linux-kernel - Re: [PATCH v3] mm/mempolicy: fix mpol_new leak in shared_policy

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <8575a7af-eacf-db06-4e48-b825f37575db@huawei.com>
Date:   Tue, 29 Mar 2022 10:13:36 +0800
From:   Miaohe Lin <linmiaohe@...wei.com>
To:     Andrew Morton <akpm@...ux-foundation.org>
CC:     <mhocko@...e.com>, <kosaki.motohiro@...fujitsu.com>,
        <mgorman@...e.de>, <linux-mm@...ck.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v3] mm/mempolicy: fix mpol_new leak in
 shared_policy_replace

On 2022/3/29 5:26, Andrew Morton wrote:
> On Sat, 26 Mar 2022 14:46:28 +0800 Miaohe Lin <linmiaohe@...wei.com> wrote:
> 
>> On 2022/3/26 8:29, Andrew Morton wrote:
>>> On Tue, 22 Mar 2022 18:43:45 +0800 Miaohe Lin <linmiaohe@...wei.com> wrote:
>>>
>>>> If mpol_new is allocated but not used in restart loop, mpol_new will be
>>>> freed via mpol_put before returning to the caller.  But refcnt is not
>>>> initialized yet, so mpol_put could not do the right things and might leak
>>>> the unused mpol_new. This would happen if mempolicy was updated on the
>>>> shared shmem file while the sp->lock has been dropped during the memory
>>>> allocation.
>>>>
>>>> This issue could be triggered easily with the below code snippet if there
>>>> are many processes doing the below work at the same time:
>>>>
>>>>   shmid = shmget((key_t)5566, 1024 * PAGE_SIZE, 0666|IPC_CREAT);
>>>>   shm = shmat(shmid, 0, 0);
>>>>   loop many times {
>>>>     mbind(shm, 1024 * PAGE_SIZE, MPOL_LOCAL, mask, maxnode, 0);
>>>>     mbind(shm + 128 * PAGE_SIZE, 128 * PAGE_SIZE, MPOL_DEFAULT, mask,
>>>>           maxnode, 0);
>>>>   }
>>>>
>>>> ...
>>>>
>>>> --- a/mm/mempolicy.c
>>>> +++ b/mm/mempolicy.c
>>>> @@ -2733,6 +2733,7 @@ static int shared_policy_replace(struct shared_policy *sp, unsigned long start,
>>>>  	mpol_new = kmem_cache_alloc(policy_cache, GFP_KERNEL);
>>>>  	if (!mpol_new)
>>>>  		goto err_out;
>>>> +	refcount_set(&mpol_new->refcnt, 1);
>>>>  	goto restart;
>>>>  }
>>>
>>> Two other sites in this file do
>>>
>>> 	atomic_set(&policy->refcnt, 1);
>>>
>>>
>>> Could we please instead have a little helper function which does the
>>> kmem_cache_alloc()+refcount_set()?> .
>>
>> There are usecases like below:
>>
>> 	struct mempolicy *new = kmem_cache_alloc(policy_cache, GFP_KERNEL);
>> 	*new = *old;
>> 	^^^^^^^^^^^^
>> 	refcount_set(&new->refcnt, 1);
>>
>> If we use helper function to do kmem_cache_alloc()+refcount_set() above, separate
>> refcount_set(&new->refcnt, 1) is still needed as old is copied to new and overwrites
>> the refcnt field. So that little helper function might not work. Or am I miss something?
>>
> 
> Hm, spose so.  I guess the helper doesn't add much in that case.
> 
> Can we please redo this on mainline?  I'm not happy with the bloat
> which refcnt_t adds and I think I'll drop
> mm-mempolicy-convert-from-atomic_t-to-refcount_t-on-mempolicy-refcnt.patch.

Will do this soon. Many thanks.

> .
>