

BPF_ALU64 | BPF_MOVSX sign extends 8-bit, 16-bit, and 32-bit  operands
into 64-bit operands.

The offset field of the instruction is used to tell the number of bit to
use for sign-extension. BPF_MOV and BPF_MOVSX have the same code but the
former sets offset to 0 and the later one sets the offset to 8, 16 or 32

The behaviour of this instruction is dst = (s8,s16,s32)src

On ARM32 the implementation uses LSH and ARSH to extend the 8/16 bits to
a 32-bit register and then it is sign extended to the upper 32-bit
register using ARSH. For 32-bit we just move it to the destination
register and use ARSH to extend it to the upper 32-bit register.

Signed-off-by: Puranjay Mohan <[email protected]>
Reviewed-by: Russell King (Oracle) <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alexei Starovoitov <[email protected]>

arch/arm/net/bpf_jit_32.c

@@ -747,12 +747,16 @@ static inline void emit_a32_alu_r64(const bool is64, const s8 dst[],
 }
 
 /* dst = src (4 bytes)*/
-static inline void emit_a32_mov_r(const s8 dst, const s8 src,
+static inline void emit_a32_mov_r(const s8 dst, const s8 src, const u8 off,
 				  struct jit_ctx *ctx) {
 	const s8 *tmp = bpf2a32[TMP_REG_1];
 	s8 rt;
 
 	rt = arm_bpf_get_reg32(src, tmp[0], ctx);
+	if (off && off != 32) {
+		emit(ARM_LSL_I(rt, rt, 32 - off), ctx);
+		emit(ARM_ASR_I(rt, rt, 32 - off), ctx);
+	}
 	arm_bpf_put_reg32(dst, rt, ctx);
 }
 
@@ -761,15 +765,15 @@ static inline void emit_a32_mov_r64(const bool is64, const s8 dst[],
 				  const s8 src[],
 				  struct jit_ctx *ctx) {
 	if (!is64) {
-		emit_a32_mov_r(dst_lo, src_lo, ctx);
+		emit_a32_mov_r(dst_lo, src_lo, 0, ctx);
 		if (!ctx->prog->aux->verifier_zext)
 			/* Zero out high 4 bytes */
 			emit_a32_mov_i(dst_hi, 0, ctx);
 	} else if (__LINUX_ARM_ARCH__ < 6 &&
 		   ctx->cpu_architecture < CPU_ARCH_ARMv5TE) {
 		/* complete 8 byte move */
-		emit_a32_mov_r(dst_lo, src_lo, ctx);
-		emit_a32_mov_r(dst_hi, src_hi, ctx);
+		emit_a32_mov_r(dst_lo, src_lo, 0, ctx);
+		emit_a32_mov_r(dst_hi, src_hi, 0, ctx);
 	} else if (is_stacked(src_lo) && is_stacked(dst_lo)) {
 		const u8 *tmp = bpf2a32[TMP_REG_1];
 
@@ -785,6 +789,24 @@ static inline void emit_a32_mov_r64(const bool is64, const s8 dst[],
 	}
 }
 
+/* dst = (signed)src */
+static inline void emit_a32_movsx_r64(const bool is64, const u8 off, const s8 dst[], const s8 src[],
+				      struct jit_ctx *ctx) {
+	const s8 *tmp = bpf2a32[TMP_REG_1];
+	const s8 *rt;
+
+	rt = arm_bpf_get_reg64(dst, tmp, ctx);
+
+	emit_a32_mov_r(dst_lo, src_lo, off, ctx);
+	if (!is64) {
+		if (!ctx->prog->aux->verifier_zext)
+			/* Zero out high 4 bytes */
+			emit_a32_mov_i(dst_hi, 0, ctx);
+	} else {
+		emit(ARM_ASR_I(rt[0], rt[1], 31), ctx);
+	}
+}
+
 /* Shift operations */
 static inline void emit_a32_alu_i(const s8 dst, const u32 val,
 				struct jit_ctx *ctx, const u8 op) {
@@ -1450,7 +1472,10 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx)
 				emit_a32_mov_i(dst_hi, 0, ctx);
 				break;
 			}
-			emit_a32_mov_r64(is64, dst, src, ctx);
+			if (insn->off)
+				emit_a32_movsx_r64(is64, insn->off, dst, src, ctx);
+			else
+				emit_a32_mov_r64(is64, dst, src, ctx);
 			break;
 		case BPF_K:
 			/* Sign-extend immediate value to destination reg */