Re: [PATCH] crypto: using size_add() for kmalloc()

[Su Hui <suhui@nfschina.com>]

On 2025/4/21 20:05, Herbert Xu wrote:
> On Mon, Apr 21, 2025 at 01:51:06PM +0800, Su Hui wrote:
>> @@ -433,7 +434,7 @@ static inline struct aead_request *aead_request_alloc(struct crypto_aead *tfm,
>>   {
>>   	struct aead_request *req;
>>   
>> -	req = kmalloc(sizeof(*req) + crypto_aead_reqsize(tfm), gfp);
>> +	req = kmalloc(size_add(sizeof(*req), crypto_aead_reqsize(tfm)), gfp);
> This is just wrong.  You should fail the allocation altogether
> rather than proceeding with a length that is insufficient.
>
> However, reqsize shouldn't be anywhere near overflowing in the
> first place.  If you're truly worried about this, you should
> change the algorithm registration code to check whether reqsize
> is sane.
>
> And that needs to wait until the algorithms are fixed to not use
> dynamic reqsizes.
Got it, thanks for your explanation.
This patch (v1 and v2) is wrong.  Sorry for the noise again.

Su Hui