Merge branch 'bpf_modify_ret'

KP Singh says:

====================
v3 -> v4:

* Fix a memory leak noticed by Daniel.

v2 -> v3:

* bpf_trampoline_update_progs -> bpf_trampoline_get_progs + const
  qualification.
* Typos in commit messages.
* Added Andrii's Acks.

v1 -> v2:

* Adressed Andrii's feedback.
* Fixed a bug that Alexei noticed about nop generation.
* Rebase.

This was brought up in the KRSI v4 discussion and found to be useful
both for security and tracing programs.

  https://lore.kernel.org/bpf/20200225193108.GB22391@chromium.org/

The modify_return programs are allowed for security hooks (with an
extra CAP_MAC_ADMIN check) and functions whitelisted for error
injection (ALLOW_ERROR_INJECTION).

The "security_" check is expected to be cleaned up with the KRSI patch
series.

Here is an example of how a fmod_ret program behaves:

int func_to_be_attached(int a, int b)
{  <--- do_fentry

do_fmod_ret:
   <update ret by calling fmod_ret>
   if (ret != 0)
        goto do_fexit;

original_function:

    <side_effects_happen_here>

}  <--- do_fexit

ALLOW_ERROR_INJECTION(func_to_be_attached, ERRNO)

The fmod_ret program attached to this function can be defined as:

SEC("fmod_ret/func_to_be_attached")
int BPF_PROG(func_name, int a, int b, int ret)
{
        // This will skip the original function logic.
        return -1;
}
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

arch/x86/net/bpf_jit_comp.c

@@ -1361,13 +1361,12 @@ static void restore_regs(const struct btf_func_model *m, u8 **prog, int nr_args,
 			 -(stack_size - i * 8));
 }
 
-static int invoke_bpf(const struct btf_func_model *m, u8 **pprog,
-		      struct bpf_prog **progs, int prog_cnt, int stack_size)
+static int invoke_bpf_prog(const struct btf_func_model *m, u8 **pprog,
+			   struct bpf_prog *p, int stack_size, bool mod_ret)
 {
 	u8 *prog = *pprog;
-	int cnt = 0, i;
+	int cnt = 0;
 
-	for (i = 0; i < prog_cnt; i++) {
 	if (emit_call(&prog, __bpf_prog_enter, prog))
 		return -EINVAL;
 	/* remember prog start time returned by __bpf_prog_enter */
@@ -1376,26 +1375,151 @@ static int invoke_bpf(const struct btf_func_model *m, u8 **pprog,
 	/* arg1: lea rdi, [rbp - stack_size] */
 	EMIT4(0x48, 0x8D, 0x7D, -stack_size);
 	/* arg2: progs[i]->insnsi for interpreter */
-		if (!progs[i]->jited)
+	if (!p->jited)
 		emit_mov_imm64(&prog, BPF_REG_2,
-				       (long) progs[i]->insnsi >> 32,
-				       (u32) (long) progs[i]->insnsi);
+			       (long) p->insnsi >> 32,
+			       (u32) (long) p->insnsi);
 	/* call JITed bpf program or interpreter */
-		if (emit_call(&prog, progs[i]->bpf_func, prog))
+	if (emit_call(&prog, p->bpf_func, prog))
 		return -EINVAL;
 
+	/* BPF_TRAMP_MODIFY_RETURN trampolines can modify the return
+	 * of the previous call which is then passed on the stack to
+	 * the next BPF program.
+	 */
+	if (mod_ret)
+		emit_stx(&prog, BPF_DW, BPF_REG_FP, BPF_REG_0, -8);
+
 	/* arg1: mov rdi, progs[i] */
-		emit_mov_imm64(&prog, BPF_REG_1, (long) progs[i] >> 32,
-			       (u32) (long) progs[i]);
+	emit_mov_imm64(&prog, BPF_REG_1, (long) p >> 32,
+		       (u32) (long) p);
 	/* arg2: mov rsi, rbx <- start time in nsec */
 	emit_mov_reg(&prog, true, BPF_REG_2, BPF_REG_6);
 	if (emit_call(&prog, __bpf_prog_exit, prog))
 		return -EINVAL;
+
+	*pprog = prog;
+	return 0;
+}
+
+static void emit_nops(u8 **pprog, unsigned int len)
+{
+	unsigned int i, noplen;
+	u8 *prog = *pprog;
+	int cnt = 0;
+
+	while (len > 0) {
+		noplen = len;
+
+		if (noplen > ASM_NOP_MAX)
+			noplen = ASM_NOP_MAX;
+
+		for (i = 0; i < noplen; i++)
+			EMIT1(ideal_nops[noplen][i]);
+		len -= noplen;
+	}
+
+	*pprog = prog;
+}
+
+static void emit_align(u8 **pprog, u32 align)
+{
+	u8 *target, *prog = *pprog;
+
+	target = PTR_ALIGN(prog, align);
+	if (target != prog)
+		emit_nops(&prog, target - prog);
+
+	*pprog = prog;
+}
+
+static int emit_cond_near_jump(u8 **pprog, void *func, void *ip, u8 jmp_cond)
+{
+	u8 *prog = *pprog;
+	int cnt = 0;
+	s64 offset;
+
+	offset = func - (ip + 2 + 4);
+	if (!is_simm32(offset)) {
+		pr_err("Target %p is out of range\n", func);
+		return -EINVAL;
+	}
+	EMIT2_off32(0x0F, jmp_cond + 0x10, offset);
+	*pprog = prog;
+	return 0;
+}
+
+static int emit_mod_ret_check_imm8(u8 **pprog, int value)
+{
+	u8 *prog = *pprog;
+	int cnt = 0;
+
+	if (!is_imm8(value))
+		return -EINVAL;
+
+	if (value == 0)
+		EMIT2(0x85, add_2reg(0xC0, BPF_REG_0, BPF_REG_0));
+	else
+		EMIT3(0x83, add_1reg(0xF8, BPF_REG_0), value);
+
+	*pprog = prog;
+	return 0;
+}
+
+static int invoke_bpf(const struct btf_func_model *m, u8 **pprog,
+		      struct bpf_tramp_progs *tp, int stack_size)
+{
+	int i;
+	u8 *prog = *pprog;
+
+	for (i = 0; i < tp->nr_progs; i++) {
+		if (invoke_bpf_prog(m, &prog, tp->progs[i], stack_size, false))
+			return -EINVAL;
 	}
 	*pprog = prog;
 	return 0;
 }
 
+static int invoke_bpf_mod_ret(const struct btf_func_model *m, u8 **pprog,
+			      struct bpf_tramp_progs *tp, int stack_size,
+			      u8 **branches)
+{
+	u8 *prog = *pprog;
+	int i;
+
+	/* The first fmod_ret program will receive a garbage return value.
+	 * Set this to 0 to avoid confusing the program.
+	 */
+	emit_mov_imm32(&prog, false, BPF_REG_0, 0);
+	emit_stx(&prog, BPF_DW, BPF_REG_FP, BPF_REG_0, -8);
+	for (i = 0; i < tp->nr_progs; i++) {
+		if (invoke_bpf_prog(m, &prog, tp->progs[i], stack_size, true))
+			return -EINVAL;
+
+		/* Generate a branch:
+		 *
+		 * if (ret !=  0)
+		 *	goto do_fexit;
+		 *
+		 * If needed this can be extended to any integer value which can
+		 * be passed by user-space when the program is loaded.
+		 */
+		if (emit_mod_ret_check_imm8(&prog, 0))
+			return -EINVAL;
+
+		/* Save the location of the branch and Generate 6 nops
+		 * (4 bytes for an offset and 2 bytes for the jump) These nops
+		 * are replaced with a conditional jump once do_fexit (i.e. the
+		 * start of the fexit invocation) is finalized.
+		 */
+		branches[i] = prog;
+		emit_nops(&prog, 4 + 2);
+	}
+
+	*pprog = prog;
+	return 0;
+}
+
 /* Example:
  * __be16 eth_type_trans(struct sk_buff *skb, struct net_device *dev);
  * its 'struct btf_func_model' will be nr_args=2
@@ -1458,12 +1582,15 @@ static int invoke_bpf(const struct btf_func_model *m, u8 **pprog,
  */
 int arch_prepare_bpf_trampoline(void *image, void *image_end,
 				const struct btf_func_model *m, u32 flags,
-				struct bpf_prog **fentry_progs, int fentry_cnt,
-				struct bpf_prog **fexit_progs, int fexit_cnt,
+				struct bpf_tramp_progs *tprogs,
 				void *orig_call)
 {
-	int cnt = 0, nr_args = m->nr_args;
+	int ret, i, cnt = 0, nr_args = m->nr_args;
 	int stack_size = nr_args * 8;
+	struct bpf_tramp_progs *fentry = &tprogs[BPF_TRAMP_FENTRY];
+	struct bpf_tramp_progs *fexit = &tprogs[BPF_TRAMP_FEXIT];
+	struct bpf_tramp_progs *fmod_ret = &tprogs[BPF_TRAMP_MODIFY_RETURN];
+	u8 **branches = NULL;
 	u8 *prog;
 
 	/* x86-64 supports up to 6 arguments. 7+ can be added in the future */
@@ -1492,28 +1619,64 @@ int arch_prepare_bpf_trampoline(void *image, void *image_end,
 
 	save_regs(m, &prog, nr_args, stack_size);
 
-	if (fentry_cnt)
-		if (invoke_bpf(m, &prog, fentry_progs, fentry_cnt, stack_size))
+	if (fentry->nr_progs)
+		if (invoke_bpf(m, &prog, fentry, stack_size))
 			return -EINVAL;
 
+	if (fmod_ret->nr_progs) {
+		branches = kcalloc(fmod_ret->nr_progs, sizeof(u8 *),
+				   GFP_KERNEL);
+		if (!branches)
+			return -ENOMEM;
+
+		if (invoke_bpf_mod_ret(m, &prog, fmod_ret, stack_size,
+				       branches)) {
+			ret = -EINVAL;
+			goto cleanup;
+		}
+	}
+
 	if (flags & BPF_TRAMP_F_CALL_ORIG) {
-		if (fentry_cnt)
+		if (fentry->nr_progs || fmod_ret->nr_progs)
 			restore_regs(m, &prog, nr_args, stack_size);
 
 		/* call original function */
-		if (emit_call(&prog, orig_call, prog))
-			return -EINVAL;
+		if (emit_call(&prog, orig_call, prog)) {
+			ret = -EINVAL;
+			goto cleanup;
+		}
 		/* remember return value in a stack for bpf prog to access */
 		emit_stx(&prog, BPF_DW, BPF_REG_FP, BPF_REG_0, -8);
 	}
 
-	if (fexit_cnt)
-		if (invoke_bpf(m, &prog, fexit_progs, fexit_cnt, stack_size))
-			return -EINVAL;
+	if (fmod_ret->nr_progs) {
+		/* From Intel 64 and IA-32 Architectures Optimization
+		 * Reference Manual, 3.4.1.4 Code Alignment, Assembly/Compiler
+		 * Coding Rule 11: All branch targets should be 16-byte
+		 * aligned.
+		 */
+		emit_align(&prog, 16);
+		/* Update the branches saved in invoke_bpf_mod_ret with the
+		 * aligned address of do_fexit.
+		 */
+		for (i = 0; i < fmod_ret->nr_progs; i++)
+			emit_cond_near_jump(&branches[i], prog, branches[i],
+					    X86_JNE);
+	}
+
+	if (fexit->nr_progs)
+		if (invoke_bpf(m, &prog, fexit, stack_size)) {
+			ret = -EINVAL;
+			goto cleanup;
+		}
 
 	if (flags & BPF_TRAMP_F_RESTORE_REGS)
 		restore_regs(m, &prog, nr_args, stack_size);
 
+	/* This needs to be done regardless. If there were fmod_ret programs,
+	 * the return value is only updated on the stack and still needs to be
+	 * restored to R0.
+	 */
 	if (flags & BPF_TRAMP_F_CALL_ORIG)
 		/* restore original return value back into RAX */
 		emit_ldx(&prog, BPF_DW, BPF_REG_0, BPF_REG_FP, -8);
@@ -1525,45 +1688,15 @@ int arch_prepare_bpf_trampoline(void *image, void *image_end,
 		EMIT4(0x48, 0x83, 0xC4, 8); /* add rsp, 8 */
 	EMIT1(0xC3); /* ret */
 	/* Make sure the trampoline generation logic doesn't overflow */
-	if (WARN_ON_ONCE(prog > (u8 *)image_end - BPF_INSN_SAFETY))
-		return -EFAULT;
-	return prog - (u8 *)image;
-}
-
-static int emit_cond_near_jump(u8 **pprog, void *func, void *ip, u8 jmp_cond)
-{
-	u8 *prog = *pprog;
-	int cnt = 0;
-	s64 offset;
-
-	offset = func - (ip + 2 + 4);
-	if (!is_simm32(offset)) {
-		pr_err("Target %p is out of range\n", func);
-		return -EINVAL;
+	if (WARN_ON_ONCE(prog > (u8 *)image_end - BPF_INSN_SAFETY)) {
+		ret = -EFAULT;
+		goto cleanup;
 	}
-	EMIT2_off32(0x0F, jmp_cond + 0x10, offset);
-	*pprog = prog;
-	return 0;
-}
+	ret = prog - (u8 *)image;
 
-static void emit_nops(u8 **pprog, unsigned int len)
-{
-	unsigned int i, noplen;
-	u8 *prog = *pprog;
-	int cnt = 0;
-
-	while (len > 0) {
-		noplen = len;
-
-		if (noplen > ASM_NOP_MAX)
-			noplen = ASM_NOP_MAX;
-
-		for (i = 0; i < noplen; i++)
-			EMIT1(ideal_nops[noplen][i]);
-		len -= noplen;
-	}
-
-	*pprog = prog;
+cleanup:
+	kfree(branches);
+	return ret;
 }
 
 static int emit_fallback_jump(u8 **pprog)
@@ -1588,7 +1721,7 @@ static int emit_fallback_jump(u8 **pprog)
 
 static int emit_bpf_dispatcher(u8 **pprog, int a, int b, s64 *progs)
 {
-	u8 *jg_reloc, *jg_target, *prog = *pprog;
+	u8 *jg_reloc, *prog = *pprog;
 	int pivot, err, jg_bytes = 1, cnt = 0;
 	s64 jg_offset;
 
@@ -1643,9 +1776,7 @@ static int emit_bpf_dispatcher(u8 **pprog, int a, int b, s64 *progs)
 	 * Coding Rule 11: All branch targets should be 16-byte
 	 * aligned.
 	 */
-	jg_target = PTR_ALIGN(prog, 16);
-	if (jg_target != prog)
-		emit_nops(&prog, jg_target - prog);
+	emit_align(&prog, 16);
 	jg_offset = prog - jg_reloc;
 	emit_code(jg_reloc - jg_bytes, jg_offset, jg_bytes);
 

include/linux/bpf.h

@@ -433,6 +433,16 @@ struct btf_func_model {
  */
 #define BPF_TRAMP_F_SKIP_FRAME		BIT(2)
 
+/* Each call __bpf_prog_enter + call bpf_func + call __bpf_prog_exit is ~50
+ * bytes on x86.  Pick a number to fit into BPF_IMAGE_SIZE / 2
+ */
+#define BPF_MAX_TRAMP_PROGS 40
+
+struct bpf_tramp_progs {
+	struct bpf_prog *progs[BPF_MAX_TRAMP_PROGS];
+	int nr_progs;
+};
+
 /* Different use cases for BPF trampoline:
  * 1. replace nop at the function entry (kprobe equivalent)
  *    flags = BPF_TRAMP_F_RESTORE_REGS
@@ -455,8 +465,7 @@ struct btf_func_model {
  */
 int arch_prepare_bpf_trampoline(void *image, void *image_end,
 				const struct btf_func_model *m, u32 flags,
-				struct bpf_prog **fentry_progs, int fentry_cnt,
-				struct bpf_prog **fexit_progs, int fexit_cnt,
+				struct bpf_tramp_progs *tprogs,
 				void *orig_call);
 /* these two functions are called from generated trampoline */
 u64 notrace __bpf_prog_enter(void);
@@ -465,6 +474,7 @@ void notrace __bpf_prog_exit(struct bpf_prog *prog, u64 start);
 enum bpf_tramp_prog_type {
 	BPF_TRAMP_FENTRY,
 	BPF_TRAMP_FEXIT,
+	BPF_TRAMP_MODIFY_RETURN,
 	BPF_TRAMP_MAX,
 	BPF_TRAMP_REPLACE, /* more than MAX */
 };
@@ -1146,6 +1156,9 @@ int bpf_prog_test_run_xdp(struct bpf_prog *prog, const union bpf_attr *kattr,
 			  union bpf_attr __user *uattr);
 int bpf_prog_test_run_skb(struct bpf_prog *prog, const union bpf_attr *kattr,
 			  union bpf_attr __user *uattr);
+int bpf_prog_test_run_tracing(struct bpf_prog *prog,
+			      const union bpf_attr *kattr,
+			      union bpf_attr __user *uattr);
 int bpf_prog_test_run_flow_dissector(struct bpf_prog *prog,
 				     const union bpf_attr *kattr,
 				     union bpf_attr __user *uattr);
@@ -1303,6 +1316,13 @@ static inline int bpf_prog_test_run_skb(struct bpf_prog *prog,
 	return -ENOTSUPP;
 }
 
+static inline int bpf_prog_test_run_tracing(struct bpf_prog *prog,
+					    const union bpf_attr *kattr,
+					    union bpf_attr __user *uattr)
+{
+	return -ENOTSUPP;
+}
+
 static inline int bpf_prog_test_run_flow_dissector(struct bpf_prog *prog,
 						   const union bpf_attr *kattr,
 						   union bpf_attr __user *uattr)

include/uapi/linux/bpf.h

@@ -210,6 +210,7 @@ enum bpf_attach_type {
 	BPF_TRACE_RAW_TP,
 	BPF_TRACE_FENTRY,
 	BPF_TRACE_FEXIT,
+	BPF_MODIFY_RETURN,
 	__MAX_BPF_ATTACH_TYPE
 };
 

kernel/bpf/bpf_struct_ops.c

@@ -320,6 +320,7 @@ static int bpf_struct_ops_map_update_elem(struct bpf_map *map, void *key,
 	struct bpf_struct_ops_value *uvalue, *kvalue;
 	const struct btf_member *member;
 	const struct btf_type *t = st_ops->type;
+	struct bpf_tramp_progs *tprogs = NULL;
 	void *udata, *kdata;
 	int prog_fd, err = 0;
 	void *image;
@@ -343,6 +344,10 @@ static int bpf_struct_ops_map_update_elem(struct bpf_map *map, void *key,
 	if (uvalue->state || refcount_read(&uvalue->refcnt))
 		return -EINVAL;
 
+	tprogs = kcalloc(BPF_TRAMP_MAX, sizeof(*tprogs), GFP_KERNEL);
+	if (!tprogs)
+		return -ENOMEM;
+
 	uvalue = (struct bpf_struct_ops_value *)st_map->uvalue;
 	kvalue = (struct bpf_struct_ops_value *)&st_map->kvalue;
 
@@ -425,10 +430,12 @@ static int bpf_struct_ops_map_update_elem(struct bpf_map *map, void *key,
 			goto reset_unlock;
 		}
 
+		tprogs[BPF_TRAMP_FENTRY].progs[0] = prog;
+		tprogs[BPF_TRAMP_FENTRY].nr_progs = 1;
 		err = arch_prepare_bpf_trampoline(image,
 						  st_map->image + PAGE_SIZE,
 						  &st_ops->func_models[i], 0,
-						  &prog, 1, NULL, 0, NULL);
+						  tprogs, NULL);
 		if (err < 0)
 			goto reset_unlock;
 
@@ -469,6 +476,7 @@ static int bpf_struct_ops_map_update_elem(struct bpf_map *map, void *key,
 	memset(uvalue, 0, map->value_size);
 	memset(kvalue, 0, map->value_size);
 unlock:
+	kfree(tprogs);
 	mutex_unlock(&st_map->lock);
 	return err;
 }

kernel/bpf/btf.c

@@ -3710,13 +3710,26 @@ bool btf_ctx_access(int off, int size, enum bpf_access_type type,
 		nr_args--;
 	}
 
-	if (prog->expected_attach_type == BPF_TRACE_FEXIT &&
-	    arg == nr_args) {
+	if (arg == nr_args) {
+		if (prog->expected_attach_type == BPF_TRACE_FEXIT) {
 			if (!t)
-			/* Default prog with 5 args. 6th arg is retval. */
 				return true;
-		/* function return type */
 			t = btf_type_by_id(btf, t->type);
+		} else if (prog->expected_attach_type == BPF_MODIFY_RETURN) {
+			/* For now the BPF_MODIFY_RETURN can only be attached to
+			 * functions that return an int.
+			 */
+			if (!t)
+				return false;
+
+			t = btf_type_skip_modifiers(btf, t->type, NULL);
+			if (!btf_type_is_int(t)) {
+				bpf_log(log,
+					"ret type %s not allowed for fmod_ret\n",
+					btf_kind_str[BTF_INFO_KIND(t->info)]);
+				return false;
+			}
+		}
 	} else if (arg >= nr_args) {
 		bpf_log(log, "func '%s' doesn't have %d-th argument\n",
 			tname, arg + 1);

kernel/bpf/syscall.c

@@ -2324,6 +2324,7 @@ static int bpf_tracing_prog_attach(struct bpf_prog *prog)
 
 	if (prog->expected_attach_type != BPF_TRACE_FENTRY &&
 	    prog->expected_attach_type != BPF_TRACE_FEXIT &&
+	    prog->expected_attach_type != BPF_MODIFY_RETURN &&
 	    prog->type != BPF_PROG_TYPE_EXT) {
 		err = -EINVAL;
 		goto out_put_prog;

kernel/bpf/trampoline.c

@@ -190,40 +190,50 @@ static int register_fentry(struct bpf_trampoline *tr, void *new_addr)
 	return ret;
 }
 
-/* Each call __bpf_prog_enter + call bpf_func + call __bpf_prog_exit is ~50
- * bytes on x86.  Pick a number to fit into BPF_IMAGE_SIZE / 2
- */
-#define BPF_MAX_TRAMP_PROGS 40
+static struct bpf_tramp_progs *
+bpf_trampoline_get_progs(const struct bpf_trampoline *tr, int *total)
+{
+	const struct bpf_prog_aux *aux;
+	struct bpf_tramp_progs *tprogs;
+	struct bpf_prog **progs;
+	int kind;
+
+	*total = 0;
+	tprogs = kcalloc(BPF_TRAMP_MAX, sizeof(*tprogs), GFP_KERNEL);
+	if (!tprogs)
+		return ERR_PTR(-ENOMEM);
+
+	for (kind = 0; kind < BPF_TRAMP_MAX; kind++) {
+		tprogs[kind].nr_progs = tr->progs_cnt[kind];
+		*total += tr->progs_cnt[kind];
+		progs = tprogs[kind].progs;
+
+		hlist_for_each_entry(aux, &tr->progs_hlist[kind], tramp_hlist)
+			*progs++ = aux->prog;
+	}
+	return tprogs;
+}
 
 static int bpf_trampoline_update(struct bpf_trampoline *tr)
 {
 	void *old_image = tr->image + ((tr->selector + 1) & 1) * BPF_IMAGE_SIZE/2;
 	void *new_image = tr->image + (tr->selector & 1) * BPF_IMAGE_SIZE/2;
-	struct bpf_prog *progs_to_run[BPF_MAX_TRAMP_PROGS];
-	int fentry_cnt = tr->progs_cnt[BPF_TRAMP_FENTRY];
-	int fexit_cnt = tr->progs_cnt[BPF_TRAMP_FEXIT];
-	struct bpf_prog **progs, **fentry, **fexit;
+	struct bpf_tramp_progs *tprogs;
 	u32 flags = BPF_TRAMP_F_RESTORE_REGS;
-	struct bpf_prog_aux *aux;
-	int err;
+	int err, total;
 
-	if (fentry_cnt + fexit_cnt == 0) {
+	tprogs = bpf_trampoline_get_progs(tr, &total);
+	if (IS_ERR(tprogs))
+		return PTR_ERR(tprogs);
+
+	if (total == 0) {
 		err = unregister_fentry(tr, old_image);
 		tr->selector = 0;
 		goto out;
 	}
 
-	/* populate fentry progs */
-	fentry = progs = progs_to_run;
-	hlist_for_each_entry(aux, &tr->progs_hlist[BPF_TRAMP_FENTRY], tramp_hlist)
-		*progs++ = aux->prog;
-
-	/* populate fexit progs */
-	fexit = progs;
-	hlist_for_each_entry(aux, &tr->progs_hlist[BPF_TRAMP_FEXIT], tramp_hlist)
-		*progs++ = aux->prog;
-
-	if (fexit_cnt)
+	if (tprogs[BPF_TRAMP_FEXIT].nr_progs ||
+	    tprogs[BPF_TRAMP_MODIFY_RETURN].nr_progs)
 		flags = BPF_TRAMP_F_CALL_ORIG | BPF_TRAMP_F_SKIP_FRAME;
 
 	/* Though the second half of trampoline page is unused a task could be
@@ -232,12 +242,11 @@ static int bpf_trampoline_update(struct bpf_trampoline *tr)
 	 * preempted task. Hence wait for tasks to voluntarily schedule or go
 	 * to userspace.
 	 */
+
 	synchronize_rcu_tasks();
 
 	err = arch_prepare_bpf_trampoline(new_image, new_image + BPF_IMAGE_SIZE / 2,
-					  &tr->func.model, flags,
-					  fentry, fentry_cnt,
-					  fexit, fexit_cnt,
+					  &tr->func.model, flags, tprogs,
 					  tr->func.addr);
 	if (err < 0)
 		goto out;
@@ -252,6 +261,7 @@ static int bpf_trampoline_update(struct bpf_trampoline *tr)
 		goto out;
 	tr->selector++;
 out:
+	kfree(tprogs);
 	return err;
 }
 
@@ -260,6 +270,8 @@ static enum bpf_tramp_prog_type bpf_attach_type_to_tramp(enum bpf_attach_type t)
 	switch (t) {
 	case BPF_TRACE_FENTRY:
 		return BPF_TRAMP_FENTRY;
+	case BPF_MODIFY_RETURN:
+		return BPF_TRAMP_MODIFY_RETURN;
 	case BPF_TRACE_FEXIT:
 		return BPF_TRAMP_FEXIT;
 	default:
@@ -409,8 +421,7 @@ void notrace __bpf_prog_exit(struct bpf_prog *prog, u64 start)
 int __weak
 arch_prepare_bpf_trampoline(void *image, void *image_end,
 			    const struct btf_func_model *m, u32 flags,
-			    struct bpf_prog **fentry_progs, int fentry_cnt,
-			    struct bpf_prog **fexit_progs, int fexit_cnt,
+			    struct bpf_tramp_progs *tprogs,
 			    void *orig_call)
 {
 	return -ENOTSUPP;

kernel/bpf/verifier.c

@@ -19,6 +19,7 @@
 #include <linux/sort.h>
 #include <linux/perf_event.h>
 #include <linux/ctype.h>
+#include <linux/error-injection.h>
 
 #include "disasm.h"
 
@@ -9800,6 +9801,33 @@ static int check_struct_ops_btf_id(struct bpf_verifier_env *env)
 
 	return 0;
 }
+#define SECURITY_PREFIX "security_"
+
+static int check_attach_modify_return(struct bpf_verifier_env *env)
+{
+	struct bpf_prog *prog = env->prog;
+	unsigned long addr = (unsigned long) prog->aux->trampoline->func.addr;
+
+	if (within_error_injection_list(addr))
+		return 0;
+
+	/* This is expected to be cleaned up in the future with the KRSI effort
+	 * introducing the LSM_HOOK macro for cleaning up lsm_hooks.h.
+	 */
+	if (!strncmp(SECURITY_PREFIX, prog->aux->attach_func_name,
+		     sizeof(SECURITY_PREFIX) - 1)) {
+
+		if (!capable(CAP_MAC_ADMIN))
+			return -EPERM;
+
+		return 0;
+	}
+
+	verbose(env, "fmod_ret attach_btf_id %u (%s) is not modifiable\n",
+		prog->aux->attach_btf_id, prog->aux->attach_func_name);
+
+	return -EINVAL;
+}
 
 static int check_attach_btf_id(struct bpf_verifier_env *env)
 {
@@ -9950,6 +9978,7 @@ static int check_attach_btf_id(struct bpf_verifier_env *env)
 		if (!prog_extension)
 			return -EINVAL;
 		/* fallthrough */
+	case BPF_MODIFY_RETURN:
 	case BPF_TRACE_FENTRY:
 	case BPF_TRACE_FEXIT:
 		if (!btf_type_is_func(t)) {
@@ -9999,6 +10028,9 @@ static int check_attach_btf_id(struct bpf_verifier_env *env)
 		}
 		tr->func.addr = (void *)addr;
 		prog->aux->trampoline = tr;
+
+		if (prog->expected_attach_type == BPF_MODIFY_RETURN)
+			ret = check_attach_modify_return(env);
 out:
 		mutex_unlock(&tr->mutex);
 		if (ret)

kernel/trace/bpf_trace.c

@@ -1266,6 +1266,7 @@ const struct bpf_verifier_ops tracing_verifier_ops = {
 };
 
 const struct bpf_prog_ops tracing_prog_ops = {
+	.test_run = bpf_prog_test_run_tracing,
 };
 
 static bool raw_tp_writable_prog_is_valid_access(int off, int size,

net/bpf/test_run.c

@@ -10,6 +10,7 @@
 #include <net/bpf_sk_storage.h>
 #include <net/sock.h>
 #include <net/tcp.h>
+#include <linux/error-injection.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/bpf_test_run.h>
@@ -143,6 +144,14 @@ int noinline bpf_fentry_test6(u64 a, void *b, short c, int d, void *e, u64 f)
 	return a + (long)b + c + d + (long)e + f;
 }
 
+int noinline bpf_modify_return_test(int a, int *b)
+{
+	*b += 1;
+	return a + *b;
+}
+
+ALLOW_ERROR_INJECTION(bpf_modify_return_test, ERRNO);
+
 static void *bpf_test_init(const union bpf_attr *kattr, u32 size,
 			   u32 headroom, u32 tailroom)
 {
@@ -160,16 +169,46 @@ static void *bpf_test_init(const union bpf_attr *kattr, u32 size,
 		kfree(data);
 		return ERR_PTR(-EFAULT);
 	}
+
+	return data;
+}
+
+int bpf_prog_test_run_tracing(struct bpf_prog *prog,
+			      const union bpf_attr *kattr,
+			      union bpf_attr __user *uattr)
+{
+	u16 side_effect = 0, ret = 0;
+	int b = 2, err = -EFAULT;
+	u32 retval = 0;
+
+	switch (prog->expected_attach_type) {
+	case BPF_TRACE_FENTRY:
+	case BPF_TRACE_FEXIT:
 		if (bpf_fentry_test1(1) != 2 ||
 		    bpf_fentry_test2(2, 3) != 5 ||
 		    bpf_fentry_test3(4, 5, 6) != 15 ||
 		    bpf_fentry_test4((void *)7, 8, 9, 10) != 34 ||
 		    bpf_fentry_test5(11, (void *)12, 13, 14, 15) != 65 ||
-	    bpf_fentry_test6(16, (void *)17, 18, 19, (void *)20, 21) != 111) {
-		kfree(data);
-		return ERR_PTR(-EFAULT);
+		    bpf_fentry_test6(16, (void *)17, 18, 19, (void *)20, 21) != 111)
+			goto out;
+		break;
+	case BPF_MODIFY_RETURN:
+		ret = bpf_modify_return_test(1, &b);
+		if (b != 2)
+			side_effect = 1;
+		break;
+	default:
+		goto out;
 	}
-	return data;
+
+	retval = ((u32)side_effect << 16) | ret;
+	if (copy_to_user(&uattr->test.retval, &retval, sizeof(retval)))
+		goto out;
+
+	err = 0;
+out:
+	trace_bpf_test_finish(&err);
+	return err;
 }
 
 static void *bpf_ctx_init(const union bpf_attr *kattr, u32 max_size)

tools/include/uapi/linux/bpf.h

@@ -210,6 +210,7 @@ enum bpf_attach_type {
 	BPF_TRACE_RAW_TP,
 	BPF_TRACE_FENTRY,
 	BPF_TRACE_FEXIT,
+	BPF_MODIFY_RETURN,
 	__MAX_BPF_ATTACH_TYPE
 };
 

tools/lib/bpf/libbpf.c

@@ -6288,6 +6288,10 @@ static const struct bpf_sec_def section_defs[] = {
 		.expected_attach_type = BPF_TRACE_FENTRY,
 		.is_attach_btf = true,
 		.attach_fn = attach_trace),
+	SEC_DEF("fmod_ret/", TRACING,
+		.expected_attach_type = BPF_MODIFY_RETURN,
+		.is_attach_btf = true,
+		.attach_fn = attach_trace),
 	SEC_DEF("fexit/", TRACING,
 		.expected_attach_type = BPF_TRACE_FEXIT,
 		.is_attach_btf = true,

tools/testing/selftests/bpf/prog_tests/fentry_fexit.c

 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2019 Facebook */
 #include <test_progs.h>
-#include "test_pkt_access.skel.h"
 #include "fentry_test.skel.h"
 #include "fexit_test.skel.h"
 
 void test_fentry_fexit(void)
 {
-	struct test_pkt_access *pkt_skel = NULL;
 	struct fentry_test *fentry_skel = NULL;
 	struct fexit_test *fexit_skel = NULL;
 	__u64 *fentry_res, *fexit_res;
 	__u32 duration = 0, retval;
-	int err, pkt_fd, i;
+	int err, prog_fd, i;
 
-	pkt_skel = test_pkt_access__open_and_load();
-	if (CHECK(!pkt_skel, "pkt_skel_load", "pkt_access skeleton failed\n"))
-		return;
 	fentry_skel = fentry_test__open_and_load();
 	if (CHECK(!fentry_skel, "fentry_skel_load", "fentry skeleton failed\n"))
 		goto close_prog;
@@ -31,8 +26,8 @@ void test_fentry_fexit(void)
 	if (CHECK(err, "fexit_attach", "fexit attach failed: %d\n", err))
 		goto close_prog;
 
-	pkt_fd = bpf_program__fd(pkt_skel->progs.test_pkt_access);
-	err = bpf_prog_test_run(pkt_fd, 1, &pkt_v6, sizeof(pkt_v6),
+	prog_fd = bpf_program__fd(fexit_skel->progs.test1);
+	err = bpf_prog_test_run(prog_fd, 1, NULL, 0,
 				NULL, NULL, &retval, &duration);
 	CHECK(err || retval, "ipv6",
 	      "err %d errno %d retval %d duration %d\n",
@@ -49,7 +44,6 @@ void test_fentry_fexit(void)
 	}
 
 close_prog:
-	test_pkt_access__destroy(pkt_skel);
 	fentry_test__destroy(fentry_skel);
 	fexit_test__destroy(fexit_skel);
 }

tools/testing/selftests/bpf/prog_tests/fentry_test.c

 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2019 Facebook */
 #include <test_progs.h>
-#include "test_pkt_access.skel.h"
 #include "fentry_test.skel.h"
 
 void test_fentry_test(void)
 {
-	struct test_pkt_access *pkt_skel = NULL;
 	struct fentry_test *fentry_skel = NULL;
-	int err, pkt_fd, i;
+	int err, prog_fd, i;
 	__u32 duration = 0, retval;
 	__u64 *result;
 
-	pkt_skel = test_pkt_access__open_and_load();
-	if (CHECK(!pkt_skel, "pkt_skel_load", "pkt_access skeleton failed\n"))
-		return;
 	fentry_skel = fentry_test__open_and_load();
 	if (CHECK(!fentry_skel, "fentry_skel_load", "fentry skeleton failed\n"))
 		goto cleanup;
@@ -23,10 +18,10 @@ void test_fentry_test(void)
 	if (CHECK(err, "fentry_attach", "fentry attach failed: %d\n", err))
 		goto cleanup;
 
-	pkt_fd = bpf_program__fd(pkt_skel->progs.test_pkt_access);
-	err = bpf_prog_test_run(pkt_fd, 1, &pkt_v6, sizeof(pkt_v6),
+	prog_fd = bpf_program__fd(fentry_skel->progs.test1);
+	err = bpf_prog_test_run(prog_fd, 1, NULL, 0,
 				NULL, NULL, &retval, &duration);
-	CHECK(err || retval, "ipv6",
+	CHECK(err || retval, "test_run",
 	      "err %d errno %d retval %d duration %d\n",
 	      err, errno, retval, duration);
 
@@ -39,5 +34,4 @@ void test_fentry_test(void)
 
 cleanup:
 	fentry_test__destroy(fentry_skel);
-	test_pkt_access__destroy(pkt_skel);
 }

tools/testing/selftests/bpf/prog_tests/fexit_test.c

 // SPDX-License-Identifier: GPL-2.0
 /* Copyright (c) 2019 Facebook */
 #include <test_progs.h>
+#include "fexit_test.skel.h"
 
 void test_fexit_test(void)
 {
-	struct bpf_prog_load_attr attr = {
-		.file = "./fexit_test.o",
-	};
-
-	char prog_name[] = "fexit/bpf_fentry_testX";
-	struct bpf_object *obj = NULL, *pkt_obj;
-	int err, pkt_fd, kfree_skb_fd, i;
-	struct bpf_link *link[6] = {};
-	struct bpf_program *prog[6];
+	struct fexit_test *fexit_skel = NULL;
+	int err, prog_fd, i;
 	__u32 duration = 0, retval;
-	struct bpf_map *data_map;
-	const int zero = 0;
-	u64 result[6];
+	__u64 *result;
 
-	err = bpf_prog_load("./test_pkt_access.o", BPF_PROG_TYPE_SCHED_CLS,
-			    &pkt_obj, &pkt_fd);
-	if (CHECK(err, "prog_load sched cls", "err %d errno %d\n", err, errno))
-		return;
-	err = bpf_prog_load_xattr(&attr, &obj, &kfree_skb_fd);
-	if (CHECK(err, "prog_load fail", "err %d errno %d\n", err, errno))
-		goto close_prog;
+	fexit_skel = fexit_test__open_and_load();
+	if (CHECK(!fexit_skel, "fexit_skel_load", "fexit skeleton failed\n"))
+		goto cleanup;
 
-	for (i = 0; i < 6; i++) {
-		prog_name[sizeof(prog_name) - 2] = '1' + i;
-		prog[i] = bpf_object__find_program_by_title(obj, prog_name);
-		if (CHECK(!prog[i], "find_prog", "prog %s not found\n", prog_name))
-			goto close_prog;
-		link[i] = bpf_program__attach_trace(prog[i]);
-		if (CHECK(IS_ERR(link[i]), "attach_trace", "failed to link\n"))
-			goto close_prog;
-	}
-	data_map = bpf_object__find_map_by_name(obj, "fexit_te.bss");
-	if (CHECK(!data_map, "find_data_map", "data map not found\n"))
-		goto close_prog;
+	err = fexit_test__attach(fexit_skel);
+	if (CHECK(err, "fexit_attach", "fexit attach failed: %d\n", err))
+		goto cleanup;
 
-	err = bpf_prog_test_run(pkt_fd, 1, &pkt_v6, sizeof(pkt_v6),
+	prog_fd = bpf_program__fd(fexit_skel->progs.test1);
+	err = bpf_prog_test_run(prog_fd, 1, NULL, 0,
 				NULL, NULL, &retval, &duration);
-	CHECK(err || retval, "ipv6",
+	CHECK(err || retval, "test_run",
 	      "err %d errno %d retval %d duration %d\n",
 	      err, errno, retval, duration);
 
-	err = bpf_map_lookup_elem(bpf_map__fd(data_map), &zero, &result);
-	if (CHECK(err, "get_result",
-		  "failed to get output data: %d\n", err))
-		goto close_prog;
-
-	for (i = 0; i < 6; i++)
-		if (CHECK(result[i] != 1, "result", "bpf_fentry_test%d failed err %ld\n",
-			  i + 1, result[i]))
-			goto close_prog;
+	result = (__u64 *)fexit_skel->bss;
+	for (i = 0; i < 6; i++) {
+		if (CHECK(result[i] != 1, "result",
+			  "fexit_test%d failed err %lld\n", i + 1, result[i]))
+			goto cleanup;
+	}
 
-close_prog:
-	for (i = 0; i < 6; i++)
-		if (!IS_ERR_OR_NULL(link[i]))
-			bpf_link__destroy(link[i]);
-	bpf_object__close(obj);
-	bpf_object__close(pkt_obj);
+cleanup:
+	fexit_test__destroy(fexit_skel);
 }

tools/testing/selftests/bpf/prog_tests/modify_return.c

+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Copyright 2020 Google LLC.
+ */
+
+#include <test_progs.h>
+#include "modify_return.skel.h"
+
+#define LOWER(x) ((x) & 0xffff)
+#define UPPER(x) ((x) >> 16)
+
+
+static void run_test(__u32 input_retval, __u16 want_side_effect, __s16 want_ret)
+{
+	struct modify_return *skel = NULL;
+	int err, prog_fd;
+	__u32 duration = 0, retval;
+	__u16 side_effect;
+	__s16 ret;
+
+	skel = modify_return__open_and_load();
+	if (CHECK(!skel, "skel_load", "modify_return skeleton failed\n"))
+		goto cleanup;
+
+	err = modify_return__attach(skel);
+	if (CHECK(err, "modify_return", "attach failed: %d\n", err))
+		goto cleanup;
+
+	skel->bss->input_retval = input_retval;
+	prog_fd = bpf_program__fd(skel->progs.fmod_ret_test);
+	err = bpf_prog_test_run(prog_fd, 1, NULL, 0, NULL, 0,
+				&retval, &duration);
+
+	CHECK(err, "test_run", "err %d errno %d\n", err, errno);
+
+	side_effect = UPPER(retval);
+	ret  = LOWER(retval);
+
+	CHECK(ret != want_ret, "test_run",
+	      "unexpected ret: %d, expected: %d\n", ret, want_ret);
+	CHECK(side_effect != want_side_effect, "modify_return",
+	      "unexpected side_effect: %d\n", side_effect);
+
+	CHECK(skel->bss->fentry_result != 1, "modify_return",
+	      "fentry failed\n");
+	CHECK(skel->bss->fexit_result != 1, "modify_return",
+	      "fexit failed\n");
+	CHECK(skel->bss->fmod_ret_result != 1, "modify_return",
+	      "fmod_ret failed\n");
+
+cleanup:
+	modify_return__destroy(skel);
+}
+
+void test_modify_return(void)
+{
+	run_test(0 /* input_retval */,
+		 1 /* want_side_effect */,
+		 4 /* want_ret */);
+	run_test(-EINVAL /* input_retval */,
+		 0 /* want_side_effect */,
+		 -EINVAL /* want_ret */);
+}
+

tools/testing/selftests/bpf/progs/modify_return.c

+// SPDX-License-Identifier: GPL-2.0
+
+/*
+ * Copyright 2020 Google LLC.
+ */
+
+#include <linux/bpf.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_tracing.h>
+
+char _license[] SEC("license") = "GPL";
+
+static int sequence = 0;
+__s32 input_retval = 0;
+
+__u64 fentry_result = 0;
+SEC("fentry/bpf_modify_return_test")
+int BPF_PROG(fentry_test, int a, __u64 b)
+{
+	sequence++;
+	fentry_result = (sequence == 1);
+	return 0;
+}
+
+__u64 fmod_ret_result = 0;
+SEC("fmod_ret/bpf_modify_return_test")
+int BPF_PROG(fmod_ret_test, int a, int *b, int ret)
+{
+	sequence++;
+	/* This is the first fmod_ret program, the ret passed should be 0 */
+	fmod_ret_result = (sequence == 2 && ret == 0);
+	return input_retval;
+}
+
+__u64 fexit_result = 0;
+SEC("fexit/bpf_modify_return_test")
+int BPF_PROG(fexit_test, int a, __u64 b, int ret)
+{
+	sequence++;
+	/* If the input_reval is non-zero a successful modification should have
+	 * occurred.
+	 */
+	if (input_retval)
+		fexit_result = (sequence == 3 && ret == input_retval);
+	else
+		fexit_result = (sequence == 3 && ret == 4);
+
+	return 0;
+}