[PATCH bpf-next v2 0/2] Introduce bpf_preempt_{disable,enable}

[PATCH bpf-next v2 0/2] Introduce bpf_preempt_{disable,enable}

[Kumar Kartikeya Dwivedi]

This set introduces two kfuncs, bpf_preempt_disable and
bpf_preempt_enable, which are wrappers around preempt_disable and
preempt_enable in the kernel. These functions allow a BPF program to
have code sections where preemption is disabled. There are multiple use
cases that are served by such a feature, a few are listed below:

1. Writing safe per-CPU alogrithms/data structures that work correctly
   across different contexts.
2. Writing safe per-CPU allocators similar to bpf_memalloc on top of
   array/arena memory blobs.
3. Writing locking algorithms in BPF programs natively.

Note that local_irq_disable/enable equivalent is also needed for proper
IRQ context protection, but that is a more involved change and will be
sent later.

While bpf_preempt_{disable,enable} is not sufficient for all of these
usage scenarios on its own, it is still necessary.

The same effect as these kfuncs can in some sense be already achieved
using the bpf_spin_lock or rcu_read_lock APIs, therefore from the
standpoint of kernel functionality exposure in the verifier, this is
well understood territory.

Note that these helpers do allow calling kernel helpers and kfuncs from
within the non-preemptible region (unless sleepable). Otherwise, any
locks built using the preemption helpers will be as limited as
existing bpf_spin_lock.

Nesting is allowed by keeping a counter for tracking remaining enables
required to be performed. Similar approach can be applied to
rcu_read_locks in a follow up.

Changelog
=========
v1: https://lore.kernel.org/bpf/20240423061922.2295517-1-memxor@gmail.com

 * Move kfunc BTF ID declerations above css task kfunc for
   !CONFIG_CGROUPS config (Alexei)
 * Add test case for global function call in non-preemptible region
   (Jiri)

Kumar Kartikeya Dwivedi (2):
  bpf: Introduce bpf_preempt_[disable,enable] kfuncs
  selftests/bpf: Add tests for preempt kfuncs

 include/linux/bpf_verifier.h                  |   1 +
 kernel/bpf/helpers.c                          |  12 ++
 kernel/bpf/verifier.c                         |  71 ++++++++-
 .../selftests/bpf/prog_tests/preempt_lock.c   |   9 ++
 .../selftests/bpf/progs/preempt_lock.c        | 135 ++++++++++++++++++
 5 files changed, 226 insertions(+), 2 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/preempt_lock.c
 create mode 100644 tools/testing/selftests/bpf/progs/preempt_lock.c


base-commit: 6e10b6350a67d398c795ac0b93a7bb7103633fe4
-- 
2.43.0

[PATCH bpf-next v2 1/2] bpf: Introduce bpf_preempt_[disable,enable] kfuncs

[Kumar Kartikeya Dwivedi]

Introduce two new BPF kfuncs, bpf_preempt_disable and
bpf_preempt_enable. These kfuncs allow disabling preemption in BPF
programs. Nesting is allowed, since the intended use cases includes
building native BPF spin locks without kernel helper involvement. Apart
from that, this can be used to per-CPU data structures for cases where
programs (or userspace) may preempt one or the other. Currently, while
per-CPU access is stable, whether it will be consistent is not
guaranteed, as only migration is disabled for BPF programs.

Global functions are disallowed from being called, but support for them
will be added as a follow up not just preempt kfuncs, but rcu_read_lock
kfuncs as well. Static subprog calls are permitted. Sleepable helpers
and kfuncs are disallowed in non-preemptible regions.

Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 include/linux/bpf_verifier.h |  1 +
 kernel/bpf/helpers.c         | 12 ++++++
 kernel/bpf/verifier.c        | 71 +++++++++++++++++++++++++++++++++++-
 3 files changed, 82 insertions(+), 2 deletions(-)

diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h
index 9db35530c878..50aa87f8d77f 100644
--- a/include/linux/bpf_verifier.h
+++ b/include/linux/bpf_verifier.h
@@ -421,6 +421,7 @@ struct bpf_verifier_state {
 	struct bpf_active_lock active_lock;
 	bool speculative;
 	bool active_rcu_lock;
+	u32 active_preempt_lock;
 	/* If this state was ever pointed-to by other state's loop_entry field
 	 * this flag would be set to true. Used to avoid freeing such states
 	 * while they are still in use.
diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index 047a21b7e4ba..84997e95ab78 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -2742,6 +2742,16 @@ __bpf_kfunc int bpf_wq_set_callback_impl(struct bpf_wq *wq,
 	return __bpf_async_set_callback(async, callback_fn, aux, flags, BPF_ASYNC_TYPE_WQ);
 }
 
+__bpf_kfunc void bpf_preempt_disable(void)
+{
+	preempt_disable();
+}
+
+__bpf_kfunc void bpf_preempt_enable(void)
+{
+	preempt_enable();
+}
+
 __bpf_kfunc_end_defs();
 
 BTF_KFUNCS_START(generic_btf_ids)
@@ -2822,6 +2832,8 @@ BTF_ID_FLAGS(func, bpf_modify_return_test_tp)
 BTF_ID_FLAGS(func, bpf_wq_init)
 BTF_ID_FLAGS(func, bpf_wq_set_callback_impl)
 BTF_ID_FLAGS(func, bpf_wq_start)
+BTF_ID_FLAGS(func, bpf_preempt_disable)
+BTF_ID_FLAGS(func, bpf_preempt_enable)
 BTF_KFUNCS_END(common_btf_ids)
 
 static const struct btf_kfunc_id_set common_kfunc_set = {
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 9715c88cc025..8312c7a0ee19 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -1434,6 +1434,7 @@ static int copy_verifier_state(struct bpf_verifier_state *dst_state,
 	}
 	dst_state->speculative = src->speculative;
 	dst_state->active_rcu_lock = src->active_rcu_lock;
+	dst_state->active_preempt_lock = src->active_preempt_lock;
 	dst_state->in_sleepable = src->in_sleepable;
 	dst_state->curframe = src->curframe;
 	dst_state->active_lock.ptr = src->active_lock.ptr;
@@ -9599,6 +9600,13 @@ static int check_func_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
 			return -EINVAL;
 		}
 
+		/* Only global subprogs cannot be called with preemption disabled. */
+		if (env->cur_state->active_preempt_lock) {
+			verbose(env, "global function calls are not allowed with preemption disabled,\n"
+				     "use static function instead\n");
+			return -EINVAL;
+		}
+
 		if (err) {
 			verbose(env, "Caller passes invalid args into func#%d ('%s')\n",
 				subprog, sub_name);
@@ -10285,6 +10293,17 @@ static int check_helper_call(struct bpf_verifier_env *env, struct bpf_insn *insn
 			env->insn_aux_data[insn_idx].storage_get_func_atomic = true;
 	}
 
+	if (env->cur_state->active_preempt_lock) {
+		if (fn->might_sleep) {
+			verbose(env, "sleepable helper %s#%d in non-preemptible region\n",
+				func_id_name(func_id), func_id);
+			return -EINVAL;
+		}
+
+		if (in_sleepable(env) && is_storage_get_function(func_id))
+			env->insn_aux_data[insn_idx].storage_get_func_atomic = true;
+	}
+
 	meta.func_id = func_id;
 	/* check args */
 	for (i = 0; i < MAX_BPF_FUNC_REG_ARGS; i++) {
@@ -11027,6 +11046,8 @@ enum special_kfunc_type {
 	KF_bpf_percpu_obj_drop_impl,
 	KF_bpf_throw,
 	KF_bpf_wq_set_callback_impl,
+	KF_bpf_preempt_disable,
+	KF_bpf_preempt_enable,
 	KF_bpf_iter_css_task_new,
 };
 
@@ -11081,6 +11102,8 @@ BTF_ID(func, bpf_percpu_obj_new_impl)
 BTF_ID(func, bpf_percpu_obj_drop_impl)
 BTF_ID(func, bpf_throw)
 BTF_ID(func, bpf_wq_set_callback_impl)
+BTF_ID(func, bpf_preempt_disable)
+BTF_ID(func, bpf_preempt_enable)
 #ifdef CONFIG_CGROUPS
 BTF_ID(func, bpf_iter_css_task_new)
 #else
@@ -11107,6 +11130,16 @@ static bool is_kfunc_bpf_rcu_read_unlock(struct bpf_kfunc_call_arg_meta *meta)
 	return meta->func_id == special_kfunc_list[KF_bpf_rcu_read_unlock];
 }
 
+static bool is_kfunc_bpf_preempt_disable(struct bpf_kfunc_call_arg_meta *meta)
+{
+	return meta->func_id == special_kfunc_list[KF_bpf_preempt_disable];
+}
+
+static bool is_kfunc_bpf_preempt_enable(struct bpf_kfunc_call_arg_meta *meta)
+{
+	return meta->func_id == special_kfunc_list[KF_bpf_preempt_enable];
+}
+
 static enum kfunc_ptr_arg_type
 get_kfunc_ptr_arg_type(struct bpf_verifier_env *env,
 		       struct bpf_kfunc_call_arg_meta *meta,
@@ -12195,11 +12228,11 @@ static int check_return_code(struct bpf_verifier_env *env, int regno, const char
 static int check_kfunc_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
 			    int *insn_idx_p)
 {
-	const struct btf_type *t, *ptr_type;
+	bool sleepable, rcu_lock, rcu_unlock, preempt_disable, preempt_enable;
 	u32 i, nargs, ptr_type_id, release_ref_obj_id;
 	struct bpf_reg_state *regs = cur_regs(env);
 	const char *func_name, *ptr_type_name;
-	bool sleepable, rcu_lock, rcu_unlock;
+	const struct btf_type *t, *ptr_type;
 	struct bpf_kfunc_call_arg_meta meta;
 	struct bpf_insn_aux_data *insn_aux;
 	int err, insn_idx = *insn_idx_p;
@@ -12260,6 +12293,9 @@ static int check_kfunc_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
 	rcu_lock = is_kfunc_bpf_rcu_read_lock(&meta);
 	rcu_unlock = is_kfunc_bpf_rcu_read_unlock(&meta);
 
+	preempt_disable = is_kfunc_bpf_preempt_disable(&meta);
+	preempt_enable = is_kfunc_bpf_preempt_enable(&meta);
+
 	if (env->cur_state->active_rcu_lock) {
 		struct bpf_func_state *state;
 		struct bpf_reg_state *reg;
@@ -12292,6 +12328,22 @@ static int check_kfunc_call(struct bpf_verifier_env *env, struct bpf_insn *insn,
 		return -EINVAL;
 	}
 
+	if (env->cur_state->active_preempt_lock) {
+		if (preempt_disable) {
+			env->cur_state->active_preempt_lock++;
+		} else if (preempt_enable) {
+			env->cur_state->active_preempt_lock--;
+		} else if (sleepable) {
+			verbose(env, "kernel func %s is sleepable within non-preemptible region\n", func_name);
+			return -EACCES;
+		}
+	} else if (preempt_disable) {
+		env->cur_state->active_preempt_lock++;
+	} else if (preempt_enable) {
+		verbose(env, "unmatched attempt to enable preemption (kernel function %s)\n", func_name);
+		return -EINVAL;
+	}
+
 	/* In case of release function, we get register number of refcounted
 	 * PTR_TO_BTF_ID in bpf_kfunc_arg_meta, do the release now.
 	 */
@@ -15439,6 +15491,11 @@ static int check_ld_abs(struct bpf_verifier_env *env, struct bpf_insn *insn)
 		return -EINVAL;
 	}
 
+	if (env->cur_state->active_preempt_lock) {
+		verbose(env, "BPF_LD_[ABS|IND] cannot be used inside bpf_preempt_disable-ed region\n");
+		return -EINVAL;
+	}
+
 	if (regs[ctx_reg].type != PTR_TO_CTX) {
 		verbose(env,
 			"at the time of BPF_LD_ABS|IND R6 != pointer to skb\n");
@@ -17006,6 +17063,9 @@ static bool states_equal(struct bpf_verifier_env *env,
 	if (old->active_rcu_lock != cur->active_rcu_lock)
 		return false;
 
+	if (old->active_preempt_lock != cur->active_preempt_lock)
+		return false;
+
 	if (old->in_sleepable != cur->in_sleepable)
 		return false;
 
@@ -17957,6 +18017,13 @@ static int do_check(struct bpf_verifier_env *env)
 					return -EINVAL;
 				}
 
+				if (env->cur_state->active_preempt_lock && !env->cur_state->curframe) {
+					verbose(env, "%d bpf_preempt_enable%s missing\n",
+						env->cur_state->active_preempt_lock,
+						env->cur_state->active_preempt_lock == 1 ? " is" : "(s) are");
+					return -EINVAL;
+				}
+
 				/* We must do check_reference_leak here before
 				 * prepare_func_exit to handle the case when
 				 * state->curframe > 0, it may be a callback
-- 
2.43.0

[PATCH bpf-next v2 2/2] selftests/bpf: Add tests for preempt kfuncs

[Kumar Kartikeya Dwivedi]

Add tests for nested cases, nested count preservation upon different
subprog calls that disable/enable preemption, and test sleepable helper
call in non-preemptible regions.

182/1   preempt_lock/preempt_lock_missing_1:OK
182/2   preempt_lock/preempt_lock_missing_2:OK
182/3   preempt_lock/preempt_lock_missing_3:OK
182/4   preempt_lock/preempt_lock_missing_3_minus_2:OK
182/5   preempt_lock/preempt_lock_missing_1_subprog:OK
182/6   preempt_lock/preempt_lock_missing_2_subprog:OK
182/7   preempt_lock/preempt_lock_missing_2_minus_1_subprog:OK
182/8   preempt_lock/preempt_balance:OK
182/9   preempt_lock/preempt_balance_subprog_test:OK
182/10  preempt_lock/preempt_global_subprog_test:OK
182/11  preempt_lock/preempt_sleepable_helper:OK
182     preempt_lock:OK
Summary: 1/11 PASSED, 0 SKIPPED, 0 FAILED

Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 .../selftests/bpf/prog_tests/preempt_lock.c   |   9 ++
 .../selftests/bpf/progs/preempt_lock.c        | 135 ++++++++++++++++++
 2 files changed, 144 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/preempt_lock.c
 create mode 100644 tools/testing/selftests/bpf/progs/preempt_lock.c

diff --git a/tools/testing/selftests/bpf/prog_tests/preempt_lock.c b/tools/testing/selftests/bpf/prog_tests/preempt_lock.c
new file mode 100644
index 000000000000..02917c672441
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/preempt_lock.c
@@ -0,0 +1,9 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <test_progs.h>
+#include <network_helpers.h>
+#include <preempt_lock.skel.h>
+
+void test_preempt_lock(void)
+{
+	RUN_TESTS(preempt_lock);
+}
diff --git a/tools/testing/selftests/bpf/progs/preempt_lock.c b/tools/testing/selftests/bpf/progs/preempt_lock.c
new file mode 100644
index 000000000000..6c637ee01ec4
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/preempt_lock.c
@@ -0,0 +1,135 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+#include <bpf/bpf_tracing.h>
+#include "bpf_misc.h"
+
+void bpf_preempt_disable(void) __ksym;
+void bpf_preempt_enable(void) __ksym;
+
+SEC("?tc")
+__failure __msg("1 bpf_preempt_enable is missing")
+int preempt_lock_missing_1(struct __sk_buff *ctx)
+{
+	bpf_preempt_disable();
+	return 0;
+}
+
+SEC("?tc")
+__failure __msg("2 bpf_preempt_enable(s) are missing")
+int preempt_lock_missing_2(struct __sk_buff *ctx)
+{
+	bpf_preempt_disable();
+	bpf_preempt_disable();
+	return 0;
+}
+
+SEC("?tc")
+__failure __msg("3 bpf_preempt_enable(s) are missing")
+int preempt_lock_missing_3(struct __sk_buff *ctx)
+{
+	bpf_preempt_disable();
+	bpf_preempt_disable();
+	bpf_preempt_disable();
+	return 0;
+}
+
+SEC("?tc")
+__failure __msg("1 bpf_preempt_enable is missing")
+int preempt_lock_missing_3_minus_2(struct __sk_buff *ctx)
+{
+	bpf_preempt_disable();
+	bpf_preempt_disable();
+	bpf_preempt_disable();
+	bpf_preempt_enable();
+	bpf_preempt_enable();
+	return 0;
+}
+
+static __noinline void preempt_disable(void)
+{
+	bpf_preempt_disable();
+}
+
+static __noinline void preempt_enable(void)
+{
+	bpf_preempt_enable();
+}
+
+SEC("?tc")
+__failure __msg("1 bpf_preempt_enable is missing")
+int preempt_lock_missing_1_subprog(struct __sk_buff *ctx)
+{
+	preempt_disable();
+	return 0;
+}
+
+SEC("?tc")
+__failure __msg("2 bpf_preempt_enable(s) are missing")
+int preempt_lock_missing_2_subprog(struct __sk_buff *ctx)
+{
+	preempt_disable();
+	preempt_disable();
+	return 0;
+}
+
+SEC("?tc")
+__failure __msg("1 bpf_preempt_enable is missing")
+int preempt_lock_missing_2_minus_1_subprog(struct __sk_buff *ctx)
+{
+	preempt_disable();
+	preempt_disable();
+	preempt_enable();
+	return 0;
+}
+
+static __noinline void preempt_balance_subprog(void)
+{
+	preempt_disable();
+	preempt_enable();
+}
+
+SEC("?tc")
+__success int preempt_balance(struct __sk_buff *ctx)
+{
+	bpf_preempt_disable();
+	bpf_preempt_enable();
+	return 0;
+}
+
+SEC("?tc")
+__success int preempt_balance_subprog_test(struct __sk_buff *ctx)
+{
+	preempt_balance_subprog();
+	return 0;
+}
+
+SEC("?fentry.s/" SYS_PREFIX "sys_getpgid")
+__failure __msg("sleepable helper bpf_copy_from_user#")
+int preempt_sleepable_helper(void *ctx)
+{
+	u32 data;
+
+	bpf_preempt_disable();
+	bpf_copy_from_user(&data, sizeof(data), NULL);
+	bpf_preempt_enable();
+	return 0;
+}
+
+int __noinline preempt_global_subprog(void)
+{
+	preempt_balance_subprog();
+	return 0;
+}
+
+SEC("?tc")
+__failure __msg("global function calls are not allowed with preemption disabled")
+int preempt_global_subprog_test(struct __sk_buff *ctx)
+{
+	preempt_disable();
+	preempt_global_subprog();
+	preempt_enable();
+	return 0;
+}
+
+char _license[] SEC("license") = "GPL";
-- 
2.43.0

Re: [PATCH bpf-next v2 0/2] Introduce bpf_preempt_{disable,enable}

[Jiri Olsa]

On Wed, Apr 24, 2024 at 03:13:13AM +0000, Kumar Kartikeya Dwivedi wrote:
> This set introduces two kfuncs, bpf_preempt_disable and
> bpf_preempt_enable, which are wrappers around preempt_disable and
> preempt_enable in the kernel. These functions allow a BPF program to
> have code sections where preemption is disabled. There are multiple use
> cases that are served by such a feature, a few are listed below:
> 
> 1. Writing safe per-CPU alogrithms/data structures that work correctly
>    across different contexts.
> 2. Writing safe per-CPU allocators similar to bpf_memalloc on top of
>    array/arena memory blobs.
> 3. Writing locking algorithms in BPF programs natively.
> 
> Note that local_irq_disable/enable equivalent is also needed for proper
> IRQ context protection, but that is a more involved change and will be
> sent later.
> 
> While bpf_preempt_{disable,enable} is not sufficient for all of these
> usage scenarios on its own, it is still necessary.
> 
> The same effect as these kfuncs can in some sense be already achieved
> using the bpf_spin_lock or rcu_read_lock APIs, therefore from the
> standpoint of kernel functionality exposure in the verifier, this is
> well understood territory.
> 
> Note that these helpers do allow calling kernel helpers and kfuncs from
> within the non-preemptible region (unless sleepable). Otherwise, any
> locks built using the preemption helpers will be as limited as
> existing bpf_spin_lock.
> 
> Nesting is allowed by keeping a counter for tracking remaining enables
> required to be performed. Similar approach can be applied to
> rcu_read_locks in a follow up.
> 
> Changelog
> =========
> v1: https://lore.kernel.org/bpf/20240423061922.2295517-1-memxor@gmail.com
> 
>  * Move kfunc BTF ID declerations above css task kfunc for
>    !CONFIG_CGROUPS config (Alexei)
>  * Add test case for global function call in non-preemptible region
>    (Jiri)

Acked-by: Jiri Olsa <jolsa@kernel.org>

jirka

> 
> Kumar Kartikeya Dwivedi (2):
>   bpf: Introduce bpf_preempt_[disable,enable] kfuncs
>   selftests/bpf: Add tests for preempt kfuncs
> 
>  include/linux/bpf_verifier.h                  |   1 +
>  kernel/bpf/helpers.c                          |  12 ++
>  kernel/bpf/verifier.c                         |  71 ++++++++-
>  .../selftests/bpf/prog_tests/preempt_lock.c   |   9 ++
>  .../selftests/bpf/progs/preempt_lock.c        | 135 ++++++++++++++++++
>  5 files changed, 226 insertions(+), 2 deletions(-)
>  create mode 100644 tools/testing/selftests/bpf/prog_tests/preempt_lock.c
>  create mode 100644 tools/testing/selftests/bpf/progs/preempt_lock.c
> 
> 
> base-commit: 6e10b6350a67d398c795ac0b93a7bb7103633fe4
> -- 
> 2.43.0
> 
> 

Re: [PATCH bpf-next v2 0/2] Introduce bpf_preempt_{disable,enable}

[patchwork-bot+netdevbpf]

Hello:

This series was applied to bpf/bpf-next.git (master)
by Alexei Starovoitov <ast@kernel.org>:

On Wed, 24 Apr 2024 03:13:13 +0000 you wrote:
> This set introduces two kfuncs, bpf_preempt_disable and
> bpf_preempt_enable, which are wrappers around preempt_disable and
> preempt_enable in the kernel. These functions allow a BPF program to
> have code sections where preemption is disabled. There are multiple use
> cases that are served by such a feature, a few are listed below:
> 
> 1. Writing safe per-CPU alogrithms/data structures that work correctly
>    across different contexts.
> 2. Writing safe per-CPU allocators similar to bpf_memalloc on top of
>    array/arena memory blobs.
> 3. Writing locking algorithms in BPF programs natively.
> 
> [...]

Here is the summary with links:
  - [bpf-next,v2,1/2] bpf: Introduce bpf_preempt_[disable,enable] kfuncs
    https://git.kernel.org/bpf/bpf-next/c/fc7566ad0a82
  - [bpf-next,v2,2/2] selftests/bpf: Add tests for preempt kfuncs
    https://git.kernel.org/bpf/bpf-next/c/3134396f1cba

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html