netdev - Re: difference in counters values between vlan and parent device

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: <4A48D191.9020903@trash.net>
Date:	Mon, 29 Jun 2009 16:37:05 +0200
From:	Patrick McHardy <kaber@...sh.net>
To:	Or Gerlitz <ogerlitz@...taire.com>
CC:	netdev@...r.kernel.org
Subject: Re: difference in counters values between vlan and parent device

Or Gerlitz wrote:
> I noticed that the rx byte counters for 8021q/vlan devices are showing lower
> quantities then the parent device. Below are three cuts of /proc/net/dev
> where one can see clearly that after running unidirectional traffic over
> the vlan the vlan TX packet/byte counters follow well the parent counters,
> where after running bidirectional traffic the parent RX byte counters are
> 30% (30M vs 22M) higher, the packet counters follow the parent.
> 
> Now, looking in net/8021q/vlan_dev.c :: vlan_skb_recv() I see this code,
> 
> 	stats->rx_packets++;
> 	stats->rx_bytes += skb->len;
> 
> I understand that skb->len as seen by the parent device may be different
> then the len seen by 8021q but I don't see how it can account to 30%.
> 
> For what it worth, this is 2.6.30
> 
> Or.
> 
> 1. both devices are fresh, just loaded the NIC driver (igb) and then setup a vlan
> 
> Inter-|   Receive                                                |  Transmit
>  face |bytes    packets errs drop fifo frame compressed multicast|bytes    packets errs drop fifo colls carrier compressed
>   eth1:       0       0    0    0    0     0          0         0    14456      55    0    0    0     0       0          0
> eth1.4001:       0       0    0    0    0     0          0         0     6434      23    0    0    0     0       0          0
> 
> 2. after running unidirectional traffic over the vlan (netperf UDP_STREAM)
> 
> Inter-|   Receive                                                |  Transmit
>  face |bytes    packets errs drop fifo frame compressed multicast|bytes    packets errs drop fifo colls carrier compressed
>   eth1:    1644      11    0    0    0     0          0         0 1136967388 17764927    0    0    0     0       0          0
> eth1.4001:    1490      11    0    0    0     0          0         0 1136959282 17764893    0    0    0     0       0          0
> 
> 3. after running bidirectional traffic over the vlan (netperf UDP_RR)
> 
> Inter-|   Receive                                                |  Transmit
>  face |bytes    packets errs drop fifo frame compressed multicast|bytes    packets errs drop fifo colls carrier compressed
>   eth1:30611926  546614    0    0    0     0          0         0 1160471756 18311532    0    0    0     0       0          0
> eth1.4001:22959330  546614    0    0    0     0          0         0 1160463650 18311498    0    0    0     0       0          0
> 

So the average packet size is 56 bytes on eth1 and 42 bytes
on eth1.4001. So the difference is 14 bytes, which is expected
since the mac header is pulled off before delivery to the
VLAN device.


--
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