Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 7 Aug 2017 08:15:14 -0400
From: Glenn Randers-Pehrson <glennrp@...il.com>
To: oss-security@...ts.openwall.com
Subject: Re: Cve issue discussion

Do memory-exhaustion bugs get a CVE?  Suppose an application is fooled
into requesting 2Gb of memory but then never uses it other than
attempting to read it, immediately hitting EOF, and cleaning up.

I'm addressing such a bug in libpng right now, in which the user
is sent a PNG file containing a tEXt chunk that claims to have a 2GB
length (but none of the 2GB data is included in the PNG).  On my
platform libpng deals with that almost instantaneously, but I think
some platforms (ASAN builds?) would actually allocate the memory
before proceeding to read the data.

Glenn


On Mon, Aug 7, 2017 at 5:47 AM, ne xo <nexo123@...look.kr> wrote:
> Hello,
>
> thank you for the reply!
>
> I chose the report at random.
>
> I'm sorry if I was offended to mention the report.
>
> Thanks.
> <http://aka.ms/weboutlook>
> ________________________________
> 보낸 사람: Agostino Sarubbo <ago@...too.org>
> 보낸 날짜: 2017년 8월 7일 월요일 오후 4:42:05
> 받는 사람: oss-security@...ts.openwall.com
> 제목: Re: [oss-security] Cve issue discussion
>
> On Monday 07 August 2017 01:03:53 ne xo wrote:
>> Hello,
>>
>>
>> I am curious about issuing CVEs.
>>
>> I can see that a "NULL pointer dereference" or a bug where the exploit has
>> not been verified also get a CVE.
>
>>
>> heap-overflows may or may not be exploitable.
>>
>>
>> It takes a lot of time to analyze the exploit and create the exploit code.
>>
>>
>> Is it right to be assigned a CVE only if it is exploitable?
>>
>>
>> Or do you think all bugs need to get a CVE?
>>
>>
>> Thanks.
>>
>> ---
>>
>> ref
>>
>> ---
>>
>> [1]http://www.openwall.com/lists/oss-security/2017/04/10/17 - NULL pointer
>> dereference
>> [2]http://www.openwall.com/lists/oss-security/2017/04/10/15 -
>> memory allocation failure
>
> Hi.
>
> Since you mentioned some issues reported by me, let me answer directly.
> For the first, it is an undefined behavior, so actually you don't see the
> crash.
> Nowadays, the undefined behavior issues do not get anymore a CVE.
>
>
> For the second, ASAN reports that the program want to use more that 64GB of
> ram to execute the process so ASAN hangs the process. In this case is up to
> the maintainer check whether there is a problem in the code or not, or it is
> expected. The better double-check would be verify what happens without ASAN.
>
> I'd like also to mention that MITRE assigns CVE after they analyze the
> reported issue, so if an issue does not deserve a CVE, MITRE probably won't
> assign accompanied by an explanation.
>
> --
> Agostino Sarubbo
> Gentoo Linux Developer

Powered by blists - more mailing lists

Your e-mail address:

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.