Skip to contentSkip to conversation

bpf: Check percpu map value size first

RFC PATCH Superseded Tao Chen 2 participants 3 messages Sep 5, 2024 – Sep 6, 2024 BPF

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Tao Chen

author
#0 lore raw

RFC PATCH 1/1 bpf: Check percpu map value size first

Percpu map is often used, but the map value size limit often ignored, like issue: https://github.com/iovisor/bcc/issues/2519. Actually, percpu map value size is bound by PCPU_MIN_UNIT_SZIE, so we can check the value size whether it exceeds PCPU_MIN_UNIT_SZIE first, like percpu map of local_storage. Maybe the error message seems clearer compared with "cannot allocate memory".

the test case we create a percpu map with large value like:

struct test_t {
        int a[100000];
};
struct {
        __uint(type, BPF_MAP_TYPE_PERCPU_ARRAY);
        __uint(max_entries, 1);
        __type(key, u32);
        __type(value, struct test_t);
} start SEC(".maps");

test on ubuntu24.04 vm:
libbpf: Error in bpf_create_map_xattr(start):Argument list too long(-7). Retrying without BTF.

---
 kernel/bpf/arraymap.c | 3 +++
 kernel/bpf/hashtab.c  | 3 +++
 2 files changed, 6 insertions(+)
kernel/bpf/arraymap.c +3 -0
index a43e62e2a8bb..7c3c32f156ff 100644
@@ -73,6 +73,9 @@ int array_map_alloc_check(union bpf_attr *attr)
 	/* avoid overflow on round_up(map->value_size) */
 	if (attr->value_size > INT_MAX)
 		return -E2BIG;
+	/* percpu map value size is bound by PCPU_MIN_UNIT_SIZE */
+	if (percpu && attr->value_size > PCPU_MIN_UNIT_SIZE)
+		return -E2BIG;
 
 	return 0;
 }
kernel/bpf/hashtab.c +3 -0
index 45c7195b65ba..16d590fe1ffb 100644
@@ -462,6 +462,9 @@ static int htab_map_alloc_check(union bpf_attr *attr)
 		 * kmalloc-able later in htab_map_update_elem()
 		 */
 		return -E2BIG;
+	/* percpu map value size is bound by PCPU_MIN_UNIT_SIZE */
+	if (percpu && attr->value_size > PCPU_MIN_UNIT_SIZE)
+		return -E2BIG;
 
 	return 0;
 }
Signature · 1 line
2.25.1
#1 lore raw

Hi,

Quoted 3 lines · Tao Chen

Percpu map is often used, but the map value size limit often ignored, like issue: https://github.com/iovisor/bcc/issues/2519. Actually, percpu map value size is bound by PCPU_MIN_UNIT_SZIE, so we

s/SZIE/SIZE

Quoted 1 line

can check the value size whether it exceeds PCPU_MIN_UNIT_SZIE first,

The same typo here.

Quoted 14 lines

like percpu map of local_storage. Maybe the error message seems clearer compared with "cannot allocate memory".

the test case we create a percpu map with large value like:

struct test_t {
        int a[100000];
};
struct {
        __uint(type, BPF_MAP_TYPE_PERCPU_ARRAY);
        __uint(max_entries, 1);
        __type(key, u32);
        __type(value, struct test_t);
} start SEC(".maps");
Quoted 3 lines

test on ubuntu24.04 vm:
libbpf: Error in bpf_create_map_xattr(start):Argument list too long(-7). Retrying without BTF.

Could you please convert it into a separated bpf selftest patch ?

Quoted 12 lines
---
 kernel/bpf/arraymap.c | 3 +++
 kernel/bpf/hashtab.c  | 3 +++
 2 files changed, 6 insertions(+)
kernel/bpf/arraymap.c +0 -0
index a43e62e2a8bb..7c3c32f156ff 100644

kernel/bpf/arraymap.c int array_map_alloc_check(union bpf_attr *attr) · line 81

Quoted 9 lines
kernel/bpf/arraymap.c +3 -0
@@ -73,6 +73,9 @@ int array_map_alloc_check(union bpf_attr *attr)
 	/* avoid overflow on round_up(map->value_size) */
 	if (attr->value_size > INT_MAX)
 		return -E2BIG;
+	/* percpu map value size is bound by PCPU_MIN_UNIT_SIZE */
+	if (percpu && attr->value_size > PCPU_MIN_UNIT_SIZE)
+		return -E2BIG;
 
 	return 0;
 }

Make sense. However because the map passes round_up(attr->value_size, 8) to bpf_map_alloc_percpu(). Is it more reasonable to check round_up(value_size) > PCPU_MIN_UNIT_SIZE instead ?

kernel/bpf/hashtab.c static int htab_map_alloc_check(union bpf_attr *attr) · line 468

Quoted 12 lines
kernel/bpf/hashtab.c +3 -0
index 45c7195b65ba..16d590fe1ffb 100644
@@ -462,6 +462,9 @@ static int htab_map_alloc_check(union bpf_attr *attr)
 		 * kmalloc-able later in htab_map_update_elem()
 		 */
 		return -E2BIG;
+	/* percpu map value size is bound by PCPU_MIN_UNIT_SIZE */
+	if (percpu && attr->value_size > PCPU_MIN_UNIT_SIZE)
+		return -E2BIG;
 

The percpu allocation logic is the same as the array map:
round_up(value_size, 8) is used.

Quoted 2 lines
kernel/bpf/hashtab.c +0 -0
 	return 0;
 }
author
#2 lore raw
Quoted 8 lines · Hou Tao

Hi,

Quoted 3 lines · Tao Chen

Percpu map is often used, but the map value size limit often ignored, like issue: https://github.com/iovisor/bcc/issues/2519. Actually, percpu map value size is bound by PCPU_MIN_UNIT_SZIE, so we

s/SZIE/SIZE

Hi Hou, thank you for your reply!
My bad, i will fix it in v2.

Quoted 17 lines
Quoted 1 line

can check the value size whether it exceeds PCPU_MIN_UNIT_SZIE first,

The same typo here.

Quoted 14 lines

like percpu map of local_storage. Maybe the error message seems clearer compared with "cannot allocate memory".

the test case we create a percpu map with large value like:

struct test_t {
         int a[100000];
};
struct {
         __uint(type, BPF_MAP_TYPE_PERCPU_ARRAY);
         __uint(max_entries, 1);
         __type(key, u32);
         __type(value, struct test_t);
} start SEC(".maps");
Quoted 5 lines
Quoted 3 lines

test on ubuntu24.04 vm:
libbpf: Error in bpf_create_map_xattr(start):Argument list too long(-7). Retrying without BTF.

Could you please convert it into a separated bpf selftest patch ?

No problem, i will add test case in v2.

Quoted 19 lines
Quoted 19 lines
---
  kernel/bpf/arraymap.c | 3 +++
  kernel/bpf/hashtab.c  | 3 +++
  2 files changed, 6 insertions(+)
kernel/bpf/arraymap.c +3 -0
index a43e62e2a8bb..7c3c32f156ff 100644
@@ -73,6 +73,9 @@ int array_map_alloc_check(union bpf_attr *attr)
  	/* avoid overflow on round_up(map->value_size) */
  	if (attr->value_size > INT_MAX)
  		return -E2BIG;
+	/* percpu map value size is bound by PCPU_MIN_UNIT_SIZE */
+	if (percpu && attr->value_size > PCPU_MIN_UNIT_SIZE)
+		return -E2BIG;
 
Quoted 4 lines

kernel/bpf/arraymap.c int array_map_alloc_check(union bpf_attr *attr) · line 81

Quoted 2 lines
kernel/bpf/arraymap.c +0 -0
  	return 0;
  }

Make sense. However because the map passes round_up(attr->value_size, 8)

Yeah, you are right, it seems better, i will add it in v2.

Quoted 17 lines

to bpf_map_alloc_percpu(). Is it more reasonable to check round_up(value_size) > PCPU_MIN_UNIT_SIZE instead ?

kernel/bpf/hashtab.c static int htab_map_alloc_check(union bpf_attr *attr) · line 468

Quoted 12 lines
kernel/bpf/hashtab.c +3 -0
index 45c7195b65ba..16d590fe1ffb 100644
@@ -462,6 +462,9 @@ static int htab_map_alloc_check(union bpf_attr *attr)
  		 * kmalloc-able later in htab_map_update_elem()
  		 */
  		return -E2BIG;
+	/* percpu map value size is bound by PCPU_MIN_UNIT_SIZE */
+	if (percpu && attr->value_size > PCPU_MIN_UNIT_SIZE)
+		return -E2BIG;
 

The percpu allocation logic is the same as the array map:
round_up(value_size, 8) is used.

ok.

Quoted 3 lines
Quoted 2 lines
kernel/bpf/hashtab.c +0 -0
  	return 0;
  }
Signature · 2 lines
Best Regards
Dylane Chen