linux-kernel - Re: [PATCH] kernel/irq/proc.c: set MAX_NAMELEN 11 instead of 10 which express the maximize size of "%d" or "%u".

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:	Sun, 06 Oct 2013 23:57:04 +0800
From:	Chen Gang <gang.chen@...anux.com>
To:	Richard Weinberger <richard@....at>
CC:	Richard Weinberger <richard.weinberger@...il.com>,
	Joe Perches <joe@...ches.com>,
	Al Viro <viro@...iv.linux.org.uk>,
	Thomas Gleixner <tglx@...utronix.de>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] kernel/irq/proc.c: set MAX_NAMELEN 11 instead of 10 which
 express the maximize size of "%d" or "%u".

On 10/06/2013 08:37 AM, Chen Gang wrote:
> On 10/06/2013 01:45 AM, Richard Weinberger wrote:
>> Am 05.10.2013 19:06, schrieb Chen Gang:
>>> On 10/06/2013 12:50 AM, Chen Gang wrote:
>>>> On 10/06/2013 12:08 AM, Richard Weinberger wrote:
>>>>> On Sat, Oct 5, 2013 at 5:46 PM, Chen Gang <gang.chen@...anux.com> wrote:
>>>>>> On 10/05/2013 11:41 PM, Joe Perches wrote:
>>>>>>> On Sat, 2013-10-05 at 23:19 +0800, Chen Gang wrote:
>>>>>>>> Theoretically, the maximize size of "%d" or "%u" is 11 (10 + '\0'), so
>>>>>>>> need set MAX_NAMELEN 11 instead of 10.
>>>>>>>
>>>>>>> %d can be negative.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> Oh, really, it is incorrect. Al Viro succeeds once.  :-(
>>>>>>
>>>>>> And I should send patch v2 for it.
>>>>>
>>>>> irq is in both register_irq_proc() and unregister_irq_proc() an unsigned int.
>>>>> Therefore %d makes not really sense. Both should use %u.
>>>>> IMHO sprintf() should also get replaced by snprintf() but that's a
>>>>> matter of taste.
>>>>>
>>>
>>> Oh, commonly, snprintf() are used for the string which can be truncated,
>>> and can not be used for the string which contents must not be truncated.
>>>
>>> In our case, the name string must be not truncated (or may not unique,
>>> theoretically), so we have to still use sprintf().
>>
>> Of course you would have to check the return value of snprintf() to detect
>> a truncation and abort...
>>
> 
> OK, thanks, that sounds reasonable to me, so I feel that's not a matter
> of taste.
> 

Hmm... does 'taste' means: "for string which can not be truncated, if
maximized length is obvious (e.g. our case 11 for "%u"), sprintf() is
more clearer than snprintf() for both readers and writers"?


BTW: when MAX_NAMELEN is defined as 128, is it suitable to check the
return value of snprintf() in register_handler_proc(), and WARN_ON() if
it is not less than 128?


Thanks.

> In my opinion, when we know the maximized length, we need always use
> s(c)nprintf instead of sprintf, if the string can be truncated, use
> scnprintf, else use snprintf and also check the return value.
> 
> sprintf is 'dangrous', need try to use s(c)nprintf instead of. We can
> scan whole kernel, I guess quite a few of sprintf may be related with
> memory overflow 'theoretically' (welcome any members to give a check).
> 
> 
> Thanks.
> 
>> Thanks,
>> //richard
>>
>>
>>
> 


-- 
Chen Gang
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/