linux-kernel - Re: [PATCH] mm: fix use-after free of page

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

Date:   Wed, 20 Jul 2022 20:38:58 +0530
From:   Charan Teja Kalla <quic_charante@...cinc.com>
To:     Michal Hocko <mhocko@...e.com>
CC:     <akpm@...ux-foundation.org>, <pasha.tatashin@...een.com>,
        <sjpark@...zon.de>, <sieberf@...zon.com>, <shakeelb@...gle.com>,
        <dhowells@...hat.com>, <willy@...radead.org>, <vbabka@...e.cz>,
        <david@...hat.com>, <minchan@...nel.org>,
        <linux-kernel@...r.kernel.org>, <linux-mm@...ck.org>,
        "iamjoonsoo.kim@....com" <iamjoonsoo.kim@....com>,
        Pavan Kondeti <quic_pkondeti@...cinc.com>
Subject: Re: [PATCH] mm: fix use-after free of page_ext after race with
 memory-offline

Thanks Michal here!!

On 7/19/2022 9:13 PM, Michal Hocko wrote:
>>>>>> diff --git a/mm/page_ext.c b/mm/page_ext.c
>>>>>> index 3dc715d..5ccd3ee 100644
>>>>>> --- a/mm/page_ext.c
>>>>>> +++ b/mm/page_ext.c
>>>>>> @@ -299,8 +299,9 @@ static void __free_page_ext(unsigned long pfn)
>>>>>>  	if (!ms || !ms->page_ext)
>>>>>>  		return;
>>>>>>  	base = get_entry(ms->page_ext, pfn);
>>>>>> -	free_page_ext(base);
>>>>>>  	ms->page_ext = NULL;
>>>>>> +	synchronize_rcu();
>>>>>> +	free_page_ext(base);
>>>>>>  }
>>>>> So you are imposing the RCU grace period for each page_ext! This can get
>>>>> really expensive. Have you tried to measure the effect?
>>> I was wrong here! This is for each memory section which is not as
>>> terrible as every single page_ext. This can be still quite a lot memory
>>> sections in a single memory block (e.g. on ppc memory sections are
>>> ridiculously small).
>>>
>> On the ARM64, I see that the minimum a section size will go is 128MB. I
>> think 16MB is the section size on ppc. Any inputs on how frequently
>> offline/online operation is being done on this ppc arch?
> I have seen several reports where 16MB sections were used on PPC LPARs
> with a non trivial size. My usual answer to that is tha this is mostly a
> self inflicted injury but I am told that for some reasons I cannot
> udnerstand this is not easy to change. So reasonable or not this is not
> all that uncommon in PPC land.
> 
> We definitely shouldn't optimize for those setups but we shouldn't make
> them suffer even more as well. Besides that it seems that a single
> rcu_synchronize per offline operation should be doable.

I too feel it is doable but the code might look to need to traverse the
sections of a memory block twice.

1) with synchronize_rcu() calling for each memory section of a memblock:
---------------------------------------------------------------------
for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
    __free_page_ext(pfn):
         ms->page_ext = NULL
         synchronize_rcu();// Called on every section.
	 free_page_ext();//free the page_ext.

2) With a single synchronize_rcu() for each offlined block:
-------------------------------------------------------
for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION) {
	__free_page_ext(pfn):
	    ms->page_ext = NULL;
}
synchronize_rcu(); // call synchronize_rcu for just once
for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
	free_page_ext(); // Free the page_ext.

Any better code you have in mind here please?

But since there are few sections the overhead of traversing them will be
definitely less compared to synchronize_rcu() per memsection.

But I really doubt if there will be a real impact making sync_rcu per
section because,(as david also mentioned and you also corrected it I
think), the concern here is for ppc where the min memblock size is 256M
with section size of 16M and there is a single offline operation on that
block but can end up in calling 16 sync_rcu() calls. Should we really
optimize this case here? If yes, I can go with the approach 2) mentioned
above. Sorry if I am really undermining the problem here.


Thanks,
Charan