[PATCH bpf-next v2 0/9] Add arena argument support to kfuncs and struct_ops

[PATCH bpf-next v2 0/9] Add arena argument support to kfuncs and struct_ops

[Kumar Kartikeya Dwivedi]

This is a continuation of patches in [0], with mostly minor changes and
reordering. The motivation is covered in that link. A major change is
moving to two tags (__arena and __arena_nullable) and moving the changes
to JIT to emit more optimized sequences.

Please see commit logs for details.

  [0]: https://lore.kernel.org/bpf/20260713024414.3759854-1-tj@kernel.org

Changelog:
----------
v1 -> v2
v1: https://lore.kernel.org/bpf/20260715220052.1590783-1-memxor@gmail.com

 * Fix documentation to only mention x86 for now. (Sashiko)
 * Move arg bitmap from insn_aux_data to kfunc descriptor. (Eduard)

Tejun Heo (9):
  bpf: Support __arena and __arena_nullable kfunc argument suffixes
  bpf: Support __arena and __arena_nullable on struct_ops arguments
  bpf, x86: JIT __arena kfunc argument rebasing
  bpf, x86: Convert struct_ops arena arguments in the trampoline
  selftests/bpf: Add kfunc __arena and __arena_nullable argument tests
  selftests/bpf: Add JIT-sequence tests for __arena kfunc arguments
  selftests/bpf: Add struct_ops __arena and __arena_nullable argument
    tests
  bpf, x86: Fix stack-passed arguments for indirect trampolines
  selftests/bpf: Test stack-passed struct_ops arena arguments

 Documentation/bpf/kfuncs.rst                  |  39 +++
 arch/x86/net/bpf_jit_comp.c                   | 129 ++++++++-
 include/linux/bpf.h                           |  25 ++
 include/linux/filter.h                        |   1 +
 kernel/bpf/bpf_struct_ops.c                   |  34 ++-
 kernel/bpf/btf.c                              |  28 +-
 kernel/bpf/core.c                             |   5 +
 kernel/bpf/trampoline.c                       |  64 +++++
 kernel/bpf/verifier.c                         |  62 ++++-
 .../selftests/bpf/prog_tests/arena_kfunc.c    |  15 +
 .../bpf/prog_tests/arena_kfunc_jit.c          |  13 +
 .../bpf/prog_tests/test_struct_ops_arena.c    |  74 +++++
 .../testing/selftests/bpf/progs/arena_kfunc.c | 260 ++++++++++++++++++
 .../selftests/bpf/progs/arena_kfunc_jit.c     |  96 +++++++
 .../selftests/bpf/progs/struct_ops_arena.c    | 115 ++++++++
 .../bpf/progs/struct_ops_arena_fail.c         |  20 ++
 .../selftests/bpf/test_kmods/bpf_testmod.c    |  95 +++++++
 .../selftests/bpf/test_kmods/bpf_testmod.h    |   6 +
 .../bpf/test_kmods/bpf_testmod_kfunc.h        |  13 +
 19 files changed, 1062 insertions(+), 32 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/arena_kfunc.c
 create mode 100644 tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c
 create mode 100644 tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c
 create mode 100644 tools/testing/selftests/bpf/progs/arena_kfunc.c
 create mode 100644 tools/testing/selftests/bpf/progs/arena_kfunc_jit.c
 create mode 100644 tools/testing/selftests/bpf/progs/struct_ops_arena.c
 create mode 100644 tools/testing/selftests/bpf/progs/struct_ops_arena_fail.c


base-commit: 4748a67f711188cbd64bc72c3df636f627858ed0
-- 
2.53.0

[PATCH bpf-next v2 1/9] bpf: Support __arena and __arena_nullable kfunc argument suffixes

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

Passing an arena pointer to a kfunc takes two steps today. There is no
arena pointer argument type, so the pointer crosses the boundary as a
bare scalar, and the kfunc then offsets it by the arena base and casts
it before it can touch the memory. Every such kfunc open-codes the same
translation.

Add the __arena and __arena_nullable argument suffixes to make this more
convenient. The kfunc declares the parameter by its real pointer type
and dereferences it directly, with the JIT rebasing the value at the
call site, rN = kern_vm_start + (u32)rN. No bounds check is needed: the
u32 offset stays within the guard-padded arena kernel mapping, and a
fault on an unpopulated page recovers through the per-arena scratch
page. A suffixed argument accepts a PTR_TO_ARENA or scalar register,
matching global subprog arena arguments.

__arena rebases unconditionally, so the kfunc never sees NULL and a
value with zero in the low 32 bits arrives as the arena base.
__arena_nullable preserves NULL for optional arguments by skipping the
rebase when the truncated value, arena offset 0, is zero. Keeping the
plain form NULL-free saves the NULL test on every call.

This patch adds the verifier side: the suffixes are recognized in
check_kfunc_args() and distilled into argument flags in the function
model stored in the kfunc descriptor. JITs retrieve the model while
emitting the call, avoiding per-call state in insn_aux_data.

JITs declare support with bpf_jit_supports_arena_args() and verification
fails with -ENOTSUPP elsewhere.

Signed-off-by: Tejun Heo <tj@kernel.org>
Co-developed-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 Documentation/bpf/kfuncs.rst | 29 +++++++++++++++++++++++++++
 include/linux/bpf.h          |  6 ++++++
 include/linux/filter.h       |  1 +
 kernel/bpf/btf.c             | 18 ++++++++++++++++-
 kernel/bpf/core.c            |  5 +++++
 kernel/bpf/verifier.c        | 39 ++++++++++++++++++++++++++++++++++++
 6 files changed, 97 insertions(+), 1 deletion(-)

diff --git a/Documentation/bpf/kfuncs.rst b/Documentation/bpf/kfuncs.rst
index c801a330aece..9809680bee43 100644
--- a/Documentation/bpf/kfuncs.rst
+++ b/Documentation/bpf/kfuncs.rst
@@ -250,6 +250,33 @@ Or::
                 ...
         }
 
+2.3.7 __arena and __arena_nullable Annotations
+----------------------------------------------
+
+Both annotations indicate that the pointer argument points into the
+calling program's arena. The JIT rebases the value at the call site so
+the kfunc receives a directly dereferenceable kernel address, subject to
+the access rules described in :ref:`BPF_kfunc_arena_access` (at most
+``GUARD_SZ / 2``, 32 KiB, past the pointer in a single unchecked access).
+
+With ``__arena`` the rebase is unconditional and the argument is never
+NULL: a value whose lower 32 bits are zero arrives as the arena base
+address (arena offset 0). The kfunc must not check the argument for NULL.
+With ``__arena_nullable`` such a value arrives as NULL instead and the
+kfunc must check before dereferencing.
+
+An example is given below::
+
+        __bpf_kfunc int bpf_process_item(struct item *item__arena)
+        {
+        ...
+        }
+
+Calling such a kfunc requires the program to use an arena map and a JIT with
+arena argument support (currently x86-64); verification fails otherwise. The
+program can pass any value without compromising the kernel. A value that does
+not point into the arena is a program bug.
+
 .. _BPF_kfunc_nodef:
 
 2.4 Using an existing kernel function
@@ -487,6 +514,8 @@ In order to accommodate such requirements, the verifier will enforce strict
 PTR_TO_BTF_ID type matching if two types have the exact same name, with one
 being suffixed with ``___init``.
 
+.. _BPF_kfunc_arena_access:
+
 2.8 Accessing arena memory through kfunc arguments
 --------------------------------------------------
 
diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 7bfc28673124..1ca4f2115fde 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -1213,6 +1213,12 @@ struct bpf_prog_offload {
 /* The argument is signed. */
 #define BTF_FMODEL_SIGNED_ARG		BIT(1)
 
+/* The argument is an arena pointer. */
+#define BTF_FMODEL_ARENA_ARG		BIT(2)
+
+/* The argument is nullable. */
+#define BTF_FMODEL_NULLABLE_ARG		BIT(3)
+
 struct btf_func_model {
 	u8 ret_size;
 	u8 ret_flags;
diff --git a/include/linux/filter.h b/include/linux/filter.h
index 32d5297c557e..36ce3403fe59 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -1183,6 +1183,7 @@ bool bpf_jit_supports_subprog_tailcalls(void);
 bool bpf_jit_supports_percpu_insn(void);
 bool bpf_jit_supports_kfunc_call(void);
 bool bpf_jit_supports_stack_args(void);
+bool bpf_jit_supports_arena_args(void);
 bool bpf_jit_supports_far_kfunc_call(void);
 bool bpf_jit_supports_exceptions(void);
 bool bpf_jit_supports_ptr_xchg(void);
diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
index 4eeeaeb69790..f54fddfb5c8e 100644
--- a/kernel/bpf/btf.c
+++ b/kernel/bpf/btf.c
@@ -7541,6 +7541,22 @@ static u8 __get_type_fmodel_flags(const struct btf_type *t)
 	return flags;
 }
 
+static u8 __get_arg_fmodel_flags(const struct btf *btf,
+				 const struct btf_param *arg,
+				 const struct btf_type *t)
+{
+	u8 flags = __get_type_fmodel_flags(t);
+
+	if (btf_param_match_suffix(btf, arg, "__arena") ||
+	    btf_param_match_suffix(btf, arg, "__arena_nullable"))
+		flags |= BTF_FMODEL_ARENA_ARG;
+	if (btf_param_match_suffix(btf, arg, "__nullable") ||
+	    btf_param_match_suffix(btf, arg, "__arena_nullable"))
+		flags |= BTF_FMODEL_NULLABLE_ARG;
+
+	return flags;
+}
+
 int btf_distill_func_proto(struct bpf_verifier_log *log,
 			   struct btf *btf,
 			   const struct btf_type *func,
@@ -7606,7 +7622,7 @@ int btf_distill_func_proto(struct bpf_verifier_log *log,
 			return -EINVAL;
 		}
 		m->arg_size[i] = ret;
-		m->arg_flags[i] = __get_type_fmodel_flags(t);
+		m->arg_flags[i] = __get_arg_fmodel_flags(btf, &args[i], t);
 	}
 	m->nr_args = nargs;
 	return 0;
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index e2076667b245..a3e1fae32eac 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -3308,6 +3308,11 @@ bool __weak bpf_jit_supports_stack_args(void)
 	return false;
 }
 
+bool __weak bpf_jit_supports_arena_args(void)
+{
+	return false;
+}
+
 bool __weak bpf_jit_supports_far_kfunc_call(void)
 {
 	return false;
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 641c3c62c1ec..f3e6190f639e 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -10847,6 +10847,16 @@ static bool is_kfunc_arg_irq_flag(const struct btf *btf, const struct btf_param
 	return btf_param_match_suffix(btf, arg, "__irq_flag");
 }
 
+static bool is_kfunc_arg_arena(const struct btf *btf, const struct btf_param *arg)
+{
+	return btf_param_match_suffix(btf, arg, "__arena");
+}
+
+static bool is_kfunc_arg_arena_nullable(const struct btf *btf, const struct btf_param *arg)
+{
+	return btf_param_match_suffix(btf, arg, "__arena_nullable");
+}
+
 static bool is_kfunc_arg_scalar_with_name(const struct btf *btf,
 					  const struct btf_param *arg,
 					  const char *name)
@@ -12056,6 +12066,35 @@ static int check_kfunc_args(struct bpf_verifier_env *env, struct bpf_call_arg_me
 
 		t = btf_type_skip_modifiers(btf, args[i].type, NULL);
 
+		if (is_kfunc_arg_arena(btf, &args[i]) || is_kfunc_arg_arena_nullable(btf, &args[i])) {
+			if (verifier_bug_if(!btf_type_is_ptr(t), env,
+					    "kfunc %s arg#%d has __arena tag on non-pointer",
+					    func_name, i))
+				return -EFAULT;
+			if (!bpf_jit_supports_arena_args()) {
+				verbose(env, "JIT does not support kfunc %s() with arena pointer arguments\n",
+					func_name);
+				return -ENOTSUPP;
+			}
+			if (!env->prog->aux->arena) {
+				verbose(env,
+					"%s arena pointer requires a program with an associated arena\n",
+					reg_arg_name(env, argno));
+				return -EINVAL;
+			}
+			if (regno < 0) {
+				verbose(env, "%s arena pointer cannot be a stack argument\n",
+					reg_arg_name(env, argno));
+				return -EINVAL;
+			}
+			if (reg->type != PTR_TO_ARENA && reg->type != SCALAR_VALUE) {
+				verbose(env, "%s is not a pointer to arena or scalar\n",
+					reg_arg_name(env, argno));
+				return -EINVAL;
+			}
+			continue;
+		}
+
 		if (btf_type_is_scalar(t)) {
 			if (reg->type != SCALAR_VALUE) {
 				verbose(env, "%s is not a scalar\n", reg_arg_name(env, argno));
-- 
2.53.0

Re: [PATCH bpf-next v2 1/9] bpf: Support __arena and __arena_nullable kfunc argument suffixes

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The `__arena_nullable` kfunc argument design uses a 32-bit truncated value of 0 as a sentinel for NULL, which collides with valid memory allocations from `bpf_arena_alloc_pages()`.
--

--- Patch [1]: [PATCH bpf-next v2 1/9] bpf: Support __arena and __arena_nullable kfunc argument suffixes ---
commit fb759ca6f84030ccde9ee83f1e83634342bdae0d
Author: Tejun Heo <tj@kernel.org>

bpf: Support __arena and __arena_nullable kfunc argument suffixes

This patch adds support for passing BPF arena pointers to kfuncs by introducing
__arena and __arena_nullable argument suffixes. The verifier processes these
suffixes and allows JITs to handle the translation (rebasing) at the call site.
__arena rebases unconditionally, while __arena_nullable preserves NULL by skipping
the rebase when the arena offset is zero.

> diff --git a/Documentation/bpf/kfuncs.rst b/Documentation/bpf/kfuncs.rst
> index c801a330aece0..9809680bee43e 100644
> --- a/Documentation/bpf/kfuncs.rst
> +++ b/Documentation/bpf/kfuncs.rst
[ ... ]
> @@ -250,6 +250,33 @@ Or::
>                  ...
>          }
>  
> +2.3.7 __arena and __arena_nullable Annotations
> +----------------------------------------------
> +
> +Both annotations indicate that the pointer argument points into the
> +calling program's arena. The JIT rebases the value at the call site so
> +the kfunc receives a directly dereferenceable kernel address, subject to
> +the access rules described in :ref:`BPF_kfunc_arena_access` (at most
> +``GUARD_SZ / 2``, 32 KiB, past the pointer in a single unchecked access).
> +
> +With ``__arena`` the rebase is unconditional and the argument is never
> +NULL: a value whose lower 32 bits are zero arrives as the arena base
> +address (arena offset 0). The kfunc must not check the argument for NULL.
> +With ``__arena_nullable`` such a value arrives as NULL instead and the
> +kfunc must check before dereferencing.

[Severity: High]
Does this __arena_nullable design cause a collision with valid allocations?

If an arena is mapped at a 4GB-aligned boundary, the user_vm_start address will
have its lower 32 bits as 0. When bpf_arena_alloc_pages() returns the first
free page at offset 0, the resulting user pointer also has its lower 32 bits
equal to 0.

If a BPF program passes this valid pointer to an __arena_nullable kfunc, won't
the JIT see the 0 value in the lower 32 bits and incorrectly pass NULL to the
kfunc instead of the valid arena base address?

> +
> +An example is given below::
[ ... ]
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 641c3c62c1ecf..f3e6190f639e4 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
[ ... ]
> @@ -12056,6 +12066,35 @@ static int check_kfunc_args(struct bpf_verifier_env *env, struct bpf_call_arg_me
>  
>  		t = btf_type_skip_modifiers(btf, args[i].type, NULL);
>  
> +		if (is_kfunc_arg_arena(btf, &args[i]) || is_kfunc_arg_arena_nullable(btf, &args[i])) {

[Severity: High]
Should check_kfunc_args() or the arena allocator ensure that offset 0 is
reserved?

If offset 0 is not reserved, the verifier will allow programs to pass a valid
offset 0 allocation to __arena_nullable arguments. The JIT will then silently
translate this to NULL, causing unexpected behavior in the kfunc since it will
assume the optional argument was not provided.

> +			if (verifier_bug_if(!btf_type_is_ptr(t), env,
> +					    "kfunc %s arg#%d has __arena tag on non-pointer",
> +					    func_name, i))
> +				return -EFAULT;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260726013105.3689867-1-memxor@gmail.com?part=1

Re: [PATCH bpf-next v2 1/9] bpf: Support __arena and __arena_nullable kfunc argument suffixes

[Kumar Kartikeya Dwivedi]

On Sun Jul 26, 2026 at 3:47 AM CEST, sashiko-bot wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [High] The `__arena_nullable` kfunc argument design uses a 32-bit truncated value of 0 as a sentinel for NULL, which collides with valid memory allocations from `bpf_arena_alloc_pages()`.
> --
>
> --- Patch [1]: [PATCH bpf-next v2 1/9] bpf: Support __arena and __arena_nullable kfunc argument suffixes ---
> commit fb759ca6f84030ccde9ee83f1e83634342bdae0d
> Author: Tejun Heo <tj@kernel.org>
>
> bpf: Support __arena and __arena_nullable kfunc argument suffixes
>
> This patch adds support for passing BPF arena pointers to kfuncs by introducing
> __arena and __arena_nullable argument suffixes. The verifier processes these
> suffixes and allows JITs to handle the translation (rebasing) at the call site.
> __arena rebases unconditionally, while __arena_nullable preserves NULL by skipping
> the rebase when the arena offset is zero.
>
>> diff --git a/Documentation/bpf/kfuncs.rst b/Documentation/bpf/kfuncs.rst
>> index c801a330aece0..9809680bee43e 100644
>> --- a/Documentation/bpf/kfuncs.rst
>> +++ b/Documentation/bpf/kfuncs.rst
> [ ... ]
>> @@ -250,6 +250,33 @@ Or::
>>                  ...
>>          }
>>
>> +2.3.7 __arena and __arena_nullable Annotations
>> +----------------------------------------------
>> +
>> +Both annotations indicate that the pointer argument points into the
>> +calling program's arena. The JIT rebases the value at the call site so
>> +the kfunc receives a directly dereferenceable kernel address, subject to
>> +the access rules described in :ref:`BPF_kfunc_arena_access` (at most
>> +``GUARD_SZ / 2``, 32 KiB, past the pointer in a single unchecked access).
>> +
>> +With ``__arena`` the rebase is unconditional and the argument is never
>> +NULL: a value whose lower 32 bits are zero arrives as the arena base
>> +address (arena offset 0). The kfunc must not check the argument for NULL.
>> +With ``__arena_nullable`` such a value arrives as NULL instead and the
>> +kfunc must check before dereferencing.
>
> [Severity: High]
> Does this __arena_nullable design cause a collision with valid allocations?
>
> If an arena is mapped at a 4GB-aligned boundary, the user_vm_start address will
> have its lower 32 bits as 0. When bpf_arena_alloc_pages() returns the first
> free page at offset 0, the resulting user pointer also has its lower 32 bits
> equal to 0.
>
> If a BPF program passes this valid pointer to an __arena_nullable kfunc, won't
> the JIT see the 0 value in the lower 32 bits and incorrectly pass NULL to the
> kfunc instead of the valid arena base address?

Yes, but it is intentional. Referencing the first page is equivalent to NULL, or
start of the arena "address space".

>
>> +
>> +An example is given below::
> [ ... ]
>> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
>> index 641c3c62c1ecf..f3e6190f639e4 100644
>> --- a/kernel/bpf/verifier.c
>> +++ b/kernel/bpf/verifier.c
> [ ... ]
>> @@ -12056,6 +12066,35 @@ static int check_kfunc_args(struct bpf_verifier_env *env, struct bpf_call_arg_me
>>
>>  		t = btf_type_skip_modifiers(btf, args[i].type, NULL);
>>
>> +		if (is_kfunc_arg_arena(btf, &args[i]) || is_kfunc_arg_arena_nullable(btf, &args[i])) {
>
> [Severity: High]
> Should check_kfunc_args() or the arena allocator ensure that offset 0 is
> reserved?
>
> If offset 0 is not reserved, the verifier will allow programs to pass a valid
> offset 0 allocation to __arena_nullable arguments. The JIT will then silently
> translate this to NULL, causing unexpected behavior in the kfunc since it will
> assume the optional argument was not provided.

It will always be left unmapped in practice.

>
>> +			if (verifier_bug_if(!btf_type_is_ptr(t), env,
>> +					    "kfunc %s arg#%d has __arena tag on non-pointer",
>> +					    func_name, i))
>> +				return -EFAULT;

[PATCH bpf-next v2 2/9] bpf: Support __arena and __arena_nullable on struct_ops arguments

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

A struct_ops callback cannot receive an arena pointer directly, so
passing one takes two steps. The pointer arrives as a bare u64 that the
callback casts, and because the two sides address the arena through
different bases it also has to be rebased by hand on the way in.

Add the __arena and __arena_nullable stub argument suffixes to make this
convenient. The callback declares the parameter as an arena pointer,
receives it as a PTR_TO_ARENA register, and dereferences it directly,
while the kernel caller just passes the natural kernel arena address
(kaddr). The trampoline converts the value while saving the arguments
into the BPF ctx, ctx[slot] = (u32)(kaddr - kern_vm_start), so the
program never sees a kernel address and nothing rewrites the ctx after
the fact. The converted value keeps the upper 32 bits clear as the JITs
require of arena pointer registers and behaves like any cast_kern'ed
arena pointer, so cast_user recovers the full user-visible address.

__arena converts unconditionally and the kernel caller must not pass
NULL. __arena_nullable preserves NULL, tested on the full 64-bit kernel
pointer, and surfaces to the verifier as PTR_TO_ARENA (but not as a
PTR_TO_ARENA | PTR_MAYBE_NULL). The reason is that PTR_TO_ARENA in
program's type state already encompasses NULL-ness, so it is not
meaningful to force a NULL check for the program.

This patch adds the generic side. bpf_tramp_collect_arena_args() derives
the conversion map from the prog's ctx_arg_info, keyed by the flattened
ctx byte offset since preceding 16-byte arguments occupy two slots. Only
the struct_ops indirect trampoline converts: it dispatches to a single
prog whose arena is fixed at generation time. Generic trampolines can
mix progs with different arenas and reject arena ctx args defensively,
which is unreachable today as only struct_ops progs carry them. Arch
trampolines that do not implement the conversion are gated out at
verification time with bpf_jit_supports_arena_args().

Signed-off-by: Tejun Heo <tj@kernel.org>
Co-developed-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 Documentation/bpf/kfuncs.rst | 10 ++++++
 include/linux/bpf.h          | 19 +++++++++++
 kernel/bpf/bpf_struct_ops.c  | 34 ++++++++++++++-----
 kernel/bpf/btf.c             | 10 ++++--
 kernel/bpf/trampoline.c      | 64 ++++++++++++++++++++++++++++++++++++
 kernel/bpf/verifier.c        | 23 ++++++++++---
 6 files changed, 143 insertions(+), 17 deletions(-)

diff --git a/Documentation/bpf/kfuncs.rst b/Documentation/bpf/kfuncs.rst
index 9809680bee43..426aa1f314b9 100644
--- a/Documentation/bpf/kfuncs.rst
+++ b/Documentation/bpf/kfuncs.rst
@@ -277,6 +277,16 @@ arena argument support (currently x86-64); verification fails otherwise. The
 program can pass any value without compromising the kernel. A value that does
 not point into the arena is a program bug.
 
+The suffixes have the same meaning on the arguments of struct_ops stub
+functions, with the conversion running in the opposite direction. The
+kernel caller passes the kernel arena address and the trampoline converts
+it while saving the arguments, so the callback receives an arena pointer
+it can dereference directly. With ``__arena`` the kernel caller must not
+pass NULL. With ``__arena_nullable`` a NULL kernel pointer arrives as NULL.
+However, there is no obligation to prove to the verifier that such a pointer is
+non-NULL before use, in-line with existing semantics of arena pointers used in
+a program (or obtained from any other source).
+
 .. _BPF_kfunc_nodef:
 
 2.4 Using an existing kernel function
diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 1ca4f2115fde..c839f039729a 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -1292,6 +1292,20 @@ struct bpf_tramp_nodes {
 	int nr_nodes;
 };
 
+/*
+ * Which 8-byte ctx slots of a struct_ops trampoline hold arena kernel
+ * pointers that save_args() converts to the arena pointer form,
+ * ctx[slot] = (u32)(kaddr - kern_vm_start).
+ */
+struct bpf_tramp_arena_args {
+	u32 slots;
+	u32 nullable_slots;	/* subset of @slots where NULL is preserved */
+	u64 kern_vm_start;
+};
+
+bool bpf_tramp_collect_arena_args(struct bpf_tramp_nodes *tnodes, u32 flags,
+				  struct bpf_tramp_arena_args *aargs);
+
 struct bpf_tramp_run_ctx;
 
 /* Different use cases for BPF trampoline:
@@ -1693,6 +1707,11 @@ struct bpf_ctx_arg_aux {
 	u32 btf_id;
 	u32 ref_id;
 	bool refcounted;
+	/*
+	 * We don't encode NULL-ness in the type for the program, but still need
+	 * to distinguish it for the purposes of telling JITs what sequence to emit.
+	 */
+	bool arena_nullable;
 };
 
 struct btf_mod_pair {
diff --git a/kernel/bpf/bpf_struct_ops.c b/kernel/bpf/bpf_struct_ops.c
index 51b16e5f5534..2757380194f8 100644
--- a/kernel/bpf/bpf_struct_ops.c
+++ b/kernel/bpf/bpf_struct_ops.c
@@ -147,6 +147,8 @@ void bpf_struct_ops_image_free(void *image)
 
 #define MAYBE_NULL_SUFFIX "__nullable"
 #define REFCOUNTED_SUFFIX "__ref"
+#define ARENA_SUFFIX "__arena"
+#define ARENA_MAYBE_NULL_SUFFIX "__arena_nullable"
 
 /* Prepare argument info for every nullable argument of a member of a
  * struct_ops type.
@@ -159,7 +161,7 @@ void bpf_struct_ops_image_free(void *image)
  * to provide an array of struct bpf_ctx_arg_aux, which in turn provides
  * the information that used by the verifier to check the arguments of the
  * BPF struct_ops program assigned to the member. Here, we only care about
- * the arguments that are marked as __nullable.
+ * the arguments that are marked as __nullable, __ref or __arena.
  *
  * The array of struct bpf_ctx_arg_aux is eventually assigned to
  * prog->aux->ctx_arg_info of BPF struct_ops programs and passed to the
@@ -175,7 +177,8 @@ static int prepare_arg_info(struct btf *btf,
 			    struct bpf_struct_ops_arg_info *arg_info)
 {
 	const struct btf_type *stub_func_proto, *pointed_type;
-	bool is_nullable = false, is_refcounted = false;
+	bool is_nullable = false, is_refcounted = false, is_arena = false;
+	bool is_arena_nullable = false;
 	const struct btf_param *stub_args, *args;
 	struct bpf_ctx_arg_aux *info, *info_buf;
 	u32 nargs, arg_no, info_cnt = 0;
@@ -226,26 +229,30 @@ static int prepare_arg_info(struct btf *btf,
 	info = info_buf;
 	for (arg_no = 0; arg_no < nargs; arg_no++) {
 		/* Skip arguments that is not suffixed with
-		 * "__nullable or __ref".
+		 * "__nullable", "__ref", "__arena" or "__arena_nullable".
 		 */
 		is_nullable = btf_param_match_suffix(btf, &stub_args[arg_no],
 						     MAYBE_NULL_SUFFIX);
 		is_refcounted = btf_param_match_suffix(btf, &stub_args[arg_no],
 						       REFCOUNTED_SUFFIX);
+		is_arena_nullable = btf_param_match_suffix(btf, &stub_args[arg_no],
+							   ARENA_MAYBE_NULL_SUFFIX);
+		is_arena = btf_param_match_suffix(btf, &stub_args[arg_no], ARENA_SUFFIX);
 
 		if (is_nullable)
 			suffix = MAYBE_NULL_SUFFIX;
 		else if (is_refcounted)
 			suffix = REFCOUNTED_SUFFIX;
+		else if (is_arena_nullable)
+			suffix = ARENA_MAYBE_NULL_SUFFIX;
+		else if (is_arena)
+			suffix = ARENA_SUFFIX;
 		else
 			continue;
 
-		/* Should be a pointer to struct */
-		pointed_type = btf_type_resolve_ptr(btf,
-						    args[arg_no].type,
-						    &arg_btf_id);
-		if (!pointed_type ||
-		    !btf_type_is_struct(pointed_type)) {
+		/* Should be a pointer to struct, or any pointer for __arena/__arena_nullable */
+		pointed_type = btf_type_resolve_ptr(btf, args[arg_no].type, &arg_btf_id);
+		if (!pointed_type || (!is_arena && !is_arena_nullable && !btf_type_is_struct(pointed_type))) {
 			pr_warn("stub function %s has %s tagging to an unsupported type\n",
 				stub_fname, suffix);
 			goto err_out;
@@ -273,6 +280,15 @@ static int prepare_arg_info(struct btf *btf,
 		} else if (is_refcounted) {
 			info->reg_type = PTR_TRUSTED | PTR_TO_BTF_ID;
 			info->refcounted = true;
+		} else if (is_arena || is_arena_nullable) {
+			/*
+			 * Both types get PTR_TO_ARENA. In verifier state,
+			 * PTR_TO_ARENA encompasses potential NULL values, but
+			 * we do not force the program to check it, or maintain
+			 * precision around it, since it has no safety implication.
+			 */
+			info->reg_type = PTR_TO_ARENA;
+			info->arena_nullable = is_arena_nullable;
 		}
 
 		info++;
diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
index f54fddfb5c8e..8ebedad7b9f0 100644
--- a/kernel/bpf/btf.c
+++ b/kernel/bpf/btf.c
@@ -6963,15 +6963,19 @@ bool btf_ctx_access(int off, int size, enum bpf_access_type type,
 		return false;
 	}
 
-	/* check for PTR_TO_RDONLY_BUF_OR_NULL or PTR_TO_RDWR_BUF_OR_NULL */
+	/*
+	 * Check for PTR_TO_RDONLY_BUF_OR_NULL, PTR_TO_RDWR_BUF_OR_NULL or
+	 * PTR_TO_ARENA (both nullable and non-nullable cases).
+	 */
 	for (i = 0; i < prog->aux->ctx_arg_info_size; i++) {
 		const struct bpf_ctx_arg_aux *ctx_arg_info = &prog->aux->ctx_arg_info[i];
 		u32 type, flag;
 
 		type = base_type(ctx_arg_info->reg_type);
 		flag = type_flag(ctx_arg_info->reg_type);
-		if (ctx_arg_info->offset == off && type == PTR_TO_BUF &&
-		    (flag & PTR_MAYBE_NULL)) {
+		if (ctx_arg_info->offset == off &&
+		    (type == PTR_TO_ARENA ||
+		     (type == PTR_TO_BUF && (flag & PTR_MAYBE_NULL)))) {
 			info->reg_type = ctx_arg_info->reg_type;
 			return true;
 		}
diff --git a/kernel/bpf/trampoline.c b/kernel/bpf/trampoline.c
index 129d07db117e..a1ad409bdde0 100644
--- a/kernel/bpf/trampoline.c
+++ b/kernel/bpf/trampoline.c
@@ -529,6 +529,53 @@ bpf_trampoline_get_progs(const struct bpf_trampoline *tr, int *total, bool *ip_a
 	return tnodes;
 }
 
+static bool bpf_prog_has_arena_ctx_arg(const struct bpf_prog *prog)
+{
+	int i;
+
+	for (i = 0; i < prog->aux->ctx_arg_info_size; i++)
+		if (base_type(prog->aux->ctx_arg_info[i].reg_type) == PTR_TO_ARENA)
+			return true;
+	return false;
+}
+
+/*
+ * Collect which ctx slots of a struct_ops trampoline hold arena kernel
+ * pointers that save_args() must convert to the arena pointer form. Only
+ * the struct_ops indirect trampoline converts: it dispatches to a single
+ * prog whose arena is known at generation time. Return false when there
+ * is nothing to convert.
+ */
+bool bpf_tramp_collect_arena_args(struct bpf_tramp_nodes *tnodes, u32 flags,
+				  struct bpf_tramp_arena_args *aargs)
+{
+	const struct bpf_prog *prog;
+	int i;
+
+	memset(aargs, 0, sizeof(*aargs));
+
+	if (!(flags & BPF_TRAMP_F_INDIRECT) ||
+	    tnodes[BPF_TRAMP_FENTRY].nr_nodes != 1)
+		return false;
+
+	prog = tnodes[BPF_TRAMP_FENTRY].nodes[0]->link->prog;
+	for (i = 0; i < prog->aux->ctx_arg_info_size; i++) {
+		const struct bpf_ctx_arg_aux *info = &prog->aux->ctx_arg_info[i];
+
+		if (base_type(info->reg_type) != PTR_TO_ARENA)
+			continue;
+		aargs->slots |= BIT(info->offset / 8);
+		if (info->arena_nullable)
+			aargs->nullable_slots |= BIT(info->offset / 8);
+	}
+	if (!aargs->slots)
+		return false;
+	if (WARN_ON_ONCE(!prog->aux->arena))
+		return false;
+	aargs->kern_vm_start = bpf_arena_get_kern_vm_start(prog->aux->arena);
+	return true;
+}
+
 static void bpf_tramp_image_free(struct bpf_tramp_image *im)
 {
 	bpf_image_ksym_del(&im->ksym);
@@ -685,6 +732,7 @@ static int bpf_trampoline_update(struct bpf_trampoline *tr, bool lock_direct_mut
 	u32 orig_flags = tr->flags;
 	bool ip_arg = false;
 	int err, total, size;
+	int kind, i;
 
 	tnodes = bpf_trampoline_get_progs(tr, &total, &ip_arg);
 	if (IS_ERR(tnodes))
@@ -695,6 +743,22 @@ static int bpf_trampoline_update(struct bpf_trampoline *tr, bool lock_direct_mut
 		goto out;
 	}
 
+	/*
+	 * Arena ctx args are converted only by the struct_ops indirect
+	 * trampoline, which dispatches to a single known prog. Generic
+	 * trampolines can mix progs with different arenas, so no conversion
+	 * is possible here. Not reachable today: only struct_ops progs get
+	 * arena ctx args and they never ride generic trampolines.
+	 */
+	for (kind = 0; kind < BPF_TRAMP_MAX; kind++) {
+		for (i = 0; i < tnodes[kind].nr_nodes; i++) {
+			if (bpf_prog_has_arena_ctx_arg(tnodes[kind].nodes[i]->link->prog)) {
+				err = -ENOTSUPP;
+				goto out;
+			}
+		}
+	}
+
 	/* clear all bits except SHARE_IPMODIFY and TAIL_CALL_CTX */
 	tr->flags &= (BPF_TRAMP_F_SHARE_IPMODIFY | BPF_TRAMP_F_TAIL_CALL_CTX);
 
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index f3e6190f639e..1494da7a7e6d 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -18770,6 +18770,7 @@ static int check_struct_ops_btf_id(struct bpf_verifier_env *env)
 {
 	const struct btf_type *t, *func_proto;
 	const struct bpf_struct_ops_desc *st_ops_desc;
+	const struct bpf_struct_ops_arg_info *arg_info;
 	const struct bpf_struct_ops *st_ops;
 	const struct btf_member *member;
 	struct bpf_prog *prog = env->prog;
@@ -18848,10 +18849,23 @@ static int check_struct_ops_btf_id(struct bpf_verifier_env *env)
 		return -EACCES;
 	}
 
-	for (i = 0; i < st_ops_desc->arg_info[member_idx].cnt; i++) {
-		if (st_ops_desc->arg_info[member_idx].info[i].refcounted) {
+	arg_info = &st_ops_desc->arg_info[member_idx];
+	for (i = 0; i < arg_info->cnt; i++) {
+		const struct bpf_ctx_arg_aux *info = &arg_info->info[i];
+
+		if (info->refcounted)
 			has_refcounted_arg = true;
-			break;
+		if (base_type(info->reg_type) == PTR_TO_ARENA) {
+			if (!bpf_jit_supports_arena_args()) {
+				verbose(env, "JIT does not support arena arguments\n");
+				return -ENOTSUPP;
+			}
+			if (!prog->aux->arena) {
+				verbose(env,
+					"arena argument of %s requires a program with an associated arena\n",
+					mname);
+				return -EINVAL;
+			}
 		}
 	}
 
@@ -18872,8 +18886,7 @@ static int check_struct_ops_btf_id(struct bpf_verifier_env *env)
 	prog->aux->attach_func_name = mname;
 	env->ops = st_ops->verifier_ops;
 
-	return bpf_prog_ctx_arg_info_init(prog, st_ops_desc->arg_info[member_idx].info,
-					  st_ops_desc->arg_info[member_idx].cnt);
+	return bpf_prog_ctx_arg_info_init(prog, arg_info->info, arg_info->cnt);
 }
 #define SECURITY_PREFIX "security_"
 
-- 
2.53.0

[PATCH bpf-next v2 3/9] bpf, x86: JIT __arena kfunc argument rebasing

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

Implement arena argument rebasing for kfunc calls on x86. R12 already
holds kern_vm_start whenever the prog has an arena, so each tagged
argument costs two instructions emitted right before the call:

  movl %eN, %eN         /* truncate, clear the upper 32 bits */
  addq %r12, %rN

A nullable argument tests the truncated value and jumps over the add:

  movl  %eN, %eN
  testl %eN, %eN
  jz    1f
  addq  %r12, %rN
1:

addq carries a REX prefix for every argument register and is always
three bytes, so the jz displacement is constant. The sequence is native
code generated after constant blinding has run on the BPF instruction
stream, so blinding never sees the rebase and needs no special handling.

bpf_jit_supports_arena_args() is not flipped yet; that happens when the
struct_ops trampoline side is in place as well.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 arch/x86/net/bpf_jit_comp.c | 50 +++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)

diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
index b2feec81e231..08013c2fcfaa 100644
--- a/arch/x86/net/bpf_jit_comp.c
+++ b/arch/x86/net/bpf_jit_comp.c
@@ -1678,6 +1678,50 @@ static int emit_spectre_bhb_barrier(u8 **pprog, u8 *ip,
 	return 0;
 }
 
+/*
+ * Rebase the __arena args of a kfunc call to arena kernel addresses,
+ * rN = kern_vm_start + (u32)rN, with R12 holding kern_vm_start. A nullable
+ * arg preserves NULL by skipping the add, tested on the truncated value as
+ * arena NULL is offset 0. Return the number of emitted bytes.
+ */
+static int emit_kfunc_arena_args(struct bpf_prog *bpf_prog,
+				 const struct bpf_insn *insn, u8 **pprog)
+{
+	const struct btf_func_model *fm;
+	u8 *prog = *pprog;
+	u8 *start = prog;
+	int i;
+
+	fm = bpf_jit_find_kfunc_model(bpf_prog, insn);
+	if (!fm)
+		return -EINVAL;
+
+	for (i = 0; i < min_t(int, fm->nr_args, MAX_BPF_FUNC_REG_ARGS); i++) {
+		u8 flags = fm->arg_flags[i];
+		u32 reg = BPF_REG_1 + i;
+
+		if (!(flags & BTF_FMODEL_ARENA_ARG))
+			continue;
+		if (WARN_ON_ONCE(!bpf_prog->aux->arena))
+			return -EINVAL;
+
+		/* mov eN, eN: truncate and clear the upper 32 bits */
+		emit_mov_reg(&prog, false, reg, reg);
+		if (flags & BTF_FMODEL_NULLABLE_ARG) {
+			/* test eN, eN; jz over the 3-byte add */
+			maybe_emit_mod(&prog, reg, reg, false);
+			EMIT2(0x85, add_2reg(0xC0, reg, reg));
+			EMIT2(X86_JE, 3);
+		}
+		/* add rN, r12 */
+		maybe_emit_mod(&prog, reg, X86_REG_R12, true);
+		EMIT2(0x01, add_2reg(0xC0, reg, X86_REG_R12));
+	}
+
+	*pprog = prog;
+	return prog - start;
+}
+
 static int do_jit(struct bpf_verifier_env *env, struct bpf_prog *bpf_prog, int *addrs, u8 *image,
 		  u8 *rw_image, int oldproglen, struct jit_context *ctx, bool jmp_padding)
 {
@@ -2583,6 +2627,12 @@ st:			insn_off = insn->off;
 			}
 			if (!imm32)
 				return -EINVAL;
+			if (src_reg == BPF_PSEUDO_KFUNC_CALL) {
+				err = emit_kfunc_arena_args(bpf_prog, insn, &prog);
+				if (err < 0)
+					return err;
+				ip += err;
+			}
 			if (priv_frame_ptr) {
 				push_r9(&prog);
 				ip += 2;
-- 
2.53.0

Re: [PATCH bpf-next v2 3/9] bpf, x86: JIT __arena kfunc argument rebasing

[bot+bpf-ci]

[-- Attachment #1: Type: text/plain, Size: 3341 bytes --]

> diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
> index b2feec81e231..08013c2fcfaa 100644
> --- a/arch/x86/net/bpf_jit_comp.c
> +++ b/arch/x86/net/bpf_jit_comp.c
> @@ -1678,6 +1678,50 @@ static int emit_spectre_bhb_barrier(u8 **pprog, u8 *ip,
>  	return 0;
>  }
>
> +/*
> + * Rebase the __arena args of a kfunc call to arena kernel addresses,
> + * rN = kern_vm_start + (u32)rN, with R12 holding kern_vm_start. A nullable
> + * arg preserves NULL by skipping the add, tested on the truncated value as
> + * arena NULL is offset 0. Return the number of emitted bytes.
> + */
> +static int emit_kfunc_arena_args(struct bpf_prog *bpf_prog,
> +				 const struct bpf_insn *insn, u8 **pprog)
> +{
> +	const struct btf_func_model *fm;
> +	u8 *prog = *pprog;
> +	u8 *start = prog;
> +	int i;
> +
> +	fm = bpf_jit_find_kfunc_model(bpf_prog, insn);
> +	if (!fm)
> +		return -EINVAL;
> +
> +	for (i = 0; i < min_t(int, fm->nr_args, MAX_BPF_FUNC_REG_ARGS); i++) {
> +		u8 flags = fm->arg_flags[i];
> +		u32 reg = BPF_REG_1 + i;
> +
> +		if (!(flags & BTF_FMODEL_ARENA_ARG))
> +			continue;
> +		if (WARN_ON_ONCE(!bpf_prog->aux->arena))
> +			return -EINVAL;
> +
> +		/* mov eN, eN: truncate and clear the upper 32 bits */
> +		emit_mov_reg(&prog, false, reg, reg);
> +		if (flags & BTF_FMODEL_NULLABLE_ARG) {
> +			/* test eN, eN; jz over the 3-byte add */
> +			maybe_emit_mod(&prog, reg, reg, false);
> +			EMIT2(0x85, add_2reg(0xC0, reg, reg));
> +			EMIT2(X86_JE, 3);
> +		}
> +		/* add rN, r12 */
> +		maybe_emit_mod(&prog, reg, X86_REG_R12, true);
> +		EMIT2(0x01, add_2reg(0xC0, reg, X86_REG_R12));
> +	}
> +
> +	*pprog = prog;
> +	return prog - start;
> +}

In v1 review, Eduard Zingerman raised a concern about the JIT-based
implementation approach, suggesting that it spreads complexity across
multiple JITs (verifier.c + 6 jits) instead of keeping the bulk of the
implementation in one place. He proposed an alternative using R0 by default
with a jit_has_dedicated_arena_register() predicate to avoid the
64-immediate move overhead while centralizing the logic in verifier.c.

Kumar responded that while this approach is reasonable for kfuncs, struct_ops
would require uglier code with reloading/translating/storing. The v2 patch
retained the JIT-based approach.

Given that this will eventually need implementation in 6 different JITs
(assuming all except s390 will gain arena support), would it be worth
reconsidering the centralized verifier.c approach? The optimization saves a
64-immediate move per callsite, but does that performance benefit justify the
increased maintenance burden of ensuring correct implementation across
multiple JIT backends?

Link: https://lore.kernel.org/bpf/bbf82d80eb0dc307d09a6c7cc33c23063a55a90f.camel@gmail.com/

> @@ -2583,6 +2627,12 @@ st:			insn_off = insn->off;
>  		}
>  		if (!imm32)
>  			return -EINVAL;
> +		if (src_reg == BPF_PSEUDO_KFUNC_CALL) {
> +			err = emit_kfunc_arena_args(bpf_prog, insn, &prog);
> +			if (err < 0)
> +				return err;
> +			ip += err;
> +		}
>  		if (priv_frame_ptr) {
>  			push_r9(&prog);
>  			ip += 2;


---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/30183383612

Re: [PATCH bpf-next v2 3/9] bpf, x86: JIT __arena kfunc argument rebasing

[Kumar Kartikeya Dwivedi]

On Sun Jul 26, 2026 at 4:22 AM CEST, bot+bpf-ci wrote:
>> diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
>> index b2feec81e231..08013c2fcfaa 100644
>> --- a/arch/x86/net/bpf_jit_comp.c
>> +++ b/arch/x86/net/bpf_jit_comp.c
>> @@ -1678,6 +1678,50 @@ static int emit_spectre_bhb_barrier(u8 **pprog, u8 *ip,
>>  	return 0;
>>  }
>>
>> +/*
>> + * Rebase the __arena args of a kfunc call to arena kernel addresses,
>> + * rN = kern_vm_start + (u32)rN, with R12 holding kern_vm_start. A nullable
>> + * arg preserves NULL by skipping the add, tested on the truncated value as
>> + * arena NULL is offset 0. Return the number of emitted bytes.
>> + */
>> +static int emit_kfunc_arena_args(struct bpf_prog *bpf_prog,
>> +				 const struct bpf_insn *insn, u8 **pprog)
>> +{
>> +	const struct btf_func_model *fm;
>> +	u8 *prog = *pprog;
>> +	u8 *start = prog;
>> +	int i;
>> +
>> +	fm = bpf_jit_find_kfunc_model(bpf_prog, insn);
>> +	if (!fm)
>> +		return -EINVAL;
>> +
>> +	for (i = 0; i < min_t(int, fm->nr_args, MAX_BPF_FUNC_REG_ARGS); i++) {
>> +		u8 flags = fm->arg_flags[i];
>> +		u32 reg = BPF_REG_1 + i;
>> +
>> +		if (!(flags & BTF_FMODEL_ARENA_ARG))
>> +			continue;
>> +		if (WARN_ON_ONCE(!bpf_prog->aux->arena))
>> +			return -EINVAL;
>> +
>> +		/* mov eN, eN: truncate and clear the upper 32 bits */
>> +		emit_mov_reg(&prog, false, reg, reg);
>> +		if (flags & BTF_FMODEL_NULLABLE_ARG) {
>> +			/* test eN, eN; jz over the 3-byte add */
>> +			maybe_emit_mod(&prog, reg, reg, false);
>> +			EMIT2(0x85, add_2reg(0xC0, reg, reg));
>> +			EMIT2(X86_JE, 3);
>> +		}
>> +		/* add rN, r12 */
>> +		maybe_emit_mod(&prog, reg, X86_REG_R12, true);
>> +		EMIT2(0x01, add_2reg(0xC0, reg, X86_REG_R12));
>> +	}
>> +
>> +	*pprog = prog;
>> +	return prog - start;
>> +}
>
> In v1 review, Eduard Zingerman raised a concern about the JIT-based
> implementation approach, suggesting that it spreads complexity across
> multiple JITs (verifier.c + 6 jits) instead of keeping the bulk of the
> implementation in one place. He proposed an alternative using R0 by default
> with a jit_has_dedicated_arena_register() predicate to avoid the
> 64-immediate move overhead while centralizing the logic in verifier.c.
>
> Kumar responded that while this approach is reasonable for kfuncs, struct_ops
> would require uglier code with reloading/translating/storing. The v2 patch
> retained the JIT-based approach.
>

Yes, it was discussed offlist, we'll move ahead with JIT-based approach.

> [...]

[PATCH bpf-next v2 4/9] bpf, x86: Convert struct_ops arena arguments in the trampoline

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

Implement the struct_ops arena argument conversion on x86. save_args()
gains the conversion map from bpf_tramp_collect_arena_args() and, as it
copies each native argument into the BPF ctx, routes a marked slot
through RAX:

  movl %esrc, %eax      /* truncate and clear the upper 32 bits */
  subl $base_lo, %eax
  movq %rax, ctx_slot

A nullable slot tests the full 64-bit kernel pointer first:

  movq  %rsrc, %rax
  testq %rax, %rax
  jz    1f
  subl  $base_lo, %eax
1:
  movq  %rax, ctx_slot

The 32-bit subtraction is sufficient since (u32)(kaddr - base) ==
(u32)kaddr - (u32)base, and it clears the upper half as the JITs require
of arena pointer registers. Stack-passed arguments already reload
through RAX, so only the subtraction (and the NULL test) is inserted
there. The marked slots are tracked with a running slot counter shared
by the register and stack branches, matching the flattened ctx offsets
in ctx_arg_info. The size probe reruns the same emission with the same
tnodes, so the image size matches by construction.

With both the kfunc and struct_ops directions implemented, flip
bpf_jit_supports_arena_args() on for x86.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 arch/x86/net/bpf_jit_comp.c | 64 +++++++++++++++++++++++++++++++++----
 1 file changed, 57 insertions(+), 7 deletions(-)

diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
index 08013c2fcfaa..dfb5335ad837 100644
--- a/arch/x86/net/bpf_jit_comp.c
+++ b/arch/x86/net/bpf_jit_comp.c
@@ -3043,12 +3043,39 @@ static int get_nr_used_regs(const struct btf_func_model *m)
 	return nr_used_regs;
 }
 
+/*
+ * Convert an arena kernel address into the arena pointer form on its way
+ * into the BPF ctx, rax = (u32)(src - kern_vm_start). A nullable arg
+ * preserves NULL, tested on the full 64-bit kernel pointer. The 32-bit
+ * subtraction both truncates and clears the upper half, so the stored
+ * value satisfies the JIT invariant for arena pointer registers.
+ */
+static void emit_arena_arg_conv(u8 **pprog, u32 src_reg, bool nullable, u32 base_lo)
+{
+	u8 *prog = *pprog;
+
+	if (nullable) {
+		if (src_reg != BPF_REG_0)
+			emit_mov_reg(&prog, true, BPF_REG_0, src_reg);
+		/* test rax, rax; jz over the 5-byte sub */
+		EMIT3(0x48, 0x85, 0xC0);
+		EMIT2(X86_JE, 5);
+	} else if (src_reg != BPF_REG_0) {
+		emit_mov_reg(&prog, false, BPF_REG_0, src_reg);
+	}
+	/* sub eax, base_lo */
+	EMIT1_off32(0x2D, base_lo);
+
+	*pprog = prog;
+}
+
 static void save_args(const struct btf_func_model *m, u8 **prog,
-		      int stack_size, bool for_call_origin, u32 flags)
+		      int stack_size, bool for_call_origin, u32 flags,
+		      const struct bpf_tramp_arena_args *aargs)
 {
 	int arg_regs, first_off = 0, nr_regs = 0, nr_stack_slots = 0;
 	bool use_jmp = bpf_trampoline_use_jmp(flags);
-	int i, j;
+	int i, j, slot = 0;
 
 	/* Store function arguments to stack.
 	 * For a function that accepts two pointers the sequence will be:
@@ -3089,6 +3116,10 @@ static void save_args(const struct btf_func_model *m, u8 **prog,
 			for (j = 0; j < arg_regs; j++) {
 				emit_ldx(prog, BPF_DW, BPF_REG_0, BPF_REG_FP,
 					 nr_stack_slots * 8 + 16 + (!use_jmp) * 8);
+				if (aargs && (aargs->slots & BIT(slot)))
+					emit_arena_arg_conv(prog, BPF_REG_0,
+							    aargs->nullable_slots & BIT(slot),
+							    (u32)aargs->kern_vm_start);
 				emit_stx(prog, BPF_DW, BPF_REG_FP, BPF_REG_0,
 					 -stack_size);
 
@@ -3096,6 +3127,7 @@ static void save_args(const struct btf_func_model *m, u8 **prog,
 					first_off = stack_size;
 				stack_size -= 8;
 				nr_stack_slots++;
+				slot++;
 			}
 		} else {
 			/* Only copy the arguments on-stack to current
@@ -3104,16 +3136,24 @@ static void save_args(const struct btf_func_model *m, u8 **prog,
 			 */
 			if (for_call_origin) {
 				nr_regs += arg_regs;
+				slot += arg_regs;
 				continue;
 			}
 
 			/* copy the arguments from regs into stack */
 			for (j = 0; j < arg_regs; j++) {
-				emit_stx(prog, BPF_DW, BPF_REG_FP,
-					 nr_regs == 5 ? X86_REG_R9 : BPF_REG_1 + nr_regs,
-					 -stack_size);
+				u32 src = nr_regs == 5 ? X86_REG_R9 : BPF_REG_1 + nr_regs;
+
+				if (aargs && (aargs->slots & BIT(slot))) {
+					emit_arena_arg_conv(prog, src,
+							    aargs->nullable_slots & BIT(slot),
+							    (u32)aargs->kern_vm_start);
+					src = BPF_REG_0;
+				}
+				emit_stx(prog, BPF_DW, BPF_REG_FP, src, -stack_size);
 				stack_size -= 8;
 				nr_regs++;
+				slot++;
 			}
 		}
 	}
@@ -3404,11 +3444,13 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *rw_im
 	struct bpf_tramp_nodes *fentry = &tnodes[BPF_TRAMP_FENTRY];
 	struct bpf_tramp_nodes *fexit = &tnodes[BPF_TRAMP_FEXIT];
 	struct bpf_tramp_nodes *fmod_ret = &tnodes[BPF_TRAMP_MODIFY_RETURN];
+	struct bpf_tramp_arena_args aargs;
 	void *orig_call = func_addr;
 	int cookie_off, cookie_cnt;
 	u8 **branches = NULL;
 	u64 func_meta;
 	u8 *prog;
+	bool has_aargs;
 	bool save_ret;
 
 	/*
@@ -3419,6 +3461,8 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *rw_im
 	WARN_ON_ONCE((flags & BPF_TRAMP_F_INDIRECT) &&
 		     (flags & ~(BPF_TRAMP_F_INDIRECT | BPF_TRAMP_F_RET_FENTRY_RET)));
 
+	has_aargs = bpf_tramp_collect_arena_args(tnodes, flags, &aargs);
+
 	/* extra registers for struct arguments */
 	for (i = 0; i < m->nr_args; i++) {
 		if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG)
@@ -3556,7 +3600,8 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *rw_im
 		emit_store_stack_imm64(&prog, BPF_REG_0, -ip_off, (long)func_addr);
 	}
 
-	save_args(m, &prog, regs_off, false, flags);
+	save_args(m, &prog, regs_off, false, flags,
+		  has_aargs ? &aargs : NULL);
 
 	if (flags & BPF_TRAMP_F_CALL_ORIG) {
 		/* arg1: mov rdi, im */
@@ -3598,7 +3643,7 @@ static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *rw_im
 
 	if (flags & BPF_TRAMP_F_CALL_ORIG) {
 		restore_regs(m, &prog, regs_off);
-		save_args(m, &prog, arg_stack_off, true, flags);
+		save_args(m, &prog, arg_stack_off, true, flags, NULL);
 
 		if (flags & BPF_TRAMP_F_TAIL_CALL_CTX) {
 			/* Before calling the original function, load the
@@ -4099,6 +4144,11 @@ bool bpf_jit_supports_stack_args(void)
 	return true;
 }
 
+bool bpf_jit_supports_arena_args(void)
+{
+	return true;
+}
+
 void *bpf_arch_text_copy(void *dst, void *src, size_t len)
 {
 	if (text_poke_copy(dst, src, len) == NULL)
-- 
2.53.0

[PATCH bpf-next v2 5/9] selftests/bpf: Add kfunc __arena and __arena_nullable argument tests

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

Add arena-argument kfuncs to bpf_testmod, which also exercises the
argument rebasing on module kfuncs, and tests covering the accepted
argument forms (arena pointer, low 32 bits as a scalar, full user
address as a scalar), the exact rebase semantics via capture kfuncs
returning the raw argument (zero low 32 bits arrive as the arena kernel
base under __arena and as NULL under __arena_nullable), a NULL round
trip through a nullable deref kfunc, five arena arguments in one call, a
mixed __arena plus __arena_nullable call exercising both bitmasks on one
call site, a kernel-side dereference of an unpopulated page recovering
through the scratch page, and the rejections (no arena in the program,
incompatible register type).

The tests run on x86-64 and skip elsewhere, as programs with
arena-tagged kfunc args fail verification where the JIT lacks support.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 .../selftests/bpf/prog_tests/arena_kfunc.c    |  15 ++
 .../testing/selftests/bpf/progs/arena_kfunc.c | 250 ++++++++++++++++++
 .../selftests/bpf/test_kmods/bpf_testmod.c    |  57 ++++
 .../bpf/test_kmods/bpf_testmod_kfunc.h        |  10 +
 4 files changed, 332 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/arena_kfunc.c
 create mode 100644 tools/testing/selftests/bpf/progs/arena_kfunc.c

diff --git a/tools/testing/selftests/bpf/prog_tests/arena_kfunc.c b/tools/testing/selftests/bpf/prog_tests/arena_kfunc.c
new file mode 100644
index 000000000000..9be19565f0d5
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/arena_kfunc.c
@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
+#include <test_progs.h>
+
+#include "arena_kfunc.skel.h"
+
+/*
+ * The test kfuncs live in bpf_testmod. Resolving kfuncs against module
+ * BTFs needs CAP_SYS_ADMIN, so run with full capabilities instead of
+ * through the verifier tests' capability-restricted runner.
+ */
+void test_arena_kfunc(void)
+{
+	RUN_TESTS(arena_kfunc);
+}
diff --git a/tools/testing/selftests/bpf/progs/arena_kfunc.c b/tools/testing/selftests/bpf/progs/arena_kfunc.c
new file mode 100644
index 000000000000..e7250c5197ab
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/arena_kfunc.c
@@ -0,0 +1,250 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
+
+#define BPF_NO_KFUNC_PROTOTYPES
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+#include "bpf_misc.h"
+#include "bpf_experimental.h"
+#include <bpf_arena_common.h>
+#include "../test_kmods/bpf_testmod_kfunc.h"
+
+struct {
+	__uint(type, BPF_MAP_TYPE_ARENA);
+	__uint(map_flags, BPF_F_MMAPABLE);
+	/* page 0 hosts the arena global, page 1 is for allocations */
+	__uint(max_entries, 2);
+} arena SEC(".maps");
+
+/*
+ * Occupies page 0 so no allocation lands at arena offset 0, which the
+ * nullable tests below must be able to tell apart from NULL.
+ */
+u64 __arena arena_pad;
+
+/* volatile to force the scalar reloads below */
+volatile u64 stash;
+
+SEC("syscall")
+__arch_x86_64
+__success __retval(0)
+int arena_arg_forms(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	u64 __arena *val;
+	u64 ret;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	/* PTR_TO_ARENA argument */
+	*val = 41;
+	ret = bpf_kfunc_arena_arg_test((u64 *)val);
+	if (ret != 41 || *val != 42)
+		return 2;
+
+	/* the low 32 bits as a scalar */
+	stash = (u32)(u64)val;
+	ret = bpf_kfunc_arena_arg_test((u64 *)stash);
+	if (ret != 42 || *val != 43)
+		return 3;
+
+	/* the full user address as a scalar */
+	stash = (u64)val;
+	bpf_addr_space_cast(stash, 1, 0);
+	ret = bpf_kfunc_arena_arg_test((u64 *)stash);
+	if (ret != 43 || *val != 44)
+		return 4;
+
+	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
+#endif
+	return 0;
+}
+
+/*
+ * Pin the rebase semantics using the capture kfuncs, which return the raw
+ * argument value: __arena rebases unconditionally, so zero low 32 bits
+ * arrive as the arena kernel base, while __arena_nullable turns them into
+ * NULL.
+ */
+SEC("syscall")
+__arch_x86_64
+__success __retval(0)
+int arena_arg_rebase(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	u64 __arena *val;
+	u64 base, off;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	base = bpf_kfunc_arena_cap_test(NULL);
+	if (!base)
+		return 2;
+
+	/* only the low 32 bits contribute */
+	stash = 0xbadc0ffe00000000;
+	if (bpf_kfunc_arena_cap_test((u64 *)stash) != base)
+		return 3;
+
+	off = (u32)(u64)val;
+	if (bpf_kfunc_arena_cap_test((u64 *)val) != base + off)
+		return 4;
+
+	if (bpf_kfunc_arena_cap_nullable_test(NULL) != 0)
+		return 5;
+
+	stash = 0xbadc0ffe00000000;
+	if (bpf_kfunc_arena_cap_nullable_test((u64 *)stash) != 0)
+		return 6;
+
+	if (bpf_kfunc_arena_cap_nullable_test((u64 *)val) != base + off)
+		return 7;
+
+	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
+#endif
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__success __retval(0)
+int arena_arg_nullable(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	u64 __arena *val;
+	u64 ret;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	*val = 41;
+	ret = bpf_kfunc_arena_nullable_arg_test((u64 *)val);
+	if (ret != 41 || *val != 42)
+		return 2;
+
+	if (bpf_kfunc_arena_nullable_arg_test(NULL) != 0xdeadbeef)
+		return 3;
+
+	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
+#endif
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__success __retval(0)
+int arena_args5(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	u64 __arena *val;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	val[0] = 1;
+	val[1] = 2;
+	val[2] = 4;
+	val[3] = 8;
+	val[4] = 16;
+
+	if (bpf_kfunc_arena_args5_test((u64 *)&val[0], (u64 *)&val[1],
+				       (u64 *)&val[2], (u64 *)&val[3],
+				       (u64 *)&val[4]) != 31)
+		return 2;
+	if (bpf_kfunc_arena_args5_test((u64 *)&val[0], (u64 *)&val[1],
+				       (u64 *)&val[2], (u64 *)&val[3], NULL) != 15)
+		return 3;
+
+	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
+#endif
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__success __retval(0)
+int arena_arg_mixed(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	u64 __arena *val;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	val[0] = 7;
+	val[1] = 5;
+
+	if (bpf_kfunc_arena_mixed_test((u64 *)&val[0], NULL) != 7)
+		return 2;
+
+	if (bpf_kfunc_arena_mixed_test((u64 *)&val[0], (u64 *)&val[1]) != 12)
+		return 3;
+
+	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
+#endif
+	return 0;
+}
+
+/* kernel-side faults on unpopulated pages recover via the scratch page */
+SEC("syscall")
+__arch_x86_64
+__success __retval(0)
+int arena_arg_unpopulated(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	u64 __arena *val;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	stash = (u64)val + PAGE_SIZE;
+	bpf_kfunc_arena_arg_test((u64 *)stash);
+
+	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
+#endif
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__failure __msg("arena pointer requires a program with an associated arena")
+int arena_arg_no_arena(void *ctx)
+{
+	bpf_kfunc_arena_arg_test((u64 *)1);
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__failure __msg("is not a pointer to arena or scalar")
+int arena_arg_bad_reg(void *ctx)
+{
+	u64 buf = 0;
+
+	/* use the arena so the program passes the arena presence check */
+	bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	bpf_kfunc_arena_arg_test(&buf);
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__failure __msg("arena pointer cannot be a stack argument")
+int arena_arg_stack(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	bpf_kfunc_arena_stack_arg_test(1, 2, 3, 4, 5, (u64 *)1);
+#endif
+	return 0;
+}
+
+char _license[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
index 30f1cd23093c..fad106b1549c 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
@@ -210,6 +210,56 @@ __bpf_kfunc void bpf_kfunc_common_test(void)
 {
 }
 
+__bpf_kfunc u64 bpf_kfunc_arena_arg_test(u64 *val__arena)
+{
+	u64 old;
+
+	old = *val__arena;
+	*val__arena = old + 1;
+	return old;
+}
+
+__bpf_kfunc u64 bpf_kfunc_arena_nullable_arg_test(u64 *val__arena_nullable)
+{
+	u64 old;
+
+	if (!val__arena_nullable)
+		return 0xdeadbeef;
+
+	old = *val__arena_nullable;
+	*val__arena_nullable = old + 1;
+	return old;
+}
+
+__bpf_kfunc u64 bpf_kfunc_arena_cap_test(u64 *val__arena)
+{
+	return (u64)val__arena;
+}
+
+__bpf_kfunc u64 bpf_kfunc_arena_cap_nullable_test(u64 *val__arena_nullable)
+{
+	return (u64)val__arena_nullable;
+}
+
+__bpf_kfunc u64 bpf_kfunc_arena_args5_test(u64 *a__arena, u64 *b__arena,
+					   u64 *c__arena, u64 *d__arena,
+					   u64 *e__arena_nullable)
+{
+	return *a__arena + *b__arena + *c__arena + *d__arena +
+	       (e__arena_nullable ? *e__arena_nullable : 0);
+}
+
+__bpf_kfunc u64 bpf_kfunc_arena_stack_arg_test(u64 a, u64 b, u64 c, u64 d, u64 e,
+						u64 *f__arena)
+{
+	return a + b + c + d + e + *f__arena;
+}
+
+__bpf_kfunc u64 bpf_kfunc_arena_mixed_test(u64 *a__arena, u64 *b__arena_nullable)
+{
+	return *a__arena + (b__arena_nullable ? *b__arena_nullable : 0);
+}
+
 __bpf_kfunc void bpf_kfunc_dynptr_test(struct bpf_dynptr *ptr,
 				       struct bpf_dynptr *ptr__nullable)
 {
@@ -723,6 +773,13 @@ BTF_ID_FLAGS(func, bpf_iter_testmod_seq_next, KF_ITER_NEXT | KF_RET_NULL)
 BTF_ID_FLAGS(func, bpf_iter_testmod_seq_destroy, KF_ITER_DESTROY)
 BTF_ID_FLAGS(func, bpf_iter_testmod_seq_value)
 BTF_ID_FLAGS(func, bpf_kfunc_common_test)
+BTF_ID_FLAGS(func, bpf_kfunc_arena_arg_test)
+BTF_ID_FLAGS(func, bpf_kfunc_arena_nullable_arg_test)
+BTF_ID_FLAGS(func, bpf_kfunc_arena_cap_test)
+BTF_ID_FLAGS(func, bpf_kfunc_arena_cap_nullable_test)
+BTF_ID_FLAGS(func, bpf_kfunc_arena_args5_test)
+BTF_ID_FLAGS(func, bpf_kfunc_arena_stack_arg_test)
+BTF_ID_FLAGS(func, bpf_kfunc_arena_mixed_test)
 BTF_ID_FLAGS(func, bpf_kfunc_call_test_mem_len_pass1)
 BTF_ID_FLAGS(func, bpf_kfunc_dynptr_test)
 BTF_ID_FLAGS(func, bpf_kfunc_nested_acquire_nonzero_offset_test, KF_ACQUIRE)
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
index c36bb911defa..42d3fccc6ff9 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
@@ -98,6 +98,16 @@ void bpf_kfunc_call_test_release(struct prog_test_ref_kfunc *p) __ksym;
 void bpf_kfunc_call_test_ref(struct prog_test_ref_kfunc *p) __ksym;
 
 void bpf_kfunc_call_test_mem_len_pass1(void *mem, int len) __ksym;
+__u64 bpf_kfunc_arena_arg_test(__u64 *val__arena) __ksym;
+__u64 bpf_kfunc_arena_nullable_arg_test(__u64 *val__arena_nullable) __ksym;
+__u64 bpf_kfunc_arena_cap_test(__u64 *val__arena) __ksym;
+__u64 bpf_kfunc_arena_cap_nullable_test(__u64 *val__arena_nullable) __ksym;
+__u64 bpf_kfunc_arena_args5_test(__u64 *a__arena, __u64 *b__arena,
+				 __u64 *c__arena, __u64 *d__arena,
+				 __u64 *e__arena_nullable) __ksym;
+__u64 bpf_kfunc_arena_stack_arg_test(__u64 a, __u64 b, __u64 c, __u64 d, __u64 e,
+				     __u64 *f__arena) __ksym;
+__u64 bpf_kfunc_arena_mixed_test(__u64 *a__arena, __u64 *b__arena_nullable) __ksym;
 int *bpf_kfunc_call_test_get_rdwr_mem(struct prog_test_ref_kfunc *p, const int rdwr_buf_size) __ksym;
 int *bpf_kfunc_call_test_get_rdonly_mem(struct prog_test_ref_kfunc *p, const int rdonly_buf_size) __ksym;
 int *bpf_kfunc_call_test_acq_rdonly_mem(struct prog_test_ref_kfunc *p, const int rdonly_buf_size) __ksym;
-- 
2.53.0

Re: [PATCH bpf-next v2 5/9] selftests/bpf: Add kfunc __arena and __arena_nullable argument tests

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The `arena_arg_stack` failure test is conditionally compiled out on compilers missing `__BPF_FEATURE_ADDR_SPACE_CAST`, causing a false test suite failure.
--

--- Patch [5]: [PATCH bpf-next v2 5/9] selftests/bpf: Add kfunc __arena and __arena_nullable argument tests ---
commit b7abeb2e41481276f90bb3144448d08e3c6e2c96
Author: Tejun Heo <tj@kernel.org>

selftests/bpf: Add kfunc __arena and __arena_nullable argument tests

This commit adds kfunc arena argument tests to bpf_testmod to exercise
argument rebasing, various accepted argument forms, and verifier rejections.

> diff --git a/tools/testing/selftests/bpf/progs/arena_kfunc.c b/tools/testing/selftests/bpf/progs/arena_kfunc.c
> new file mode 100644
> index 0000000000000..e7250c5197ab4
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/progs/arena_kfunc.c

[ ... ]

> +SEC("syscall")
> +__arch_x86_64
> +__failure __msg("arena pointer cannot be a stack argument")
> +int arena_arg_stack(void *ctx)
> +{
> +#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
> +	bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
> +	bpf_kfunc_arena_stack_arg_test(1, 2, 3, 4, 5, (u64 *)1);
> +#endif
> +	return 0;
> +}

[Severity: Medium]
Does this conditional compilation cause the test suite to fail on older
compilers missing __BPF_FEATURE_ADDR_SPACE_CAST?

Because this test is marked with __failure, the test runner expects the
verifier to reject it. If the macro is not defined, the program compiles
to just return 0. The verifier will successfully load the empty program,
causing the test runner to report a failure because the expected rejection
did not occur.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260726013105.3689867-1-memxor@gmail.com?part=5

Re: [PATCH bpf-next v2 5/9] selftests/bpf: Add kfunc __arena and __arena_nullable argument tests

[Kumar Kartikeya Dwivedi]

On Sun Jul 26, 2026 at 3:41 AM CEST, sashiko-bot wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] The `arena_arg_stack` failure test is conditionally compiled out on compilers missing `__BPF_FEATURE_ADDR_SPACE_CAST`, causing a false test suite failure.
> --
>
> --- Patch [5]: [PATCH bpf-next v2 5/9] selftests/bpf: Add kfunc __arena and __arena_nullable argument tests ---
> commit b7abeb2e41481276f90bb3144448d08e3c6e2c96
> Author: Tejun Heo <tj@kernel.org>
>
> selftests/bpf: Add kfunc __arena and __arena_nullable argument tests
>
> This commit adds kfunc arena argument tests to bpf_testmod to exercise
> argument rebasing, various accepted argument forms, and verifier rejections.
>
>> diff --git a/tools/testing/selftests/bpf/progs/arena_kfunc.c b/tools/testing/selftests/bpf/progs/arena_kfunc.c
>> new file mode 100644
>> index 0000000000000..e7250c5197ab4
>> --- /dev/null
>> +++ b/tools/testing/selftests/bpf/progs/arena_kfunc.c
>
> [ ... ]
>
>> +SEC("syscall")
>> +__arch_x86_64
>> +__failure __msg("arena pointer cannot be a stack argument")
>> +int arena_arg_stack(void *ctx)
>> +{
>> +#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
>> +	bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
>> +	bpf_kfunc_arena_stack_arg_test(1, 2, 3, 4, 5, (u64 *)1);
>> +#endif
>> +	return 0;
>> +}
>
> [Severity: Medium]
> Does this conditional compilation cause the test suite to fail on older
> compilers missing __BPF_FEATURE_ADDR_SPACE_CAST?
>
> Because this test is marked with __failure, the test runner expects the
> verifier to reject it. If the macro is not defined, the program compiles
> to just return 0. The verifier will successfully load the empty program,
> causing the test runner to report a failure because the expected rejection
> did not occur.

I think the goal is to not fail compilation, this is fine as a tradeoff.

[PATCH bpf-next v2 6/9] selftests/bpf: Add JIT-sequence tests for __arena kfunc arguments

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

Pin the exact rebase sequences the JITs emit for __arena and
__arena_nullable kfunc arguments with __jited assertions on x86-64: the
unconditional truncate-and-add, the nullable test-and-skip variant, and
all five argument registers in one call, which also covers the
REX-prefixed encoding of r8 on x86. The capture kfuncs take the argument
without dereferencing, so only the emitted code is under test.  The
tests skip without LLVM disassembler support.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 .../bpf/prog_tests/arena_kfunc_jit.c          | 13 +++
 .../selftests/bpf/progs/arena_kfunc_jit.c     | 96 +++++++++++++++++++
 2 files changed, 109 insertions(+)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c
 create mode 100644 tools/testing/selftests/bpf/progs/arena_kfunc_jit.c

diff --git a/tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c b/tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c
new file mode 100644
index 000000000000..2359cde24c45
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c
@@ -0,0 +1,13 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
+#include <test_progs.h>
+#include "arena_kfunc_jit.skel.h"
+
+/*
+ * Runs with full capabilities: resolving module kfunc ksyms requires
+ * CAP_SYS_ADMIN, which rules out the capability-restricted runner.
+ */
+void test_arena_kfunc_jit(void)
+{
+	RUN_TESTS(arena_kfunc_jit);
+}
diff --git a/tools/testing/selftests/bpf/progs/arena_kfunc_jit.c b/tools/testing/selftests/bpf/progs/arena_kfunc_jit.c
new file mode 100644
index 000000000000..84747611725f
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/arena_kfunc_jit.c
@@ -0,0 +1,96 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
+
+/*
+ * Verify the JIT-emitted rebase sequences for __arena and __arena_nullable
+ * kfunc arguments. The capture kfuncs take the argument without
+ * dereferencing it, so these tests pin only the emitted code.
+ */
+#define BPF_NO_KFUNC_PROTOTYPES
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+#include "bpf_misc.h"
+#include "bpf_experimental.h"
+#include <bpf_arena_common.h>
+#include "../test_kmods/bpf_testmod_kfunc.h"
+
+struct {
+	__uint(type, BPF_MAP_TYPE_ARENA);
+	__uint(map_flags, BPF_F_MMAPABLE);
+	__uint(max_entries, 1);
+} arena SEC(".maps");
+
+/* volatile to force the scalar reloads below */
+volatile u64 stash;
+
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+
+SEC("syscall")
+__arch_x86_64
+__jited("...")
+__jited("	movl	%edi, %edi")
+__jited("	addq	%r12, %rdi")
+__jited("...")
+__jited("	callq	{{.*}}")
+__success
+int arena_arg_jit_rebase(void *ctx)
+{
+	stash = (u64)bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	bpf_kfunc_arena_cap_test((u64 *)stash);
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__jited("...")
+__jited("	movl	%edi, %edi")
+__jited("	testl	%edi, %edi")
+__jited("	je	{{.*}}")
+__jited("	addq	%r12, %rdi")
+__success
+int arena_arg_jit_nullable(void *ctx)
+{
+	stash = (u64)bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	bpf_kfunc_arena_cap_nullable_test((u64 *)stash);
+	return 0;
+}
+
+SEC("syscall")
+__arch_x86_64
+__jited("...")
+__jited("	movl	%edi, %edi")
+__jited("	addq	%r12, %rdi")
+__jited("	movl	%esi, %esi")
+__jited("	addq	%r12, %rsi")
+__jited("	movl	%edx, %edx")
+__jited("	addq	%r12, %rdx")
+__jited("	movl	%ecx, %ecx")
+__jited("	addq	%r12, %rcx")
+__jited("	movl	%r8d, %r8d")
+__jited("	testl	%r8d, %r8d")
+__jited("	je	{{.*}}")
+__jited("	addq	%r12, %r8")
+__success
+int arena_arg_jit_args5(void *ctx)
+{
+	u64 __arena *val;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	val[0] = 1;
+	val[1] = 2;
+	val[2] = 4;
+	val[3] = 8;
+	val[4] = 16;
+
+	bpf_kfunc_arena_args5_test((u64 *)&val[0], (u64 *)&val[1],
+				   (u64 *)&val[2], (u64 *)&val[3],
+				   (u64 *)&val[4]);
+	return 0;
+}
+
+#endif /* __BPF_FEATURE_ADDR_SPACE_CAST */
+
+char _license[] SEC("license") = "GPL";
-- 
2.53.0

Re: [PATCH bpf-next v2 6/9] selftests/bpf: Add JIT-sequence tests for __arena kfunc arguments

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Missing test dependency check for `bpf_testmod.ko` in `test_arena_kfunc_jit()`.
--

--- Patch [6]: [PATCH bpf-next v2 6/9] selftests/bpf: Add JIT-sequence tests for __arena kfunc arguments ---
commit 127efe77c3180fef987aeb8dfccbc04666cf9935
Author: Tejun Heo <tj@kernel.org>

selftests/bpf: Add JIT-sequence tests for __arena kfunc arguments

This commit adds JIT-sequence tests for __arena and __arena_nullable
kfunc arguments on x86-64. It pins the exact rebase sequences the JITs
emit to verify correct code generation.

> diff --git a/tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c b/tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c
> new file mode 100644
> index 0000000000000..2359cde24c45d
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/prog_tests/arena_kfunc_jit.c
> @@ -0,0 +1,13 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
> +#include <test_progs.h>
> +#include "arena_kfunc_jit.skel.h"
> +
> +/*
> + * Runs with full capabilities: resolving module kfunc ksyms requires
> + * CAP_SYS_ADMIN, which rules out the capability-restricted runner.
> + */
> +void test_arena_kfunc_jit(void)
> +{
> +	RUN_TESTS(arena_kfunc_jit);
> +}

[Severity: Medium]
Since this test relies on kfuncs from bpf_testmod, should there be a check
for env.has_testmod before running the tests?

If the tests are run in an environment where bpf_testmod cannot be loaded,
this will fail with a verifier kfunc resolution error instead of skipping
gracefully.

For example, adding:

    if (!env.has_testmod) {
        test__skip();
        return;
    }

would allow the test runner to skip gracefully.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260726013105.3689867-1-memxor@gmail.com?part=6

[PATCH bpf-next v2 7/9] selftests/bpf: Add struct_ops __arena and __arena_nullable argument tests

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

Add test_arena and test_arena_nullable members to bpf_testmod_ops3 with
arena-tagged stub arguments and kfuncs that forward a caller-provided
pointer to them. The kfuncs take arena-tagged arguments, so each round
trip exercises both conversion directions end to end: the kfunc receives
a kernel arena address and the trampoline converts it back to an arena
pointer for the callback.

The non-nullable callback dereferences its argument with no NULL branch
and captures the raw ctx value, which the trigger program compares
against the arena offset of the passed object, pinning the exact
(u32)(kaddr - kern_vm_start) conversion. The nullable callback verifies
that only a true kernel NULL arrives as NULL. Failure coverage: a
program with no arena is rejected when it loads. The tests run on x86-64
and skip elsewhere, as the programs fail verification where the JIT
lacks arena argument support.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 .../bpf/prog_tests/test_struct_ops_arena.c    | 74 +++++++++++++++
 .../testing/selftests/bpf/progs/arena_kfunc.c | 14 ++-
 .../selftests/bpf/progs/struct_ops_arena.c    | 94 +++++++++++++++++++
 .../bpf/progs/struct_ops_arena_fail.c         | 20 ++++
 .../selftests/bpf/test_kmods/bpf_testmod.c    | 24 +++++
 .../selftests/bpf/test_kmods/bpf_testmod.h    |  3 +
 .../bpf/test_kmods/bpf_testmod_kfunc.h        |  2 +
 7 files changed, 229 insertions(+), 2 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c
 create mode 100644 tools/testing/selftests/bpf/progs/struct_ops_arena.c
 create mode 100644 tools/testing/selftests/bpf/progs/struct_ops_arena_fail.c

diff --git a/tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c
new file mode 100644
index 000000000000..b1e39defa38a
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_arena.c
@@ -0,0 +1,74 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
+#include <test_progs.h>
+
+#include "struct_ops_arena.skel.h"
+#include "struct_ops_arena_fail.skel.h"
+
+#if defined(__x86_64__)
+/*
+ * Attach callbacks with __arena and __arena_nullable arguments and drive
+ * them through the bpf_testmod_ops3_call_test_arena*() kfuncs.
+ */
+static void arena_arg(void)
+{
+	LIBBPF_OPTS(bpf_test_run_opts, topts);
+	struct struct_ops_arena *skel;
+	struct bpf_link *link = NULL;
+	int err;
+
+	skel = struct_ops_arena__open_and_load();
+	if (!ASSERT_OK_PTR(skel, "struct_ops_arena__open_and_load"))
+		return;
+
+	link = bpf_map__attach_struct_ops(skel->maps.testmod_arena);
+	if (!ASSERT_OK_PTR(link, "attach_struct_ops"))
+		goto out;
+
+	err = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.trigger),
+				     &topts);
+	ASSERT_OK(err, "test_run");
+	ASSERT_EQ(topts.retval, 0, "trigger_retval");
+
+out:
+	bpf_link__destroy(link);
+	struct_ops_arena__destroy(skel);
+}
+
+/*
+ * A program with no arena cannot attach to a member with an __arena
+ * argument.
+ */
+static void arena_arg_fail(void)
+{
+	struct struct_ops_arena_fail *skel;
+
+	skel = struct_ops_arena_fail__open_and_load();
+	if (ASSERT_ERR_PTR(skel, "struct_ops_arena_fail__open_and_load"))
+		return;
+
+	struct_ops_arena_fail__destroy(skel);
+}
+#endif
+
+/*
+ * Serialized because it attaches the singleton bpf_testmod_ops3, which
+ * test_struct_ops_private_stack also attaches; registering it twice fails
+ * with -EEXIST.
+ */
+void serial_test_struct_ops_arena(void)
+{
+	/*
+	 * Arena struct_ops arguments need JIT support, currently x86-64 only.
+	 * Elsewhere verification fails with "JIT does not support arena
+	 * arguments", so the programs cannot even load.
+	 */
+#if defined(__x86_64__)
+	if (test__start_subtest("arena_arg"))
+		arena_arg();
+	if (test__start_subtest("arena_arg_fail"))
+		arena_arg_fail();
+#else
+	test__skip();
+#endif
+}
diff --git a/tools/testing/selftests/bpf/progs/arena_kfunc.c b/tools/testing/selftests/bpf/progs/arena_kfunc.c
index e7250c5197ab..15d48151797d 100644
--- a/tools/testing/selftests/bpf/progs/arena_kfunc.c
+++ b/tools/testing/selftests/bpf/progs/arena_kfunc.c
@@ -235,16 +235,26 @@ int arena_arg_bad_reg(void *ctx)
 	return 0;
 }
 
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST) && \
+	defined(__BPF_FEATURE_STACK_ARGUMENT)
 SEC("syscall")
 __arch_x86_64
 __failure __msg("arena pointer cannot be a stack argument")
 int arena_arg_stack(void *ctx)
 {
-#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
 	bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
 	bpf_kfunc_arena_stack_arg_test(1, 2, 3, 4, 5, (u64 *)1);
-#endif
 	return 0;
 }
+#else
+SEC("syscall")
+__arch_x86_64
+__description("arena_arg_stack: not supported, dummy test")
+__success
+int arena_arg_stack(void *ctx)
+{
+	return 0;
+}
+#endif
 
 char _license[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/progs/struct_ops_arena.c b/tools/testing/selftests/bpf/progs/struct_ops_arena.c
new file mode 100644
index 000000000000..40c856a748d2
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/struct_ops_arena.c
@@ -0,0 +1,94 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
+
+#define BPF_NO_KFUNC_PROTOTYPES
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+#include "bpf_experimental.h"
+#include <bpf_arena_common.h>
+#include "../test_kmods/bpf_testmod.h"
+#include "../test_kmods/bpf_testmod_kfunc.h"
+
+char _license[] SEC("license") = "GPL";
+
+struct {
+	__uint(type, BPF_MAP_TYPE_ARENA);
+	__uint(map_flags, BPF_F_MMAPABLE);
+	/* page 0 hosts the arena globals, page 1 is for allocations */
+	__uint(max_entries, 2);
+} arena SEC(".maps");
+
+/* also associates the callbacks with the arena */
+u64 __arena arena_touch;
+/* raw value of the last __arena ctx argument, captured by test_arena_cb */
+u64 __arena cb_ptr_val;
+
+SEC("struct_ops/test_arena")
+int test_arena_cb(unsigned long long *ctx)
+{
+	u64 __arena *ptr = (u64 __arena *)ctx[0];
+
+	arena_touch++;
+	cb_ptr_val = ctx[0];
+	*ptr += 1;
+	return 0;
+}
+
+SEC("struct_ops/test_arena_nullable")
+int test_arena_nullable_cb(unsigned long long *ctx)
+{
+	u64 __arena *ptr = (u64 __arena *)ctx[0];
+
+	arena_touch++;
+	if (!ptr)
+		return 0xbee;
+	*ptr += 1;
+	return 0;
+}
+
+SEC(".struct_ops.link")
+struct bpf_testmod_ops3 testmod_arena = {
+	.test_arena = (void *)test_arena_cb,
+	.test_arena_nullable = (void *)test_arena_nullable_cb,
+};
+
+SEC("syscall")
+int trigger(void *ctx)
+{
+#if defined(__BPF_FEATURE_ADDR_SPACE_CAST)
+	u64 __arena *val;
+	int ret;
+
+	val = bpf_arena_alloc_pages(&arena, NULL, 1, NUMA_NO_NODE, 0);
+	if (!val)
+		return 1;
+
+	*val = 41;
+	ret = bpf_testmod_ops3_call_test_arena((u64 *)val);
+	if (ret)
+		return 2;
+	if (*val != 42)
+		return 3;
+
+	/*
+	 * The callback must have seen exactly (u32)(kaddr - kern_vm_start),
+	 * which is the arena offset of val with the upper 32 bits clear.
+	 */
+	if (cb_ptr_val != (u32)(u64)val)
+		return 4;
+
+	ret = bpf_testmod_ops3_call_test_arena_nullable((u64 *)val);
+	if (ret)
+		return 5;
+	if (*val != 43)
+		return 6;
+
+	/* NULL survives the nullable kfunc and the trampoline as NULL */
+	ret = bpf_testmod_ops3_call_test_arena_nullable(NULL);
+	if (ret != 0xbee)
+		return 7;
+
+	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
+#endif
+	return 0;
+}
diff --git a/tools/testing/selftests/bpf/progs/struct_ops_arena_fail.c b/tools/testing/selftests/bpf/progs/struct_ops_arena_fail.c
new file mode 100644
index 000000000000..1c0ec727d637
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/struct_ops_arena_fail.c
@@ -0,0 +1,20 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2026 Meta Platforms, Inc. and affiliates. */
+
+#include <vmlinux.h>
+#include <bpf/bpf_helpers.h>
+#include "../test_kmods/bpf_testmod.h"
+
+char _license[] SEC("license") = "GPL";
+
+/* No arena in the program: attaching to test_arena must be rejected. */
+SEC("struct_ops/test_arena")
+int test_arena_no_arena(unsigned long long *ctx)
+{
+	return 0;
+}
+
+SEC(".struct_ops.link")
+struct bpf_testmod_ops3 testmod_arena_fail = {
+	.test_arena = (void *)test_arena_no_arena,
+};
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
index fad106b1549c..48ce36a6b82f 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
@@ -370,9 +370,21 @@ static int bpf_testmod_test_4(void)
 	return 0;
 }
 
+static int bpf_testmod_ops3__test_arena(u64 *ptr__arena)
+{
+	return 0;
+}
+
+static int bpf_testmod_ops3__test_arena_nullable(u64 *ptr__arena_nullable)
+{
+	return 0;
+}
+
 static struct bpf_testmod_ops3 __bpf_testmod_ops3 = {
 	.test_1 = bpf_testmod_test_3,
 	.test_2 = bpf_testmod_test_4,
+	.test_arena = bpf_testmod_ops3__test_arena,
+	.test_arena_nullable = bpf_testmod_ops3__test_arena_nullable,
 };
 
 static void bpf_testmod_test_struct_ops3(void)
@@ -391,6 +403,16 @@ __bpf_kfunc void bpf_testmod_ops3_call_test_2(void)
 	st_ops3->test_2();
 }
 
+__bpf_kfunc int bpf_testmod_ops3_call_test_arena(u64 *ptr__arena)
+{
+	return st_ops3->test_arena(ptr__arena);
+}
+
+__bpf_kfunc int bpf_testmod_ops3_call_test_arena_nullable(u64 *ptr__arena_nullable)
+{
+	return st_ops3->test_arena_nullable(ptr__arena_nullable);
+}
+
 struct bpf_testmod_btf_type_tag_1 {
 	int a;
 };
@@ -795,6 +817,8 @@ BTF_ID_FLAGS(func, bpf_testmod_ctx_create, KF_ACQUIRE | KF_RET_NULL)
 BTF_ID_FLAGS(func, bpf_testmod_ctx_release, KF_RELEASE)
 BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_1)
 BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_2)
+BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_arena)
+BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_arena_nullable)
 BTF_ID_FLAGS(func, bpf_kfunc_get_default_trusted_ptr_test);
 BTF_ID_FLAGS(func, bpf_kfunc_put_default_trusted_ptr_test);
 BTF_KFUNCS_END(bpf_testmod_common_kfunc_ids)
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h
index 863fd10f1619..c367ec856776 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h
@@ -106,6 +106,9 @@ struct bpf_testmod_ops2 {
 struct bpf_testmod_ops3 {
 	int (*test_1)(void);
 	int (*test_2)(void);
+	/* Used to test arena pointer arguments. */
+	int (*test_arena)(u64 *ptr);
+	int (*test_arena_nullable)(u64 *ptr);
 };
 
 struct st_ops_args {
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
index 42d3fccc6ff9..ff0d3894d7af 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
@@ -121,6 +121,8 @@ u32 bpf_kfunc_call_test_static_unused_arg(u32 arg, u32 unused) __ksym;
 #endif
 
 void bpf_testmod_test_mod_kfunc(int i) __ksym;
+int bpf_testmod_ops3_call_test_arena(__u64 *ptr__arena) __ksym;
+int bpf_testmod_ops3_call_test_arena_nullable(__u64 *ptr__arena_nullable) __ksym;
 
 __u64 bpf_kfunc_call_test1(struct sock *sk, __u32 a, __u64 b,
 				__u32 c, __u64 d) __ksym;
-- 
2.53.0

[PATCH bpf-next v2 8/9] bpf, x86: Fix stack-passed arguments for indirect trampolines

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

save_args() reads stack-passed arguments relative to rbp assuming two
return addresses sit between the saved rbp and the arguments, which
holds when the trampoline is entered through the fentry call from a
traced function. An indirect trampoline is called through a function
pointer, so only the caller's return address is on the stack and the
arguments start at rbp + 16, not rbp + 24. Every stack-passed argument
of a struct_ops callback with more than six argument slots is read one
slot off.

This has gone unnoticed because no in-tree struct_ops member passes
arguments on the stack. The jmp-entry form already accounts for having
a single return address; treat BPF_TRAMP_F_INDIRECT the same way.

Fixes: 473e3150e30a ("bpf, x86: allow function arguments up to 12 for TRACING")
Cc: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 arch/x86/net/bpf_jit_comp.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
index dfb5335ad837..4b349ae35ebf 100644
--- a/arch/x86/net/bpf_jit_comp.c
+++ b/arch/x86/net/bpf_jit_comp.c
@@ -3075,6 +3075,7 @@ static void save_args(const struct btf_func_model *m, u8 **prog,
 {
 	int arg_regs, first_off = 0, nr_regs = 0, nr_stack_slots = 0;
 	bool use_jmp = bpf_trampoline_use_jmp(flags);
+	int stack_args_off = (use_jmp || (flags & BPF_TRAMP_F_INDIRECT)) ? 16 : 24;
 	int i, j, slot = 0;
 
 	/* Store function arguments to stack.
@@ -3106,16 +3107,16 @@ static void save_args(const struct btf_func_model *m, u8 **prog,
 			/* copy function arguments from origin stack frame
 			 * into current stack frame.
 			 *
-			 * The starting address of the arguments on-stack
-			 * is:
-			 *   rbp + 8(push rbp) +
-			 *   8(return addr of origin call) +
-			 *   8(return addr of the caller)
-			 * which means: rbp + 24
+			 * The arguments on-stack start above the saved rbp
+			 * and the return addresses: two return addresses
+			 * (origin call and caller) when the trampoline is
+			 * entered through the fentry call, so rbp + 24, and
+			 * a single one when it is entered with a jmp or
+			 * called indirectly, so rbp + 16.
 			 */
 			for (j = 0; j < arg_regs; j++) {
 				emit_ldx(prog, BPF_DW, BPF_REG_0, BPF_REG_FP,
-					 nr_stack_slots * 8 + 16 + (!use_jmp) * 8);
+					 nr_stack_slots * 8 + stack_args_off);
 				if (aargs && (aargs->slots & BIT(slot)))
 					emit_arena_arg_conv(prog, BPF_REG_0,
 							    aargs->nullable_slots & BIT(slot),
-- 
2.53.0

Re: [PATCH bpf-next v2 8/9] bpf, x86: Fix stack-passed arguments for indirect trampolines

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The patch fixes the stack-passed argument offset for indirect struct_ops trampolines on x86, but leaves the exact same logical bug unfixed in the ARM64, RISC-V, and LoongArch BPF JIT compilers.
--

--- Patch [8]: [PATCH bpf-next v2 8/9] bpf, x86: Fix stack-passed arguments for indirect trampolines ---
commit 1625042935fc87b38acf83c8304433545d672f35
Author: Tejun Heo <tj@kernel.org>

bpf, x86: Fix stack-passed arguments for indirect trampolines

This patch fixes the calculation of stack-passed arguments for indirect
trampolines in the x86 BPF JIT compiler. It introduces stack_args_off
to properly handle the number of return addresses on the stack depending
on how the trampoline was called.

> diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
> index dfb5335ad8374..4b349ae35ebf9 100644
> --- a/arch/x86/net/bpf_jit_comp.c
> +++ b/arch/x86/net/bpf_jit_comp.c
> @@ -3075,6 +3075,7 @@ static void save_args(const struct btf_func_model *m, u8 **prog,
>  {
>  	int arg_regs, first_off = 0, nr_regs = 0, nr_stack_slots = 0;
>  	bool use_jmp = bpf_trampoline_use_jmp(flags);
> +	int stack_args_off = (use_jmp || (flags & BPF_TRAMP_F_INDIRECT)) ? 16 : 24;

[Severity: High]
This is a pre-existing issue, but does this same logical bug need to be fixed
for ARM64, RISC-V, and LoongArch? 

Looking at arch/arm64/net/bpf_jit_comp.c:save_args(), there is a hardcoded
offset that unconditionally assumes two frames were pushed:

    soff = 32; /* on stack arguments start from FP + 32 */

Similarly, in arch/riscv/net/bpf_jit_comp64.c:store_args(), there is a
hardcoded offset of 16 which skips the first 16 bytes of caller arguments for
struct_ops:

    emit_ld(RV_REG_T1, 16 + (i - RV_MAX_REG_ARGS) * 8, RV_REG_FP, ctx);

And arch/loongarch/net/bpf_jit.c:store_args() shares the same hardcoded offset:

    emit_insn(ctx, ldd, LOONGARCH_GPR_T1, LOONGARCH_GPR_FP,
                  16 + (i - LOONGARCH_MAX_REG_ARGS) * 8);

If a struct_ops callback with more than 8 arguments (or 6 on some
architectures) is executed on these platforms, could it receive out-of-bounds
stack memory instead of the actual arguments?

>  	int i, j, slot = 0;
>  
>  	/* Store function arguments to stack.
[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260726013105.3689867-1-memxor@gmail.com?part=8

Re: [PATCH bpf-next v2 8/9] bpf, x86: Fix stack-passed arguments for indirect trampolines

[Kumar Kartikeya Dwivedi]

On Sun Jul 26, 2026 at 3:51 AM CEST, sashiko-bot wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
>
> Pre-existing issues:
> - [High] The patch fixes the stack-passed argument offset for indirect struct_ops trampolines on x86, but leaves the exact same logical bug unfixed in the ARM64, RISC-V, and LoongArch BPF JIT compilers.
> --
>
> --- Patch [8]: [PATCH bpf-next v2 8/9] bpf, x86: Fix stack-passed arguments for indirect trampolines ---
> commit 1625042935fc87b38acf83c8304433545d672f35
> Author: Tejun Heo <tj@kernel.org>
>
> bpf, x86: Fix stack-passed arguments for indirect trampolines
>
> This patch fixes the calculation of stack-passed arguments for indirect
> trampolines in the x86 BPF JIT compiler. It introduces stack_args_off
> to properly handle the number of return addresses on the stack depending
> on how the trampoline was called.
>
>> diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
>> index dfb5335ad8374..4b349ae35ebf9 100644
>> --- a/arch/x86/net/bpf_jit_comp.c
>> +++ b/arch/x86/net/bpf_jit_comp.c
>> @@ -3075,6 +3075,7 @@ static void save_args(const struct btf_func_model *m, u8 **prog,
>>  {
>>  	int arg_regs, first_off = 0, nr_regs = 0, nr_stack_slots = 0;
>>  	bool use_jmp = bpf_trampoline_use_jmp(flags);
>> +	int stack_args_off = (use_jmp || (flags & BPF_TRAMP_F_INDIRECT)) ? 16 : 24;
>
> [Severity: High]
> This is a pre-existing issue, but does this same logical bug need to be fixed
> for ARM64, RISC-V, and LoongArch?
>
> Looking at arch/arm64/net/bpf_jit_comp.c:save_args(), there is a hardcoded
> offset that unconditionally assumes two frames were pushed:
>
>     soff = 32; /* on stack arguments start from FP + 32 */
>
> Similarly, in arch/riscv/net/bpf_jit_comp64.c:store_args(), there is a
> hardcoded offset of 16 which skips the first 16 bytes of caller arguments for
> struct_ops:
>
>     emit_ld(RV_REG_T1, 16 + (i - RV_MAX_REG_ARGS) * 8, RV_REG_FP, ctx);
>
> And arch/loongarch/net/bpf_jit.c:store_args() shares the same hardcoded offset:
>
>     emit_insn(ctx, ldd, LOONGARCH_GPR_T1, LOONGARCH_GPR_FP,
>                   16 + (i - LOONGARCH_MAX_REG_ARGS) * 8);
>
> If a struct_ops callback with more than 8 arguments (or 6 on some
> architectures) is executed on these platforms, could it receive out-of-bounds
> stack memory instead of the actual arguments?
>

Will be followed up on separately for each arch.

>>  	int i, j, slot = 0;
>>
>>  	/* Store function arguments to stack.
> [ ... ]

[PATCH bpf-next v2 9/9] selftests/bpf: Test stack-passed struct_ops arena arguments

[Kumar Kartikeya Dwivedi]

From: Tejun Heo <tj@kernel.org>

Add a test_arena_stack member with eight leading scalar arguments so the
arena pointer is passed on the stack.

The callback validates the first and last scalar ctx slots before
dereferencing the pointer in ctx[8]. This exercises the indirect
trampoline stack layout and arena conversion together, and prevents a
regression where stack arguments are read one slot late.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
---
 .../selftests/bpf/progs/struct_ops_arena.c    | 21 +++++++++++++++++++
 .../selftests/bpf/test_kmods/bpf_testmod.c    | 14 +++++++++++++
 .../selftests/bpf/test_kmods/bpf_testmod.h    |  3 +++
 .../bpf/test_kmods/bpf_testmod_kfunc.h        |  1 +
 4 files changed, 39 insertions(+)

diff --git a/tools/testing/selftests/bpf/progs/struct_ops_arena.c b/tools/testing/selftests/bpf/progs/struct_ops_arena.c
index 40c856a748d2..ba04c73d8d96 100644
--- a/tools/testing/selftests/bpf/progs/struct_ops_arena.c
+++ b/tools/testing/selftests/bpf/progs/struct_ops_arena.c
@@ -46,10 +46,24 @@ int test_arena_nullable_cb(unsigned long long *ctx)
 	return 0;
 }
 
+SEC("struct_ops/test_arena_stack")
+int test_arena_stack_cb(unsigned long long *ctx)
+{
+	u64 __arena *ptr = (u64 __arena *)ctx[8];
+
+	arena_touch++;
+	/* pin the slot layout: the leading args fill ctx[0]..ctx[7] */
+	if (ctx[0] != 1 || ctx[7] != 8)
+		return 0xbad;
+	*ptr += 1;
+	return 0;
+}
+
 SEC(".struct_ops.link")
 struct bpf_testmod_ops3 testmod_arena = {
 	.test_arena = (void *)test_arena_cb,
 	.test_arena_nullable = (void *)test_arena_nullable_cb,
+	.test_arena_stack = (void *)test_arena_stack_cb,
 };
 
 SEC("syscall")
@@ -88,6 +102,13 @@ int trigger(void *ctx)
 	if (ret != 0xbee)
 		return 7;
 
+	/* the arena pointer is stack-passed into the trampoline here */
+	ret = bpf_testmod_ops3_call_test_arena_stack((u64 *)val);
+	if (ret)
+		return 8;
+	if (*val != 44)
+		return 9;
+
 	bpf_arena_free_pages(&arena, (void __arena *)val, 1);
 #endif
 	return 0;
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
index 48ce36a6b82f..9e268fd69a93 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.c
@@ -380,11 +380,19 @@ static int bpf_testmod_ops3__test_arena_nullable(u64 *ptr__arena_nullable)
 	return 0;
 }
 
+static int bpf_testmod_ops3__test_arena_stack(u64 a, u64 b, u64 c, u64 d,
+					      u64 e, u64 f, u64 g, u64 h,
+					      u64 *ptr__arena)
+{
+	return 0;
+}
+
 static struct bpf_testmod_ops3 __bpf_testmod_ops3 = {
 	.test_1 = bpf_testmod_test_3,
 	.test_2 = bpf_testmod_test_4,
 	.test_arena = bpf_testmod_ops3__test_arena,
 	.test_arena_nullable = bpf_testmod_ops3__test_arena_nullable,
+	.test_arena_stack = bpf_testmod_ops3__test_arena_stack,
 };
 
 static void bpf_testmod_test_struct_ops3(void)
@@ -413,6 +421,11 @@ __bpf_kfunc int bpf_testmod_ops3_call_test_arena_nullable(u64 *ptr__arena_nullab
 	return st_ops3->test_arena_nullable(ptr__arena_nullable);
 }
 
+__bpf_kfunc int bpf_testmod_ops3_call_test_arena_stack(u64 *ptr__arena)
+{
+	return st_ops3->test_arena_stack(1, 2, 3, 4, 5, 6, 7, 8, ptr__arena);
+}
+
 struct bpf_testmod_btf_type_tag_1 {
 	int a;
 };
@@ -819,6 +832,7 @@ BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_1)
 BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_2)
 BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_arena)
 BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_arena_nullable)
+BTF_ID_FLAGS(func, bpf_testmod_ops3_call_test_arena_stack)
 BTF_ID_FLAGS(func, bpf_kfunc_get_default_trusted_ptr_test);
 BTF_ID_FLAGS(func, bpf_kfunc_put_default_trusted_ptr_test);
 BTF_KFUNCS_END(bpf_testmod_common_kfunc_ids)
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h
index c367ec856776..33f2af5b7085 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod.h
@@ -109,6 +109,9 @@ struct bpf_testmod_ops3 {
 	/* Used to test arena pointer arguments. */
 	int (*test_arena)(u64 *ptr);
 	int (*test_arena_nullable)(u64 *ptr);
+	/* enough leading args to force @ptr onto the stack on x86 and arm64 */
+	int (*test_arena_stack)(u64 a, u64 b, u64 c, u64 d, u64 e, u64 f,
+				u64 g, u64 h, u64 *ptr);
 };
 
 struct st_ops_args {
diff --git a/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
index ff0d3894d7af..1a72d0fda53c 100644
--- a/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
+++ b/tools/testing/selftests/bpf/test_kmods/bpf_testmod_kfunc.h
@@ -123,6 +123,7 @@ u32 bpf_kfunc_call_test_static_unused_arg(u32 arg, u32 unused) __ksym;
 void bpf_testmod_test_mod_kfunc(int i) __ksym;
 int bpf_testmod_ops3_call_test_arena(__u64 *ptr__arena) __ksym;
 int bpf_testmod_ops3_call_test_arena_nullable(__u64 *ptr__arena_nullable) __ksym;
+int bpf_testmod_ops3_call_test_arena_stack(__u64 *ptr__arena) __ksym;
 
 __u64 bpf_kfunc_call_test1(struct sock *sk, __u32 a, __u64 b,
 				__u32 c, __u64 d) __ksym;
-- 
2.53.0