lkml.org 
[lkml]   [2015]   [Mar]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 4/4] netfilter: Fix format string of nfnetlink_queue proc file
Am 13.03.2015 um 14:53 schrieb Pablo Neira Ayuso:
> On Fri, Mar 13, 2015 at 02:43:54PM +0100, Richard Weinberger wrote:
>> Am 13.03.2015 um 13:15 schrieb Pablo Neira Ayuso:
>>> On Fri, Mar 13, 2015 at 12:31:16PM +0100, Richard Weinberger wrote:
>>>> The printed values are all of type unsigned integer, therefore use
>>>> %u instead of %d. Otherwise an user can face negative values.
>>>>
>>>> Fixes:
>>>> $ cat /proc/net/netfilter/nfnetlink_queue
>>>> 0 29508 278 2 65531 0 2004213241 -2129885586 1
>>>> 1 -27747 0 2 65531 0 0 0 1
>>>> 2 -27748 0 2 65531 0 0 0 1
>>>
>>> I guess you want to access stats on dropped packets.
>>
>> Correct. :)
>>
>>> I prefer if you extend nfnetlink_queue to provide statistics through
>>> nfnetlink_queue, so you don't have to manually parse this text-based
>>> /proc entry and we can deprecate this interface. That shouldn't have
>>> been there in first place.
>>
>> You mean statistics via netlink attributes? I can add that!
>
> Add a new NFQNL_CFG_CMD_STATS command to request the statistics. If
> NLM_F_DUMP is set, then we'll basically provide the full list of
> instances. Otherwise, in case you want to retrieve stats for a
> specific netlink socket, you can use the netlink portID as index.
> And you'll have to add attributes for this new command, yes.

This was my plan. Thanks for the pointer!

>> But I think we should also fix the format string of the proc file
>> as the fix is easy and non-intrusive.
>
> Unfortunately we don't know how many people are relying on that
> output, I prefer to remain conservative and provide a proper netlink
> interface for this.

I understand your concerns but an application which is able to parse positive
and negative numbers can also parse pure positives.
Just made a small test application, glibc's %d in sscanf() can also deal with UINT_MAX.
And I don't expect that applications to check whether the returned values from
/proc/net/netfilter/nfnetlink_queue are between INT_MIN and INT_MAX.

That said, I'd have assumed that an user would report negative values as plain kernel bug.

Thanks,
//richard


\
 
 \ /
  Last update: 2015-03-13 15:41    [W:0.047 / U:11.664 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site