Skip to contentSkip to conversation

bpf: Fix off-by-one boundary validation in arena direct-value access

PATCH Superseded Junyoung Jang 6 participants 8 messages Apr 26 – May 9 1 message not in the archive BPF

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): 2026042614-scowling-hankie-f84b@gregkh

#1 lore raw

PATCH 1/1 bpf: Fix off-by-one boundary validation in arena direct-value access

BPF_MAP_TYPE_ARENA accepts BPF_PSEUDO_MAP_VALUE offsets at exactly the end of the arena mapping (off == arena_size). The boundary check in arena_map_direct_value_addr() uses > instead of >=, which incorrectly allows a one-past-end pointer to be accepted.

Change the condition to >= to correctly reject offsets that fall outside the valid arena user_vm range.

---
 kernel/bpf/arena.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/bpf/arena.c +1 -1
index 802656c6fd3c..49a8f7b1beef 100644
@@ -511,7 +511,7 @@ static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32
 {
 	struct bpf_arena *arena = container_of(map, struct bpf_arena, map);
 
-	if ((u64)off > arena->user_vm_end - arena->user_vm_start)
+	if ((u64)off >= arena->user_vm_end - arena->user_vm_start)
 		return -ERANGE;
 	*imm = (unsigned long)arena->user_vm_start;
 	return 0;
Signature · 1 line
2.43.0
#2 lore raw
This commit fixes a bug but lacks a Fixes: tag. The commit being fixed appears to be:
Quoted 1 line

= to correctly reject offsets outside the valid arena user_vm range.

This commit fixes a bug but lacks a Fixes: tag. The commit being fixed appears to be:

Fixes 5a2f726fe6d0 ("adding ci files")

Should this tag be added to help with backporting and tracking the original issue?

---
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/24960599139

#3 lore raw

PATCH v2 1/1 bpf: Fix off-by-one boundary validation in arena direct-value access

In reply to a message not in the archive

BPF_MAP_TYPE_ARENA accepts BPF_PSEUDO_MAP_VALUE offsets at exactly the end of the arena mapping (off == arena_size). The boundary check in arena_map_direct_value_addr() uses > instead of >=, which incorrectly allows a one-past-end pointer to be accepted.

Change the condition to >= to correctly reject offsets that fall outside the valid arena user_vm range.

Fixes 317460317a02 ("bpf: Introduce bpf_arena.")
---
 kernel/bpf/arena.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/bpf/arena.c +1 -1
index 802656c6fd3c..49a8f7b1beef 100644
@@ -511,7 +511,7 @@ static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32
 {
 	struct bpf_arena *arena = container_of(map, struct bpf_arena, map);
 
-	if ((u64)off > arena->user_vm_end - arena->user_vm_start)
+	if ((u64)off >= arena->user_vm_end - arena->user_vm_start)
 		return -ERANGE;
 	*imm = (unsigned long)arena->user_vm_start;
 	return 0;
Signature · 1 line
2.43.0
#4 lore raw
Sashiko AI review found 1 potential issue(s): - [Medium] Data race and bounds check bypass in arena_map_direct_value_addr() due to lockless reads --

Sashiko AI review found 1 potential issue(s):
- [Medium] Data race and bounds check bypass in arena_map_direct_value_addr() due to lockless reads --

commit a9b5d3016475abf10be02a320dc627c95b306d92
Author: Junyoung Jang <[email protected]>

bpf: Fix off-by-one boundary validation in arena direct-value access

This commit fixes an off-by-one error in the arena map boundary check where offsets exactly at the end of the mapping were incorrectly accepted.

kernel/bpf/arena.c static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32 · line 514

Quoted 10 lines
kernel/bpf/arena.c +1 -1
index 802656c6fd3c..49a8f7b1beef 100644
@@ -511,7 +511,7 @@ static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32
 {
 	struct bpf_arena *arena = container_of(map, struct bpf_arena, map);
 
-	if ((u64)off > arena->user_vm_end - arena->user_vm_start)
+	if ((u64)off >= arena->user_vm_end - arena->user_vm_start)

Could this evaluation be vulnerable to a data race?

This function reads arena->user_vm_end and arena->user_vm_start without holding arena->lock or using READ_ONCE().

Concurrently, a user-space process calling mmap() on the arena map FD can invoke arena_map_mmap(), which updates these fields under arena->lock. Because user_vm_start is updated first, a concurrent reader here might observe the new user_vm_start (a high address) and the old user_vm_end (which might be 0).

In that scenario, arena->user_vm_end - arena->user_vm_start would underflow to a large 64-bit value, which would allow arbitrary offsets to bypass this bounds check.

Does this logic require synchronization or READ_ONCE() annotations to ensure the verifier's boundary validation remains effective?

Quoted 4 lines
kernel/bpf/arena.c +0 -0
 		return -ERANGE;
 	*imm = (unsigned long)arena->user_vm_start;
 	return 0;

}

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
#5 lore raw
Quoted 34 lines · sashiko-bot

Sashiko AI review found 1 potential issue(s):
- [Medium] Data race and bounds check bypass in arena_map_direct_value_addr() due to lockless reads --

commit a9b5d3016475abf10be02a320dc627c95b306d92
Author: Junyoung Jang <[email protected]>

bpf: Fix off-by-one boundary validation in arena direct-value access

This commit fixes an off-by-one error in the arena map boundary check where offsets exactly at the end of the mapping were incorrectly accepted.

kernel/bpf/arena.c static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32 · line 514

Quoted 10 lines
kernel/bpf/arena.c +1 -1
index 802656c6fd3c..49a8f7b1beef 100644
@@ -511,7 +511,7 @@ static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32
 {
 	struct bpf_arena *arena = container_of(map, struct bpf_arena, map);
 
-	if ((u64)off > arena->user_vm_end - arena->user_vm_start)
+	if ((u64)off >= arena->user_vm_end - arena->user_vm_start)

Could this evaluation be vulnerable to a data race?

This function reads arena->user_vm_end and arena->user_vm_start without holding arena->lock or using READ_ONCE().

Concurrently, a user-space process calling mmap() on the arena map FD can invoke arena_map_mmap(), which updates these fields under arena->lock. Because user_vm_start is updated first, a concurrent reader here might observe the new user_vm_start (a high address) and the old user_vm_end (which might be 0).

Quoted 3 lines

In that scenario, arena->user_vm_end - arena->user_vm_start would underflow to a large 64-bit value, which would allow arbitrary offsets to bypass this bounds check.

This seems like a valid concern, though it is unrelated to the fix here which seems correct. Can you sent a followup patch using an arena lock guard in map_direct_value_addr? I don't see any issues with taking the lock since the method is only used at verification time.

Quoted 8 lines

Does this logic require synchronization or READ_ONCE() annotations to ensure the verifier's boundary validation remains effective?

Quoted 4 lines
kernel/bpf/arena.c +0 -0
 		return -ERANGE;
 	*imm = (unsigned long)arena->user_vm_start;
 	return 0;

}

#6 lore raw
Quoted 36 lines · Emil Tsalapatis
Quoted 34 lines · sashiko-bot

Sashiko AI review found 1 potential issue(s):
- [Medium] Data race and bounds check bypass in arena_map_direct_value_addr() due to lockless reads --

commit a9b5d3016475abf10be02a320dc627c95b306d92
Author: Junyoung Jang <[email protected]>

bpf: Fix off-by-one boundary validation in arena direct-value access

This commit fixes an off-by-one error in the arena map boundary check where offsets exactly at the end of the mapping were incorrectly accepted.

kernel/bpf/arena.c static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32 · line 514

Quoted 10 lines
kernel/bpf/arena.c +1 -1
index 802656c6fd3c..49a8f7b1beef 100644
@@ -511,7 +511,7 @@ static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32
 {
     struct bpf_arena *arena = container_of(map, struct bpf_arena, map);
 
-    if ((u64)off > arena->user_vm_end - arena->user_vm_start)
+    if ((u64)off >= arena->user_vm_end - arena->user_vm_start)

Could this evaluation be vulnerable to a data race?

This function reads arena->user_vm_end and arena->user_vm_start without holding arena->lock or using READ_ONCE().

Concurrently, a user-space process calling mmap() on the arena map FD can invoke arena_map_mmap(), which updates these fields under arena->lock. Because user_vm_start is updated first, a concurrent reader here might observe the new user_vm_start (a high address) and the old user_vm_end (which might be 0).

Quoted 8 lines
Quoted 3 lines

In that scenario, arena->user_vm_end - arena->user_vm_start would underflow to a large 64-bit value, which would allow arbitrary offsets to bypass this bounds check.

This seems like a valid concern, though it is unrelated to the fix here which seems correct. Can you sent a followup patch using an arena lock guard in map_direct_value_addr? I don't see any issues with taking the lock since the method is only used at verification time.

No. sashk is wrong here.
In general I suggest to ignore sashk's medium and low suggestions. 'critical' are often correct.
'high' should be considered, but acting on them needs thought.

reviewer Reviewed-by on v2 1/1
#7 lore raw
In reply to Junyoung Jang
Quoted 11 lines · Junyoung Jang

BPF_MAP_TYPE_ARENA accepts BPF_PSEUDO_MAP_VALUE offsets at exactly the end of the arena mapping (off == arena_size). The boundary check in arena_map_direct_value_addr() uses > instead of >=, which incorrectly allows a one-past-end pointer to be accepted.

Change the condition to >= to correctly reject offsets that fall outside the valid arena user_vm range.

Fixes 317460317a02 ("bpf: Introduce bpf_arena.")

---

Quoted 16 lines
 kernel/bpf/arena.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/bpf/arena.c +1 -1
index 802656c6fd3c..49a8f7b1beef 100644
@@ -511,7 +511,7 @@ static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32
 {
 	struct bpf_arena *arena = container_of(map, struct bpf_arena, map);
 
-	if ((u64)off > arena->user_vm_end - arena->user_vm_start)
+	if ((u64)off >= arena->user_vm_end - arena->user_vm_start)
 		return -ERANGE;
 	*imm = (unsigned long)arena->user_vm_start;
 	return 0;
#8 lore raw
In reply to Junyoung Jang
This patch was applied to bpf/bpf.git (master) by Alexei Starovoitov <[email protected]>:

Hello:

This patch was applied to bpf/bpf.git (master)
by Alexei Starovoitov <[email protected]>:

Quoted 9 lines · On Mon, 27 Apr 2026 02:25:05 +0900 you wrote:

BPF_MAP_TYPE_ARENA accepts BPF_PSEUDO_MAP_VALUE offsets at exactly the end of the arena mapping (off == arena_size). The boundary check in arena_map_direct_value_addr() uses > instead of >=, which incorrectly allows a one-past-end pointer to be accepted.

Change the condition to >= to correctly reject offsets that fall outside the valid arena user_vm range.

[...]

Here is the summary with links:
- [bpf,v2] bpf: Fix off-by-one boundary validation in arena direct-value access

    https://git.kernel.org/bpf/bpf/c/3ac1a467e376

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html