[PATCH bpf-next v2] bpf: reject kfunc calls that overflow insn->imm

[PATCH bpf-next v2] bpf: reject kfunc calls that overflow insn->imm

[Hou Tao]

Now kfunc call uses s32 to represent the offset between the address
of kfunc and __bpf_call_base, but it doesn't check whether or not
s32 will be overflowed, so add an extra checking to reject these
invalid kfunc calls.

Signed-off-by: Hou Tao <houtao1@huawei.com>
---
v2:
 * instead of checking the overflow in selftests, just reject
   these kfunc calls directly in verifier

v1: https://lore.kernel.org/bpf/20220206043107.18549-1-houtao1@huawei.com
---
 kernel/bpf/verifier.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index a39eedecc93a..fd836e64b701 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -1832,6 +1832,13 @@ static struct btf *find_kfunc_desc_btf(struct bpf_verifier_env *env,
 	return btf_vmlinux ?: ERR_PTR(-ENOENT);
 }
 
+static inline bool is_kfunc_call_imm_overflowed(unsigned long addr)
+{
+	unsigned long offset = BPF_CALL_IMM(addr);
+
+	return (unsigned long)(s32)offset != offset;
+}
+
 static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16 offset)
 {
 	const struct btf_type *func, *func_proto;
@@ -1925,6 +1932,12 @@ static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16 offset)
 		return -EINVAL;
 	}
 
+	if (is_kfunc_call_imm_overflowed(addr)) {
+		verbose(env, "address of kernel function %s is out of range\n",
+			func_name);
+		return -EINVAL;
+	}
+
 	desc = &tab->descs[tab->nr_descs++];
 	desc->func_id = func_id;
 	desc->imm = BPF_CALL_IMM(addr);
-- 
2.29.2

Re: [PATCH bpf-next v2] bpf: reject kfunc calls that overflow insn->imm

[Yonghong Song]

On 2/8/22 4:33 AM, Hou Tao wrote:
> Now kfunc call uses s32 to represent the offset between the address
> of kfunc and __bpf_call_base, but it doesn't check whether or not
> s32 will be overflowed, so add an extra checking to reject these
> invalid kfunc calls.
> 
> Signed-off-by: Hou Tao <houtao1@huawei.com>
> ---
> v2:
>   * instead of checking the overflow in selftests, just reject
>     these kfunc calls directly in verifier
> 
> v1: https://lore.kernel.org/bpf/20220206043107.18549-1-houtao1@huawei.com
> ---
>   kernel/bpf/verifier.c | 13 +++++++++++++
>   1 file changed, 13 insertions(+)
> 
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index a39eedecc93a..fd836e64b701 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -1832,6 +1832,13 @@ static struct btf *find_kfunc_desc_btf(struct bpf_verifier_env *env,
>   	return btf_vmlinux ?: ERR_PTR(-ENOENT);
>   }
>   
> +static inline bool is_kfunc_call_imm_overflowed(unsigned long addr)
> +{
> +	unsigned long offset = BPF_CALL_IMM(addr);
> +
> +	return (unsigned long)(s32)offset != offset;
> +}
> +
>   static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16 offset)
>   {
>   	const struct btf_type *func, *func_proto;
> @@ -1925,6 +1932,12 @@ static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16 offset)
>   		return -EINVAL;
>   	}
>   
> +	if (is_kfunc_call_imm_overflowed(addr)) {
> +		verbose(env, "address of kernel function %s is out of range\n",
> +			func_name);
> +		return -EINVAL;
> +	}
> +
>   	desc = &tab->descs[tab->nr_descs++];
>   	desc->func_id = func_id;
>   	desc->imm = BPF_CALL_IMM(addr);

Thanks, I would like to call BPF_CALL_IMM only once and keep checking 
overflow and setting desc->imm close to each other. How about the 
following not-compile-tested code

	unsigned long call_imm;

	...
	call_imm = BPF_CALL_IMM(addr);
	/* some comment here */
	if ((unsigned long)(s32)call_imm != call_imm) {
		verbose(env, ...);
		return -EINVAL;
	} else {
		desc->imm = call_imm;
	}

Re: [PATCH bpf-next v2] bpf: reject kfunc calls that overflow insn->imm

[Hou Tao]

Hi,

On 2/9/2022 12:57 AM, Yonghong Song wrote:
>
>
> On 2/8/22 4:33 AM, Hou Tao wrote:
>> Now kfunc call uses s32 to represent the offset between the address
>> of kfunc and __bpf_call_base, but it doesn't check whether or not
>> s32 will be overflowed, so add an extra checking to reject these
>> invalid kfunc calls.
>>
>> Signed-off-by: Hou Tao <houtao1@huawei.com>
>> ---
>> v2:
>>   * instead of checking the overflow in selftests, just reject
>>     these kfunc calls directly in verifier
>>
>> v1: https://lore.kernel.org/bpf/20220206043107.18549-1-houtao1@huawei.com
>> ---
>>   kernel/bpf/verifier.c | 13 +++++++++++++
>>   1 file changed, 13 insertions(+)
>>
>> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
>> index a39eedecc93a..fd836e64b701 100644
>> --- a/kernel/bpf/verifier.c
>> +++ b/kernel/bpf/verifier.c
>> @@ -1832,6 +1832,13 @@ static struct btf *find_kfunc_desc_btf(struct
>> bpf_verifier_env *env,
>>       return btf_vmlinux ?: ERR_PTR(-ENOENT);
>>   }
>>   +static inline bool is_kfunc_call_imm_overflowed(unsigned long addr)
>> +{
>> +    unsigned long offset = BPF_CALL_IMM(addr);
>> +
>> +    return (unsigned long)(s32)offset != offset;
>> +}
>> +
>>   static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16
>> offset)
>>   {
>>       const struct btf_type *func, *func_proto;
>> @@ -1925,6 +1932,12 @@ static int add_kfunc_call(struct bpf_verifier_env
>> *env, u32 func_id, s16 offset)
>>           return -EINVAL;
>>       }
>>   +    if (is_kfunc_call_imm_overflowed(addr)) {
>> +        verbose(env, "address of kernel function %s is out of range\n",
>> +            func_name);
>> +        return -EINVAL;
>> +    }
>> +
>>       desc = &tab->descs[tab->nr_descs++];
>>       desc->func_id = func_id;
>>       desc->imm = BPF_CALL_IMM(addr);
>
> Thanks, I would like to call BPF_CALL_IMM only once and keep checking overflow
> and setting desc->imm close to each other. How about the following
> not-compile-tested code
>
>     unsigned long call_imm;
>
>     ...
>     call_imm = BPF_CALL_IMM(addr);
>     /* some comment here */
>     if ((unsigned long)(s32)call_imm != call_imm) {
>         verbose(env, ...);
>         return -EINVAL;
>     } else {
>         desc->imm = call_imm;
>     }
call BPF_CALL_IMM once is OK for me. but I don't think the else branch is
unnecessary and it make the code
ugly. Can we just return directly when found that imm is overflowed ?

        call_imm = BPF_CALL_IMM(addr);
        /* Check whether or not the relative offset overflows desc->imm */
        if ((unsigned long)(s32)call_imm != call_imm) {
                verbose(env, "address of kernel function %s is out of range\n",
                        func_name);
                return -EINVAL;
        }

        desc = &tab->descs[tab->nr_descs++];
        desc->func_id = func_id;
        desc->imm = call_imm;




> .

Re: [PATCH bpf-next v2] bpf: reject kfunc calls that overflow insn->imm

[Yonghong Song]

On 2/8/22 10:20 PM, Hou Tao wrote:
> Hi,
> 
> On 2/9/2022 12:57 AM, Yonghong Song wrote:
>>
>>
>> On 2/8/22 4:33 AM, Hou Tao wrote:
>>> Now kfunc call uses s32 to represent the offset between the address
>>> of kfunc and __bpf_call_base, but it doesn't check whether or not
>>> s32 will be overflowed, so add an extra checking to reject these
>>> invalid kfunc calls.
>>>
>>> Signed-off-by: Hou Tao <houtao1@huawei.com>
>>> ---
>>> v2:
>>>    * instead of checking the overflow in selftests, just reject
>>>      these kfunc calls directly in verifier
>>>
>>> v1: https://lore.kernel.org/bpf/20220206043107.18549-1-houtao1@huawei.com
>>> ---
>>>    kernel/bpf/verifier.c | 13 +++++++++++++
>>>    1 file changed, 13 insertions(+)
>>>
>>> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
>>> index a39eedecc93a..fd836e64b701 100644
>>> --- a/kernel/bpf/verifier.c
>>> +++ b/kernel/bpf/verifier.c
>>> @@ -1832,6 +1832,13 @@ static struct btf *find_kfunc_desc_btf(struct
>>> bpf_verifier_env *env,
>>>        return btf_vmlinux ?: ERR_PTR(-ENOENT);
>>>    }
>>>    +static inline bool is_kfunc_call_imm_overflowed(unsigned long addr)
>>> +{
>>> +    unsigned long offset = BPF_CALL_IMM(addr);
>>> +
>>> +    return (unsigned long)(s32)offset != offset;
>>> +}
>>> +
>>>    static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16
>>> offset)
>>>    {
>>>        const struct btf_type *func, *func_proto;
>>> @@ -1925,6 +1932,12 @@ static int add_kfunc_call(struct bpf_verifier_env
>>> *env, u32 func_id, s16 offset)
>>>            return -EINVAL;
>>>        }
>>>    +    if (is_kfunc_call_imm_overflowed(addr)) {
>>> +        verbose(env, "address of kernel function %s is out of range\n",
>>> +            func_name);
>>> +        return -EINVAL;
>>> +    }
>>> +
>>>        desc = &tab->descs[tab->nr_descs++];
>>>        desc->func_id = func_id;
>>>        desc->imm = BPF_CALL_IMM(addr);
>>
>> Thanks, I would like to call BPF_CALL_IMM only once and keep checking overflow
>> and setting desc->imm close to each other. How about the following
>> not-compile-tested code
>>
>>      unsigned long call_imm;
>>
>>      ...
>>      call_imm = BPF_CALL_IMM(addr);
>>      /* some comment here */
>>      if ((unsigned long)(s32)call_imm != call_imm) {
>>          verbose(env, ...);
>>          return -EINVAL;
>>      } else {
>>          desc->imm = call_imm;
>>      }
> call BPF_CALL_IMM once is OK for me. but I don't think the else branch is
> unnecessary and it make the code
> ugly. Can we just return directly when found that imm is overflowed ?
> 
>          call_imm = BPF_CALL_IMM(addr);
>          /* Check whether or not the relative offset overflows desc->imm */
>          if ((unsigned long)(s32)call_imm != call_imm) {
>                  verbose(env, "address of kernel function %s is out of range\n",
>                          func_name);
>                  return -EINVAL;
>          }
> 
>          desc = &tab->descs[tab->nr_descs++];
>          desc->func_id = func_id;
>          desc->imm = call_imm;

Sure. Your above change looks good. My change is just
an illustration :-).

> 
> 
> 
> 
>> .
> 
>