netdev - Re: [PATCH iproute2] ss: Show more info (ring,fanout) for packet socks

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]

Message-ID: <CA+FuTSfzZCwz-FAiq_K3Sq2ALR7rnziWaXq_27Q8OKBAyGJ0gA@mail.gmail.com>
Date:	Thu, 14 May 2015 16:39:01 -0400
From:	Willem de Bruijn <willemb@...gle.com>
To:	Vadim Kochan <vadim4j@...il.com>
Cc:	Network Development <netdev@...r.kernel.org>
Subject: Re: [PATCH iproute2] ss: Show more info (ring,fanout) for packet socks

On Thu, May 14, 2015 at 4:21 PM, Vadim Kochan <vadim4j@...il.com> wrote:
> On Thu, May 14, 2015 at 03:54:41PM -0400, Willem de Bruijn wrote:
>> On Thu, May 14, 2015 at 7:09 AM, Vadim Kochan <vadim4j@...il.com> wrote:
>> > On Wed, May 13, 2015 at 05:10:38PM -0400, Willem de Bruijn wrote:
>> >> > +               }
>> >> > +               if (has_fanout) {
>> >> > +                       uint16_t type = (fanout >> 16) & 0xffff;
>> >>
>> >> type can be modified by flags in the upper 8 bits. Better
>> >> mask those out and report them separately.
>> > Kernel puts only id & type in diag message, flags are stored in separated
>> > member of 'struct packet_sock'.
>>
>> I don't mean pdi_flags, but
>>
>>   PACKET_FANOUT_FLAG_ROLLOVER
>>   PACKET_FANOUT_FLAG_DEFRAG
> sure, I understood what you meant.
>>
>> which are passed as part of this u32 by the process during setsockopt
>> PACKET_FANOUT.
>>
> OK, its better to look into the kernel:
>
> net/packet/af_packet.c
>
> static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
> {
>         struct packet_fanout *f, *match;
>         u8 type = type_flags & 0xff;
>         u8 flags = type_flags >> 8;
>
>         ...
>
>         match->id = id;
>         match->type = type;
>         match->flags = flags;
>
>         ...
> }
>
> As I see this function is called from setsockopt, takes type_flags and splits it into type & flags
> and store them separately in struct packet_fanout.
>
> Then what I see in:
>
> net/packet/diag.c:
>
>
> static int pdiag_put_fanout(struct packet_sock *po, struct sk_buff *nlskb)
> {
>         int ret = 0;
>
>         mutex_lock(&fanout_mutex);
>         if (po->fanout) {
>                 u32 val;
>
>                 val = (u32)po->fanout->id | ((u32)po->fanout->type << 16);
>                 ret = nla_put_u32(nlskb, PACKET_DIAG_FANOUT, val);
>         }
>         mutex_unlock(&fanout_mutex);
>
>         return ret;
> }
>
> I see that to put flags here it needs to use po->fanout->flags and mix
> them into u32 val, but instead I see that it uses only id & type.
>
> Please correct me if I am wrong.

Thanks for the details. You are correct. Flags are not mixed in, like
are in the getsockopt case. In that case, type can only be 8 bits.

>
> Thanks,
>> >
>> >>
>> >> > +
>> >> > +                       printf("\n\tfanout(");
>> >> > +                       printf("id:%d,", fanout & 0xff);

ID can be 16 bits.

>> >> > +                       printf("type:0x%x", type);
>> >> > +                       if (type == 0)
>> >> > +                               printf("(hash)");
>> >> > +                       else if (type == 1)
>> >> > +                               printf("(lb)");
>> >> > +                       else if (type == 2)
>> >> > +                               printf("(cpu)");
>> >> > +                       else if (type == 3)
>> >> > +                               printf("(rollover)");
>> >> > +                       else if (type == 4)
>> >> > +                               printf("(random)");
>> >> > +                       else if (type == 5)
>> >> > +                               printf("(qm)");
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html