riscv, bpf: Fix broken BPF tail calls

The BPF JIT incorrectly clobbered the a0 register, and did not flag
usage of s5 register when BPF stack was being used.

Fixes: 2353ecc6 ("bpf, riscv: add BPF JIT for RV64G")
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20191216091343.23260-2-bjorn.topel@gmail.com

arch/riscv/net/bpf_jit_comp.c

@@ -120,6 +120,11 @@ static bool seen_reg(int reg, struct rv_jit_context *ctx)
 	return false;
 }
 
+static void mark_fp(struct rv_jit_context *ctx)
+{
+	__set_bit(RV_CTX_F_SEEN_S5, &ctx->flags);
+}
+
 static void mark_call(struct rv_jit_context *ctx)
 {
 	__set_bit(RV_CTX_F_SEEN_CALL, &ctx->flags);
@@ -596,7 +601,8 @@ static void __build_epilogue(u8 reg, struct rv_jit_context *ctx)
 
 	emit(rv_addi(RV_REG_SP, RV_REG_SP, stack_adjust), ctx);
 	/* Set return value. */
-	emit(rv_addi(RV_REG_A0, RV_REG_A5, 0), ctx);
+	if (reg == RV_REG_RA)
+		emit(rv_addi(RV_REG_A0, RV_REG_A5, 0), ctx);
 	emit(rv_jalr(RV_REG_ZERO, reg, 0), ctx);
 }
 
@@ -1426,6 +1432,10 @@ static void build_prologue(struct rv_jit_context *ctx)
 {
 	int stack_adjust = 0, store_offset, bpf_stack_adjust;
 
+	bpf_stack_adjust = round_up(ctx->prog->aux->stack_depth, 16);
+	if (bpf_stack_adjust)
+		mark_fp(ctx);
+
 	if (seen_reg(RV_REG_RA, ctx))
 		stack_adjust += 8;
 	stack_adjust += 8; /* RV_REG_FP */
@@ -1443,7 +1453,6 @@ static void build_prologue(struct rv_jit_context *ctx)
 		stack_adjust += 8;
 
 	stack_adjust = round_up(stack_adjust, 16);
-	bpf_stack_adjust = round_up(ctx->prog->aux->stack_depth, 16);
 	stack_adjust += bpf_stack_adjust;
 
 	store_offset = stack_adjust - 8;